o
-.. _
-u‘..
' o
-
.
.--.. -
._,
o
a... -
a»
u
I
Q
. .

 

 

 

 

-
u. . ‘
'>. .~
5
.
-.. -
..._'-.‘

_ > ‘
a.....,l..“
.' . .-
u. ‘

.... Q-

o
y . .
I -...' .w.
.
.. . ‘
.‘g§_. .. -
\7 , .
. o
. .."‘ .
.l _
Q. -
s ”.“>
‘.
.-_.
._‘-' .. .. .
....
c,‘ -
0“
‘-
~_~ ‘."'.
.. .
H... ~ A
- u
‘s
w'.‘
. ..-
\“\ -v' .
.~‘-r
v._ .
. ~ “'
.‘s. ‘ _ -
. I
-,.--J
o
In. .
v
u.‘ '.‘
c
Q \o‘
t .1
'.
..-
'.

- --
Q‘.‘- ‘
- ..

ls“
§ ..4
'.
"
r. . ‘
A
.Q .. 1'
'I
D. .
§
an.‘ F.
o ."
.1
.. 7
§-- ‘
‘0

ABSTRACT

METROPOLITAN DOMINANCE AND THE PERSISTENCE OF THE
URBAN-RURAL FERTILITY DIFFERENTIAL: A DISTRIBUTIVE
APPROACH TO THE STUDY OF FACTORS AFFECTING URBAN-
RURAL FERTILITY IN THE UNITED STATES, 1960

by Rodger R. Rice

This dissertation focuses on factors associated with
the urban—rural fertility differential in the United States.
The parameters of the problem are presented in the form of
requisites for current differential fertility research.
Differential fertility analysis requires a causal framework.
Fertility is social group behavior explanable within an
ecological framework. Prediction of convergence of urban-
rural fertility levels does not necessitate termination of
research on this differential but does indicate need for
alternative approaches. Independent of convergence there
remains variation within urban and rural fertility which
requires explanation. Either a different set of factors
affect urban and rural fertility or the same factors exert
a different effect on urban and rural fertility. The tradi-
tional aggregative approach is rejected in favor of the dis-
tributive. Multiple regression analysis applied to inter-
community fertility variation permits a comparative analysis

of residential differential fertility patterns.

 

 

 

 

 

 

 

‘ v"
,,
..
..ru
--- ‘1‘
_-- ‘ a
-. -
o
-.- ""'
.
. -1 V— .
.91. *-‘< ’
"Q
av Ooh---
u.-.- ‘--‘
. v
.u._ . --
o....-.. -
.. ..
I .‘---_-
“‘- Q- .-
b
"h .- u ,
.~-.IIv
a “!i -
.-
"" u-»-
. . .
w.
‘ .. Q--..
-.
'0... ,
-. -.
-‘ '2...-
-.,_ ‘ v~
'. -"-_-
c-..
.. fl. .
‘-
-
..
.r
..
u“
‘i. .
O
.
.
R
.. ‘--
~fi
. .-
‘-
'0
u
‘ .--‘
V . s
‘ -\.
in
--~
._'_
‘.~;
Q
g
‘0 '-'..
.Q
‘Q
~ .
fl'c‘
‘ ~
5
'v.
s ..
‘Q
l--
'\
.-
. -
.
. .‘
'0 "a.
..
vs_
‘-
O
o
-
I
‘ .
‘s
-
.
~
~..
‘\
.
s
5
§
7
-
D‘ -‘
. .
Q
-.
Q
‘»
n'q
.,_ .
a-
,‘ §
-‘ .
Q
.
"t
..¢ .
I
H
.. Iv
I.§
I.- -
‘s
.
U
" J
‘. t
O
- n
'0.
‘Q
-
Q
\
.
,
._ «
'-
.
s
.
. .
5 I‘-
fl
'
n
. §
. I
Q, ‘_
‘ .
,
"
n-‘ ‘
q.
, -
‘
~

Rodger R. Rice

A review of empirical studies provides evidence (1)
tlurt no previous study had implemented these requisites but
(2) that contrasting patterns of differential fertility

annong residential categories are plausible. In view of this

‘A/ ‘l

\v/ .. ~-"

sand the conclusion that urban dominance theory logically
lnypothesizes the blurring of urban-rural differences, urban
dominance theory is rejected and metropolitan dominance
theory accepted as the theoretical framework by which to
explain contrasting patterns of differential fertility among
residential categories and to generate hypotheses for testing.

Seven hypotheses are derived from metropolitan dominance

theory:

1. Community social structure is a function of metro—
politan dominance; metropolitan dominance manifests
a different impact on community social structure in
urban and rural hinterlands.

2. Fertility is a function of community social structure
and metropolitan dominance; community social struc-
ture and metropolitan dominance manifest a different
impact on fertility in urban and rural hinterlands.

3. Urban and rural hinterland fertility is not only a
function of metropolitan dominance, but also of con—
ditions of its immediate locality.

4. Fertility is more a function of metropolitan domi—
nance in the urban hinterland, but more a function
of local community social structure in the rural
hinterland.

5. Metropolitan dominance is more important in account—
ing for variation in community social structure and
fertility in both urban and rural hinterlands in the
more metropolitan geographic divisions compared with
the less metropolitan geographic divisions.

6. In the more metropolitan geographic divisions metro-
politan dominance is more important in accounting
for variation in urban and rural hinterland fertility

u‘u
:—
.~

 

 

 

Rodger R. Rice

than local community social structure; in less metro-
politan geographic divisions local community social
structure is more important in accounting for varia—
tion in fertility than metropolitan dominance.

7. The impact of community social structure and metro-
politan dominance on fertility manifest fewer dif—
ferences when comparing the same type of hinterland
communities (urban or rural) on an interdivisional
basis than when comparing different types of hinter-

land communities (urban vs. rural) on an intradivi-
sional basis.

Basic unit of analysis is the residential component
of a county. Fertility, the dependent variable, is opera—
tionalized as the cumulative fertility ratio. Operational-
izing metropolitan dominance required deriving for all
counties in the nation a numerical value reflecting distance
from and size of a dominating metropolitan center. Community
social structure is represented by eight empirical variables:
employment of farmers and farm managers, farm laborers and
foremen, education, family income, female income, female
employment, ever—married females age 15—44 who were age
15-24 and 25-34. Analysis was limited to the white popula—
tion of conterminous United States. A multiple regression
analysis was performed for each of the residential catego-
ries at the national and nine divisional levels.

Analysis of statistical results provided confirmation
of all hypotheses but the fourth. Confirmation of six of the
hypotheses suggests the value of metropolitan dominance the-
ory as a framework by which to explain the differential

impact of factors in urban and rural hinterlands.

 

METROPOLITAN DOMINANCE AND THE PERSISTENCE OF THE
URBAN-RURAL FERTILITY DIFFERENTIAL: A DISTRIBUTIVE
APPROACH TO THE STUDY OF FACTORS AFFECTING URBAN-

RURAL FERTILITY IN THE UNITED STATES, 1960

BY

Rodger R: Rice

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Sociology

1967

 

! ’
c9319?
_:-ao«é€

PREFACE

Amateurs should be cautioned
not to rush headlong into re-
gression studies involving many
variables. Some peOple think
there is magic in the collection
of vast amounts of data--that by
some alchemy multiple regression
will yield authentic information
from careless measurements on
heterogeneous material. The fact
is that hazards increase with the
extent and complexity of the in-
vestigation.... However, do not
be deterred if you have well taken
measurements on carefully chosen
material, and if you have definite
questions whose answers lie in the
methods to be used. As compared
to the labor of getting the data
the calculation of regression
statistics is easy.

George W. Snedecor, Statistical
Methods (Ames: Iowa State Univer-
sity Press, 1956), p. 434.

 

ii

ACKNOWLEDGMENTS

"Giving birth“ is a universal human experience.
IDemographers treat it objectively, mothers live it subjec-
tively. But "giving birth" is not the experience of mothers
alone; it has happened to me. What mother could have ever
survived a conception—parturition period of over five years?
It happened to me. Intercourse, conception, "morning sick-
ness," "depression," "labor pains," etc., they happened to
lne. I am not knocking mothers, believe me, for without them
this study would not have been possible. Retrospection at
this point, however, tells me that "giving birth" is an ex-
perience that takes on numerous forms in addition to child-
‘birth. Producing a dissertation is one of these forms.

Yes, this is my baby, but many have contributed to
its birth. Without them, an inevitable miscarriage. I wish
to give special acknowledgment to Dr. J. Allan Beegle who
served both as impregnator and obstetrician. An amazing
feat? An amazing person, as many of my fellow graduate
students (maternity mates?) do resound. I am appreciative
of his fertile mind, his patience, his gentleness, and his
‘untiring interest in my case. Indeed, he is a good father
and a good doctor all rolled into one.

His professional staff is to be recognized and
appreciated as well. I wish to thank the good doctors,
J} W. Artis, Harvey Choldin, and'Walter E. Freeman of the
.Department of Sociology and Charles Press, Chairman of the
Department of Political Science, for their fruitful comments
and maternal care. Without their professional concern, an
inevitable stillbirth.

Finally, to my life-long midwife and handholder,
Ruth, and to her cherubic attendants, Sheri and Mark, how
can I repay you for your unrelinquished vigilance and per—
petual anticipation? As birth ends with the renewing of
family ties, I ask only that you become my family once again.
(And thank you too, Babar, for your companionship. Our
family circle will be smaller without you.)

iii

 

fig.

—-_...
a- -
o
«- o
.—

Q
.

—
-c

u
‘v-
9
on
\

PREFACE

TABLE OF CONTENTS

.ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . .

LIST OF

LIST OF

Chapter

I.

II.

TABLES . . . . . . . . . . . . . . . . . . .

APPENDICES . . . . . . . . . . . . . . . . .

INTRODUCTION TO THE PROBLEM . . . . . . . . .

Fertility as a Social Phenomenon . . . . .
Fertility as an Important Demographic
Variable . . . . . . . . . . . . . . . .
Differential Fertility . . . . . . . . . .
The Goldberg Hypothesis . . . . . . . . . .
The Causal Approach to Differential
Fertility . . . . . . . . . . . . . . . .
Basic Design of the Study: Distributive
Approach . . . . . . . . . . . . . . . .
Summary: Requisites of Needed Research on
the Urban-Rural Fertility Differential .
Origin and Organization of the Study . . .

REVIEW OF RELEVANT LITERATURE . . . . . . . .

Criteria Employed to Select Empirical
Studies for Intensive Review . . . . . .
Intensive Review of Selected List of
Empirical Studies . . . . . . . . . . . .
General Characteristics of Studies . .
Study Design of Empirical Studies . . .
Summary of Findings, by Variable . . .
Resumé of Empirical Propositions from
Intensive Review of Empirical Studies . .
Distance . . . . . . . . . . . . . . .
Agricultural Occupation . . . . . . . .
Education . . . . . . . . . . . . . . .
Income . . . . . . . . . . . . . . . .
Female Employment . . . . . . . . . . .
Age Composition . . . . . . . . . . . .
Conclusion . . . . . . . . . . . . . . . .

iv

Page
ii
iii
vi

xvi

10
12
46

56
6O

74
76

79

82

89
96
98
101

136
137
138
139
140
141
141
142

 

 

Chapter

III. THEORETICAL FRAMEWORK FOR STUDY OF
DIFFERENTIAL FERTILITY . . . . . . . .

Status of Differential Fertility Theory
Critical Guidelines for a Theoretical
Design for the Study of Differential
Fertility . . . . . . . . . . . . . .
Theoretical Framework . . . . . . . . .
Urban Dominance Theory . . . . . .
Metropolitan Dominance Theory . . .
Summary: Hypotheses Derived from
Metropolitan Dominance Theory . . . .

IV. METHODOLOGICAL PROCEDURES . . . . . . . .

Conceptual Framework: Specification of
Variables . . . . . . . . . . . . . .
Community Social Structure . . . .
Metropolitan Dominance . . . . . .
Community Fertility Behavior . . .
Levels of Comparison . . . . . . .

Statistical Framework: Specification of

Statistical Techniques . . . . . . .
Multiple Regression Model . . . . .
Beta Coefficients . . . . . . . . .

Coefficient of Multiple Determination

Zero-Order Correlation Coefficients
Partial Correlation Coefficients .
Multiple-Partial Correlation
Coefficients . . . . . . . . . .
Statistical Tests . . . . . . . . .

V. ANALYSIS OF DATA: NATION AND DIVISION .

Hypothesis 1 . . . . . . . . . . . . .

Hypothesis 2

Hypothesis 3

Hypothesis 4

Hypothesis 5 . . . . . . . . . . . . .

Hypothesis 6

Hypothesis 7

Summary of Findings . . . . . . . . . .

VI. REFLECTIONS AND IMPLICATIONS FOR FURTHER
RESEARCH . . . . . . . . . . . . . . .

BIBLIOGRAPHY . O O O O O O O O O O O C O O O O C

Page

147

148

150
165
174
191

224

227

227
228
231
237
241

245
246
248
250
251
253

254
256

262

262
272
289
298
322
332
335
345

354

360

~44

0‘.--

 

. I.
.. .-
. ....
. ,,
..
.. o‘-
.

.

 

D -

 

Table

LIST OF TABLES

Number and percentage change of children under

5 years old per 1,000 white women 20 to 44

years old, by nation and divisions, urban and

rural: 1800 to 1840 and 1910 to 1960 . . .

Percentage change of number of children under

5 years old per 1,000 white women 20 to 44
years old, by nation and divisions, urban:
1800 to 1840 and 1910 to 1960 . . . . . . .

Percentage change of number of children under

5 years old per 1,000 white women 20 to 44
years old, by nation and divisions, rural:
1800 to 1840 and 1910 to 1960 . . . . . . .

Percentage of total population urban for
conterminous United States and divisions:
1800 to 1840 and 1910 to 1960 . . . . . . .

Absolute differences in urban-rural number of

children under 5 years old per 1,000 white

women 20 to 44 years old, by nation and divi—

sions: 1800 to 1840 and 1910 to 1960 . . .

Urban as a proportion of rural number of
children under 5 years old per 1,000 white
women 20 to 44 years old, by nation and
divisions: 1800 to 1840 and 1910 to 1960 .

Rank order correlations of percentage of total
population urban with absolute differences in

urban-rural fertility ratios and urban to

rural fertility ratios, for conterminous United
States and divisions: 1800-1840 and 1910-1960

Number of children ever born per 1,000 ever-
married white women by residence and age for
conterminous United States and divisions,

1960 C O O O O O O O O O C O O O O O C O O 0

vi

Page

19

23

23

24

29

29

33

37

 

'Iable

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Percentage distribution of urban, rural-non-
farm and rural-farm parts of counties by
children ever born per 1,000 ever-married
white women age 15—44 for conterminous United
States and divisions: 1960 . . . . . . . . .

Urban-rural components of decline in number

of children under 5 years old per 1,000 white

women 20 to 44 years old, by divisions: 1810—
1940 . . . . . . . . . . . . . . . . . . . . .

Urban—rural movement between 1935 and 1940

of native white women 15 to 49 years old, by
marital status and number of own children
under 5 years old in 1940 . . . . . . . . . .

Distribution of the white civilian population
18 years of age and over, by residence, farm
or nonfarm birthplace, United States, 1958 . .

Summary table of fertility studies selected
for intensive review . . . . . . . . . . . . .

Summary table of relationships of fertility
to independent variables based on fertility
studies selected for intensive review . . . .

Tabular summary of empirical propositions by
independent variable and residence category:
expected relationships of fertility to
selected independent variables and expected
relative importance of independent variables
in determining fertility variation within
residence groups . . . . . . . . . . . . . . .

Distribution of population in the United
States and divisions, by metropolitan status
1960 O O C O O O O O O O O O O O O O O O O O O

Zero-order correlation of metropolitan domi-
nance and employment in agricultural occupa—
tions (urbanization) for conterminous United
States and divisions, by residence: 1960 . .

Zero—order correlation of metropolitan domi-
nance and socio-economic status for contermi-
nous United States and divisions, by residence:
1960 O O O O O O O O O O O O O O O O O O O O 0

vii

Page

44

50

52

54

90

102

143

220

264

266

Table

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Page

Zero-order correlation of metropolitan domi—

nance and wife's alternative opportunities

for conterminous United States and divisions,

by residence: 1960 . . . . . . . . . . . . . . 268

Zero-order correlation of metropolitan domi—

nance and demographic age structure for

conterminous United States and divisions,

by residence: 1960 . . . . . . . . . . . . . . 270

Zero-order correlation of fertility and metro-
politan dominance for conterminous United
States and divisions, by residence: 1960 . . . 273

Zero-order correlation of fertility and employ—

ment in agricultural occupations (urbanization)

for conterminous United States and divisions,

by residence: 1960 . . . . . . . . . . . . . . 275

Zero-order correlation of fertility and socio-
economic status for conterminous United States
and divisions, by residence: 1960 . . . . . . . 278

Zero-order correlation of fertility and wife's
alternative opportunities for conterminous

United States and divisions, by residence:

1960 . . . . . . . . . . . . . . . . . . . . . . 281

Zero-order correlation of fertility and demo—
graphic age structure for conterminous United
States and divisions, by residence: 1960 . . . 283

Coefficients of multiple determination (R2)

of fertility and variables in multiple regres-

sion equations for conterminous United States

and divisions, by residence: 1960 . . . . . . . 287

Comparison 3f coefficient of multiple deter-
mination (R ) and multiple— partial correlation
coefficient (r2 ) of fertility and community

social structure controlling for metropolitan
dominance for conterminous United States and
divisions, by residence: 1960 . . . . . . . . . 292

Comparison of coefficient of multiple deter-
mination (R2) and multiple— partial correlation
coefficient (r2 ) of fertility and employment

in agricultural occupations (urbanization)
controlling for metropolitan dominance for
conterminous United States and divisions, by
residence: 1960 . . . . . . . . . . . . . . . . 293

viii

 

.
v-4.-. .

 

-
-

o

‘ o

. u.

‘0».

 

 

:-
O~
\ .'
.‘os
n
'Qr
§-. -
o.--
‘0....
-
.‘-‘
‘Q-
‘ D—
a...
.
...~
.—I.
.
q»...
- §
‘
h't-
b .
-_.‘
...
'\
'2'.
§
,
s...
‘v.
~
-...‘
fl- -
n
-‘c‘
h.
u_.:
-._
-‘I
g.‘
g
§"
§..
a.-
'l
'A.‘
‘
'Q
_ .
h- a
.
i
v.‘
‘F
I
:-
C
.
§‘.
§..
A
.
h..
5-.
”V
Q
."‘
I...
-.~
I" .
‘.
s
-
‘fl.
~
.
A
b.‘
5-.
Ԥ
-
n
0'.
D- u
Ԥ
§
.
‘0-
's
n
A‘-
..i
“‘
§

 

O.
.-
~
-
—-
a-
5-
a.
o.
~.
‘.
h-
u
-v-
u
‘
u.
‘.I'
--
§.
"
.o
.-
‘6
\
~~
o-
.._
'u
“s
a
s
'.
d~
~._
’.'
‘.
‘
”A
5
s
a"
'-
'—
...
-
"s
‘
5
.
-._
'u
‘.
a
7
‘Q.
A
g
c
a,‘
'-
a
y
‘.
-
s
§
4..
‘0
‘.
c

Table

29.

30.

31.

32.

33.

34.

35.

36.

Comparison of coefficient of multiple deter-
mination (R2) and multiple—partial correlation
coefficient (r2) of fertility and socio—economic
status controlling for metropolitan dominance
for conterminous United States and divisions,

by residence: 1960 . . . . . . . . . . . . . .

Comparison of coefficient of multiple deter—
mination (R2) and multiple-partial correlation
coefficient (r2) of fertility and wife's
alternative opportunities controlling for
metropolitan dominance for conterminous United
States and divisions, by residence: 1960 . . .

Comparison of coefficient of multiple deter-
mination (R2) and multiple—partial correlation
coefficient (r2) of fertility and demographic
age structure controlling for metropolitan
dominance for conterminous United States and
divisions, by residence: 1960 . . . . . . . . .

Beta coefficient (B-) and rank of independent
variable by relativg importance in accounting
for variation in fertility for New England
division, by residence: 1960 . . . . . . . . .

Beta coefficient (B-) and rank of independent
variable by relativg importance in accounting
for variation in fertility for Middle Atlantic
division, by residence: 1960 . . . . . . . . .

Beta coefficient (B') and rank of independent
variable by relative importance in accounting
for variation in fertility for East North
Central division, by residence: 1960 . . . . .

Beta coefficient (B-) and rank of independent
variable by relative importance in accounting
for variation in fertility for West North
Central division, by residence: 1960 . . . . .

Beta coefficient (B-) and rank of independent
variable by relativg importance in accounting
for variation in fertility for South Atlantic
division, by residence: 1960 . . . . . . . . .

ix

Page

294

295

296

306

308

309

311

312

Table Page

37. Beta coefficient (B-) and rank of independent
variable by relative importance in accounting
for variation in fertility for East South
Central division, by residence: 1960 . . . . . 314

38. Beta coefficient (B-) and rank of independent
variable by relative importance in accounting
for variation in fertility for West South
Central division, by residence: 1960 . . . . . 315

39. Beta coefficient (B-) and rank of independent
variable by relative importance in accounting
for variation in fertility for Mountain
division, by residence: 1960 . . . . . . . . . 316

40. Beta coefficient (B-) and rank of independent
variable by relative importance in accounting
for variation in fertility for Pacific
division, by residence: 1960 . . . . . . . . . 318

41. Average rank among divisions of relative
importance of independent variables in
accounting for variation in fertility mea—
sured by Beta coefficients, by residence:
1960 . . . . . . . . . . . . . . . . . . . . . . 320

42. Zero-order correlation coefficients (r2) of
metropolitan dominance and community social
structure for divisions by level of metro—
politanization, urban: 1960 . . . . . . . . . . 324

43. Zero-order correlation coefficients (r2) of
metropolitan dominance and community social
structure for divisions by level of metropol-
itanization, rural-nonfarm: 1960 . . . . . . . 327

44. Zero-order correlation coefficients (r2) of
metropolitan dominance and community social
structure for divisions by level of metropol—
itanization, rural—farm: 1960 . . . . . . . . . 328

45. Rank of metropolitan dominance by relative
importance for variation in fertility based
on Beta coefficients and zero-order correla—
tion coefficient of metropolitan dominance
and fertility for divisions by level of
metropolitanization, by residence: 1960 . . . . 330

... ._ ..u ~q ~.« ‘4‘ . . ._. . . . .v
r. L. .o . . . . o . :. .. o. o . I.
... u .1. :. . . ..1 . . v. .4.
n . . On . . .H — ... .. — ... _u .. ..
a.“ o. .. .n. I. o‘. o. p“ .x. .. n. ... .3
. . ..u ..H
n. .. o.

 

Table Page

46. Average rank of independent variable by rela—
tive importance in accounting for variation
in fertility measured by Beta coefficients
by level of metropolitanization of divisions
and by residence: 1960 . . . . . . . . . . . . 334

47. Summary of results for multiple comparison
tests among residential sectors of contermi—
nous United States divisions . . . . . . . . . . 338

48. Summary of results for multiple comparison
tests between divisions of conterminous
United States by residential sector . . . . . . 342

49. First-order partial correlation of fertility
and employment in agricultural occupations
(urbanization) controlling for metropolitan
dominance for conterminous United States and
divisions, by residence: 1960 . . . . . . . . . 372

50. First-order partial correlation of fertility
and socio—economic status controlling for
metropolitan dominance for conterminous United
States and divisions, by residence: 1960 . . . 373

 

51. First-order partial correlation of fertility
and wife's alternative opportunities control-
ling for metropolitan dominance for contermi-
nous United States and divisions, by residence:
1960 . . . . . . . . . . . . . . . . . . . . . . 374

 

52. First-order partial correlation of fertility
and demographic age structure controlling for
metropolitan dominance for conterminous United
States and divisions, by residence: 1960 . . . 375

 

53. Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for urban
part of county--conterminous United States:
1960 . . . . . . . . . . . . . . . . . . . . . . 376

.54“ Results of the analysis of factors influencing
number of children evern born per 1,000 ever—
married white females, age 15—44, for rural-
nonfarm part of county--conterminous United
States: 1960 . . . . . . . . . . . . . . . . . 377

 

 

 

po-

 

..
U
.

-~vv~.-‘.

was.--

,

“~V- V':V

-

.‘U. ‘Q

.
n. .

Jr
Q... g .

-
.‘¢~ ..
u... ‘
o5u-‘i‘
—-- ‘v _

.-
“1...... _
"Vv.. _
- b.
~‘..~_‘

.
r... g.
-
~"' .-
.
. Q
"- ..
-_.~ ‘
..._.-_
. - V
"*.L_.
--,"‘ .
‘ I
h“"§-
-‘-.~'—

.
"~~.-‘_
‘0
In
-..- .
- n
'l: L
..‘_-‘~
.-—. '
.._ K“
~:".- .

.
.‘-‘~_
-
‘zv- ,
‘ n. _ _
....
.
.

.

4:-
...‘ .
- o
."
"\ D-
l ‘
‘ _‘_\
."~,
.n‘--
‘_'v..
“n.‘-, .
_“s -
h“ -.
-~‘

. -‘
I.

4r

.-

' c
-

"Q .
n‘. “
._h“ \
. ‘.

p
"u.--.

' 3
~.-’ '
.. _'
.‘.. ‘

¢_ .

-
-..

a. ‘ ~
—.‘~

'1‘.
~‘

s -.
~§_ o

...
s
I ~
.5:
'y.‘ C

“‘:.

‘-
n. 'r
n. .

Ԥ
“

-V‘
g ‘
Ҥ .

‘v
- a
‘.

_
.Ԥ .

b

u‘
.v I
a,“
“'
v
0
.

v-_ ‘
‘-:. ..
C

s‘:
5

\
I, ._
-‘.. ‘-
‘
~_
~...'-’,.
‘Q ..
~-
. O
b‘-"‘
Qfi“ ’;
‘- v‘
r.
‘
‘4’
.U

Table Page

55. Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural-
farm part of county——conterminous United States:
1960 . . . . . . . . . . . . . . . . . . . . . . 378

56. Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for urban
part of county-—New England division: 1960 . . 379

57. Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural-
nonfarm part of county-—New England division:
1960 . . . . . . . . . . . . . . . . . . . . . . 380

58. Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural-
farm part of county--New England division:
1960 . . . . . . . . . . . . . . . . . . . . . . 381

59. Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15—44, for urban
part of county-~Middle Atlantic division:
1960 . . . . . . . . . . . . . . . . . . . . . . 382

 

60. Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural—
nonfarm part of county-—Middle Atlantic divi—
sion: 1960 . . . . . . . . . . . . . . . . . . 383

 

 

61. Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15—44, for rural—
farm part of county—-Middle Atlantic division:
1960 . . . . . . . . . . . . . . . . . . . . . . 384

62. Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for urban
part of county-~East North Central division:
1960 . . . . . . . . . . . . . . . . . . . . . . 385

xii

 

 

 

 

 

. O

a
in. ~
.

or-
Q. . -
-
'_:
' ‘~-
. Q”
I~~*
—-,

. r

a-
-v.

5“ .

‘-
'4'

.-

.*~.

. "-

 

\
u-
r
~.
_-
I
._
‘N
v
9‘
”a
'.
--
t
o
-

 

 

!

(I

'tl

l('

(H

‘1

 

 

Table

63.

64.

65.

66.

67.

68.

69.

70.

Page

Results of the analysis of factors influencing

number of children ever born per 1,000 ever-

married white females, age 15—44, for rural-

nonfarm part of county-~East North Central

division: 1960 . . . . . . . . . . . . . . . . 386

Results of the analysis of factors influencing

number of children ever born per 1,000 ever—

married white females, age 15-44, for rural-farm
part of county-—East North Central division:

1960 . . . . . . . . . . . . . . . . . . . . . . 387

Results of the analysis of factors influencing

number of children ever born per 1,000 ever-

married white females, age 15-44, for urban

part of county--West North Central division:

1960 . . . . . . . . . . . . . . . . . . . . . . 388

Results of the analysis of factors influencing

number of children ever born per 1,000 ever—

married white females, age 15—44, for rural-

nonfarm part of county--West North Central

division: 1960 . . . . . . . . . . . . . . . . 389

Results of the analysis of factors influencing

number of children ever born per 1,000 ever-

married white females, age 15-44, for rural-

farm part of county--West North Central divi—

sion: 1960 . . . . . . . . . . . . . . . . . . 390

Results of the analysis of factors influencing

number of children ever born per 1,000 ever-

married white females, age 15—44, for urban

part of county—~South Atlantic division: 1960 . 391

Results of the analysis of factors influencing

number of children ever born per 1,000 ever-

married white females, age 15—44, for rural—

nonfarm part of county--South Atlantic divi—

sion: 1960 . . . . . . . . . . . . . . . . . . 392

Results of the analysis of factors influencing

number of children ever born per 1,000 ever—

married white females, age 15-44, for rural-

farm part of county—-South Atlantic division:

1960 . . . . . . . . . . . . . . . . . . . . . . 393

xiii

-u

w.
1.. .
w...

 

 

~A

Table

71.

72.

73.

74.

75.

76.

77.

78.

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for urban
part of county-—East South Central division:
1960 . . . . . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for rural-
nonfarm part of county-—East South Central
division: 1960 . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural—
farm part of county--East South Central divi-
sion: 1960 . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for urban
part of county-—West South Central division:
1960 . . . . . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for rural-
nonfarm part of county-—West South Central
division: 1960 . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for rural-
farm part of county--West South Central divi-
sion: 1960 . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15-44, for urban
part of county--Mountain division: 1960 . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever
married white females, age 15-44, for rural-
nonfarm part of county-—Mountain division:
1960: . . . . . . . . . . . . . . . . . . . .

xiv

Page

394

395

396

397

398

399

400

401

c
.‘v‘ .n-‘
ono. .,.
a, ‘
Jr
a.-. .
~
‘0‘. .‘
1-. ~
o u-..._
- -In'
.-
0-“--.
-«".~ ‘
‘ a
-‘...‘ .
.
0.,.. --
‘-.. V.
o
. I
.-. .‘
~_‘ ~
’ “ o-._
.--...'
“-5..-.
--"‘.

- D
h-..-‘--
a
D‘.."-
"-0.-“

., -

dr
Q-..
‘ I

‘R

‘~

Q

_‘ ‘
.u,‘.-‘
I.

-. ‘V
“is..- "
h- ~
4"...
_-.‘-‘ .
n
.u

-" .-
-§

 

 

 

Table

79.

80.

81.

82.

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15—44, for rural—
farm part of county—-Mountain division:

1960 . . . . . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15-44, for urban
part of county—-Pacific division: 1960. . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever—
married white females, age 15—44, for rural-
nonfarm part of county—~Pacific division:

1960 . . . . . . . . . . . . . . . . . . . .

Results of the analysis of factors influencing
number of children ever born per 1,000 ever-
married white females, age 15—44, for rural—
farm part of county--Pacific division: 1960 .

XV

Page

402

403

404

405

 

vm
.._

\-

v

Appendix

A.

 

 

“A

LIST OF APPENDICES

Page
DETAILED RESULTS OF MULTIPLE REGRESSION
EQUATIONS (See Tables 49—82, List of
Tables) . . . . . . . . . . . . . . . . . . 371
CHECKLIST FOR SELECTING STUDIES FOR
INTENSIVE REVIEW . . . . . . . . . . . . . . 406

xvi

 

 

 

 

 

.
m": -.
- ‘ '
..-__~_ -
-_ .~ ‘
‘ d.. -
-..¢‘.m.~
’ .
;-‘...- ‘-~
_
\.~ ~
-..- “‘
.-. -- .
‘. _. ‘
..‘.-- -
. «
..~:§.-.‘. ‘
- n.
b‘
‘.-.‘_--
c
-
.- ‘-
\
._~ -
‘-
T Q
~.- ‘ ‘- ‘
~-
.. c.‘ ‘
. ..
. .-
.. - g‘ . -
~.. ‘ 7
‘~ ‘~‘
~‘v_‘.
J1 ~ .~ ‘
-- .~
I _
.
. - .
. - “H
. -‘~ ~ - ‘ ‘-
~d“‘
‘.
.‘ ‘ V
.- -.‘
n. w.
I o
.1
-.
I 'z-
. ‘ .
.-\. ‘_-.
‘d-
‘ a-
\
- .
‘.‘
‘0
.
..‘.
‘-
. ~ g
u‘ “3".
_.‘ . -
-_A
.
\~
‘
s|. I
. ‘
~ u‘N
- -'
._-‘
.
‘~
~\
s‘ ‘ .
-
.‘

 

 

 

CHAPTER I

INTRODUCTION TO PROBLEM

Fertility as a Social Phenomenon

 

Fertility is a common but complex phenomenon. It
affects many aspects of a society, is affected by the same,
and captures the interest of many levels of society. It
affects and is reflected in industrial manpower, family
structure, educational facilities, housing, ad nauseam.
Fundamentally it is society's recruitment process.

Its relevance is so extensive that no discipline can
claim it in its entirety. Ryder has commented that "the
fields of learning which have been most immediately con—
cerned with and instrumental in the understanding of fertil-
ity are sociology, biology, and, to a lesser extent, econom-
ics, anthropology and psychology. No science concerned with
man has ignored or could properly ignore the 'facts of life.'"1

In this study it will be argued that fertility is to

a large extent a social phenomenon, i.e., though it is a

1N. B. Ryder, "Fertility," in Philip Hauser and Otis
IMMfley Duncan, The Studyyof Population (Chicago: University
ofcnucago Press, 1959), p. 400.

 

..
...-_- 0' .

o

.,. ‘fibbv-
- — .—

--- 0---.

 

--
c-

.
‘7’. ...
' O;

o
.
.
,
.
I
a

It
1'

-- -.
.-- .. . _ -
‘ -

.,

\
..-.-_-__
..,,_- --
.~.-‘ ~ .
"9..

D '-

 

-
n.
.'v-
_ -
‘V‘.--
Y
-.
.~ . '
--_
..y‘>v
-
-
O
-
.
-.
~v

.

. -.

.-‘ _

t_‘

u.
.

u"-
“
‘O

“_u
-

-V

s ‘-,

‘Q-

._.
.s

--
-b

‘ l
.,.
h ‘

’

_ ..

~-

.~

F“
s
--.
‘

a '-
.‘-

-
“-..

-‘\--
“

_ .

---.
~.\.
‘

‘i
--‘C
s o
‘ Q ‘
“.~
~
.--
n‘ ‘ .
a
‘ u
‘
‘-
. n.
‘.
. .
‘Q Q s
" ~ -
~
I
"v
_ »
-§ 9
i
u
.
.I. ‘
\“s.
-‘, ‘
‘~

 

 

biological fact, social factors play a significant role in
the determination of the biological result. Considering the
absence of modern contraceptive devices and controls, it is
quite likely that biological factors are more important in
determining fertility levels in primitive and underdevelOped
societies. However, Lorimer2 and Davis and Blake3 have sug-
gested the possibility that even in underdevelOped areas
fertility levels are influenced significantly by social and
cultural factors. Davis and Blake write:

A striking feature of underdeveloped areas is that vir—
tually all of them exhibit a much higher fertility than
do urban-industrial societies. This well-documented but
insufficiently analyzed fact is known to be connected
with profound differences in social organization as
between the two types of society, and is, therefore,
significant for the comparative sociology of reproduc-
tion. The clarity and importance of the contrast, how-
ever, should not be allowed to obscure the equally
important fact that underdeveloped areas themselves
differ markedly in social organization, and that these
differences appear to bring about variations in fertil-

ity.q

Previously with respect to urban-industrial societies there

 

was a "respectable body of opinion to the effect that group

differences in fertility reflected differences in biological

2Frank Lorimer, Culture and Human Fertility (New
York: International Documents Service, Columbia University
Press, 1955).

 

3Kingsley Davis and Judith Blake, "Social Structure
amiFértility: {An Analytical Framework," Economic Develop-
33m; and Cultural Change, Vol. IV (1956), pp. 211-35

 

4Ibid., p. 211 (italics mine).

 

 

 

capacity to reproduce,"5 a view advocated as recently as the
1920's and 30's by such leading demographers as Gini6 and
Pearl.7 This Opinion has since subsided in the shadow of
the sudden upsurge in fertility levels of urban populations.
Beginning with the Indianapolis Study many efforts
have been made to relate psychological factors to fertility
differences, but seemingly with little success. Kiser and
Whelpton's own evaluation of the Indianapolis Study's effort
in this direction suggests the insignificance of such factors.

The Indianapolis Study presents a challenge to learn the
reasons for the overriding influence of socio-economic
status. There is good reason to believe that it is not
socio—economic status per se but rather the underlying
attitudes and psychological characteristics of these
classes that account for the fertility behavior. And
yet, whereas characteristic patterns of fertility dif-
ferentials are found consistently in classifications by
socio-economic status, most classifications by psycho-
logical characteristics within socio—economic groups
fail to show such patterns.BI

h1a.critica1 review of the Indianapolis Study, Hauser and

Duncan assail the authors for not appearing "to entertain

5Clyde V. Kiser, "Differential Fertility in the

United States," in National Bureau of Economic Research,
Demographic and Economic Change in Develgped Countries
(Princeton: Princeton University Press, 1960), p. 77.

_ 6C. Gini, "The Cyclical Rise and Fall of Population,"
JJIPOpulation, Harris Foundation Lectures (Chicago: Univer-
sity of Chicago Press, 1929).

7Raymond Pearl, The Biology of Population Growth
(NEW York: Knopf, 1925).

. 8C. V. Kiser and P. K. Whelpton, "Resume of the
Inxhanapolis Study of Social and Psychological Factors

Affecting Fertility," Population Studies, Vol. VII (1953),
P- 08.

 

 

 

..
a
. ‘ v ...-
2.0-"-
I'-
' r“ 'N' .
' o h . ‘ ~
5 ..u
.0.
' Or' ‘ r
'. a... . ..
- ~
‘.1
.U-D'
'
-0” -'
t.y-:- ~|-'. ~-
. ba-
‘-.'
4.1

o
A. ".

...... _
.. . u -d.
.. -__ v-
.
' .
""- ~q-— .

s u
""-'v w-..u—

' a
"‘ ’- a- -......_

v
K ' I .
n. .' ‘.-.~ ~
.
n. s..gg '._ r ‘

....‘.-.‘- .
o

 

 

U. _ .
- , a
-" u. -:
.-‘ ‘. “ J
b 'n n- -. _
h _ ‘
9" .. _~ .' v
.0
. --
.‘. ._'.
. ‘ ‘
"o .-‘.
.
-‘. v. .
-
‘°“.¢.,_“
.._
‘ . ‘..‘-c -. - v-
‘."I . h
I..-...‘-'
H.-

. ..' -
.‘h’ 5‘: “ s ,.
sue-” ‘ .

-_ --
“n
-- u '0 ‘ I.
" ' h v ,
0 I.
. '0 .
..
-‘-‘_
‘I - ‘
‘J "‘c
‘:.~
..
a
4
‘
'C
t. ..
V ‘Q‘-
.
a.
u _ g
.i
“- o “.
v
a |
"-§n

 

smriously the supposition that psycho-social variables actu-
ally are not very important or useful as eXplanatory factors.
nlfact, their own comments on their results provide a re-
markable example of maintaining a hypothesis tenaciously in
the face of rather consistently negative evidence."9

Studies conducted since the Indianapolis Study have contin-
Lmd to discover consistently low correlations between psy-
chological variables and fertility, even when fertility is
Immsured as "desired size of family," a measure which would

smem to be at a level more relevant to a psychological

analysis.lo

Another important consideration is the extent to
“mich intentional family limitation practices explain dif—
ferences in fertility levels of various subgroups of modern
Society. .A United Nations publication in an extensive

review of the economic and social factors affecting fertility

9P. Hauser and O. D. Duncan, The Study of Population

(Chicago: University of Chicago Press, 1959), p. 99.

10See as examples of such studies: R. Gutman and I.
Bender, "Some Sources of Variation in Family Size of College
<haduates," Milbank Memorial Fund Quarterly, XXXV (July,
1957), 287-301; C. F. Westoff, P. C. Sagi, and L. W. Kelly,
'Tertility Through Twenty Years of Marriage," American
§Qgiological Review, XXIII, No. 5 (October, 1958), 549-56;
C.EK Westoff, R. G. Potter, Jr., P. C. Sagi, and E. G.
Dfishler, Family Growth in Metropolitan America (Princeton:
ITinceton University Press, 1961); C. F. Westoff, R. G.
Enter, Jr., and P. C. Sagi, The Third Child (Princeton:
Iminceton University Press, 1963); and David Goldberg, "Some
Ikment Developments in.American Fertility Research," in
National Bureau of Economic Research, Demographic and Eco-
RgflgkChange in Developed Countries (Princeton: Princeton
Ikuversity Press, 1960), pp. 137-51.

 

 

 

 

 

 

 

.
‘ O
.’.-...o 5".
' -- - .1- h -
"nu--5... 5
. . -
or: -;Q up; c '
... 0......‘ .‘v d

 

 

" *. . 1‘
'7‘ -
-.- _...‘
..‘.... .‘~
'h u- _
”'--.~ .1.‘

 

 

.
...
- , "u.

.-
-"‘..~

 

 

 

 

 

.__‘
- a.
'1 ~ .‘Q
"' \ _ 1
A...
~-... \
H ‘A_
v
. .
.,_
~-_. H“
.._ -
‘ d
O."-
I; -
'-. “.-
s§ -. ._
o._ v:
‘H,.
.. .
s ..-'
‘.~: ~
I. \«a
-' n1
.
--.‘-.:
. . ‘
" V” :. -
‘ ‘.-\ .
-.- ~
., .
."
H- ..
"‘~ ~ ‘
. 'V
‘¢ a u
--
. e _‘
.
.\ u .
Q" ‘1‘-
1
. 5“: 2.
s

 

 

 

IEPOItS that "in recent years, there is almost universal

agreement that the major part, if not all (italics mine), of

 

the decline in family size has been brought about by the

practice of family limitation. . . . Another relevant find—
ing of the existing studies is that the differences between
the size of families of different socio—economic, religious
and other groups can be eXplained largely by differences in

11

the extent and effectiveness of family limitation." While

several studies, since the initial finding of the Indianap—
cflis Study, have shown a direct relation between socio-
emonomic status and family planning,12 they have never found
a one-to-one relationship, and quite often the correlation
has'been quite low. It is assumed here that family planning
and limitation is essentially the equalization of number of
Children wanted to number of children born. A study of
Westoff, Mishler, and Kelly demonstrated the difficulty of
Emedicting actual fertility from responses of preferred

family size at an individual level. They found that though

11United Nations, The Determinants and Consequences
EELPopulation Trends (New York: United Nations Department
Cf Social Affairs, Population Division, 1953), pp. 75—76.

12See, for example, C. V. Kiser and P. K. Whelpton,
'Tertility Planning and Fertility Rates by Socio-Economic
Ehatus," Milbank Memorial Fund Quarterly, XXVII (April, 1949),
1W$244; R. Freedman, P. K. Whelpton, and.A. A. Campbell,
QEEgly Planning, Sterility and Pppulation Growth (New York:
D’chraw-Hill, 1959) (GAF Study); also Westoff, Potter, Sagi,
andkfishler, op. cit.; and‘Westoff, Potter, and Sagi,
02.cit. (Princeton Study).

 

 

the estimate from the initial response to number of children
Cksired was fairly accurate in predicting the average number
cm children of the whole group twenty years later, the indi—
xddual correlation between preferred and actual size of
family is quite low. In View of the fact that the near
ayfiyalence of desires and behavior for the total group was
armt result involving an averaging-out of various factors,
the authors arrive at the conclusion that "completed fertil—
ity is a deceptively simple net result of an extremely com—

;flex series of antecedents which includes the process of

family building itself."13

Furthermore, in connection with family limitation,
11m widespread use of contraceptive methods should be noted.
Ihlsupport of this Campbell has written concerning the
Princeton Study results that:

We have discovered from preliminary tabulations that the
proportion of couples who have used contraception has
increased during the past five years. In 1955, the pro—
portion was 70 percent. In 1960, among comparable
couples (metropolitan, white, wives 18 to 39 years old,
husband present), it is about 80 percent. The propor-
tion has increased for every age group. We think that
this rise means that more couples have all or most of
the children they want, and are using contraceptions to
prevent or delay the births of additional children. We
know that average number of children ever born has
increased for each age group between 1955 and 1960. The
rise in the proportion of contraceptors is probably a
result of that increase . . . [and that] this increase

13C. F. Westoff, E. G. Mishler and E. L. Kelly,

'Tmeferences in Size of Family and Eventual Fertility Twenty
Yfins.After," American Journal of Sociology, LXII (March.
1957), 494 .

.3 2. v. ..
.: .: 1. .: ..
... .. ._. ... 1
. .1 .1 .. :.
. I. ... ..
.,. l. .

..
.. .
2. I.
._ . .
. ...
. ..
2.
. . .

~
\

'-

 

...

 

in average number of children is voluntary; it is not

due to the failure to use contraception or to careless-

ness in its use.14
Considering the increasingly widespread use of contracep—
tives and the low correlation between preferred and actual
size of family, it seems plausible that perhaps family
limitation is not the only explanatory factor of fertility
levels. That is, family limitation per se does not deter—
nune fertility levels but there are more ultimate causes
vmich affect fertility perhaps indirectly through family
limitation. Several studies have argued that such "causes”
are attitudinal; however, this study will argue that fertil-
ity levels, with or without the use of contraceptives, are
to be explained by social factors.

References in support of this approach are numerous.

Only a few will be cited at this point to support the notion
CT the significance of fertility as a social phenomenon.
Ifitagawa and Hauser note the importance of the social in
Contrast with the biological and the consequent conjunction
of demography with sociology. To them the study of fertil-
ity transcends matters of immediate demographic concern,
Ednce fertility is a function of the social milieu in which
it occurs, even though it possesses important biological
(KMponents. "This basic premise, which in a large measure

14Arthur A. Campbell, "Design and Scope of the 1960

Study‘of Growth of American Families," in C. V. Kiser,
Egfiparch in Family Planning (Princeton: Princeton Univer-
31ty Press, 1960) , pp. 176 and 182.

... .Q'-
~

L...o--~- '4‘- .’
:-.._: .:“oiur.

m. ...- up

.
--.-.--§-‘ '.
' _. ..
_- ub-s..-..~o a.

—
—O'--~v~r .Ir
-\—-
. .v.--.~.. -.
.” ‘F ODr- -o
’- ‘
-
--.. -c..-. _
'; .,.-.,- A”
.,. ‘ 8‘.
"'v _ ‘1...
4. . . O
.I .‘
.- v
.6. b . ..'
“ Q»-

 

-.
~“- A
u.‘ ,1 ‘v
C. ‘_ '
v"‘<
"u.-“
v _. _- ‘ .
‘-.-‘ “ .o"“
s-..
l “-0-.
.n..-
. -
n. - -“~p.s ‘
“Cc... -
-.~t“-
.‘u
a. -
.
~s. -V‘V-fi,
~C
5
-""*‘~.,
.
“c o
v.
_ .
~-.
~-. ..
. -bt‘. k
a
.
m.
V "~
- v.
. . -
s. _- _
“‘u‘ A. ‘
~
a. ‘.‘
-.

 

‘I
O
-.
i
q
..
s
M.

   

accounts for the interest of the sociologist in demography
and the capture by him in academic United States of most of
the teaching and research in demography, is well supported
kw research findings-~those dealt with here (in this study)
and in other studies."15 Finally, Hauser, in another arti-
cle, shows concern over the fact that most efforts to under—
stand fertility behavior have been restricted to considera—
tion of demographic, personal, and social psychological
\mriables, and have virtually ignored cultural and social
organizational factors. In view of the fact that studies of
lumen reproductive behavior, up to this time, have accounted
fin very low proportions of variance in birth rates or other
measures of fertility, Hauser writes that:

the failure of the above-mentioned studies to eXplain a
greater proportion of variance in fertility is attribut-
able to the fact that they have ignored the culturalog-
ical in the sense in which Durkheim treated this subject.
Durkheim's insistence that the "social fact" is anterior
and exterior to the individual while exerting constraint
upon his behavior may contain the clue to an important
missing ingredient in fertility research. That is, the
pr0posed design of research (which Hauser outlines in
his article) is based on the assumption that fertility
behavior is in large measure dependent upon the social
milieu, and that changes in fertility behavior neces-
sarily involve social change. Or, put in another way,
knowledge of the person's attitudes, values, and motiva—
tion cannot be expected to account for differences in
fertility behavior out of their cultural context; and
consequently, changes in fertility behavior cannot be

. 15Evelyn M. Kitagawa and Philip M. Hauser, "Trends
UlDifferential Fertility and Mortality in a Metropolis—-
(Hficago," in E. Burgess and D. Bogue, Contributions to Urban
&Xfiology (Chicago: University of Chicago Press, 1964),
P-60.

 

produced through efforts to change attitudes, values,

or motivation, except in the context of changes in the

social order.16

From such statements it seems legitimate to assume

that fertility behavior is largely a social phenomenon. If
the biological, psychological, or attitudinal are to be rec—
cmnized at all within a sociological framework, they must be
viewed as "intermediate" variables with respect to their
influence on fertility. However, these variables are not
considered essential to this research. In summary, this
study is not concerned with biological factors as they
affect fertility behavior. Nor is it a study of how psycho—
Iknical characteristics, such as feelings of economic inse-
curity, or attitudes, such as preferred number of children,
impinge upon reproductive behavior. It appears that a
fruitful and necessary approach is within the realm of the

Ekmiological. This is, then, a sociological study of fer-

tility.

16Philip M. Hauser, "On Design for EXperiment and
Research in Fertility Control," in C. V. Kiser, Research in
Efipilprlanning (Princeton: Princeton University Press,
1960), pp. 464-465.

 

~'-“ 4-..-1

(I)
'II
‘I.

f
l-'

I
(n
H'
1

. \ .
s

 

 

10

Fertilipy as an Important
Demographic Variable

 

 

Demographic analysis, according to Duncan and Hauser,
is the study of the components of population change.17
According to demographers, these basic components of popula—
tion processes are fertility, mortality, and migration.
Kingsley Davis comments, "it is clear that any factor influ—
enming the number of people must operate through one or more
cm'the variables mentioned. In no other way can a pOpula—
tion be changed."18 Fertility, then, is a very basic ele—
ment to demography and a very important source of growth for
any‘population.

In View of recent occurrences in fertility rates,
there is an increasing realization that the problematic
factor in population growth today is the fertility rate.19
In discussing the relative importance of the "components of
Fmpulation change," Freedman indicates for the United States
that:

Migration, deaths, and births are vital factors deter-

mining the rate at which a country grows. It seems
unlikely that immigration will be large enough in the

17Philip M. Hauser and Otis Dudley Duncan, The Stugy

<DfPOpulation (Chicago: University of Chicago Press, 1959),
P.33.

 

l8Kingsley Davis, The Human Society (New York:
bhcmillan, 1949), p. 552.

19Ronald Freedman, "The Sociology of Human Fertility,"
Qggpnt Sociology; X-XI, No. 2 (1961-62), 35; and‘W. H. Gra-
PlLL C. V. Kiser and P. K. Whelpton, The Fertility of Amer-
EWHIWomen (New YOrk: Wiley, 1958), p. 288.

 

 

 

 

 

-o-sa- ‘ "
..-
--o-d-' -
I | . ‘
c
a. ‘(S '
o .
._.--.~--‘
a
on " ‘ ‘ .
0- \ .
0..--» .u
. n
0,. a..- _-
.-
.H. .o .
-.-
'- ._ --
.u.-.
—
_. . .‘
~-~v~ -na rv
.‘ - hf
-‘IV’CU b..- ~~
- .
"1‘holq - .
. - _ O»?
-.v...a --- _
u ‘ ' fi
". - ~.-.. ,
‘ .
-..-- ~"‘ --
~.-‘- \
1 u
_v.. v- _
""Ou-.—- . ‘-
o . V
, ' u.
0.. .‘ '
“ . ' F —
.-
c ---' -. ‘
..._ a
. ‘0 .
,-
'.-. _ ... u
- v. -
. _‘ .
- ‘
.z - '--- --
'-'.._ x '—
. ~. -~
. .
.11 i
.
. ‘ ' 'I II ‘
..‘--.o f“
. “—
u..-
-.‘* ‘ ‘
.. \‘.‘~ ~
.-
‘Q ‘
‘-'. Q
. ~ :_ -
"w _ u“._‘
-“ -1
-01-
-"
‘ - v-
.. §
- h ..
‘ .‘\-_ '-
‘»
Ԥ
I'-
. -.A
K ...'- _ -
b-.. .A
I'-
. -.
I.
"‘1 “-I.
o ‘ ‘
“.5.
s-
‘ u
.. .
u. --
' A
5 a
.‘s h. 5‘
‘~ ‘ ~
' Q"
"-o
P 1
‘-
\ .
\- ~ ~
“. \“~‘
'4‘ N
«J‘,
("A‘
.. -
.
“ ’I.
I 9.. ‘s

 

11

next few decades to affect substantially America's popu-
lation increase. Our death rate is already low and
unlikely to vary greatly in the near future, unless
there is a disastrous war or some other catastrophe.

The dynamic force in our population growth is the birth
rate.AU

Before the "baby boom” of the 1940's, most demOgraphers
eXpected fertility rates to continue their decline and even—
tually stabilize at some low level. Such forecasts of a
stationary population were refuted by the rebound in fertil-
ity following WOrld War II. On this basis, it can be stated
that of the three components of population change, fertility
is perhaps the most volatile and least predictable. Conse-
cnmntly, "demographers have begun to turn to more broadly
lmsed social research on fertility for better predictions of
this element in pOpulation growth."21 The necessity of
Social research on fertility is further supported by Hutchin-
Son, when he says, "my feeling of caution arises from a
strong conviction that fertility is a variable quantity,
capable of adjusting itself, and rather quickly, to changes
in the socio-economic environment, rather than being ruled
wholly by its own internal dynamics."22 Similarly Davis

20Ronald Freedman, P. K. Whelpton and A. A. Campbell,

Epmily Planning, Sterility and Population Growth (New YOrk:
karaw-Hill, 1959), p. l (italics mine).

21Ronald Freedman, "The Sociology of Human Fertility,"

Op. cit., p. 35.

22Edward P. Hutchinson, "Comment on Ryder's Paper,"

thational Bureau of Economic Research, Demographic and
.Qfimpmic Change in Developed Countries (Princeton: Prince-
t<mUniversity Press, 1960), p. 132.

 

 

 

 

 

 

,-
. .q.-. A-
..- ---.-v v-
.
I O
..-- 9‘ p- -
- -- .
.- c .- 1
o
——__
0--
«Pa.
' ,..-
.... __ -
_ -
- . h‘
"“’ O. v .v
. ,.
.. .- -_ _
v“-.- -: 0

 

- §
.l ~II~"'
'-~--- .
_
‘~
.‘ ’
.. ; ~ .“- - .
"'-»-
u
..Q G
u 'v‘
~~:._ R. ‘r .
“‘ J- _
‘ -
3..
- , .
O. -'. fig
‘Q.. 5
A.
3" .
- -

‘A
§.-“__--. s
v~~ \

--~-:
.-
‘j‘
. ~-. .
.. s
.- a ‘—
_ -‘w‘
‘ .‘h
\

 

 

 

‘
u. .
.
s.. "
‘Q
.
.‘
§ ‘~_~
‘\. I.
V
. c
O
u.
.-
1. ‘5'.- .
\- .
.--‘ L-
‘ ‘d ‘F
.
Q
-
;~ .
K; °-
‘
‘o: ‘-
..‘.. -
... Q
'-
t.-
.
.‘ -
‘s ‘
.~
v
C‘. ‘
‘~‘
~. A
&
-
g
o
c
‘9
g s
—
Q
«I
.9
u
‘.

  

12

asserts that "whenever the demographer pushes his inquiry to

the point of asking why the demographic processes behave as

they do, he enters the social field."23

 

What is called for in these brief references, there-
fore, is the nexus of sociology and demography. While fer—
tility is one of the primary cornerstones of demography, it
is necessary to seek explanation of fertility variation with-
in a sociological framework. This, then, is basically the
(ksign of this study. Whereas the dependent variable is
fertilipy behavior, the independent or eXplanatory variables
sue essentially social structural in nature. In that this
study attempts to establish connections between a demographic
phenomenon and its social setting, its scope is both demo—

graphic and sociological.24

Differential Fertilipy

 

In the traditional sense, differential fertility
IEfers simply to the unequal reproduction of subclasses of
the population based on such variables as nativity, color,
residence, income, occupation, education, etc. In a sense
13m problem confronted in this thesis is a continuation of
23Kingsley Davis, op. cit., p. 552.

24Wilbert Moore, "Sociology and Demography," in P.
Pbuser and O. D. Duncan, The Stugy of POpulation, op. cit.,

Ifih 832—51. Includes many good references to the social
Correlates of fertility .

 

u
,'--.o .
..

. ,.uu.~.
0—.-

. ---:_v-
a. _‘. 4
' .. .,_~.
I.

n

.
‘1‘". 0‘
"'Z.‘ .. ..

 

" ‘ ‘por

v - .—

‘V -,,. 5.__-

’v- -.__. .
.

"‘ -_.,,. .

 

   

 

.‘—.
- _ h .-
.»_.—o _‘.".
-
.._ .
L..""uo.-
- -
5.."‘....-‘
'4'.~-.. _
.y. .ugh‘_-‘
-.'.~..
.- . ‘
~
. . .-
'0
..‘.
- _§'~
‘-~..: -
c..“-
’ V
-.
«A-
.‘ -
‘ ‘- v‘.‘
V‘.‘
.
'0. v.
5 u. _ . .-
ha
ow‘.“ ‘
6, 'Z -
u ‘.'v-»
0..
. ‘.. h
n.-.
‘
. Q
I» ~~_~.
~_~ ~
I K
.R.‘_ .
r
'_§- a a
‘Q
I
\
‘_‘
u
§ . ’
“.‘
C
I
' ,
.
s
‘c . F.
.'u .‘d‘

 

 

 

 

13

'this traditional interest, although the nature of the prob—
lem is essentially different. Primarily, this study seeks
aanswers to questions concerning differences in rural and
11rban fertility.

Since the urban-rural fertility differential is to
kae the central focus of this thesis, it is appropriate at
tihis point to document past patterns and the current status
c>f this differential. The best known and most soundly docu—
nnented generalization for the United States with respect to

Ciifferential fertility is the long-term continual decline of

tihe magnitude of the urban-rural fertility differential.25

(Stabill, Kiser and Whelpton note that, while the urban-rural
(iifferential in fertility is among the oldest and best known
<>f demographic phenomena, it has narrowed considerably in
tihe United States. In commenting on the outlook for fertil—
Iity differentials they state:

It seems likely . . . that the long—range trend will be
toward continued narrowing of group differences in fer-
tility. The differences between rural and urban areas
with respect to style of life are being lessened by

25A few examples of this documentation are W. H.
(Srabill, C. V. Kiser, and P. K. Whelpton, The Fertility of
{Emerican WOmen, op. cit.; W. H. Grabill, "The Fertility of
‘the United States Population," in Donald J. Bogue, Egg
liopulation of the United States (Glencoe: The Free Press,
.1959), pp. 288-324; C. V. Kiser, "Differential Fertility in
'the United States," op. cit.; C. V. Kiser, "Changes in Fer-
‘tility by Socio-Economic Status during 1940-50," Milbank
bdemorial Fund Quarterly, XXXIII (October, 1955), 393-429;
C!.Eh Westoff, "Differential Fertility in the United States:
.1900 to 1952," American Sociological Review, XIX (October,
1954), 549-61; and Bernard Okun, Trends in Birth Rates in
;the United States Since 1870 (Baltimore: The Johns Hopkins
Press, 1958) .

 

 

 

'.A Q

.u-I-u-

u... .—
.' ~-
.

.
.... ,.
n
0.! h

II“

".an _.
~Q--....

o...
c- .
"..-A‘ ‘
..
."~~ '
§.—.
.-.d ‘
...

 

 

I

- a
. u
':

Q
‘ §
.-
, ‘—
‘. ~ ..

P .-

‘u

n
v.
.- “0
‘. ~‘c
s
u '
_ s
\
'.‘ -
~ .
u_~ ‘ ‘
n
‘\

-
5y . "
.' .
¥I

‘:' a
I l
.
.l. .
_.

 

 

 

 

14

reduction in the relative size of the farm population,
by improvements in highways and means or transportation,
and by television, radio, and movies. . . . Selective
factors alone probably will continue to account for
appreciable urban-rural differences in fertility, but,
in general, the outlook is for reduction in the magni—
tude of these differentials.26

Several plausible explanations have been proposed
:ftm the contraction of the traditional urban-rural fertility
(iifferential: (l) the spread of contraceptive practice
tzhrough all strata of the population, thus, virtually elim-
jgnating the differential use of contraception as a basis for
(Sifferential fertility; (2) the high degree of consensus of
E: large majority of Americans in an ideal family size rang—
:ing from two to four;27 and (3) the blurring of class dif—
:ferences in the United States as the working class takes

(on many middle class characteristics and the function of

26W. H. Grabill, C. V. Kiser, and P. Whelpton, The
liertilipy of American Women, op. cit., p. 378.

27Judith Blake, "Ideal Family Size Among White Amer-
:icans: A Quarter of a Century of Evidence," Demography, III,
ISO. 1 (1966), 154—73 (Blake says, "the two-to—four child
Itange has encompassed the ideals of approximately 80 to 90
19ercent of men and women since the middle of the 1930's").
ESee also David Goldberg, "Fertility and Fertility Differen—
‘tials: Some Observations on Recent Changes in the United
EStates" in M. C. Sheps and J. C. Ridley, Public Health and
IPopulation Changp_(Pittsburgh: University of Pittsburgh
ZPress, 1965), pp. 131-32; Ronald Freedman, David Goldberg,
and Doris Slesinger, "Current Fertility Expectations of
Idarried Couples in the United States," Population Index,
JXXIX (October, 1963), 366-91; and R. Freedman, D. Goldberg,
and.D. Slesinger, "Fertility EXpectations in the United
EStates: 1963," Population Index, XXX (April, 1964), 171-75.

 

 

 

 

.0...— ‘

 

 

‘-

\
.
u
-x
‘\
~
~
o
u. .
c
.‘ 1
§
' -
\'
\
~
.\
y
1
n
\

 

 

 

 

 

 

15

(rhildren and the family become more similar in the different

ssocial strata of the population.28

With respect to urban—rural differences in general
rnany sociologists and demographers have forecast the even—
trual demise of most urban—rural differences on the basis of
xvhich one can infer the eventual eclipse of the fertility
ciifferential. For example, Robin Williams asserts:

Were it possible to make a systematic comparison of all
major aspects of American social life in 1900 and in the
1960's, we would probably find a consistent decrease in
the sharpness of differentiation between and among major
social statuses, categories and collectivities. Rural-
urban differences clearly are less. Class differentials
are less obvious or sharp. Occupational status differ—
ences are blurred, eSpecially between manual and non—
manual jobs. . . . Regional distinctiveness, in spite
of temporary resurgence in situations of conflict,
gradually diminishes.

Ieran-rural disparities will disappear as our society
lbecomes more dominantly urban as Comhaire and Cahnman com-

Inent:

The industrial society in which we live is urban through
and through, especially in the United States, where the
farmer is a business man who keeps a sharp eye on domes—
tic and world markets, applies scientific methods in

28Kurt Mayer, "Fertility Changes and Population

I?orecasts in the United States,” Social Research, XXVI,
ISO. 3 (Autumn, 1959), 347-66.
29RObin M. Williams, Jr., "American Society in Tran-
Sition: Trends and Emerging Development in Social and Cul-
'tura1 Systems," in James H. Copp, Our Changing Rural Sociepy:
lierspectives and Trends (Ames: Iowa State University Press,
51964), pp. 23-24. See also Norman Ryder, "Variability and
(Zonvergence in the American Population," Phi Delta Kappan,
XLI (June, 1960), 379-383.

 

 

 

 

 

 

 

 

16

seeding and feeding, owns a car and a television set,
and has his wife and daughter dressed according to the
latest fashion. . . . Ecologically speaking, the Amer-
ican farmer does not live in the city, yet his ways are
citified. He is g; the city even though he is not lp'
the city.30

The question might be stirring in the reader's mind
eat this point whether it is even profitable to investigate
iiurther the urban—rural differential in fertility, whether
i:t is profitable to proceed with this very study, given the
irnevitable convergence of urban—rural fertility levels? The
Etnswer must be affirmative, though certainly new approaches
21nd techniques of analysis must be investigated and tried.

(Ioldberg justifies the continuation of research in this area

Vihen he states:

With the exception of certain types of historical data,
American fertility patterns are probably better docu-
mentated than in any other country. To document a
pattern, however, does not explain it. On the whole,
our understanding_of fertility_differentials is negli—
gible. Moreover, what we thought we knew in the past,
the relationships we took for granted, are being seri—
ously questioned by some recent research. In fact, the
most exciting research on differentials during the past
few years has either negated what was thought to be true
in the past or has found differentials to be increasing

precisely in those areas for which contraction had been
predicted.31

 

3OJean Comhaire and Werner J.Cahnman, How Cities

(grey-r (Madison, New Jersey: The Floram Park Press, 1959) ,

 

31David Goldberg, "Fertility and Fertility Differen—

‘tials: Some Observations on Recent Changes in the United
EStates," op. cit., p. 120.

..... » I

.u- -.-o

. .

..- con
- o-

..D- ..<
o

.w r. ‘
., .
.~-

.-.
~‘ C.-
"‘ 0,.
a. _ -
.
..__
i .
‘.- _

 

'G..

h._

 

 

 

 

‘ . ‘-
._ _‘
'-.

.
~
.;

.~ ‘-
-. ~ -
‘\

‘

~ -5 -
‘- s
s

‘.

. {.7
‘ .
C‘.

u

\
¢

. '.‘
.t ‘

‘u

'.

.‘_ ‘
‘\
. .

\ ~u

I!
."~
\

5“.

‘ §
. \

. §‘
.. .
‘_ c

.

..

. '.

_ §

\ "

s “
. 1
0

~ .

 

 

17

Fertility is a complex phenomenon. Convergence of
fertility levels does not ring the death knell of differen-
tial fertility analysis, but understandably indicates that
11ew relationships are perhaps develOping. As fertility
Ipatterns change, our approach to analysis must also change.
(31d techniques are not adequate or sensitive enough to per-
mrit the intensive scrutiny necessary to sufficiently under—
srtand and eXplain complex fertility behavior. With near
exxhaustive documentation of fertility differentials past and
crurrent, we have discovered much in terms of fertility pat-
tuerns, but more questions than answers have emerged. At-
t:empts to eXplain fertility behavior must continue; new
qnaestions must be formulated; different approaches and tech-
Iliques must be eXplored. Besides, urban-rural fertility
<3ifferences still have not completely disappeared,32 though
tlue'magnitude of the fertility differential is diminishing.
Answering such questions as what factors are associated with
lxrban fertility, what factors are associated with rural fer-
tlility, and are these sets of factors dissimilar, may help
‘10»better understand current urban and rural fertility
:Levels and to judge more accurately the future status of the

‘eran—rural fertility differential. To set the backdrop for

32See Leo F. Schnore, "The Rural-Urban Variable: An
Ieranite's PerSpective," Rural Sociology, XXXI, No. 2 (June,
1966), 131-43; and J. Allan Beegle, "Social Structure and
(Ihanging Fertility of the Farm Population," Rural Sociology,
XXXI, No. 4 (December, 1966), 415—27.

 

 

 

18

this attempt to investigate the why of residential fertility
variation, let us ponder past and current trends of the
urban-rural fertility differential within the United States.
Table 1 presents the trends in urban and rural fer—
tility ratios, measured as the number of children under 5
years old per 1,000 white women 20 to 44 years old, for the
period 1800 to 1960, for only the white population of the
United States and its geographic divisions. The table indi—
cates quite clearly that, although rural fertility ratios
have persistently maintained a higher level than urban
ratios, for the nation and all divisions the fertility
ratios of the rural population have kept pace with those of
the urban population in a pattern of gradual diminution. In
fact, when measured in absolute decline, the rural popula—
tion reveals the greater declines in fertility levels. On
a national basis both the urban and rural ratios of children
under 5 years old dropped between 1810 and 1840 by about 200
clhildren per 1,000 women (rural 195; urban 199) . Between
1840 and 1910, however, the decline amounted to 352 in the
rural population and only 232 for the urban. Again between
1910 and 1940 the urban ratio fell by 158 children per 1,000
Women, the rural ratio by 231. The divisions for these same
Periods reflect a similar pattern of greater absolute reduc—
tion of rural fertility ratios, although for the period
1910-1940 the more urban divisions (New England, Middle

Atlantic, East North Central) reverse this pattern.

ab. 0" . u
n 0 - ~§ - . s u a -
v I .5. - s o
Q .50.
Q Q . . Q -
-c I
n 0.! u.-
- uua

‘ul.u I I 1 , .. I .0: \l a... I.|o .- u 1 o I . - t .
I‘Quo- I I
on 0...! > D O I c 0
II: I
O: on
I I
I .n.
I. at.
. ". uh
I .v‘ .I.
.0...) .0
I
at
u v
I 0.. 1"
ll 0.00....
In: - n
O...
L .D‘v
on. .-
\ .-.~I ,
I.....,.
D a...
u.
(lu.ean‘

1.9

.mmlnm .dd .Amoma .umnEouoo .ommUALO mo muwmuw>asa .xmoHowoom wo ucoeuummoo

 

 

 

.COMumuuommac .o.£m pmzmaandmsdv :mmumum pouwcb msu ca moocouwwuan xuaaauumm cons: aw mucuomm can become: .wazmmm .w cmuasm ”mousom
.cowuwcwmmp swan: GAO so Ummmmc

... ... ... ... ... ... ... m.oml m.oau N.MHI ... ... ... m.mml m.oHI m.vml ¢.mml o.mml o.hHI o.val ovma
looma

... ... ... ... v.o¢I o.vml v.mvl m.mm| 0.5m: o.vml m.mml h.m¢I H.¢v| o.nvl ¢.0ml v.mm| o.oml ~.m~l a.mml o.Hm1 onH
Iovma

¢.H~| N.h~l m.mal 0.0m: A.Nm1 m.mml o.mml h.¢~l H.hm1 m.mm| m.mml ~.mml 0.0m: h.oml ~.~v| h.mml ¢.Hml n.HNI h.~mu m.mml oema
loama

h.m~H ~.~o h.mm o.mm m.mm m.v~ m.~m H.m H.mo m.va n.mHH h.Hm h.ooa m.ov h.ooa m.hm H.mm «.05 m.¢oa o.mm wWMM

chmcu mmmucwuuom

... ... ... ... ... ... ... moh.H How mom.H ... ... ... o¢m.H «mm omm.a hum oNH.H m¢m mam.d coma
... ... ... ... hmh 5mm.H mem.a Hon.H 0mm h¢m.~ ... oam.~ om~.H mon.H emm ¢¢M.H mew mno.~ com mmm.H oama
... ... ... ... mom -m.u mmo.a mmm.a Hmm oam.a ... mmo.H mmo.a 0H0.H New mm~.H won Nmm Hmm wh~.H omwa
... ... ... ... hum mo¢.H mom mmm.H hob mo~.a HmH.H moh.H cam ¢m¢.H NNh ooa.d VHO Hmm won omH.H OMmH
... ... ... ... mew mmv.H amm ¢Nv.H ohh mmH.H mow Hmv.H H¢m Hm~.H Han woo.H «on com Hon VMH.H o¢ma
com ovw 00¢ on com hum owe mum mow vow one 005 one who mmfi omo mo¢ mom mwv th Gama
wen moo Orv how mwc mwm Ace pew mme Hmm ode dab mme moo How omo oom Noe ADV «eh owma
won hon awe NAB oav mmh ¢H¢ awn Hov wen mom vao 00¢ moo wmm omm ha¢ va mmm mmo OMmH
mmm 90¢ woe new men Hmm mmm mew mom mom fimm mmm own mmm omN hmc HNM mev Han Hmm owed
mnv «no one fimb ~¢m mom emv own One who ¢Hm Now How mum Nme 0mm owe «do mhv Mho comma
mom new mom hob avm «on Now ems mew Now omm mon Hon «mo avv moo ¢m¢ mNo omv mmo omma
mmo amp Neh mmm omo wmn wow wow mmm Ame moo 0H0 new mmn ehm omb wmw mmh wmo hes Gama

 

Gonna Hausa Gonna Hausa Gonna Hausa :mnuD Hound Gonna Amusm cons: Hausa Gonna amusm Gonna amusm :mnHD Hausa can“: Hausa hum»

 

Dauaomm :«mu::0: Huuucoo Hmuunoo oaucmHu< Hmuucoo amuucmo Uwucwuut pamamcm moumum
susom amok nusom unmm nuaom nuuoz yum: nuuoz ummm mdccwz 3oz nouwab

 

 

 

coma 0» oaoa can ovma 0» coma «Amunu can swan:
.m:0anw>av cam ceauu: an .UHO numma vv 0» om soap: wads: coo.a hoe 0H0 numah m Moos: couoaaso no mucosa oomusoouom can uunisz

 

 

 

.H canes

 

 

 

20

In terms of percentage change, however, the urban
population reflects a greater rate of decline since 1800.
Table 1 indicates for the periods 1800-1840, 1840—1910, and
1910-1940 rates of decline nationally for the rural popula—
tion of 14.0 percent, 31.0 percent, and 29.5 percent respec-
tively and for the urban population 17.0 percent, 33.1 per-
cent and 32.7 percent respectively. For the nation and its
divisions, generally the period from 1800 to 1940 is one of
long term decline for both rural and urban fertility,
although Tables 2 and 3 reveal exceptions to this pattern
in various divisions for specific decades. More will be
said concerning these later, but here it is worth contrast-
ing the percentage change in urban and rural fertility
ratios in the geographic divisions for the two broad periods
of 1840-1910 and 1910—1940 (see Table l) . For the former
Eperiod, fertility of pggal populations declined at faster
rates'than urban in the more urbanized divisions (New
England, Middle Atlantic, East North Central, West North
Central). The more rural divisions indicate an opposite
trend. Between 1910 and 1940 in the more urbanized divisions,
the gr_bap_ fertility ratios dropped at a faster rate than the
rural (New England, Middle Atlantic, East North Central,
South Atlantic). The other divisions, becoming increasingly
urban, resemble the pattern of the more urbanized divisions
0f the previous period, 1840-1910. This indicates an appar-

ent lag on the part of the more rural divisions, but a rough

 

......ao-aw \‘ ‘
. .-
£.'L.-o.vo~ -'. "
I v
,_ _ _..Q .- 1“.
‘-:. - ‘ ~ '
.-' “u‘ -.o- s
.
.... ..- QAVOV‘ ‘

...- ...-l ~g..‘-~‘\.

-
.u -
-v. -.--y ~ A.
. . a _‘
.' -budb ‘.
o -

 

-
- -..-,
-...

.- ‘1
'* ~cv.

 

 

a
f"

l"

 

 

(I)
i

 

21

Eissociation of urbanization and fertility decline obtains.
(These data suggest a pattern of cycles in terms of the eXpan—
ssion and contraction of urban—rural fertility levels. In
eaarly stages of urbanization rural fertility ratios decline
faster than urban ratios, contracting the differential. As
Ilrbanization progresses, urban fertility ratios reduce
:faster than rural ratios, eXpanding the differential. The
foeriod 1940-60, a period of increasing fertility ratios for
130th urban and rural populations, indicates a contraction,

tout by urban fertility levels increasing faster than rural

 

Ilevels. Table 1 shows that in the more urbanized divisions
(New England, Middle Atlantic, East North Central, West
liorth Central, South Atlantic, Pacific) and for the nation
635 a whole, urban fertility levels increased at an extremely
:fast pace, closing the gap between urban-rural fertility.
i?able 6, which measures urban fertility ratios as a propor—
tlion of rural fertility ratios, reflects this pattern of
GEXpanding-contracting cycles for all divisions. As a result
<>f the "baby boom" era of 1940-60, urban-rural fertility
(iifferentials for all divisions and the nation are closer
‘to unity than ever before in their history. This pattern
(of varying rates of decline and increase for urban and rural
fertility levels in the same division suggests that perhaps
<iifferent sets of factors, in addition to urbanization, are

<Dperative in creating these fertility variations.

 

 

 

 

 

o
.
..
H...
o
-.
cu.
.-
...
.-.
..u
.-.
..
'-~|
..I
a.
‘o
‘-
-_
u.‘ll§
--__‘
s
.4
‘-
.-
_ - ‘1
‘ I
I.
'n .-
u.‘-
5,
.‘
. ,_
.
.. ‘
'\
.‘ _-‘
~
V“
94
s‘.
.. .‘
‘Q
_.'s
\
~
§
‘-
I i
I“ ’
~ I
\.
~
- ~‘
t .‘
‘.
a
‘s
. I
‘s
R .
‘ .
-\
.
§
. ._
. n
.h
v -
0-.
n
.
K o
. A
.
~
~~
I .-
‘ u‘
. .
- .
~
\
‘.
. “_
..
s
, . -
.
.‘.
. \
a
' Q
.
A.
. -_
~ ~
- .
,.5
§
‘\

22

Sultan S. Hashmi to some extent has studied these
same materials.33 Although his focus was urban fertility,
many of his findings and conclusions can be comparatively
applied to rural fertility. For example, the introductory
comment to his article says:

This study undertakes to measure differences in
fertility within the urban population and to learn what
the major factors underlying the observed differences
are. This has been a comparatively neglected aspect of
demography. Usually urban fertility is contrasted with
rural fertility, without appreciating that there may be
great internal variations in raEes of childbearing with-
in urban and rural aggregates.3

 

Hashmi supports this proposal by proceeding to demonstrate
that the correlation between fertility decline and progres-
Sive urbanization is only a very rough one. Tables 2 and 3
,present rates of change in urban and rural white fertility
ratios since 1800 by succeeding decades for the geographic
<3ivisions of the United States. Table 4 provides the per-

cent urban of these divisional populations. A careful

33Tables 1, 2, 3, and 4 were reproduced from Dr.
Hashmi's unpublished Ph.D. dissertation: "Trends and Fac-
tors in Urban Fertility Differences in the United States"
(Chicago: Department of Sociology. University of Chicago,
1962). This dissertation has also been published in abridged
form under the title "Factors in Urban Fertility Differences
in the United States," in Ernest W. Burgess and Donald J.
Bogue, Contributions to Urban Sociology (Chicago: University
cfi'Chicago Press, 1964), pp. 42—58. I wish to express my
appreciation here to Dr. Hashmi for the stimulating ideas he
has presented from the reading of his dissertation and arti-
cle. To a certain extent I consider my thesis a continuation
of the investigation which Dr. Hashmi began.

34Sultan S. Hashmi, "Factors in Urban Fertility Dif-
ferences in the United States," ibid., p. 42 (second italics
Inine).

.0.
-. noa.-‘

a I OUooll I?!

. o--‘ v
.... .L

III-..-

...lllllll i

g. - .
.. . R Q\
-. ..-

.. . ....

...l.- .<

.. ...-....

A...)

I... 0"

...-...

1.-
I ululdlc

.‘

u I
I. 0---!!!

-. .o .
.p-.o
‘.n—v.

...-o.“.-‘
10.2
a... Q

1"..ul as v

23

4

 

 

 

 

 

 

.mm .o .0” magma .AN00H .ommuflco 00 xufimuo>flco .hwoaofioom Mo unmauummmo .GOwumuummmfic

.n.nm wonmwfinamczv gmwumum vmuficp «Lu cfl mmucmummmwo >uwafluumm cmnua cw wuouUMh 0cm monouas .wsnmmm .0 :muasm "ouuSOm
.cowuwnwwmc swans 0H0 so Momma.
... ... ... 0.0 I m.~ I ... m.5 I 0.0 «.0 I 0.0 0H0HI000H
... ... ~.N I 0.m I 5.~ I 0.0 I m.0 I H.0 I 0.HHI 0.0 I 0N0HI0H0H
... ... 0.0HI 0.0 I 5.5 I H.H «.0 I 0.0HI 0.0HI 0.0 I omwalomwd
... ... N.~ 0.0 I 0.N I 0.0HI 0.maI 0.0 I 0.0 I 0.0 I 000HI000H
... ... 0.0mI m.0mI 0.0NI 5.00I 0.50I 0.0mI ~.0~I 0.HmI 0H0HI000A
0.0 I 0.0 I 0.0HI «.0 I 0.0 I 0.0 I 0.0 I 0.0 0.0 0.0 I 0~0HI0H0H
0.0HI 0.HHI ~.~HI 5.5 I 0.NHI 0.maI 0.0 I «.maI H.0HI 0.HHI om0HI0~0H
H.0 I 5.0 I n.0HI 0.5aI 0.0HI 0.~HI 0.HHI 0.N~I H.0HI 0.0HI 000HI000H
0.0m n.5H 0.0a H.~H 0.0a 0.00 0.5m 0.00 H.0m «.mu 000HI000H
~.0H 0.0a 5.0 0.H I 0.0 N.0H 0.0H 0.0m 0.00 0.Ha «000HI000H
0.NH 0.NH 0.0 5.m I 0.H I 5.0a 0.0a «.ma 0.0m H.0 000HI000~
Dammumm cwmunscx Hmuucwo Hmuucao uwuzmau< Hmuucmo Hmuucoo oquMHut unadmcm mmuwum coduum
susom amok nusom uwmm nusom :uuoz ummz nuuoz ummm manna: 302 cmuwcn

coma 0» oHOH cam oewa o» coma "Hanan .mcoflmfi>«w
0cm cowum: an .OHO mumma 00 Cu 0N :mEos 000:3 000.a “on 0H0 mumwx 0 Moons :muvawzu mo uonfiac mo moansu ommucoouwm .m manna
... ... ... ... 5.0 ... ... 0.0 ~.~ 0.0 camanomH
... ... H.0H ~.0HI 0.0 I ... 5.0HI 0.0 I 0.0 I 5.5 I 0~0HI0H0H
... ... m.H 0.0NI 0.~HI ... H.0HI 0.0HI 0.0HI 0.0HI 000HI0~0H
... ... 0.0 I 0.0 I 0.0 m.00I 0.5 I 0.H I 0.0 I 0.H I 000HI000H
... ... 0.00I 0.00I 0.50I 0.00I H.00I 0.00I 0.0NI H.00I 0H0HI000H
0.0 I 0.0 5.HdI 0.0 I 0.0 I n.~ I «.0 ~.H 0.0 0.0 0~0HI0H0H
0.AHI 0.0I 0.5 I H.0 I 0.NHI m.NHI 0.5HI 0.0mI 0.0HI 0.5HI 000HI0~0H
0.5I 0.0I 0.0HI 0.0HI 0.00I ~.HHI 0.0HI 0.0nl 0.0NI 0.0HI 000HI000H
0.00 0.00 0.00 0.00 0.50 0.00 0.00 0.H0 0.H0 0.00 000HI000H
0.Nm H.5N 0.0m m.m~ 0.00 0.00 0.50 0.Nm 0.00 m.Nm a000HI000H
m.0~ 5.0m 0.m~ m.HN 0.0m 0.00 0.00 «.00 0.Hm 0.0a 000HI000H

Ouuwowm canvases Hmuucoo Hauuaou caucmau< Hmuucoo Hmuucoo oauamuut ucsdmnm uouuum noduom
nusom you: suaom uumm £9900 suuoz no»: :quZ 9000 Genoa: 302 vuuwap

000a 0» 0H0~ can 000A 00 000A “sunk: .mco«o«>«c

0:0 c0090: an .vao uuuo0 00 0» 0m :0393 wads} 000.H “am one munch 0 Moon: couvauso no moans: no omcanu amoucoouom .m «Haws

24

.00 manme .AN00H .oomoflzu mo xuwmum>flca .xmoaofioom 00 ucmfiuummmo .COMumuuwmmwc
.0.:m uwnmwanaacdv :moumum cmuflcb 0:0 :0 mmucmummwflo xuwaauumm gangs :0 muouumm 0cm mccwu9= .wfinwmm .m acadnw "mousom

.cowuwcwwmc away: 000 :0 00mmm¢

 

... ... ...

 

0.0 ... ... N.0A N.0 H.0 000A

... ... N.NN 0.0 0.0 ... 0.0 0.HH 0.00 0.5 0HOH
... ... N.0H 0.0 0.0 ... N.a 0.HH 0.00 N.5 0NOH
... ... 5.0M 0.H 0.0 0.0 0.N 0.0H 0.0a 0.0 000d
... ... 0.0m H.N 5.5 0.0 0.0 H.0H 0.0M 0.0a 000a
0.00 0.00 0.NN 5.0a 0.00 0.00 5.00 0.05 0.05 5.00 000A
0.00 0.00 0.0m 0.0N 0.H0 5.50 0.00 0.05 0.05 N.H0 0NOH
0.00 0.00 0.00 H.0N 0.00 0.H0 0.00 5.55 0.55 N.00 000A
0.00 5.00 0.00 0.00 0.00 0.00 0.00 0.05 H.05 0.00 000A
0.00 n.00 H.00 0.00 0.00 0.00 0.00 0.05 0.05 0.00 #0000
0.05 0.00 0.00 H.00 0.00 0.~0 5.00 0.00 «.05 0.00 0000
H.H0 H.50 5.50 0.00 0.50 0.00 0.05 0.00 0.05 0.00 000A
UMMHUmm camucsoz Hmuucmo Hmuucoo caucmau< amuucoo Hmuucmo Uuucmaud usuamcu nuumum Mama

£9500 anus £u500 uamm zusom nuuoz 9003 auuoz ummu mavcfiz 302 vouwsp

 

0000 cu 0H0H 0:0 000A 0» 000A .nco«n«>«0 was amumum couanb aaonwauuucoo ham swans Gawunnamom Havou we omoucuouum .0 manna

, a .
'Ou:'t:l\n «0 5n; ‘5

’ ‘0‘
‘*..‘.“§\ 9‘ s
.
. _ - ~
‘0» ..-.
.';: " '- 5 _J
““O b§~ ‘..~ v‘
. D ‘
"‘ ‘I-a .‘Vb ‘ ‘

 

 

n-0y--'~_ . r > ‘
_ - ‘
s. .0._,___ V. ..-
. O
. ‘
....\ . ‘ ‘
- ‘ -. 0‘.
.....-.-. -“ -
‘v- . ,
' 0-,. - . ‘
- . ~
.... "-As. .
.... -
u “- --.v: 'v
h~-‘-_. __ v.-- --
-..- . .
‘ ~"" r-.. .
- t -..‘v‘ .5
. O

 

 

 

 

 

'O- ‘
: --v“ v ,
...
". -
..n - ‘0 -. - .
._.
C -. y- ‘.
,“ ‘ ‘
.
..-._- .-_ .2». ‘ ‘
--_._‘-
~‘I-
. _ '
-._.- _v.. ..
h‘ '
5““
0.-
-.:, ~~ _
w.‘ o "v.
- _‘ .
._‘
.
..‘ \ "—-.-
.‘ ‘ :v- .
.. “.- ..
‘°..
~- I
“V". .5-
" - ‘c .2...
5." . r
u ‘ .
"n ~‘A ._ _
"l n ‘
."s. .
v_.
.
~§~ “A“.
“ U . ‘.
~ '~--; ' - -
-‘v‘
...
. ’
.. ,
.. .
me -~»-.
1; ...- _
s~‘ ,
II
.‘I'
'\ ;. .
..- s. ‘ . ‘
'0
.\C‘ .'
‘, .
0
’§
‘- 'v- .
a -.. .
.‘,‘ \'
c
s. ..
N .
‘- .
5
“' .1 '-
‘b
.
.-
u. ..
~ Q
‘-. -u‘;
“ V
" ’av '
§ -
Ԥ -<
h.‘
V.
Iv
0“...
‘ 0
. b
‘0
.
‘1.
s
...-0‘
‘5
'-
s
"
,..-
0‘. -.
- a‘ '.
‘\ 0‘
.V.
‘I H‘
\
‘Nv‘
" v
I‘.
. .
‘.~Dp
.\ --
‘ ‘
V‘ "
§
.l
‘I
...;
.‘ ‘
'u\ \
'1 *.
.
a, s "
‘s

25

comparison of these tables will reveal certain irregular-
ities in the usually hypothesized association of urbaniza—
tion and fertility decline. The "baby boom" era of 1940—60
is perhaps the most obvious descrepancy. Urban and rural
fertility ratios both have increased significantly since
1940, while the nation and all divisions have continued to
‘become more urban. Likewise, between 1910 and 1920 there
“ms very rapid urbanization of the population, but during
this period there were slight increases in the urban fertil—
ity ratio of the nation and four divisions (New England,
Bfiddle Atlantic, East North Central, Mountain), and in the
rural fertility ratios of two divisions (New England, Middle
.Atlantic). During some of the very early decades of the
1800's urban and rural fertility fell quite rapidly, even
‘though the extent of urbanization was quite modest. Between
1810 and 1820 there was no increase in the proportion of
urban population, rather there was a slight decrease, but
both the urban and the rural fertility ratios decreased
substantially. During the decade 1930-1940 little increase
occurred in the proportion of the population urban, but
declines in fertility ratios were quite significant for both
urban and rural populations.

From the scrutiny of these figures the fact becomes
clear that over the past century and a half in the nine geo—
graphic divisions of the nation there is not merely 222
pattern of the relationship of urbanization and fertility

trends, but there is evidence of four. According to Hashmi

26

1:?rese are (l) fertility decrease with increasing urbaniza-
t:itnn (2) fertility decrease with decreasing urbanization,
(13) fertility increase with increasing urbanization and

(4) fertility increase with decreasing urbanization.35 The
:Eirst of these patterns is the expected inverse relationship
()f fertility and urbanization. It is found frequently in
130th the urban and rural populations of the divisions. The
ssecond pattern is associated with the economic depression of
t:he 1930's. With decreases in urbanization for 1930—1940,
ifour divisions (New England, Middle Atlantic, East North
(Zentral, Pacific) indicate a parallel decrease in both urban
21nd rural fertility ratios. This pattern occurred also for
130th ratios in the West South Central division between 1840—
31910 and the Middle Atlantic between 1810-1820 and for the
Irural fertility ratio in the West South Central between 1810—
1820. The third pattern is best exemplified by the "baby
boom" period since 1940. The upsurge of both urban and

rural fertility ratios in this period is a direct contradic-
tion of the expected pattern of decreasing fertility with
Lubanization. A similar pattern, though less spectacular

and widespread, occurred among some of the divisions during
the 1910—1920 decade, again for both urban and rural fertil—
ity. Fertility levels indicate accretion during moderate

urbanization between 1810-1820 in the urban population for

 

3SIbid., p. 44.

27

'tlne nation and New England, Middle Atlantic, and South
Zthlantic divisions and the rural population for the nation
zxnd Middle.Atlantic division. In addition there were a few
:isolated cases which occurred in the 19th century for urban
Eind rural fertility levels. The fourth pattern, decreasing
:Eertility and increasing urbanization, is rare. Omitting
czases which occurred between 1940 and 1950 because of the
jLnadequacy of the old definition of urban leaves the single
zincident of urban fertility in the West South Central divi-
ssion for 1810—1820. On the basis of these findings Hashmi
czoncludes that "urban fertility has not responded unilater-
Eilly to progressive urbanization at any time in the past."36
st far as that goes, neither has rural fertility reflected
a.unilateral pattern of change with urbanization. These
fluctuations in urban and rural fertility suggest that fac-
tors other than urbanization are operative within both pop-
ulations. These other factors must be identified and in-
vestigated in order to better understand fertility varia—
tions and trends for both urban and rural areas.

Another line of analysis of the historical changes
in the urban-rural fertility differential is to consider
independently the magnitude of the gap between the fertility
ratios of the residential populations. Are urban and rural

fertility levels approaching unity and, if so, what has been

 

36Ibid., p. 44.

 

...‘s-n

4--...v

-.~‘ q
n -
o-I' . .
a n

‘n .
-

 

.. ,
... _‘_
" nu-
-... .
~i
-~¢'- .
I'. ‘
I .“
u \
...-
I.
t -“-
-.

 

I
II

 

 

28

‘tlue pattern of change in closing the gap and what is the
crurrent magnitude of this gap? One measure of this gap is
cibtained by considering the absolute differences between
Ilrban and rural fertility levels at given points in time.
TPable 5 provides this information by decade for the nation
21nd the geographical divisions. The table shows that for
tzhe nation as a whole there were greater urban—rural differ-
eances in fertility during the period 1800-1840 than in the
I;eriods 1910-1940 and 1940-1960. Although there was a
\Midening trend from 1810 to 1830, as urban fertility ratios
rdeclined by more points than rural ratios, absolute differ-
eances in urban—rural fertility have narrowed considerably
<3ver the years. This trend continued through 1960 in spite
<3f the increases in both urban and rural fertility ratios
during the period 1940-1960. The largest gap between urban—
rural fertility appears in 1830, an absolute difference of
481, and the smallest in 1960 with a gap of only 111.

The same pattern as found in the national figures
applies also to the divisions, with some deviations, of
course. Nevertheless all divisions but New England have
reached the point of smallest absolute differences between
urban and rural fertility ratios by 1960. Through 1940 New
England and Middle Atlantic, the more progressively urbaniz-
ing divisions, maintained the lowest absolute differences
between urban and rural fertility. By 1960 they represent

the largest differences among the divisions, the other

299

.H manme ca co>flm moflumu >uflawuumu co comma :oHumuamEou "woudom

.cowuwcwmwc :mnud 0H0 :0 commma

 

 

... ... ... ... mo. ... ... «0. m5. #0. coma
... ... 5v. m5. mo. ... Vb. mm. mh. mo. Gama
... ... hm. no. no. ... mm. mm. om. mo. ONmH
... ... 00. mm. mo. 00. Ho. 09. Nb. ow. omwd
... ... hm. co. mm. mv. mo. Ah. vb. N0. ova
mm. mm. mm. Hm. vm. cm. ch. 05. mm. cm. chH
hm. mm. ¢m. Nm. cm. cm. m5. VF. mm. mm. ONmH
00. cm. hm. mm. Vm. mm. 00. mo. hh. cm. OmmH
Ac. mm. mm. am. am. cc. Ho. mo. Nb. on. cvma
mn. mm. hh. mm. mm. mh. Nb. N5. 05. Ah. «cmmH
mm. co. ch. mm. mm. vs. ms. N5. 55. Nb. cmmH
*0. mm. Nc. mm. mm. om. mm. cm. Vm. mm. omma
wawumm cwmucsoz Hmuucwu Hmuucoo Uwucmaum Hmuucmo Hmuucmo Uwucmaum panacea moumum Ham»
nuaom ummz Lusom ummm nusom :uuoz umoz nuuoz ummm macpqz 3oz couwcb

 

coma 0» chH can cvma 0» coma uncofimu>fic can sedan:
>n .GHO whack we Ou cN :oEO3 «UMSB ccc.H umm pao muwo» m “macs coupawno mo Hones: Hmuau mo COMHHOQOHQ m an cmnub .c manna

 

 

... ... ... ... fiom ... ... hmv mmN fihv ocmH
... ... 0mm mmm Adv ... cmv ON¢ VMN mN¢ OHmH
... ... cmc QVm va ... 5mm mmm mod mfiv ONQH
... ... wmm moo N¢¢ NNm ehm wbm hMN Hmv Ommd
... ... mvc mow ma¢ cub cm¢ mmN ch mm¢ cfimH
cmN vvm m5? mm¢ oc¢ wmm NON mmH 0c mam OHmH
mmN hmm mum mov mmm mmN mmH th NOH MFN ONmH
HON VQN mam 5cm m¢m m¢N ch ¢cN «NH chN Ommd
mmH mMN mVN mam HON vHN bON Aha NNH OGN cvmd
vhH one HOH cNN hNN me mmH Goa QNH «ca acmma
NwH NBA and NMN mHN mma HmH med med mad cmmd
mm FHA mm mm Na baa ccH wVH mHH HHH coma
vaudumm canvases Hauucoo Hmuucoo ofiusmaum Hmuucoc Hmuucoc caucmaum pamamcm aouaum ummw
susom um03 :usom uunm gusom :uuoz umo3 suuoz ummm macnwz 302 Quads:

 

coma 0» cam” can oemfl ou coma "mcoaua>«c can no«ua: an
.cdo anon» v¢ cu cw COED: 09%;: ccc.a Hon vac chaos m umpsn couoawzu mo amass: Hmuauucmnun :4 noucououuao ovaaownd .m canoe

>.. ..-.-
. \
flu.-.y..~

'n-‘IO 9:

,- -

.‘.>.I-. l
a

o
'4. 1.0-..

,..
. -~-v~—q-

- .. ..
......-.___,

 

 

 

0
"-v
-a
I‘..
In...
'u-.
‘.5 —
0 .
v..-
._'~t ~-
. .-
-_,_‘ §
‘
n
‘D
.n~
h "
».-

 

q...
u
. -,‘
... ~
.
‘u
‘-
‘.
‘u
o.( n. §
. '- : '
~
5
. .
. ~..
tl A
_ b ‘
-.fi -__

 

 

 

‘Q
.
~.' I
.‘~.

‘0 - -
-"_\ P.
'5

'L
i
.
‘S'N.
-. h’.‘
‘ b
h.
.
’0
‘7
.: .
. s
..
‘ r
..
‘O
-

 

 

 

..
.
‘I "\
l Arv‘
‘
v
.
_ ,
n -.-
I \ '
i .
v ..
.
'I
I
‘| a
K.
-.
‘ S
Q" ‘
..
.
-
._
\
‘ II
._ _
.\ a“
s-\ .
v _
~
3 .
.
.
‘\ C
. a
.. .
A
J“
. _.
.
.-
\ .

 

3O

éLivisions having dwindled their differences to lower levels.
Generally speaking, however, Table 5 tends to show clearly
izhat during the 19th century there were larger urban—rural
<iifferences in the fertility ratios of the divisions and

'that these differences have gone through considerable

<fleclines. Even during the last two decades, when fertility

:ratios were increasing both in the urban as well as the
:rural populations, the absolute urban—rural fertility dif—
ference continued to decrease.

A slightly different picture, however, is obtained
KNhen one compares urban—rural fertility differences in terms
Ci relative proportions. Table 6 reflects these differences
Tby divisions since 1800 by the measurement of urban fertil—
ity as a proportion of rural fertility. For the nation as
a whole early in the 19th century rural and urban fertility
levels were much closer to unity than in the first four
decades of the 20th century. In 1810 the urban ratio was
68 percent as large as the rural ratio and in 1940 the urban
ratio (though much lower in actual level than in the 19th
century) was only 56 percent as large as the rural ratio
(which was also much lower than the 19th century levels).
By 1960 the urban ratio has come closer to unity with the
rural ratio than any previous decade. The same pattern
applies to the various divisions. Formerly the more urban
divisions (New England, Middle Atlantic, East North Central)

maintained urban-rural fertility levels closer to unity than

 

s--_

-.

"

 

 

31

t:1fi1e:more rural divisions, though this has been reversed by
1.9 60 (lowest proportions are now in New England, Middle
At. lantic, Pacific).

In comparison with absolute differences of urban—
Ifllral fertility, which tends to reveal a straight-line pat-
1:earn of decline from extremely large to extremely small
Clifferences, the pattern reflected by urban fertility as a
pn:oportion of rural fertility tends to be a U—shaped curve,
VVith extremely large values appearing at both ends of the
tzime period. The comparison of the results of these two
ruethods of measuring the urban—rural fertility differential
1135 something to say concerning the necessity of investigat-
iJng factors determinative of urban and rural fertility inde—
pmandent of each other, i.e., the necessity of separate
iJnvestigation of the internal variation of both urban and
IHJral fertility levels.

Hashmi in his analysis considered only one measure
0f the magnitude of the urban—rural fertility differential:
abSolute differences. He concluded that the fact that this
djifference continued to decrease during the 1940—1960 fer—
tility upsurge proves more or less I'that urbanization alone
d0es not determine either the level or the direction of

37

Change in fertility rates." In order to support this

IPrOposition, I have attempted to extend the detail of his

 

 

37Ibid., p. 45.

32

airilalysis by the use of rank-order correlation. Table 7

E>J:‘esents two sets of rank—order correlations for the nation

211:1d its divisions. The first set of correlations indicates

t1118 level of association between changes in percent urban
EiIld the changes in absolute differences in urban—rural fer—

‘tdility ratios from 1800 to 1960. The second set of correla-

tlions indicates the level of association between percent

xxrban and the magnitude of the urban—rural differential

Ineasured by relative prOportion. Hashmi concluded that

11rbanization and the decline of absolute differences of urban

21nd rural fertility ratios were highly related. Table 7

loears this conclusion out. The nation indicates a very high
Irank—order correlation of -.93 while all the divisions also

Skumlhigh negative correlations of varying degree. Hence,

Ifiashmi is correct when he states that urbanization is only

iassociated with the long-term decline of absolute differences

7between urban and rural fertility but not the level or direc-

tion.

 

The measure of urban fertility as a prOportion of

rural fertility, it can be said, is more sensitive to the

lfiivel and direction of fertility change. For example, this

measure does respond to the change in direction of fertility

1eVels during the 1940—1960 "baby boom." If, however, urban

andrural fertility ratios had not increased for 1940—1960,
but had declined, it would have been possible for the size
0f the absolute difference between the two ratios to remain

'the same, quite independent of the direction of change in the

33

\ frzable 7. Rank order correlations of percentage of total
population urban with absolute differences in
urban—rural fertility ratios and urban to rural
fertility ratios; for conterminous United States
and divisions: 1800-1840 and 1910-1960

 

‘ Area

rho of rho of
% Urban and % Urban and
Absolute Urban
Differences in As a Proportion
Urban—Rural White of Rural White
Fertility Ratios Fertility Ratios

 

IJnited States -.93 -.02
E New England —.68 +.27
Ddiddle Atlantic —.87 +.30
IEast North Central -.83 +.30
West North Central -.98 +.64
South Atlantic —.94 -.05
East South Central -.71 —.03
West South Central —.92 +.55
Mountain —l.00 +.99
Pacific —l.00 +1.00

 

 

<3Omputations of rank-order correlations based on Tables 4, 5

and 6. Formula used:

rho = l —

 

7‘”

6ZD2

 

N(N2 — 1)

 

 

.\
v

 

34

‘ITEELtios themselves. In comparison the measure of relative
{patroportion of urban to rural fertility ratios would have
situown a lower value if both ratios had continued to decline
I:<agardless of the fact that absolute difference could have
Iceemained the same. Now looking at Table 7, it is to be
rusted that the rank order correlations between the relative
Exroportion measure of the rural—urban fertility differential
Etnd.percent urban are comparatively lower than correlations
(of absolute differences. For the nation as a whole the rank-

<3rder correlation is close to nil and even in the wrong

 

 

(iirection when considering relative proportion. This second
sset of correlations suggests that urbanization determines

/ nuostly the absolute differences of urban-rural fertility,
knat is a poor predictor of the relative proportion that

‘ Ilrban fertility is of rural fertility and, for that reason,

)

‘ (a poor predictor of the level and direction of change in

(fertility rates. I have not contracted Hashmi's original

 

PrOposition but merely added more weight to his argument.

/ On the basis of these types of information, then, Hashmi

C. One ludes :

This simple historical review of the fluctuations,
divergences, and wide inter area . . . differences in
urban fertility rather effectively contradicts the notion
that urbanization per se is an important determinant of
fertility. Instead, we should look upon urban fertility
as a phenomenon which must be studied independently and
eXplained by other variables. In other words, we must
look inside it and try to isolate factors that have a
more or less persistent relationship to fertility.

38Ibid., p. 45.

 

 

 

...-

 

 

(l'
D
n

 

35

2%. lthough Hashmi singles out only urban fertility for such
:Estudy, this conclusion must equally apply to the study of
:Irural fertility. It is possible that a different set of
ifactors influence rural fertilipy than does urban fertilipy,
gar that the same factors have a different impact on rural

ggertilipy than on urban fertilipy, These are possibilities

 

gin spite of the probable convergence of urban and rural
fertility levels. Fortunately there exists already some
empirical support that, though levels of fertility are con-
'verging, the pattern of differential fertility within urban
and rural populations is different. Goldberg, using data
from the Indianapolis study of fertility, compared urbanites

and farm migrants. He states:

If the average number of children among farm migrants

is about equal to that of urbanites, it does not neces—
sarily follow that the differential fertilityppatterns

of the migrants will be identical with or even similar

to those of the urbanites. In Indianapolis we find that
family size is the same for urbanites and couples with
some farm background, but that within each of these
groups there is a contrasting relationship between socio—
economic variables and fertility.

 

5Phough our study will deal with contrasting residence groups
of the national population, while urban and rural fertility
levels are nearly equal, it may be that within residence

groups there exist "contrasting relationships between socio-

economic variables and fertility," or for that matter

39David Goldbert, "Another Look at the Indianapolis
Fertility Data," Milbank Memorial Fundpguarterly, XXXVIII
(January, 1960), 27 (italics mine).

0".'.

O‘“" i

'3'. . P
‘ -

..-.-

.....V

u--"-.

 

 

~
---
.-~. ’
..~- "'
.
.
.0
\
.~.'
V...
.-
u
0--
...

‘9
U.

 

36

c:=ontrasting relationships for other than socio-economic

‘urariables as well. The discussion of these conclusions
1::oncerning urban and rural fertility levels suggest clearly
tzhe necessity of a comparative study of factors associated
Vvith the internal variation of urban and rural fertility
Ilevels in the United States. Such a study would be one
step beyond the contribution of Hashmi. The intent of my
study, then, is to look at the internal differential fertil—
ity pattern of each residence group and to compare the pat—
terns of each group for differences and similarities.

To this point we have concentrated on trends in
the urban—rural fertility differential and some patterns of
change have been established. Looking "within" the fertil-
ity differential pattern of each residence population, how-
ever, will not involve the explanation of trends but the
eXplanation of variation of fertility within residential
populations at one point in time. This one point in time
for our present study will be 1960 for three reasons: (1)
the difficulty of collecting comparable data for a histor—
ical or trend analysis of differential fertility, (2) the
accessibility and wealth of data available from the 1960
Census and (3) the urgency of constantly increasing our
understanding of current fertility. Though in our trend
analysis of urban-rural fertility 1960 data were included,
it is expedient to take a more detailed look at the urban-

rural fertility differential for 1960. Table 8 presents for

37

 

.mmouc cw nmmumo:oz mdmcoo .moflumad Hausa .ucm>um zuwwx cam .mamomm cmaad .b .hmzmnumm mama "muusom

 

H.H~ o.h~ H.md s.m~ ¢.om m.m~ 0mm.~ mm~.~ mm~.H Hum.~ mos.~ mo¢.H mmo.m eom.~ o~m.~ oauaoam
m.n~ o.- m.oH m.m~ ~.H~ o.o~ ooh.~ omm.~ on.H ~Sm...” oom.~ o~m.~ msv.n had.n omm.~ camucso:
m.- m.v~ 5.6H o.s~ v.mH m.H~ nmm.~ oov.~ ¢-.a maa.m omn.~ a~¢.a m-.m wom.~ mmv.H anuunuo canon yum:
o.~m m.m~ m.m~ m.mm o.- n.m~ Ho¢.~ om~.~ mHH.H mmq.m mom.~ Hmm.~ oom.m smo.~ mm~.H Hmuusoo nusom ammo
m.m~ v.vH ~.n~ m.oe v.- ¢.¢H mh~.~ «NH.~ mmo.a wmm.~ mwe.~ ao~.a mo~.m mom.~ mm~.H Unusuaua nunom
m.nH m.m~ o.mH n.s~ o.- ~.H~ mam.~ mmv.~ mom.a Hmo.m omn.~ mmm.a mm~.m mpm.~ mm¢.~ Hauucoo nuuoz um»:
¢.hd o.m~ $.5H o.Hm m.¢~ m.na vmv.m vmm.~ ma~.H maa.~ mum.~ om¢.~ mm~.n nmm.~ omv.a Hauunoo suuoz «mam
~.na m.o~ m.m~ m.av H.mm 5.0m mn~.~ emo.~ omo.~ moo.~ mme.~ vmm.a HH~.m mam.~ -¢.~ ufiucwaua manna:
s.m o.o~ m.¢~ o.am m.~m m.om Hm¢.~ oo~.~ oma.~ ome.~ om¢.~ mam.a oo~.m mam.~ oom.~ camamcm 302
o.o~ ~.¢H o.o~ m.mm ¢.m~ o.o~ woe.~ oo~.~ mBH.H mom.~ oam.~ ohm.H ~o~.m nwm.~ woe.” nuumum amass:

 

¢¢lmm #mlmN vNImH fivimm VMImN lemH ¢¢Imm enlmN GNImH ¢¢lmm ¢MImN VNImH ¢¢Imm GmlmN VNImH

 

nmnub mowmuxm mama cmnub mcwmoxm oumm swan: EHmtn:OZtamusm Bumhldmuflm
BumhInOZUHmusm unmouom Bummldmusm unwouwm

 

we: can wocmcwmmm an uumm mumawuuom ouwsz.o>quHSEbo

 

 

Goad .ndOfl0dhav
can aouuum cuuucb uuonwauoucou wow mom can oococwmou >8 c0203 over: poauumaluubm ccc.H hon anon uu>o suuwddsu HO Hun-52 .0 means

 

 

 

 

 

 

 

38

t:he nation and its nine divisions in 1960 the cumulative

 

‘Nhite fertility rate, i.e., the number of children ever born
jper 1,000 ever-married white women, for all three residence
categories (rural—farm, rural-nonfarm, urban) by broad age
divisions of the women's reproductive period (15-24, 25-34,
35-44). In contrast with the trend analysis fertility data
discussed previously, here we are (1) using a different mea—
sure of fertility, children ever born, which is actually a
measure of average number of children per married white
woman, (2) including data for fertility levels for three
residential categories rather than the simple dichotomy of
rural-urban, and (3) inserting a control for age of women.
It is of Special import to consider at this point cumulative
fertility rates by division, residence and age categories
for 1960, since these are the products of the differential
fertility patterns which will be investigated in this study.
When we consider internal variation of fertility levels for
residence groups later, the cumulative fertility rate will
be the measure employed.

Table 8 not only exhibits a consistent and sub-
stantial rural-urban fertility differential for 1960, but
also sizeable intra-residence group differences among the
divisions. Among all three age groups the number of chil-
dren ever born per 1,000 ever-married rural-farm white women
in the nation as a whole ranged from 20 to 36 percent above

that for urban white women. For rural-nonfarm white women

 

 

 

 

39

tzhis rate ranged from 14 to 21 percent above urban white
vvqmen. In comparing rural—farm and rural—nonfarm white
fertility rates, however, the differences are small (rural-
farm rates ranging from 3 to 12 percent above rural—nonfarm
fertility rates). The pattern for the nation, then, is one
of high cumulative fertility rates for the rural—farm popu-
lation, intermediate for rural-nonfarm, and low for urban.
Furthermore, fertility rates are more similar for the two
rural residence groups than either rural residence group is
to the urban group. For white women ages 35 to 44 the dif—
ference in rural-farm and urban fertility rates is 854,
rural—nonfarm and urban 495, and rural—farm and rural—
nonfarm 359.

The pattern of high rural-farm, intermediate
rural-nonfarm and low urban fertility is repeated with few
exceptions by the nine divisions of the nation in 1960.

The exceptions to this pattern are all instances where the
rural-nonfarm fertility level exceeded the rural—farm, but
in no case does the urban level exceed that of either rural
pOpulation, farm or nonfarm. The magnitude of the fertil—
ity differential between white rural-farm and urban resi-
dence groups for women age 15 to 24 ranges from 14 percent
in the South Atlantic divisions to 31 percent in the Middle
Atlantic above the fertility rate for white urban women at
this age. In this age group and women ages 25 to 34, there

is some indication that the rural-farm and urban fertility

.-+-

.v-‘

v~-

|

9.

's

‘A

v

..v

«I'

(1’

(I!

 

 

 

A—o—A-‘HfiA

 

4o

differential is largest in the more urban divisions and
leaast in the less urban divisions. Rural-farm fertility
leavels for women 25 to 34 range in excess of urban fertility
firom a high of 35 percent in the Middle Atlantic to a low of
153 percent in the West South Central. For the age group of
35 to 44, an age group of women for whom childbearing is
near completion and provides some indication of complete
fertility levels for the p0pulation, rural—farm fertility
exceeds urban fertility within a range of 41 percent for
Middle Atlantic and South Atlantic to 26 percent in the
Mountain division. To conclude this discussion on the
rural-farm and urban differential, it might be said that
insofar as the age groups of married women represent current
and completed fertility, there is evidence of a continuation
of shrinking differentials. However, considering that the
more urban divisions reveal the largest gaps between rural—
farm and urban fertility, as the nation becomes progressively
more urban, the differential should eXpand rather than con-
tract. This same pattern is supported by previously dis-
cussed trend data and the impact of upsurging fertility
levels during 1940-1960 on the urban—rural differential.

The size of the fertility differential between
rural-nonfarm and urban fertility, when compared among age
cohorts of women, seems also to show a continuation toward
convergence, although the variation among divisions is not

as extreme as in the rural-farm populations of the divisions.

 

 

 

 

 

41

Threea exceptions are New England, Middle Atlantic, and East
North.Central divisions which are highly urbanized. The
differential seems to be eXpanding in these divisions.
Rural-nonfarm fertility for women 15 to 24 ranges from

15 percent above the urban rate in New England and South
Atlantic to 24 percent in the Middle Atlantic and for women
25 to 34 from a low of 10 percent above urban in New England
to 18 percent in the East South Central. Rural-nonfarm fer-
tility for women 35 to 44 exceeds the urban level by only 9
percent in New England and 33 percent in the East South
Central division. With reSpect to the rural-nonfarm and
urban pOpulations it is more difficult to generalize that
the more urban divisions reflect fertility differentials
than the less urban divisions since the ranges are rela-
tively small.

There is considerable amount of variation within
residence groups when comparisons are made among divisions.
White women residing in the Mountain division, for each age
and residence group, show higher fertility levels than com-
parable white women in any other division. Within the
rural-farm pOpulation the average number of children per
white married woman 15 to 24 ranges from 1.6 in the Mountain
division to 1.2 in the South Atlantic. This average for
women 35 to 44 ranges from 3.5 in the Mountain division to

3.0 for Pacific. For rural-nonfarm women 15 to 24 the

 

average number of children ranges from a low of 1.3 in South

Atlantic to a high of 1.5. For women 35 to 44 New England

 

 

 

42

and Ddiddle Atlantic divisions are low with an average of

2.7 children, Mountain is high at 3.3. For urban women

15 to 24 Middle Atlantic, South Atlantic and East South
Central all have an average of 1.1 children per white
married woman and Mountain has the high average of 1.3.
Urban completed fertility levels range from a low of 2.3 for
Middle.Atlantic, South Atlantic, and Pacific, to a high of
2.8 in the Mountain division. Hence, not only is there
Continuing urban-rural fertility differential existing among
all divisions of the United States, but there is also a con-
siderable amount of variation in fertility levels within the
same residence group among the various divisions due par-
tially to the varying degrees of urbanization which exists
among these divisions. These data suggest that if one is

to investigate the internal variation of fertility within
residence groups, the analysis should control for divisional
variation of fertility levels at the same time.

The fact that this study intends to look "within"
the fertility pattern of each residence group, for the
nation and for each geographic division, raises the question
of the homogeneity of fertility levels in each residence
group. A look "within" assumes that there is variation of
fertility levels to be eXplained or accounted for. If com—
plete homogeneity existed within residence groups, i.e., if
all rural-farm women produced very close to the same number

of children within the same division, there could be no

 

 

 

43

expljanation of variation for the rural-farm population.
However homogeneity does not exist for fertility levels of
residence groups even when controlling for divisional
effects. Table 9 is a rough attempt to portray and facil—
itate the visualization of variation that does exist among
Cumulative fertility levels (measured as number of children
ever-born per 1,000 ever-married white women age 15 to 44)
for residential parts of counties in conterminous United
States and the nine divisions. The already established
fattern of the tendency of high fertility levels in rural—
farm areas, intermediate in rural—nonfarm, and low in urban
areas is easily observed in the table, as well as the fact
that among the various divisions there are considerable
differences in the pattern and extent of variation of fer-
tility levels within each of the residence groups. Never-
theless what this table attempts to establish is the fact
that there is a significant amount of variation of fertility
levels among residential parts of counties respectively even
within each of the divisions. Although we are not using a
specific measure of variation, percentage distributions can
give a rough indication of clustering or scattering of fer-
tility levels within rural—farm, rural—nonfarm and urban
pOpulations. An eyeball analysis of Table 9 tells us that
fertility rates of urban parts of counties have a greater
tendency to "bunch up" than rural-nonfarm and rural-farm

rates. In other words, there is relatively less variation

 

 

 

,_
....
~.
.-
.
...
v. .
. ,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

44

Table 9. Percentage distribution of urban, rural-nonfarm, and rural-farm parts of counties Ly
children ever born per 1,000 ever-married white women age 15-44 for conterminous
United States and divisions: 1960

Children Ever Born per 1,000 Ever-Married White Women 15-44

 

 

 

Less Than 3,500
2,000 2,000-2,499 2,500-2,999 3,000-3,499 or Over Total

NO. % N0. % NO. % NO 0 % NO. % NO. ‘7’:
URBAN
United States 407 19.2 1,387 65.2 303 14.3 23 1.1 5 0.2 2,124 100
N9W’England 0 0.0 55 93.2 4 6.8 0 0.0 0 0.0 59 100
Middle Atlantic 23 16.1 118 82.5 2 1.4 0 0.0 0 0.0 143 100
East North Central 21 5.7 286 77.5 60 16.3 2 0.5 0 0.0 369 100
West North Central 37 10.8 213 61.8 90 26.2 4 1.2 0 0.0 344 100
South Atlantic 152 40.6 217 58.1 5 1.3 0 0.0 0 0.0 374 100
East South Central 101 44.7 117 51.8 8 3.5 8 0.0 0 0.0 226 100
West South Central 66 19.1 215 62.1 55 15.9 6 1.7 4 1.2 346 100
Mountain 3 2.0 72 47.0 67 43.8 10 6.5 1 0.7 153 100
Pacific 4 3.6 94 85.5 12 10.9 0 0.0 0 0.0 110 100
RURAL-NONFARM
thuted States 59 2.0 1,385 46.2 1,313 43.8 202 6.8 37 1.2 2,996 100
New England 0 0.0 40 60.6 24 36.4 1 1.5 1 1.5 66 100
Mimfle Atlantic 2 1.4 97 68.3 43 30.3 0 0.0 0 0.0 142 100
East North Central 2 0.5 182 41.8 236 54.3 15 3.4 0 0.0 435 100
West North Central 6 1.0 211 35.5 325 54.6 51 8.6 2 0.3 595 100
South Atlantic 27 4.9 358 65.5 138 25.2 24 4.4 0 0.0 547 100
East South Central 8 2.2 198 54.9 123 34.1 24 6.6 8 2.2 361 100
west South Central 6 1.3 187 41.0 223 48.9 29 6.4 11 2.4 456 100
Mountain 7 2.7 58 22.1 126 47.8 57 21.7 15 5.7 263 100
Pacific 1 0.8 54 41.2 75 57.2 1 0.8 0 0.0 131 100
RURAL-FARM
United States 17 0.6 481 17.7 1,438 53.1 618 22.8 158 5.8 2,712 100
New England 0 0.0 12 20.7 27 46.6 18 31.0 1 1.7 58 100
Middle Atlantic 2 1.6 27 21.3 71 55.8 25 19.7 2 1.6 127 100
East North Central 0 0.0 53 12.7 225 53.8 106 25.4 34 8.1 418 100
West North Central 1 0.2 71 11.8 321 53.2 173 28.7 37 6.1 603 100
South Atlantic 8 1.8 127 28.1 238 52.7 63 13.9 16 3.5 452 100
East South Central 2 0.6 70 20.1 187 53.5 70 20.1 20 5.7 349 100
West South Central 2 0.5 88 22.2 202 50.8 86 21.7 19 4.8 397 100
Mountain 1 0.5 15 7.7 91 47.0 58 29.9 29 14.9 194 100
Pacific 1 0.9 18 15.8 76 66.6 19 16.7 0 0.0 114 100

 

p‘

nob‘

 

.nv-"'
, -

...d'l

 

...‘aoOQ

--..~ I

no...— 0

y.

.
‘k—

(I)

‘Vfl

 

 

 

45

hazaccount for within urban fertility. In comparison,
ruralrfhrm fertility reveals the largest amount of relative
variation among residence groups. It is interesting to note
that even within the urban residence category, the divisions
Which have the smallest variation are those which are more
urbanized. New England, Middle Atlantic and Pacific divi—
sions have 93 percent, 83 percent, and 85 percent of their
urban parts of counties reSpectively falling within the
Cumulative fertility range of 2,000 to 2,499. Over half

(51 percent) of the urban parts of counties in the Mountain
division have fertility rates of over 2,500. Although rural—
farm fertility levels reveal the wider variation than each

of the other residence groups, the distribution patterns for
rural—farm fertility seems to be more consistent among the
divisions than urban or rural-nonfarm fertility. Although
the amount of variation may have something to say about the
success of accounting for urban vis—a-vis rural-farm fertil-
ity fluctuations, given the same number of variables, the
point to be stressed here is that there remains a consider-
able amount of variation to be accounted for within residence

groups of the United States.

C. L

N. ..

n...,

*--.vo

 

. .
. 5p.
-.--A-:
'.
.._~
...
‘:c6 ‘
..11. .
‘ n
- Q
‘ I
-,..
...,
“- —
Q
-.

 

h. .
..-
.
‘\.‘,
u
0‘
4"
~ ‘
.0
u
‘u. .
t‘ -
‘5
‘u
‘n .
. ‘-“

 

fi

..
<'. .

\- ‘
s

"
q

.t
‘

.
a
.

v - ‘

... v-
‘I

... ‘

-

... :
~h

\

.

. ‘ ~

.

-- .
a c

.

a--~
I - -
2.

~

.

.__

. -.
s- .

‘s

-

\‘.

I ’.
._ ,

. .
-
¥

.-

p

n
.. 5

u '-

..

.

U. ‘ I
. .
§

\
\

.-

.A.‘

Q. '

\ e—
t

' v
. O

“. U

y r
‘\
I
lu'.
‘

 

 

 

 

46

The Goldberg Hypothesis

 

In concluding this discussion of trends and current
status of the urban—rural differential in fertility, it
would not be complete without taking into account a confound-
ing issue that has cropped up in recent studies.40 We may
call this the "Goldberg hypothesis," although others have
contributed to the analysis of this problem as well. The
Goldberg hypothesis emanates from the recent finding that
"farm background" over and above the more traditional "cur-
rent residence" variables (urban, rural—nonfarm, rural-farm)
is a significant determinant of the rural-urban fertility
differential. Briefly the hypothesis states that the
inverse relationship of fertility and socio-economic status
found in urban areas is the product of a large proportion of
farm migrants which are diSproportionately concentrated in
the lower socio—economic status categories of the urban

pOpulation. The Freedmans estimated that in 1952 more than

 

40The major studies include David Goldberg, "The
Fertility of Two Generation Urbanites," Population Studies,
XII (March, 1959), 214—22; David Goldberg, "Another Look at
the Indianapolis Fertility Data," 0p. cit.; Ronald Freedman
and Deborah Freedman, "Farm-Related Elements in the Nonfarm
POpulation," Rural Sociology; XXI (March, 1956), 50—61;
Ronald Freedman and Doris P. Slesinger, "Fertility Differen-
tials for the Indigenous Non—Farm Population of the United
States," ngulation Studies, XV (November, 1961), 161-73;
and Otis Dudley Duncan, "Farm Background and Differential
Fertility," Demography, II (1965), 240—49. Also see a much
earlier discussion of this problem in T. J. Woofter, "Trends
in Rural and Urban Fertility Rates," Rural Sociology, XIII
(March, 1948), 3-9.

 

 

 

 

. .-v

a
“vi...
6... ._
'
...
' ...
-
'2'~ .
on. ‘
' s

D‘.'-

'u.. ,

c
n
' .
‘
-“
I
..-
.. .-
. --’
.
‘u.
'1.»
..-
it...
Q.
‘5
. ~~

.
n..~

‘-

Ԥ

 

 

 

 

47

twicre as many farm-reared adults were living off the farm
aScnthe farm in the United States and that one of every
three adults living in a nonfarm place was reared on a

farm.41

With reSpect to a number of variables the Freedmans
found that the farm-reared pOpulation revealed relatively
distinct distribution patterns in comparison with persons of
the nonfarm pOpulation with no farm eXperience and that "the
farm-reared have come into the nonfarm economy relatively
poorly prepared from an educational point of view . . .

[and] have tended to fill relatively low-status jobs and

to earn low incomes."42 Goldberg argues that the study of
Luban fertility differentials based on current residence
categories will be complicated by the presence of rural
elements in the urban population. He finds that the inverse
relationship of fertility and socio—economic status is char—
acteristic only of the rural migrants in the urban popula-
tion, but not of the indigenous urban pOpulation.43 Fertil-
ity behavior of farm migrants, then, is much different than

indigenous urbanites in that farm migrants on the average

had a significantly larger number of children than indigenous

 

41Ronald Freedman and Deborah Freedman, op. cit.,

42Ibid., p. 54.

43David Goldberg, "The Fertility of Two Generation
Urbanites," 0p. cit., and "Another Look at the Indianapolis
Fertility Data," 0p. cit.

...

 

 

. ,4]

 

 

 

 

48

Imbariites. This finding suggests that even within the urban
residence category there are urban-rural differences which
may complicate or mask over the relationship potentially
characteristic of an indigenous urban population.

These findings are especially relevant to our study
Which will attempt to investigate the differential fertility
patterns existent within each of the residence groups.
Basically what this implies is that the fertility differen-
tial patterns in the residence categories will not be as
disshnilar as they potentially could be because of the con-
fbunding effect of rural migration to urban and rural-non-
farm areas. Except for Special surveys, in studies based
on census data, which is the case of our study, it is impos-
sible to differentiate within the residence categories
between farm and nonfarm background. Since no measure of
farm background is available it will be impossible to con-
trol for the effects of migration between residence groups
on internal fertility patterns.

However, this does not render the present study
invaluable or unreliable. There are perhaps many controls
which should be made in such a study but for which data are
lacking. Furthermore, there are reservations which could be
stated at this point concerning the hypothesized effects of
migration between residence categories on fertility patterns.
Hashmi criticized the Goldberg hypothesis on four counts:

(1) two-generation urbanites have a very different religious,

 

 

 

 

49

Iaativity, ethnic and educational background than more recent
Inigrants and these were not fully controlled in making his
comparisons; (2) fertility rates have fallen in rural areas
as rapidly and as far as in urban areas, lagging only by a
decade or two behind their urban rates; (3) there is no
sociological rationale for such a persistence; rural culture
is neither homogeneous nor especially resistent to change;
(4) the recent "baby boom" has not been shown to be an out-
break of rurality among urban pOpulations, but something
which has been most pronounced among the most urbanized
segments of the population.44 In addition, Grabill, Kiser
and Whelpton provide some indication of the impact of the
rural to urban shift of the population on the decline of
national fertility levels from 1810 to 1940. Their table is
reproduced here as Table 10. Generally for the nation the
shift of the pOpulation to urban areas accounts for only 20
percent of the changes in national fertility and much less
for changes in fertility levels of the divisions. This
table seems to suggest that changes in urban and rural fer—
tility per se account for most of the change in national and
divisional fertility levels, but what factors account for
changes within rural and urban fertility? We have already
established the necessity of considering the internal pat-

terns of fertility within residence populations.

 

44Sultan S. Hashmi, "Trends and Factors in Urban Fer-

tility Differences in the United States," op. cit., p. 126.

 

50

 

 

 

 

 

 

Table 10. Urban-rural components of decline in number of
children under 5 years old per 1,000 white women
20 to 44 years old, by divisions: 1810—1940
Percent Distribution
Absolute Decline Due To
Decline
in Rural Decline Decline
Children to Urban in in
Per 1,000 Shift of Urban Rural
Area Women Total Population Ratio Ratio
United Statesa 890 100.0 20.2 23.8 56.0
New England 705 100.0 17.0 33.5 49.5
Middle Atlantic 969 100.0 20.4 30.7 48.9
E. N. Central 1,314 100.0 17.3 25.1 57.6
w. N. Centralb 1,379 100.0 3.9 26.9 69.2
South Atlantic 861 100.0 16.3 18.6 65.1
E. S. Central 1,161 100.0 9.9 15.3 74.9
W. S. Central 909 100.0 15.4 14.4 70.2

 

 

aIncludes the Mountain and Pacific Division in 1940 but not
in 1810 when they were nonexistent.

bThere was a nonexistent urban population in the West North
Central Division in 1810 which of course had an indetermi—
nate 0/0 ratio of children to women. It was necessary to
assign some value to the ratio. The rural ratio was
assigned.

Source: Wilson H. Grabill, Clyde V. Kiser, P. K. Whelpton,
The Fertility of American Women (New YOrk: John
Wiley, 1958), p. 19.

 

 

 

 

51

In the same monograph Grabill, Kiser and Whelpton
reproduce a table (Table 11) reflecting the fertility levels
of the various types of migrants between residence cate—
gories for 1940. These authors conclude that "in general,
among ever-married women the fertility ratios of migrants
tend to be intermediate between the fertility ratios of non-
movers in the host area and the origin area."45 A closer
look at the table indicates that actually the migrants
resemble the fertility level of the host area more than the
area of origin. For example, rural—farm to urban migrants
show a fertility ratio of 407, rural-farm nonmovers 587, and
urban nonmovers 334. Rural—farm to urban migrants have a
fertility ratio differing from urban nonmovers by only 73
points, but by a 180 point difference when compared with
rural—farm nonmovers. This is true when considering rural-
nonfarm to urban migrants and rural—farm to rural-nonfarm
migrants. The Goldberg hypothesis, on the basis of these
findings, should be altered to take into consideration the
fact that rural migrants in urban areas do not closely
resemble in fertility behavior the population of their migra-
tion origin. Since migrants tend to be selective in their
fertility behavior, it might be concluded that rural migrants

may have approached urban differential fertility patterns

 

45Wilson H. Grabill, Clyde V. Kiser, and P. K.
Whelpton, 0p. cit., p. 102.

 

 

 

 

 

Table 11.

52

Urban-rural movement between 1935 and 1940 of

native white women 15 to 49 years old, by marital
status and number of own children under 5 years

old in 1940

 

 

Ever Married
Women 15 to

Own Children
Under 5 Per
1,000 Women
Ever Married

 

 

 

 

49 Years Old Standardized
Area and Mobility Status (000's) For Age
Same House, 1935 and 1940
Urban in 1940 3,205 334
Rural nonfarm in 1940 1,368 457
Rural farm in 1940 1,677 587
IntracounpyyMovers
(Interchange not tabulated)
Urban in 1940 6,026 369
Rural nonfarm in 1940 1,958 501
Rural farm in 1940 1,583 643
Migrants Between Counties
Urban to urban 1,358 314
Rural nonfarm to urban 345 344
Urban to rural nonfarm 553 400
Rural farm to urban 152 407
Rural nonfarm to rural nonfarm 273 464
Urban to rural farm 165 470
Rural farm to rural nonfarm 131 520
Rural nonfarm to rural farm 80 548
Rural farm to rural farm 294 659

 

Source:

Wilson H. Grabill, Clyde V. Kiser,
The Fertility of American Women (New York:

 

Wiley, 1958), p. 101.

and P. K. Whelpton,
John

53

already previous to their actual migration to urban areas or
that they acquired urban patterns soon after their arrival
in urban areas. Furthermore, it might be inferred that due
to a selective process in rural to urban migration the
effect of this migration on urban fertility differentials
may not be as great as hypothesized and that there still
will be considerable differences in internal residential
differential fertility patterns in Spite of the employment
of current residence categories.

Another consideration with respect to the impact of
rural-urban migration on fertility patterns has to do with
the magnitude of this migration by 1960 and thereafter. The
massive farm-to-city migration of previous decades is no
longer possible given the shrinking size of the rural, and
eSpecially the rural—farm, population of the United States.
While farm to urban migration will no doubt continue, those
with farm backgrounds can have little numerical significance
in the future. Bogue Speaking of farm migration and urban
growth in 1950 said, "the rural population has diminished to
a point where it can no longer be the major source of supply
of urban growth. If cities are to grow in the future,
natural increase probably must contribute by far the major

I 46 I 0
share of the increase." Hence, an "indigenous urban

 

46Donald J. Bogue, "Urbanism in the United States,
1950," American Journal of Sociology. LX (March, 1955), 478.

 

54

population" may not be as far off as the Goldberg hypothesis
seems to imply. Table 12 provides data for the percentage
distribution of farm-born persons in the white farm and non-
farm populations 18 years of age and over for the United
States in 1958. To me these data suggest that the Size of
the farm—born pOpulation in nonfarm areas is not consider-
ably large, hence, nonfarm fertility patterns will not be

greatly influenced by the presence of farm—born elements.

Table 12. Distribution of the white civilian population 18
years of age and over, by residence, farm or non—
farm birthplace, United States, 1958

 

 

 

 

 

Residence and Farm or Percentage
Nonfarm Birthplace POpulation Distribution
Total 98,014,000 100.0
Farm-born 22,199,000 22.6
Nonfarm—born 74,743,000 76.3
Not reported 1,072,000 1.1
Farm Residents 10,621,000 100.0
Farm-born 8,109,000 76.4
Nonfarm-born 2,512,000 23.6
Nonfarm Residents 86,321,000 100.0
Farm—born 14,090,000 16.3
Nonfarm-born 72,231,000 83.7

 

Source: Leo Schnore, "The Rural-Urban Variable: An Urban-
ite's Perspective," Rural Sociology, XXXI, No. 2
(June, 1966), 138.

 

55

A final comment with respect to the Goldberg hypoth—
esis pertains to the problem of investigating fertility
behavior, which is the product of past eXperienceS, by the
use of variables which reflect only current status of the
pOpulation. This is a significant contribution of the
writers connected with the Goldberg hypothesis. However,
the problem not only pertains to current versus past res—
idence, but to other variables employed as well in differ-
ential fertility analysis. Duncan, who found that both farm
background and educational attainment have a significant
effect on fertility, points out that educational attainment
(compared with such variables as income, occupation, and
female labor force participation) is perhaps a more useful
variable because it more accurately reflects the socio-
economic situation of the couple at or before the time
family growth was in process.47 It would be sound advice to
attempt to build into the analysis of internal residential
fertility variation, variables which tend to reflect past
status rather current. This procedure could perhaps correct
somewhat for the absence of a farm background variable in

the analysis.

 

47Otis Dudley Duncan, "Farm Background and Differen-
tial Fertility," 0p. cit., p. 242.

...

..~\

56

The Causal Approach to
Differential Fertilipy

 

 

Westoff, in an article concerning the changing focus
of differential fertility research,48 classified the develop-
ment of fertility differential studies along three lines.
First is the descriptive empirical studies which attempt to
establish the nature of the relationships and to confirm
their stability. Repetition of these studies served to
measure time trends. However, Westoff declares that "these
descriptive studies have been indespensable in defining the
subject but, nevertheless, are only preliminary to the
equally important task of ascertaining the causal complexes

49

involved." The second line of focus is classified as the

evaluative approach. This approach predicted that the so—

 

called "best" elements in society would die out because of
under-reproduction. The third line, the causal approach, is
the attempt to establish causes for differential fertility.
Westoff hails the Indianapolis Study as "the first major
study to test empirically substantive hypotheses which raise

50 In Spite of the fact that the India-

the question 'why.'"
napolis Study tested constructed hypotheses, it is question—

able that the study could be classified "causal," since "the

 

48C. F. Westoff, "The Changing Focus of Differential
Fertility Research: The Social Mobility Hypothesis," Milbank
Memorial Fund Quarterly, XXXI (January, 1953), 24-38.

491bid., p. 25. 501bid., p. 25.

 

 

a .

...

 

 

57

twenty—three hypotheses of the Indianapolis Study were not
bound together by an integrating theory or organizing prin-
ciple."51 Unfortunately, a majority of differential fertil—
ity studies fall within the "descriptive empirical" category,
many of these drawing upon census data.

Considering the demand for the determination of the
causal complexes of differential fertility, the present
study is causal in sc0pe, in that it will begin from a theo—
retical framework and will test hypotheses in Spite of the
fact that it will draw upon census data. The necessity of
such a study takes its cue from comments directed by Robert
Gutman toward a recent study of differential fertility.52
Gutman sees a methodological deficiency in a good number of
fertility differential studies, especially those based on
census data, in that they tend to regard the groups whose
fertility differences are being studied as "discrete uni—
verses of facts." Summary measures are computed to describe
what is going on in each of these universes with regard to
fertility and then these measures are compared. To correct
this Gutman calls for an approach "based on the assumption

that the different subgroups of the population are really

 

51C. V. Kiser, "Methodological Lessons of the
Indianapolis Fertility Study," Eugenics Quarterly, III (1956),
152-56; also see P. Hauser and O. D. Duncan, op. cit., p. 96.

52Robert Gutman, "Comment on Kiser's Paper," in
National Bureau of Economic Research, Demogpaphic and Eco—
nomic Change in Develpped Countries (Princeton: Princeton
University Press, 1960), pp. 113-116.

 

 

58

samples drawn from a single universe of phenomena."53 Gut—

man's eXpanded comments are as follows:

It is only when we come to regard the study of differen—
tial fertility as a means of approaching the larger
question of the causes of fertility variation in the
pOpulation as a whole that the assumption of samples
drawn from a single universe becomes relevant. For in
order to answer this question, it is essential to know
not only that there p33 fertility differences by social
group, but also to know the magnitude of these differ-
ences and the direction in which they are moving. What
is really crucial for understanding the role of occupa—
tional, educational, nativity, residential, and racial
factors as determinants of fertility is to ask what
proportion of the total variance of fertility is the
result of the differences in fertility between partic—
ular groups.

 

As I have indicated, there are studies in which the
importance of this question has been recognized. But
there are also many in which it has not, especially
studies based on census data. . . . How often do we
come across a statement which indicates the amount of
the total variance which can be explained in terms of a
particular coefficient of correlation? How often do we
find statements which tell us what proportion of the
total variance in a population is the consequence of
group differences among the dimensions studied and what
proportion is the result of differences within these
groups? Even less often, and in the case of the numer-
ous studies based on census data, never at all?54

 

In response to this justifiable indictment, the present

study will apply multiple regression analysis to census data.
Such application will not merely answer the charge of method—
ological deficiency, but consequently should supply a signif-
icant contribution to the analysis and eXplanation of differ-

ential fertility. A multiple regression analysis will not

 

53Ibid., p. 113.

54l§iQ-. pp. 114-115 (italics mine).

59

only permit the determination of the proportion of variance
in fertility levels explained by the independent variables,
but also will determine the relative importance of the vari-
ables included in the analysis in eXplaining fertility
variance.

In conventional differential fertility studies the
researcher is limited with respect to the number of inde—
pendent variables that can be related or cross—tabulated
with the dependent variable. In the past, the researcher
has usually correlated independent variables one at a time
with fertility, although in some cases a limited number of
control variables were inserted. For example, in the exhaus-
tive study by Grabill, Kiser, and Whelpton55 which appeared
as recently as 1958, the authors investigate fertility dif-
ferentials under individual chapters, such as, residence,
occupation, education, and other socio—economic factors.

The problem of the intercorrelation of the socio—economic
factors frequently used in conventional differential fertil-
ity studies is described in the following quotation:

In studying the various factors which differentiate
the fertility of one group from another, it must be
remembered that such factors are frequently closely
inter-connected. For example, the variations of family
size with income or occupation are closely linked
because persons with high income are usually in certain

occupational groups, and persons with small incomes with
other occupational groups. Thus to say that fertility

 

55W. H. Grabill, C. V. Kiser, and P. K. Whelpton,
The Fertility of American Women, op. cit.

6O

varies with income and with occupation is to some extent
merely6to describe the same phenomenon in two different
ways.

The employment of the statistical technique of multi—
ple regression analysis, under certain assumptions, should
minimize both the problem of inter-correlation and the lim-
ited number of independent variables which can be handled at
one time. Essentially, multiple regression analysis permits
the manipulation of several independent variables at the
same time while holding constant the effects of all indepen-
dent variables except the one which is of concern. Thus,
this study is "causal" in its intent, makes use of census
data, and employs multiple regression analysis.

Basic Design of the Study:
Distributive Apppoach

 

 

The object of this study is the investigation of

inter-communipy_fertility differentials within the residen—

 

tial sectors of conterminous United States. From this will
flow a comparative analysis of residential differential fer-
tility patterns. The point to be emphasized is that the
unit of analysis for this study is the "community," which
when Operationalized is equivalent to the residential compo-
nent of a county. It is to be noted that this approach to
the study of fertility is in complete contradistinction to

the large-scale, contemporary studies of fertility which

 

56United Nations, Op. cit., p. 85.

61

claim the individual respondent as their unit of analysis.
A partial rationale for our approach is that fertility is
largely a social phenomenon and that fertility measures are
themselves group measures. Under such an assumption the
level of analysis required must be sociological. In sug-
gesting the need for comparative studies of "groups of
peOple" rather than "person," Bogue asserts that:

It would seem that procedures for making comparative
studies of groups of peOple must constitute an important
part of social science methodology. The fact is that
this is a badly neglected aSpect. Present concern is
largely with studies designed to analyze data for indi—
viduals. Hence, it is essential that techniques for
comparative group analysis be devised. This would
include not only practicable solutions to the problems
of data-handling, but also an ppprOpriate point of view
from which to design studies{57

 

 

In terms of definition, a community is both a social
entity and a territorial entity. Hawley has defined the
community as "the structure of relationships through which
a localized pOpulation provides its daily requirements."58
In this study community and community structure are to be
considered interchangeable. Furthermore, it is assumed that
the social environment of the community, or community struc-

ture, effects measurable influences on the behavioral pat-

terns of the peOple interacting within that community.

 

57Donald Bogue and D. Harris, Comparative Population
and Urban Research via Multiple Regpession and Covariance
Analysis (Oxford, Ohio: Scripps Foundation, Miami Univer-
sity, 1954), p. l (italics mine).
58Amos Hawley, Human Ecology.(New York: Ronald,
1950), p. 180.

 

 

62

In the design of this study the variables employed are to be
considered indices of the social structure of the community.
To Operationalize the concept of community, however,
for the purposes of this study the census definitions of
rural-farm, rural—nonfarm, and urban will be accepted. The
census classifies the pOpulation of a county on the basis of
residence in an urban place or in a rural area. Hence, the
"urban community" constitutes all persons living within
urban areas as defined by the census within a Specific
county. Persons living in the rural area of a county are
further classified into rural-farm and rural—nonfarm. Rural
residents to be classified as rural—farm must live on a
place of 10 acres or more from which the sale of farm prod-
ucts amounted to $50 or more in 1959, or on places of less
than 10 acres from which sales of farm products amounted to
$250 or more in 1959. The rural-nonfarm population of a
county, then, is a residual which remains only after the
rural—farm and urban pOpulation have been identified.59
Consequently, to accept these definitions is to admit one
set of criticisms aimed at the validity of these concepts;

but to further equate these definitions to respective

 

59For a more detailed discussion of the definition
of residence categories see U.S. Bureau of the Census, U._S_.
Census of ngfllation: 1960. gpneral and Social Economic
Characteristics, United States Summary, Final Report PC (1)—
1C (Washington, D.C.: Government Printing Office, 1962),
pp. Vii—viii.

 

 

 

63

"communities" must admit even another set of weaknesses.
But, of course, it is difficult to find any Operational
definition of "community" without deficiencies. The problem
is well stated by Jonassen when he says:
Thus the student of communities in modern urban soci-
eties is faced with problems of delimitation and over—
lapping boundaries of community systems no matter what
type of community_unit he chooses to analyze. Recogniz-
ing that these problems cannot be completely resolved,
the term "community" may be used as a generic term to
designate types of social systems whose component parts
are spatially contingent.6O
The unit of analysis for this study, then, remains the
residential part of a county. The Spatial eXpression of
this Operationalized definition of community facilitates the
application of the distributive approach in this study.

The urban-rural differential in fertility has been
well-documented, almost to the point of contributional
sterility, or at least monotony. Such documentation has
lost any utility other than providing a description of the
pattern of fertility trends, requiring simply the comparison
of fertility levels measured at different points of time.
The study of the urban-rural differential fertility, indeed
the study of differential fertility in general, is in need

of alterations in the conventional design of such studies.

In short, the "aggregative" approach needs to be complemented

 

 

6OChristen T. Jonassen, "Functional Unities in

Eighty-Eight Community Systems," American Sociological
Society, XXVI (June, 1961), 400.

 

 
 
 

64

by studies based on the "distributive" approach. This is
not to say that the "distributive" approach is a new
approach. Rather, in view of the point of progress of urban-
rural differential fertility research, it seems greater con—
tributions will be made in future studies if a "distributive"
design is employed. The present study employs a "distribu-
tive" design with hopes of illustrating this point.
The distributive approach is best understood when
contrasted with the aggregative approach. Bogue, who is
a leading promoter of this approach and has recognized the
need to apply it to demographic data, differentiates the two
basic approaches in the following manner:
The pOpulation of a whole country may be studied in two
ways--as the residents of a single area universe or as
the residents of a Single area universe or as the resi-
dents of a congeries of sub-universes of which each sub-
universe has a particular location in space. The first
approach, the "aggregative," emphasizes the whole; the
second approach, the "distributive," emphasizes the

parts. These two approaches are complementary, since
each answers a class of questions that the other cannot.

61
It is assumed in the distributive approach, when

applied to the present study, that fertility rates vary

among the areal subunits chosen. In this case, the subarea

units are residential components of counties. The basic unit

of analysis, therefore, is areal or Spatial in nature. It

may be said, then, that among urban parts of counties in the

 

lDonald Bogue, "POpulation Distribution," in Hauser
and Duncan, The Study of ngulation (Chicago: University of
Chicago Press, 1959), p. 383.

65

United States some are relatively high, others low, and
others intermediate. What initially was a Single fertility
rate for the nation is now the weighted average of the rates
of the subareas, e.g., the urban parts of counties. Bogue
labels this "internal diversity" or "place variance."62 If
this variation of fertility rates is nonrandom, i.e., per—
sistently higher in some areas than in others, there must be
ascertainable reasons for this diversity. It is assumed
that eXplanation or interpretation of the particular distri-
bution of a population event, such as the distribution of
fertility levels among the urban, rural-nonfarm, or rural-
farm components of counties in the United States, lies in
the differential composition and/or the differential environ—
mental conditions of the subarea population.
EXplanation rather than description should be the

objective of the distributive approach, although this has
not been the case in many pOpulation distribution studies.
Thus, such a study should not be directed toward the question
of ppy_variations in fertility occur, but ypy they occur in a
sociological sense. To "break through" the descriptive
phase Bogue suggests that:

The factors that produce given pOpulation events can be

ascertained only by a broad comparative analysis, such

as observation of the variation of the events and attri-

butes in a number of different areas and observation of

which characteristics that vary among the areas covary

with the pOpulation event. This requires that a hypoth-
esis be formulated about what Specific appects of the

 

 

 

62Ibid.

 

66

environment are related to the_population events. If
a given factor varies independently of the pOpulation
event (does not covary with it), it may be presupposed
not to be an eXplanatory factor for the deviation of
the local area from the nation average.63

The present study, therefore, will attempt to devise hypoth—
eses to determine why selected compositional and environmen-
tal factors should be found to systematically correlate with
fertility variation among residential sectors of counties.
Thus, the design of the distributive approach is "causal"
only in the fact that causes are inferred from a theoretical
framework, not from the discovery of a statistical relation-
Ship.

This statement is a brief discussion of the distribu-
tive approach to fertility. Interestingly, Bogue himself
has presented an illustrative research program for a dis—
tributional analysis of fertility. In this brief illustra-
tion he states:

It has been demonstrated that much interarea variation
persists when age and color composition are controlled,
and much work has been devoted to Showing that fertility
differentials exist among various occupational, income,
religious, social and other groups. But as yet there
has been no measure of how much variation in fertility
each factor eXplains when all others are controlled or
of how much variation remains when all these factors are
considered simultaneously. . . . This would call for a
multiple-variable distribution analysis of fertility

measures for the white and nonwhite pOpulation separate—
ly, controlled for age and marital status.

 

63gb_iél.. p. 390 (italics mine).

64Ibid., p. 398.

67

In a limited sense, the present study attempts to perform
this task with the use of multiple regression analysis. The
analysis is focused only on the white population of the
United States. Age is controlled by the inclusion of age
variables as independent variables in the regression equa—
tion. Marital status is partially controlled in that the
dependent variable, fertility measured as the number of

children ever—born per 1,000 ever-married women age 15 to

 

44, is measured only for the ever~married segment of women
of the child-bearing age span.

Limitations of the distributive approach are to be
noted. First, the character of the subarea boundaries are
arbitrary in one sense, but fixed in another sense. They
are arbitrary in that there is an infinite variety of ways
by which the subareas of a given territory, such as the
United States, can be divided. They are fixed in that cen—
sus data are collected from predetermined political bound—
aries, such as the county, which may not necessarily conform

65

to "natural" boundaries. There are two major schools of

 

65This limitation is not only inherent in the dis—
tributive approach but also pertains to any correlation
analysis employing demographic data for spatial units.
Hagood and Price allude to this problem as well as four
other problems associated with applying correlation analysis
to demographic data: unequal size of units, choice of order
of demographic unit, tendency for demographic characteristics
not to be normally distributed, and lack of independence of
observations. Margaret J. Hagood and Daniel 0. Price,
Statistics for Sociologists (rev. ed.; New York: Holt,
Rinehart, and Winston, 1952), pp. 350-55.

 

68

thought among distribution analysts represented in the
"homogeneous area" vs. the "nodal area" argument. It will
be necessary only to distinguish the two schools inasmuch

as they relate to this particular problem. The former main—
tains that the areal units should be of maximum internal
homogeneity; the latter argues for the maximization of rela-
tionships and, therefore, that areas are to be delimited on
the basis of functional interrelationships. This latter
school is represented by metropolitan regionalism and the
distance gradient is employed as a major device for studying
the internal structure of nodal areas. Bogue calls for a
distributional analysis which takes account of both concepts
simultaneously.66 The present study at a simplified level
could be considered an attempt to include both notions. The
influence of metropolitan regions is assumed by the inclu—
sion of a measure of ecological distance from metropolitan
centers as defined by the census. On the other hand, and
without contradiction of the nodal area hypothesis, it is
assumed that variation between counties with respect to
residential sectors is greater than within counties. But,
of course, a county is not necessarily a homogeneous unit,
and, therefore, the criterion of maximizing internal homo—
geneity is not really carried out. The "nodal area" approach
is emphasized by the employment of a metropolitan dominance

framework.

 

66Ibid., pp. 394-5.

69

A second limitation of the distributive approach is
the contiguity problem. Briefly, this problem arises from
the fact that areal units are not independent of each other
but are contiguous. In other words, counties Situated close
to each other are more likely to be similar in their charac-
teristics than are counties which are some distance apart or
grouped together at random.67 However, the problem may be
somewhat resolved considering the number of variables
employed in the multiple regression analysis. As Bogue
states, "the error may be expected to decline as the number
of variables considered simultaneously is increased."68

A third limitation of distributional analysis is the
ecological correlation problem. Briefly, it is the inability
to generalize from findings based on areal units of observa-
tion to the individuals contained in those areal units.69
Responses to Robinson's criticisms have tended to reduce the

severity of the problem by developing meaningful interpreta-

tions of ecological relationships.70 Duncan hints at the

 

67Otis D. Duncan, Ray P. Cuzzort, and Beverly Duncan,
Statistical Geography_(Glencoe, Illinois: Free Press, 1961),
p. 129.

68

 

Donald Bogue, op. cit., p. 397.
69W. S. Robinson, "Ecological Correlations and the
Behavior of Individuals," American Sociological Review, XV,
1950, 351-57.

70Leo A. Goodman, "Some Alternatives to Ecological
Correlation," American Sociological Review, LXIV, 1959,
610-25; Leo A. Goodman, "Ecological Regression and Behavior
of Individuals," American Sociolpgical Review, XVIII, 1953,
663-4; L. A. Goodman and W. H. Kruskal, "Measures of

 

 

 

70

potential contribution from a nexus of ecological and socio-
logical analysis when he asserts that ". . . areal differen—
tials are significant in their own right. There is even
sociological basis for supporting that such differentials
may reflect factors influencing demographic phenomenon that
would not come to light in studying individual characteris—
tics solely."7l

In a recent study72 Duncan illustrates the effect of
ecological factors on fertility over and above the effects
of socio-economic variables measured on an individual level.
Applying multiple classification analysis to selected vari-
ables from the 1941 Indianapolis Household Survey, Duncan
determined the net effect of median census tract rent,
dwelling—unit rent of the couple, wife's education, hus-
band's education, wife's age at marriage, spouses' region
of birth, and tenure on fertility. The study concludes that
(l) the areal classification of rent levels produces fertil—

ity variations which are partially independent ofanuiadditive

to those due to the classification of individual dwelling

 

Association for Cross—Classification," Journal of the
American Statistical Association, XLIX, 1953, 732—64; and

O. D. Duncan and Beverly Davis, "An Alternative to Ecological
Correlation," American Sociological Review, XVIII, 1953,
665-66.

71Otis Dudley Duncan, "Human Ecology and Population
Studies," in P. Hauser and O. D. Duncan, The Study of Popula-
tion (Chicago: University of Chicago Press, 1959), p. 693.

72Otis Dudley Duncan, "Residential Areas and Differ-
ential Fertility," Eggenics Quarterly, XI, June, 1964, 82-89.

 

 

 

 

.0

.4

cu

 

71

units by rent and that (2) analysis in which an areal clas—
sification is examined Simultaneously with Several individ—
ual classifications of socio-economic characteristics sug-
gest that areal differentials in fertility may not be com-
pletely reducible to the areally clustered effects of some
conventional individual variables. Hence, the analysis of
the effects of ecological variables on fertility behavior is
worthy of separate investigation as it contributes to addi-
tional understanding of fertility variation.

Hashmi, who employed the distributive approach to
study fertility variation among census tracts in Chicago,
also argues for the significance of ecological correlations
in themselves in the following passage:

For the present study, the criticisms of Robinson
are only partially relevant. It should be remembered
that a birth rate is an attribute of a population as a
group and is intended to imply nothing about the behav—
ior of individual couples within the population. In
other words, a birth rate is an average of the behavior
of groups having high or low fertility. In fact, it is
exactly in this "ecological" sense that birth rates have
been interpreted in the past. . . . The study of fer—
tility rates within census tracts and a correlation of
the social and economic tract characteristics that are
associated with them is therefore of interest and funda—
mental importance for its own sake. . . . In fact, the
total environmental "climate" or socio—economic context
within which low or high fertility takes place may be a
much more important goal of demographic research than
the development of pgobability statements that apply to
individual couples.

 

73Sultan S. Hashmi, "Trends and Factors in Urban
Fertility Differences in the United States," op. cit.,
pp. 196—7.

 

72

Hashmi further argues that, nevertheless, inferences can be
made from ecological correlation to the individual if inter-
preted properly and with due caution. A high ecological

correlation does suggest that on the average individual

 

couples tend to behave in the direction indicated by the
ecological correlation, unless there are specific interven—
ing variables operating. However, these unknown intervening
variables can be controlled through partial correlation.
Hence, by holding constant these disturbing factors, the
probability of the conformity of the two levels of correla-
tions is increased.74 However it is to be emphasized that
the present study will not attempt to "bridge the gap"
between these levels of correlation analysis, Since the
ecological approach in nexus with the sociological will be
quite sufficient to provide the interpretation of variation
of fertility levels among residential components of counties.
Since this study deals with a group measure as the dependent
variable, it must likewise provide its analysis at the same
level.

To summarize, the basic design of the present study
is distributive in character. The variation in fertility
that exists among residential sectors of counties in the
United States will be explained by variation in selected

social environmental or community structural factors.

 

74Ibid., pp. 197—8.

73

Community and residential component of county are to be con—
sidered interchangeable in this study. Since interest lies
in why fertility levels fluctuate among communities, hypoth-
eses will be constructed concerning expected relationships
between community attributes and community fertility levels.
Color, age, and marital status will be controlled to some
extent. Finally, this study will not stop merely with the
intensive study of urban, rural-nonfarm, and rural-farm
fertility. The eventual objective is to compare systemati-
cally the different fertility patterns of each residential
sector. The distributive approach is not new and several
distributive studies of fertility within residential groups
have been completed.75 However, these studies have tended
to be highly descriptive, except for Hughes, Hashmi, and

Andarawewa, and limited in scope in terms of the territory

 

75Some examples of studies employing the distribu—
tive approach are E. de S. Brunner and J. H. Kolb, Rural
Social Trends (New Yerk: McGraw-Hill, 1933), Ch. V, "Rural
and Urban Relationships," pp. 111—19; P. K. Whelpton, "Geo-
graphic and Economic Differentials in Fertility," Annals,
CLXXXVIII (November, 1936), 37—55; W. S. Thompson and N. E.
Jackson, "Fertility in Rural Areas in Relation to Their
Distance from Cities, 1930," Rural Sociology, V (June, 1940),
143-62; C. M. Rosenquist and A. H. Schafft, "Differential
Fertility in Rural Texas," Rural Sociology, XII (March,
1947), 21—26; 0. D. Duncan, "Fertility of the Village Popula-
tion in Pennsylvania, 1940," Social Forces, XXVIII (March,
1950), 304-9; R. B. Hughes, Jr., "Human Fertility Differen—
tials: The Influence of Industrial-Urban Development on
Birth Rates," Population Review, III (July, 1959), 58—69;
S. S. Hashmi, 0p. cit.; and A. Andarawewa, "An Economic
Analysis of Fertility Differentials among Rural-Farm Commu-
nities in the United States in 1960" (unpublished Ph.D. dis—
sertation, Department of Agricultural Economics, Michigan
State University, 1964).

 

 

 

 

74

on which the study focuses, except for the Andarawewa study,
which is a sister study to the present one and has drawn
upon the same data, though it focuses only on the rural-farm
sector of the United States. Furthermore, not one of these
studies has attempted a comparative analysis of all resi-
dence groups based upon the results of a distributive analy-
sis. The characteristic of these studies has been to dwell
only on one residence group where the distributive approach
has been applied.

Summary: Requisites of Needed Research

on the Urban-Rural Fertilipy
Differential

 

 

In View of how much has been written to this point
regarding the nature of the basic problem of this thesis, it
might be profitable to recollect the main points of this
chapter. The primary purpose of this chapter has been to
specify in detail the type of research in differential fer-
tility that is needed today. Indications have been scattered
throughout the chapter, of course, of the optimistic expecta—
tion that this particular study will be able to meet these
needs. In a sense, then, one may consider several of the
points presented and supported in this chapter as providing
the basic requisites of currently needed research on the
urban-rural fertility differential. In another sense, how-
ever, one could view these requisites as a resumé of the

definitive characteristics of the design of the present

 

 

 

study.
follows:

1.

10.

75

These requisites presented in concise form are as

Fertility is the problematic factor in population
growth today and therefore requires intensive
research.

Fertility is primarily social behavior, not attitude.

Fertility is a socio-demographic phenomenon in large
measure dependent upon the social milieu.

The current approach to differential fertility is a
causal approach, involving the testing of empiri—
cally substantive hypotheses which raise the ques—
tion "why?"

Fertility is group behavior explanable at the eco-
logical or areal level.

In spite of predictions of the eventual convergence
of urban and rural fertility levels, urban and rural
fertility must be studied independently with an eye
toward looking inside each type of fertility behav-
ior to isolate factors that have a determinative
relationship to fertility.

The test must be made of the possibility that a
different set of factors influence rural fertility
viS-a-vis urban fertility or that the same factors
have a different effect on rural fertility than on
urban fertility.

There exists a considerable amount of variation in

fertility levels of communities within all residence
groups and such variation is explanable by variation
of the socio-environmental attributes of communities.

Census data can and Should be employed to facilitate
the discovery of correlates of residential fertility
variation because they lend themselves well to the
construction of empirical measures and the applica-
tion of statistical techniques of analysis on an
interval scale.

Multiple regression is employed as a very useful
statistical technique which can handle several vari—
ables simultaneously, determine the direction and
relative importance of the independent variables in
explaining fertility variation, and estimate the
proportion of the variation in fertility explained
by the independent variables.

76

11. The problem of determining correlates of residential
fertility variation and the employment of multiple
regression analysis are embraced within the broader
scope of the distributive approach to fertility
analysis as the major design of the present study.
The distributive approach is cross—sectional rather
than longitudinal.

12. The unit of analysis employed in this study is the
residential component of a county, interchangeable
with the term "community."

13. Controls are included for color, age, and marital
status.

14. Since the study is basically comparative of the
correlates of urban and rural fertility variation,
there must be an attempt to systematically contrast
the results of distributional analysis of the
residential categories.

Origin and Ogganization of the Study

 

The present study is an outgrowth of a larger study
conducted by Dr. J. Allan Beegle, Department of Sociology,
and Dr. Dale E. Hathaway, Department of Agricultural Econom—
ics, at Michigan State University. The larger study will
appear as one of the 1960 Census Monograph series. The data
on which the larger study is based, as well as this thesis,
are essentially drawn from the 1960 Census of Population
statistics on social and economic characteristics of persons
enumerated on the basis of a 25 percent sample of the popula-
'tion of the United States. A truncated version of the Cen-

saus magnetic tape, from which the Bureau of Census published

77

76 the third

its General Social and Economic Characteristics,
volume in the PC (1) series, was obtained for use in prepar—
ing the monograph by finances granted by the Social Science
Research Council. The statistical analyses were carried out
by the Armour Research Corporation of the Illinois Institute
of Technology in Chicago on a Remington-Rand Univac 1105
Computer. All variables included in this study were also
included in the original Census tape with the exception of
the metropolitan dominance variable.77

Since in this chapter the nature of the problem of
this dissertation has been established and the requisites of
a design to study this problem have been posited, a brief
description of the organization of what is to follow is now
appropriate. Chapter II contains a systematic review of
differential fertility studies. First, a broad overview of
the range of differential fertility studies is presented
followed by an intensive review of a select group of empir—
ical studies. A list of criteria is employed to determine

the studies out of an innumerable host of differential fer-

tility studies which are considered most relevant to the

 

76U.S. Bureau of the Census, United States Census
gfi'Population: 1960 General Social and Economic Characteris-
tics, PC (1), 1C (Washington, D.C.: Government Printing
Office, 1962) .

77For details as to how this variable was operation-
alized for insertion on the original census tape see pp. 231-
236 of this dissertation.

 

78

present study. This review attempts to summarize the find-
ings of these selected empirical studies in terms of the
relationships between fertility behavior and the independent
variables. Chapter III establishes metropolitan dominance
theory as the theoretical framework by which hypotheses are
generated. Urban dominance theory is rejected in favor of
metropolitan dominance theory as the more meaningful frame-
work by which to investigate urban-rural fertility differen-
tial patterns. Chapter IV sets forth the methodological
procedures employed to test the theoretical hypotheses
derived from metropolitan dominance theory. Both the con—
ceptual and statistical frameworks of this study are pre—
sented. Chapter V contains the main body of the study, the
analysis of fertility data at two territorial levels:
national and divisional. Finally, Chapter VI contains some
reflections about the findings of the analysis chapter and
implications for further research in the area of differen—

tial fertility.

CHAPTER II
REVIEW OF RELEVANT LITERATURE

Fertility has been and continues to be a popular
Object for empirical study, as the number of fertility
studies would attest. The quantity Of such studies renders
impossible the task Of a complete review Of the literature
dealing with fertility within the space of a single chapter.
There are a few excellent sources to which the reader is
directed for information concerning a general overview,
codification, and/or evaluation of the development Of the

study of fertility.1

 

lCharles Westoff, "The Changing Focus of Differen-
tial Fertility Research: The Social Mobility Hypothesis,"
Milbank Memorial Fund Quarterly, XXXI (January, 1953), 24-5.
This article presents a simple, but historically oriented
classification Of differential fertility studies up to the
Indianapolis Study. Several bibliographic references are
cited for each category of his scheme. Bernard Okun, Trends
in Birth Rates in the United States Since 1870 (Baltimore:
Johns Hopkins Press, 1958), Part III. Okun suggests a six—
fold classification Of fertility studies on the basis of
methodological approach with illustrations included. United
Nations, Department of Social Affairs, The Determinants and
Igpnsequences of Population Trends (New York: United Nations,
1953), Ch. V, "Economic and Social Factors Affecting Fertil-
ity," pp. 71—97. This source gives a well documented review
of factors which have been employed in studies of fertility
trends and differentials. The references cited, however,
luave a very heavy international flavor. Ronald Freedman,
'“The Sociology of Human Fertility: A Trend Report and
Bilfliography," Current Sociology, X-XI, No. 2 (1961-62),
Sup. 35-119. Freedman offers a very thorough review of the
satudy of fertility in the post—war period. The publication

79

80

Two functions of a review of literature of fertility
studies are of immediate interest: (1) to assist the
researcher in evaluating the contribution of his particular
study to the study of fertility in general and (2) to pro—
vide a basis for the construction of meaningful propositions
with respect to the problem at hand. In the first chapter
we have considered the general problem of urban—rural fertil-
ity differences. It was established that there is consider-
able variation within urban and rural fertility which remains
to be explained. Furthermore, it was asserted that a major
question to be further investigated is whether fertility
differentials within the urban and rural populations reflect

different patterns. The Objective of this chapter, then, is

 

includes a 636 item bibliography. In addition, due to the
appearance of several large-scale studies Of factors affect—
ing fertility beginning with the Indianapolis Study, numer-
ous articles have appeared which take the form of either
progress reports or evaluations of such studies. See espe—
cially David Goldberg, "Some Recent Developments in American
Fertility Research," in National Bureau of Economic Research,
Demographic and Economic Change in Developed Countries
(Princeton: Princeton University Press, 1960), pp. 137-51;
and Ronald Freedman, "American Studies of Family Planning
and Fertility: A Review of Major Trends and Issues," in
Clyde V. Kiser (ed.), Research in Family Planning_(Princeton:
Princeton University Press, 1962), pp. 211—27. A partial
list of additional articles on the large-scale studies
includes: C. V. Kiser, "Exploration of Possibilities for
New Studies of Factors Affecting Size of Family," Milbank
,Mpmorial Fund Quarterly, XXXI (1953), 436-80; C. V. Kiser
and P. K.‘Whe1pton, "Resume of the Indianapolis Study of
Scmial and Psychological Factors Affecting Fertility,"
Pcmulation Studies, VII (1953), 95—110; C. V. Kiser, E. G.

 

 

 

jMishler, C. F. Westoff, and R. G. Potter, Jr., "Development
of Plans for a Social Psychological Study of the Fertility
<3f Two-Child Families," Population Studies, X (July, 1956),
‘43-52: C. V. Kiser, "Methodological Lessons of the

81
to determine whether there is any empirical support for the
hypothesis Of contrasting urban—rural fertility differential
patterns. By reviewing several carefully selected empirical
studies of fertility pertaining to the urban-rural contrast,
empirical propositions can be extracted which will either
support or reject the general hypothesis of contrasting
urban—rural fertility differentials. If these propositions
suggest the possibility of contrasting patterns, then we
shall not only feel justified to continue the investigation
of this problem, but such empirical propositions establiShed
by previous research will serve in addition as guidelines by

which to consider theoretical frameworks which can increase

 

Indianapolis Fertility Study," Eugenics Quarterly, III
(September, 1956), 152—56; C. V. Kiser and P. K. Whelpton,
"Social and Psychological Factors Affecting Fertility:
XXXIII. Summary of Chief Findings and Implications for
Future Studies," Milbank Memorial Fund Quarterly, XXXVI
(July, 1958), 282-329; C. V. Kiser, W. H. Grabill, and J.
Schacter, "Plans for the APHA Monograph on Fertility in the
1960 Census Period," in.Emerging Techniques in Population
Research (New York: Milbank Memorial Fund, 1963), pp. 82-
101; Milbank Memorial Fund, Current Research in Human Fertil—
_i£y (New York: Milbank, 1955); Milbank Memorial Fund,
Thirtprears of Research in Human Fertility: Retrospect and
Prospect (New York: Milbank, 1959); P. K. Whelpton, "Fertil-
ity and Fecundity," in Needed Popplation Research (Lancaster:
The Science Press Printing Company, 1938), pp. 40—62; P. K.
Whelpton and R. Freedman, "A Study of the Growth of American
Families," American Journal of Sociology, LXI (May, 1956),
595-601; C. F. Westoff, E. G. Mishler, R. G. Potter, Jr.,

and C. V. Kiser, "A New Study of American Fertility, Social
and Psychological Factors," Eugpnics Quarterly, II (December,
1955), 229-337 C. F. Westoff, R. G. Potter, Jr., and P. C.
Sagi, "Some Selected Findings of the Princeton Fertility
Study: 1963," Demography, I (1964), 130—35; and George F.
Mair (ed.), Studies in Popglation (Princeton: Princeton
University Press, 1949), especially Section V, "Future
Course of Research in Fertility," pp. 143—69.

 

 

 

82

the depth of analysis of rural-urban differential fertility

patterns.

Criteria Employed to Select Empirical
Studies for Intensive Review

 

 

Not all fertility studies of an empirical nature
bear on the particular problem considered in this thesis.
Since the more parsimonious procedure would be to review
only those studies which have immediate bearing upon the
problem, I adopted Six specific criteria by which to deter—
mine the relevancy of any given study to my problem. I
decided that in the process of reviewing fertility studies,
if a study failed to meet any one of the six criteria, it
was to be omitted from further consideration. The six cri-
teria which were adopted are:

l. The study must be empirical in nature, i.e., must
attempt to describe, establish, or explain the
existence, direction, degree, and/or nature of the
relationship of some independent variables with fer-
tility behavior. In other words, it must attempt to

explain fertility differences with reference to
selected independent variables.

 

2. The study must use a measure Of fertility as the
dependent variable(s). Studies which focus only on
such measures as "expected fertility," "fertility
planning," "desired size of family," etc., will not
be included. Interest lies only in fertility behav—
ipp, not attitudes, values, aspiration, or expecta—
tions related to fertility. Studies which combine
these two aspects of fertility in their general
focus may be reviewed.

 

3. The study must employ at least one of the following
variables as the major independent (explanatory,
causal, etc.) variable: occupation, female employ-
ment, education, income (family, female, etc.), age
of women, ecological distance.

 

 

83

4. The study must employ data obtained for some areal
segment(s), or at least representative of that areal
segment, e.g., residence, region, areal sample, cen—
sus tract, etc., and/or the entire population of the
conterminous United States.

 

5. The study should add or contribute information re—
garding differences between rural and urban fertil-
ity levels.

6. The study must focus on primarily white fertility or
total fertility where it can be assumed that the
white population is a major component of the total
population or sample.

 

The intention Of the first criterion was to insure
that the design of studies to be reviewed were comparable to
that of the present study. It omitted articles dealing with
the phenomenon of fertility which were not empirical. The
first criterion also eliminated studies which employed fer—
tility as an independent variable, or studies which dealt
only with the major processes of population change, i.e.,
studies in formal demography, but which failed to relate any
extraneous variables to these processes.

The second criterion tended to eliminate from con—

sideration studies that did not treat fertility behavior as

 

a dependent variable. Since the appearance of the Indianap—
olis Study there has appeared a new "twist" to fertility
studies: the investigation of "family planning," desired

family size," and the like, as the key to understanding

.fi- ..-

 

84

differential fertility.2 These studies often employ fertil-

ity attitudes rather than fertility behavior as their depen—

 

dent variable. These studies have been eliminated from
review since the factors which affect attitudes toward fer-
tility planning are not necessarily the same as those
affecting actual fertility.

The third criterion was most effective in reducing
the number of studies for review. The present study in—
cludes six general categories of explanatory variables to
be related with fertility behavior. These categories, of
course, were determined largely by the availability of types
of census data included in the census tape. Nevertheless,

the variables selected for analysis had been employed

 

2A partial list Of studies employing dependent vari—
able of fertility attitudes, such as, "desired family Size,"
"family planning," etc.: C. F. Westoff, E. G. Mishler, and
E. L. Kelly, "Preferences in Size of Family and Eventual
Fertility Twenty Years After," American Journal of Sociology,
LXII (March, 1957), 491—97; C. F. Westoff, P. C. Sagi, and
E. L. Kelly, "Fertility Through Twenty Years of Marriage; A
Study in Predictive Possibilities," American Sociological
Review, XXIII (October, 1958), 549-56: R. Freedman, P. K.
Whelpton, and A. A. Campbell, Family Planning, Sterility and
_Pppulation Growth (New YOrk: McGraw-Hill, 1959); C. A.
Yeracaris, "Differentials in the Relationship between Values
and Practices in Fertility," Social Forces, XXXVIII (December,
1959), 153-58; C. F. Westoff, R. G. Potter, Jr., P. C. Sagi,
and E. G. Mishler, Family Growth in Metrppolitan America
(Princeton: Princeton University Press, 1961); C. F.
Westoff, R. G. Potter, Jr., and P. C. Sagi, The Third Child:
A Study in Prediction of Fertility (Princeton: Princeton
University Press, 1963); D. Goldberg, H. Sharp, and R.
Freedman, "The Stability and Reliability of EXpected Family
Size Data," Milbank Memorial Fund_guarterly, XXXVII (October,
1959), 369—85; and Jeanne Ridley, "Number of Children
Expected in Relation to Non-Familial Activities of the Wife,"
Milbank Memorial Fund Quarterly, XXXVII (July, 1959), 277-96.

 

 

 

 

 

 

85

traditionally in other fertility studies. In spite of this
several fertility studies still failed to meet this crite—
rion. For such studies not even one of the six classes of
variables employed in the present study was also employed

in that investigation.3

 

3A partial list of studies omitted from review
because of failure to meet the third criterion: D. G.
Marshall, "The Decline in Farm Family Fertility and Its
Relationship to Nationality and Religious Background," Rural
Sociology. XV (March, 1950), 42-49; R. Freedman, P. K.
Whelpton, and J. W. Smit, "Socio-Economic Factors in Reli—
gious Differentials in Fertility," American Sociological
Review, XXVI (August, 1961), 608—14; Erich Rosenthal,
"Jewish Fertility in the United States," Eugenics Quarterly,
VIII (December, 1961), 198—217; H. E. Brooks and F. J. Henry,
"An Empirical Study of the Relationships of Catholic Prac—
tice and Occupational Mobility to Fertility,“ Milbank Memo-
rial Fund Quarterly, XXXVI (July, 1958), 222-81; P. C. Glick,
"Inter-Marriage and Fertility Patterns among Persons in
Major Religious Groups," Eugenics Quarterly, VII (March,
1960), 31—38; W. S. Thompson, "Differentials in Fertility
and Levels of Living in the Rural Population of the U. S.,"
American Sociolpgical Review, XIII (October, 1948), 516-34;
Margaret Hagood, "Changing Fertility Differentials among
Farm-Operator Families in Relation to Economic Size of Farm,"
Rural Sociology. XIII (December, 1948), 363—73; E. M. Kita-
gawa, "Differential Fertility in Chicago, 1920—1940,"
American Journal of Sociology, LVIII (March, 1953), 481—92;
A. J. Mayer and C. Klapprodt, "Fertility Differentials in
Detroit: 1920—50," Population Studies, IX (November, 1955),
148—58; E. G. Flittie, "Fertility and Mortality in the Rocky
Mountain West," American Sociolpgical Review, XXII (April,
1957), 189-937 H. Y. T'ien, "A Demographic Aspect of Inter-
state Variations in American Fertility, 1800-1860," Milbank
Memorial Fundlgparterly, XXXVII (January, 1959), 49—59;
W. Bash, "Changing Birth Rates in Developing America: New
York State, 1840-1875," Milbank Memorial Fund Qpprterly, XLI
(April, 1963), 161-82; and Sidney Goldstein and Kurt Mayer,
"Residence and Status Differences in Fertility," Milbank
Memorial Fund Quarterly, XLIII (July, 1965), 291-310.

 

 

 

 

 

 

 

 

 

 

 

 

86

Since the design of the present study calls for the
study of fertility differentials on an areal distribution
basis, it was necessary for comparability that studies to be
reviewed be based on or represent some areal segment(s) Of
the United States (Criterion 4). Thus, fertility studies
based on samples which were representative of some distin-
guishable territory or areal unit, such as, a region, a city,
a group Of cities, etc., were acceptable. The fourth crite-
rion tended to eliminate two types of studies: those based
on data from survey samples which were not areal samples4
and those based on data collected from populations outside

conterminous United States.5

 

4A partial list Of studies based on non—areal sam-
ples or populations: R. Gutman and I. Bender, "Some Sources
of Variation in Family Size of College Graduates," Milbank
Memorial Fund Quarterly, XXXV (July, 1957), 287-301: C. F.
Westoff, P. C. Sagi, and E. L. Kelly, "Fertility through
Twenty Years of Marriage: A Study in Predictive Possibil—
ities," American Sociological Review, XXIII (October, 1958),
549—56; P. Lauriat, "Marriage and Fertility Patterns of
College Graduates," Eugenics Quarterly, VI (September, 1959),
171-79; G. S. Becker, "An Economic Analysis of Fertility,"
in National Bureau of Economic Research, Demographic and
Economic Change in Developpd Countries (Princeton: Prince—
ton UniVersity Press, 1960), pp. 209-31; B. Pasamanick, S.
Dinitz, and H. Knoblock, "Socio—Economic and Seasonal Vari-
ation in Birth Rates,“ Milbank Memorial Fundpguarterly,
XXXVIII (July, 1960), 248-54; and E. D. Baltzell, "Social
MObility and Fertility within an Elite Group," Milbank Memo-
lglal Fund_guarterly, XXXI (October, 1953), 412—20.

5A partial list of studies based on populations out—
side conterminous United States: N. Keyfitz, "Differential
Fertility in Ontario: Application of Factorial Design to
Demographic Problem," population Studies, VI (November, 1952),
123-34; J. Berent, "Relationship between Family Sizes of Two
Successive Generations," Milbank Memorial Fund Qparterl ,
XXXI (January, 1953), 39—50; N. Keyfitz, "A Factorial

 

 

 

 

 

 

87

The fifth criterion was included to assure that the
studies to be reviewed would provide at least some informa-
tion on why urban and rural fertility levels differ, i.e.,
what in the community social structure of these residence
groups produces differences in the level of fertility.

Very few studies failed to meet this requirement since many
studies make accidental or implicit recognition of residence
in their designs and analyses. In a majority of cases the
decision was arbitrary as to a study's contribution in this
respect. Moreover, many studies have tended to focus on

urban fertility, Since, according to the transitional theory

 

Arrangement of Comparisons of Family Size," American Journal
Of Sociology, LVIII (March, 1953), 470-80; R. McGinniS,
"Similarity of Background Characteristics and Differential
Fertility," Social Forces, XXXIV (October, 1955), 67-72:

D. Wrong, "Trends in Class Fertility in Western Nations,"
Canadian Journal of Economics and Political Science, XXIV
(May, 1958), 216—29; H. Y. T'ien, "The Social Mobility Fer-
tility Hypothesis Reconsidered: An Empirical Study,"
American Sociological Review, XXVI (April, 1961), 247—57;

I. Adelman, "An Econometric Analysis of Population Growth,"
American Economic Review, LIII (June, 1963), 314—339; H.
Gille, "An International Survey of Recent Fertility Trends,"
in National Bureau of Economic Research, Dempgraphic and
Economic Change in Developed Countries (Princeton: Prince-
ton University Press, 1960), pp. 17—34; R. Hill, J. M.
Stycos and K. W. Back, The Family and Population Control
(Chapel Hill: University of North Carolina Press, 1959);

J. M. Stycos, Family and Fertility in Puerto Rico (New YOrk:
Columbia University Press, 1955); Paul Hatt, Backgrounds of
Human Fertility in Puerto Rico (Princeton: Princeton Univer—
sity Press, 1952); W. Stys, "The Influence of Economic Condi—
tions on the Fertility of Peasant Women," gppulation Studies,
XI (November, 1957), 136—48; D. Wrong, "Class Fertility Dif-
ferentials in England and Wales," Milbank Memorial Fund
Quarterly, XXXVIII (January, 1960), 37-47: and David M. Heer
and Elsa Turner, "Areal Difference in Latin American Fertil-
ity," Population Studies, XVIII (March, 1965), 279-82.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88

of fertility, urban fertility is the level and pattern to-
ward which the fertility levels in the high fertility rural
areas are moving. Thus, interest in fertility research has
been directed mostly toward urban fertility, especially with
respect to fertility planning among urban couples, rather
than rural fertility.6

Finally, the sixth criterion eliminated from consid-
eration studies dealing with nonwhite fertility.7 In the
case of studies contrasting white and nonwhite fertility, if
they met all the previous criteria, only the white data were

considered in the review.

 

6For example, the following large-scale studies of
urban fertility: R. Freedman, P. K. Whelpton, and A. A.
Campbell, Family Planning, Sterility and Pppulation Changg
(New YOrk: McGraw-Hill, 1959); D. F. Westoff, R. G. Potter,
Jr., P. C. Sagi, and E. G. Mishler, Familinrowth in Metro-
politan America (Princeton: Princeton University Press,
1961; C. V. KiSer, Group Differences in Urban Fertilipy
(Baltimore: Williams and Wilkins, 1952); and C. Kiser and
P. Whelpton, Social and Ppychological Factors Affecting
Fertility (New York: Milbank Memorial Fund, 1943-1954).

7J. E. Dodson, "The Differential Fertility of the
Negro Population, Houston, Texas, 1940—1950," Milbank Memo-
rial Fund_guarterly, XXXV (July, 1957), 266-79; C. V. Kiser,
"Fertility Trends and Differentials among Nonwhites in the
United States," Milbank Memorial Fund Quarterly, XXXVI
(April, 1958), 149-97; A. Lee and E. Lee, "The Future Fer—
tility of the American Negro," Social Forces, XXXVII (March,
1959), 228-31; E. Lee and A. Lee, "The Differential Fertil—
ity of the American Negro," American Sociological Review,
XVII (August, 1952), 437—47; and J. S. Vandiver, "The Repro—
duction Pattern of the Rural Negroes of the Yazoo-Mississippi
Delta," Social Forces, XXIX (October, 1950), 78-84.

 

 

 

 

 

 

89

Intensive Review of Selected List
of Empirical Studies

 

 

It is the Objective of this intensive review Of
empirical studies selected on the basis Of the Six criteria
listed above to facilitate the construction of propositions
relevant to our problem. Following an extensive survey of
the literature dealing with fertility by which a sizeable
bibliography was compiled, thirty-one empirical studies
were selected upon the application of our criteria. These
studies were thereupon submitted to a battery of questions
by which Specific characteristics and basic findings were
abstracted from the body of the study. The format of this
schedule is exhibited in Appendix B. Considering the now
manageable number Of studies to be reviewed, it was possible
to transfer this information from the schedule to a summary
table as presented in Tables 13 and 14.

The intensive review requires three stages. First,
some general statements are made to indicate the range in
characteristics Of these selected studies. Second, consid-
eration is granted to the designs of these studies as they
compare with the design of this fertility study as proposed
in the previous chapter. Finally, an attempt is made to
summarize the findings of these 31 studies inasmuch as they

are relevant to this particular study of fertility.

9()

.312: .803 .056. > .Nfifiuom
Hahn-l :5an .mowuuo loam 08000.3 HHS—n. Ou sawumaox 5.. 0002 Hanna 5 hawuwuumb; 5003006 .m .2 can comma-03a. .m .Bu

 

 

.wmlou gonad £0305 HHHBOOBU .3052 ...»uzmvuom 5 000:0.uouuwc 0030.. .fiou0ouoz .3 .mo

 

.97: .....n3 $000055 35.830 .3005 2.333000 5 083500003 33:30 05 62058000.. 63306. .s .00

..IIIIIIIIII .34: .00 .0030
...-30300 0300 0:0 ~38... .> :8 £83 Juan—130009. 3.qu 30.: 0000.2. 13000 102. .58. .m .6 0:0 005500 .0 00 .00

.0963 .133 40005 x .Nauouuama veal.

Hum: gal 0.93.33 gun :0 3 095.00 game—50o 0.3.333 no auuawuuonm 08,—... .gdnmm .x can floanouon .8 .mn

 

.210 .233 £305 29 030300000 Hmuaunaunum

003.0! 0.3 an HEB—8b 0.0030 300% 00 5.3004 haundmuuom Hauusouuuuwca 5.3000002 .3 3:00.“ can u0¥0wuu§mm .m 003mm

 

03.30 0000." 00.3.3 0»
id gum 804 2%....- HH 95000 .095 03030.0 mucus—006 use
00.— m in! 320 A55: H0 000M... 50000 on 5 Hanna one." 000000 came 0009005. .0

 

g 0.3% 035 one
mung JUNE—Hal 00895 DOE-Mm

 

 

IE- .Eu) 8.." Hum EU lanes 03.30 0095.3: c 0095 cmma h0g0 madman HE 0mm." 050000002 .m
00.3w". 00000.30
$1.3 00!: out.) 084 0500 now «use 0083 0:0
Hum m Henna 00.6.3.5 15090.3 nowuwu cuoauuoc c 008.5 omma 150 came n053.00 omen 6000.305. .4»
£5. 0 be .000 80; 003560 00
won a 3g 0006.350 0030 oceansouuae
"0000 4.55 00300000 030 88.9... 520 883
«mvlcu 5'0! Hon cu H095 25006 000A m hi8 .0330 ha a a 3mm: 053—
095.36 «300..— humflgah I953 004015.303 ca smash one." 000000 nmmd was Mean—:3 .m

 

passion-:0 00:0
lush Ben—.7259: .ouud coc.ccm

 

 

 

 

 

 

 

 

00.300! No 0.000% ccc ..n 0350a 00 cocoon no 090000 cam." loam gaaum
won on 000:0 a! 0» nuu 100.00: 00.330 000.3000 mm 0095 cam." 800.8 300.00 one." use s 0000002 .~
.000 2. “Honda
ous—3300 .3000! 0000 saunas—mom 02.6.3
.00an 30.00093 0350: no 02300 00 08.03 00430 A02 000000 0.8." Baum 050000002 can
sum—.00 .0035 con son a I952 0.03308 mm .0085 000 008.5 3S .5000 0.3-3m on: noxguuaocam ..—
000008. 333000 0300.0 0000000003 .0 .0 00305300 0300.200 008 ....o 03:00 0303 . 309304
«0 000.000 «00.: 95.5 0900 IA?—
033000 0000:0000 0000300.. “0.000% 000»

  

 

 

 

IIIIIIIII'IIIII'IIIIIIIIIIFI‘IF

.3000 030035 000 0000300 03050 333—00 00 033 in .fl 032.

 

 

9].

.muqom
..omma .00002. HHH>xx .uouuom Hmfloom =.oema .000000000000 00 0000000000 ommaaw> 0:0 no 000000000: .000000 000000 mauoo

 

 

 

 

.¢¢~ImmH .Amema .kumav HH>XX sNHu0uum0o 000m HawHOBoz xcmnawz =.m0umum DMEO:OOMI0w00m ha moumm mu“
Infiuuuh 000 mcwncmam auwawuu0m .xH "auwaauu0m mcwuu0wm< muouumm Hmowm0H0n0>0m 000 a0w00m: .noumaon3 .x .0 0:0 “onfix .> .00

.omnaw .Ahema .noumz. HHx .Nmmmmwmom wmuam ..mmxoa 00090 :0 000000000 000000000000: .0000000 .0 00>H< 000 0000000000 .2 00000

 

.omlamn .Amema .xa0nv Hxx .Hw000000o 000k amuuoamz xcmnaaz =.0«H0Q000«00H :« 00Hm000 0uaazl0>wumz mm0.a¢
maoad anaeauu0m H0000000MMflo .H “hudafiuu0h mcwuuomu< muouomm HmoamOHonohmm 000 Hmaoom: .000Hx .> .U 000 Goumdunz .& .mn

.an¢ma .mawxawz 000 memfiaaws "ouOEMuamm. waaauu0m 00805 0% 00000H0uwwo m0ouw .uonwx .> owhnUm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aoumum
«0 .000 .aaw> m\Hv
welma 00803 ooo.a 0>wu0 .0000 :w .mom oom.m Eumucoz
000 m 00000 :0ucauzo Inauumwo H0000 m0mmaaw> oeo Iamu0m o¢mH 000000 omad 000000o .HH
.0000 .000 unflauum .0H«H
0000000 0005 >000 0000A
0w 0000000 .00000000 00 00
00000 .800: .o« H0000 ouw3
.00300000 .mmlnnmu 0000005
..0000 .0uwa3l0>wuac 0>uumm . 0H030m h00um 000AH0£3
.00Hn000 cad 00m :00 noumm< mdaommcmwucu 0000: Head mdaommcmwwcH 000a wan u0udx .OH
0>wu0
ecuma 00303 ooo.H Inwuuuao 00x09 oumwasum 0:0
000 m 000:0 coucmwnu 0 no 00000000 0mm Eummnamu0m cema 000000 h0ma 0000000000 .0
m0 A ,
H0000 0ww3 .vowuumalouao m
.00033100w000 .00000000 0>w00 _ %0>H0m 0H0£000om nu0nux
.vvnma 00003 ooa 000 0M0 Inauuuwn waaommqawqu cmnuo Hews awaoamcuacau m¢md 000 noumaoaz .0
000000 m :w .000 m~
00000000 "H0000 «.m.D 00000 0H0500
.mmmn 000% you vfilma 00>«3 0>wumm no .000000 000000 H0000 Hanan a0>H0m
ooo.~ 00A onuuan obaq Ioummd ma 0“ 00%»«0 mm 00am .00009 mmon cannon H000dumz N¢mH uuouwx .5
000000: hudaduuoh amw000 0000000000m .m .D conunmnnlfi 00000HAOU 0000 no 00000» voznad Auvuonu0<
m0 aunoauom H0000 @0000 0000 In0m
ounnwun> 0000:0000 , 00000u00¢ 609000» H00»

 

 

 

 

 

moscwucouuunu 00000

S92

.m~¢lmmn .Ammwa .quOuoov
Huxxx .Nauuuumna‘unam Hmwuoauz stand“: 3.0mmauceoa mcuuac asumum uwaocoomIOMUOm ha huaawuuoh cw momcmnus .uoawx .> .oo

 

 

.mmuaoa .Ammma .awum<v
HHHxxx .qumuuaac vcam Haauoaoz xcmnaaz =.mwma .xuow 3oz .xuusoo cauwvuz :« huwaduuom Hmwuzuuoumwnz .aumm .m aaoflcoxu

 

 

ll. .HOImvm .A¢mma .uonbuoo.
xHx .3mw>0m AmowmmaowOOm cmuwuoad =.NmmH ou oooa unmumum Guyana 0:» ca %uwawnuuh Hmwucwumuuung .uwouuo3 .m uoaumnoo

 

 

.nmlmhd

 

 

.Ammmd .Haummv Hbe .3U«>om HaofimoHofiuom nwuauaafl :.uoumum mouficb on» ad auwawuuum can noduamnuoo- .Hoxcfin .2 uuunbmn

an‘ .mmuaoa .Awmma .auunxv HH>ux
.GOHumauound anuauuwuwum nmoauofid 0:» no Managed =.m0umum umuwcs ozu cw mamwucwuomuflo can uvcmus huwaauuomg .kuux .> .om

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

onH a»«
melma c0803 ouusz 0>wumm mm can o¢ma Iawuuom no .Afim .
cofiuumauuu>o ooo.H you man uoumo¢ moumum conga: .mza .cmnua omma Hmaoomm "anacou mmma ououax .oa
ouas3lo>fiumc .mmsuoo
.00 0mm Cu cauuvumocmum 0>wuma . .uwm cw ~\H u0>o =c36u: usmcmo
.nobw3 ooa uom mmo loumm4 .xucaoo cowaomz can amusm mama oumum xupw 30a mama . Gamma .mH
meuma c0803 ooo.a Mom
m uooc: convawcu ".oaom
“wouuameoo. +m¢ .A.uuao. o¢uoom~ maunoo
¢vlma :0503 ooo.a awn m>wumm hfima “>m>unw
m noon: coucawso “@5000 louvma mwumum oouaco amusulcmnua Nmma .mom ucmuuno «mad unwoumoz .¢H
ucumoum unanuan
.oco vawuume .ouag3lo>wumc o>wumm Gama
molow :0303 ooo.H uvm mmo noumma .m.D «0 mcowmou w Hmunulnmnub 0¢¢H asmcoo «mmn \mwuxcun .ma
mmlow
:08 cahonEo ooo.H uom Hmuau
m nova: nouvaunu «@5000 uow :ofiumo o¢¢a unncmo
meumd :oeo: magnuma ooo.a o>aumm ufiamaw 050» head can .nmmd >o>num
mom m uucca cuuoawno Ioummc .m.D Hmuoa ".m.D Hauoa ¢vma .mom unouuso «nod vacuum .NH
ousnmat audaauuom cognac voucommumom .m .D couanmsmam vououaaoo mama uo cannon conuud Aavuosusd
uo mucoemom gamut aaouu mama lash
cannuun> unaccomon oucocanom Auvummw F.uuma J;

 

 

 

 

 

 

 

coscaucoollna manna

533

.mn =~umsouo o aocoom can amfloom uaououuwa a“ exam maaanm. .xH .go ..mmaa .Haum
-3980: 3:8» 33: 5396 833: om can 322% 5:52 25m .2388 .4 .4 EB 53325 .x .m .55..on 333m.

   

 

 

.emnuomu .mm .Ammmn .uuoum ovum any .oouacau. nouaum
eunuch on» no coduau 0m «:9 .oamom .b nausea aw 3.:o«umHsmom aounum vouwcb 059 no huaauuu0h ugh: .Hduanuu .m nonduzu

 

 

 

Ammmu .aoawt uxuow 3oz. :0203 snowuua< no Nuwawuuoh och .coumaosz .x .m can .uouwx .> .0 .Hdwnuuo .m couawsn

IIIIIIIIIIIII: .mhlws ..mmma
.uonbuuov >Hxxx .aoouom Amwuom =.ommauo~ma .owcm>amw::um .csbuwwuuoz cw auaawuuom no nauouuum mcamcmnu: .auouucuou hoaowmn

 

 

 

 

 

 

 

 

 

 

 

 

 

wagonnaau
anfiuv mmma ha o>wumm “mamamn >o>u=u a coumdosx
unuuan no nonlsa omuu0>< loummc .Aonm .mvuv .m.D Buuuacunub mmma admin: ~40 mama .anflcuouh .o~
can ooo.a mom m noon:
nouuddsu :30 «NmmH oloucn
cusp) ooo.a Hum m anon:
caucaanu n30 umvmu maoucu
.mvlmv coin: ooo.a Mom mmma ho>usm
muo ..uacu .ewlma nob“: o>wumm meme .mom acouuao
ooo.a non nun «.mauoo uoummt nmumum cuuflcb Hausuncanua cmma cannon mmmu unawnmuo .ma
213 can
gulp: oa souuaano cauuu
"oaoucu flush ”couaounccaun
muo a “sea a» souuawno
Ouumu ”Hmmn .uca .uauuuni
.mvuma moan: ooo.u hon m
nova: coucaaso «mvma oaou m¢ma
law “.uoum .nnsn .coauuae «mma >o>u=m
ouno ..ocaun .ouwg: ‘mvumu oamu .acm uaouuao groumaugz
cosy: ooo- non mun ".m.q o>gunm ma can cum” on» gun“:
nu scab: a ..uauu ..maouo uoumm< noumum oouwcb .mzm .canup ommu uuucoo mama .Huunmuo .ma
omma .o¢m~ .onma .onma
.uua cannuo woodwon .uh n
a“ vegan on: 0&3 cusp: ou abuunm .oanzm no nmsou
anon nouuuwso .oc omnuo>4 Ioummd than .czounauuoz sank: omucuma menace: nuuwn mnmn acaouQOAoo .ha
ous-cox huuaauuom amwuon . coucoooumom .m .D coma-asmafl vouuonaoo «van no woman» con-«A Auvuonuud
no nuaaamom anou¢ nacho auon lava
cannuua> ucooconoo oucacunom Anvuonh an.»
fllllllrlll

 

 

 

 

 

 

 

 

 

 

 

 

 

UGQGdUBOOIInH OHAGH

94

..ooma
.oowunma .nm .Aooma

.muascmn. HHH>xxx . auouu

 

.muoum haamuo>anb aououcaum

.mmlawa
=.nmumum nouacb ecu mo acaumaamom Bummlcoz msocomavca Gnu new mamaucmuuwuao huaaauuumg

.Aaoma

 

 

m

.umAEo>ozv >x .moaoaum acaumaamdm
.uomaamoam .9 van amacooum camcom

.wmlmw

ouch amauOEaz xcmnaaz =.mumo huaaauuum maaommcmavca 0:» an xboq yonuocdg .muunvaou va>mav

mo smouam amcoaumz :a :.mumo mamcou mmumum omuaca on» Ca huaaauuwh amaucouuwuan: .moammam mucmz can muammnm wuunoamv

.maalhh .mm .aooma .mmmum huamuu>acb Gouwucaum

 

can oasmmummeua .suumonom uaaocoom mo smwusm anneaumz ca =.muumum cuuacb on» Ca kuaaauumm amauaououwana .uonwx .> .00

=.uoumm Suuam no unnamoao>on cmnuDIaMauuuavna mo oocmsawca one

unamaunouuumaa muaaauuom unsung

.molmm ..mmma

.ha9b. HHH .3ba>om ceauaaa om
..uh .uonmnm .m wanna

A

.mmlvan .ammma .noumzv Hax .moavaum coaumas om :.wouacmnub coaumuucoulo39 mo huaaauuom one: .muonvaow ua>mom

“couoocaumv mwauucaou 00m0a0>oo :a ummvno oaaonoum can uasmmummmun .nuumonom uaeocoum

 

"cououaaumv noauuasoo vom0a0>oo ca ummmru uaacaoom

 

 

 

 

 

 

 

 

 

ammo. m2 .mmlma .voauuME
.0uas3 .umn uo>w was ova; o>auum nucmumafi .m.D now mamamm uvmcauoam
.unuuan mo koala: omauo>4 Ioummd nmumuw wouacb Bumuucmnub mmma ho>uam m<w aoma cum unaccouh .ww
aoucmumafl.
ucsoamxoan,
annoy cannon non cone m>aua sham u> nova magnum auaum
Iaanu mo Honla: omnuu>4 Inauuuan maaommamavca tawny: .com N a¢ma maaomwcmavca coma omuonuaoo .mN
uaoaoum cannuan .uuco
voauuufl .mbauac .aunua obaumm manna .wmuaum «amine amao
.onucfi soap) ooo.a you mun Ioummd amuuaoo nuuoz sank: oema Iona nonnamu owma uaoammam .vm
ecuma coco: ooo.a yam «mma mo>usm
mun "mopusm “mvuma gnu, hmma .mom unouuao
23 21.3 333 8o; mom «>38 mm 33
m anon: couvaanu “cancoo Ioummd nouuum omuasb .mzm .cmnnb omma nancou owma uuooax .mm
aa>aav anon Hobo noun luau
Iaasu «ho>usm um¢lma :0303 aha»: a.oo :aouaaa «mob Ianuau can amma oamflnm
“on mac nouvaaso «nancoo Inauunan luau. oonuoccoa .mom aauoa emma hofihsm aauaou mmma {wwwamnu .NN
ucoaoun unanua: A.umal uaouuoa
ice ova: :33 ~95 :85 gauum vacuuoo luau w> .385 .950: no can
Iaano no Honlnfi omnugt Ioummfi amuaaomouuut .souluv 58.5 mmlumma ulna .893 no: mmma mahogaou .a~
ous-ual huaaauMOh nuance unaccounted .m .D wouannznau vouuaaaoo dune no cannon uanuaa anuubnufld
no nunulmom anon: macho anon Invm
canuaunb unaccomoo ouauoauom anyway»

 

 

 

 

 

 

 

 

vosaduBOOIIMH Danna

 

uni”

€35

.amoma .uwnfiw>ozv aa>xx .NaaEMh ofiu can UNMauumz no annuaob a.>uaaauumm new «Hausa Mo ceaumamm

.oml¢a¢ .hmoma .GGShV HHHA .3oa>0m oaeocoom :muauuafl =.>uaaauuom ou maumum oaaocoum mo coaumamm 0:9: .Gmavomum nmuonoo

.mmumv .Avwma

.mmlmm

 

.mVIowm .amomav Ha .Nnmmummsoo :.>uaaauuwm amaucmumwmaa can ccdoumxumm Sham: .nmucan .n mauo

.ma

.A¢oma

.mmmum ommoano mo >uamu0>aab

 

.mmoum ommuaso mo huamuu>acb

nommuanuv macaoaoom cmnub
mmwmusm .m :a =.ommuanollmaaomouuuz 5 Ga auaamuuoz can auaaauuom amaunoummuao ca mvawuag

.malmOm

«:9: .u::& .m maaaacmo

c

 

.uomfimm .2

 

mnu ca moocmumuuao auaaauuwm :mnub ca muouomm: COamuo> evocaunm omam mwm
.coaumuummmao .Q.£m nonmaanamcsv :moumum wouacb onu Ca moucmumuuao huaaauuwm cmnub Ca muouomm new mvwouaa .Hfinmmm .m :muaamm

.amoma

A

on mnoauanauucou .onmom .G can
.m can m3mmmuax .2 .mo

"ommuanov NmOaOaUOm cmnuD ou macauanauucoo .msmom .a can mmmmusm .m ca =.muumum cwuacb
.ommuano mo >uamu¢>acb .amOaanom no ucoauummon

 

 

 

 

 

 

 

 

 

Axuaaauuou ta
mwumameouv ucmmoum ocmn M
Ina: £ua3 .muco umauume
.¢¢Imm moon .coEo3 muaLS 0>au mwumum omuaCD mmmum . amamfimm amcoau
uwm anon uo>m cmuoaano Immoummd mo mmmu< omuacmnub nwuacmnub ooma m Inc may mamcoo mwma oucsx .am
.muaaauuou omuoam _
IEOOV awlm¢ 0mm .0wa3 o>aus ..xomm Gummy m A.m500 ooo.mv ho>
non anon um>m :muoaaco unauumaa mmumum couacb Eummlaummcoz Noma WIWuSm .mom ucmuufiu moma mcmunzn .0m
a
oumocmum .eelma c0203 _
caauumEIua>u ooo.a uwm
mnuuan "oumu .OOauumE
uv¢lma c0803 ooo.a umm m>aumm ouomnmm new
msuuam noun» .uuuw amuoa noumm< ommoaso :mQuD omlomma mumc uomuu msmcmo «oma memmuax .mm
Annoy mmma m>au
an unuuan wbaa uo umnssz Ianauuman muumum vmuaca swan: mmma oamamm >m>usm m<u moma mmweumuum .mm
_. m0:
oumu :0a00500ummu muoum o>au " Guam: omwuano Mom
canauumucmunlommauumz Isnauumaa mmu< .ouuuz omwoaso :mnuD omma p wumc uumuu unncou Noma Mancunm .hN
unsmmoz huaaauuom :mauoo coucamoumom .m .D vmnammnmEm umuomaaoofl‘ name no ooufiom conuaa anvuonunfl
mo nucufimmm amwu< macho mama . Inflm
mannaum> unaccomon mucooanom amvummw H Mama

 

 

 

 

 

 

 

coacaucoouuma manna

96

General Characteristics of Studies

It is to be emphasized that the purpose of the elim—
ination of studies from review was to focus concern upon
studies most relevant to our problem. In spite of the
reduction of studies, there remains a considerable range in
certain characteristics of these studies. For example, pub—
lication dates range from 1930 to 1965 with some tendency to
be proportionately more in the post-war period. It is felt
that the more current studies based on contemporary data are
more relevant to the present study based on 1960 census date.
The range of the periods in which data were collected is
perhaps of more importance than publication date. Eight
studies are based on data collected before 1940, nine in the
period 1940-49, eight within the 1950-54 period, four during
1955-59, and only two based on data collected since 1960.
There are, of course, other studies of fertility based on
data collected since 1960, but most of these have not met
the criteria requirements of this review. If nothing more
such information suggests the need for a more "up—to—date"
differential fertility study based on the 1960 census.

With respect to variations in the source of data, of
the thirty-one selected studies, 21 are analyses of data pro—
vided by census. However, three studies (Table 13:1,2,24)
employed census data obtained from special samples. Six
studies analyzed data from the intercensal Current Popula-
tion Surveys (Table 13:12,l4,18,l9,23,30). Of the ten inves—

tigations not based on census data, nine are specially

97

conducted sample surveys (Table l3:5,7,8,10,20,21,25,26,28)
and one is based on collected birth records (Table 13:17).
Since the present study is also an analysis of census data,
it is favorable that a majority of the review investigations
are census analyses.

Because of the nature of the sources of data avail-
able for differential fertility analysis, several overlap—
pings occur in terms of the data employed. Two studies
(Table 13:1,2) are based on the same special sample obtained
from the 1910 census. The Indianapolis Study contributes
three studies in this review (Table l3:8,lO,25). Several
studies analyzed data from the special census publications
of 1940 and 1950 dealing with differential fertility (Table
13:12,13,l4,l6,18,19,23,24). The appearance of the results
of the Current Population Surveys 1947, 1949, 1952, 1957 and
1962, contributed to a number of analyses (Table 13:12,14,18,
19,23,30). Three studies were outcomes of the Growth of
American Families Study of 1955 (Table 13:20,26,28) and two
analyses employed Chicago census tract data of 1950 (Table 13:
27,29). In spite of such "overlappings" studies based on the
same source of data tended to vary with respect to the
aspects of the data emphasized as well as other characteris-
tics of the study. For example, there is a wide range of
difference as to the areal segment of the United States rep—
resented by the data analyzed. The two most common areal
unites in fertility analysis, however, are the national

level and a city or metropolitan area level.

 

 

.....—

98

Finally, these studies vary with respect to the
measure of fertility employed. Twenty of the 31 analyses
endeavored to explain variation of children ever born or a
cumulative birth rate measure, comparable to that employed
in the present study. Among these studies, however, there
is a vast difference in the details of the particular fer—
tility measure. For example, some qualifications applied
to cumulative fertility measure were "once married women,"
"husband present," variations in child—bearing age span,
native-white women, standardization, etc. The second most
frequently employed measure was the fertility ratio, at

least some variation of it.

Study Design of Empirical Studies

The distinction of whether a study is "descriptive"
or "causal" is sometimes a very thin line. Essentially if
a study was designed to test specific hypotheses, it was
considered "causal." Nine of the 31 studies were described
as such (Table 13:15,21,22,25,26,27,28,30,31), and all of
these have been published almost within the last decade.

Another distinction on study design is relevant: ag—
gregative vs. distributive studies. There were nine studies
employing a distributive approach to the analysis of fertil—
ity variation (Table 13:4,6,9,11,22,25,27,28,30). These
studies generally applied correlation measures, simple, par-
tial, and multiple, to the analysis of fertility. Interest—

ingly there is not a one-to—one correlation of causal design

99

and distributive design. The early distributive studies
were largely descriptive (Table 13:4,6,9,ll). If these two
aspects of study design are correlated among the studies
selected, there remain only five studies which apply both a
causal design and a distributive framework (Table 13,22,25,
27,28,30) and of these only two employ an areal distributive
approach (Table 13:22,27). These two analyses of fertility
differentials, while they employ a design similar to that of
the present study, are extremely limited as to the territo-
rial scope they represent. Hughes concentrated on Tennes—
see's total population and rural farm population, employing
the county as his unit of analysis; Hashmi focuses on the
Chicago Metropolitan area, with the census tract as the unit
of analysis. While Hashmi claims to explain "urban fertil—
ity differences in the United States," generalization from
Chicago to the urban sector of the nation is quite tenuous.
In view of these facts, an accurate assessment of the next
step in the study of differential fertility should employ
both a causal and areal distributive approach on a territo—
rial level much more extensive than the state of metropol—
itan area. As has been stated, the present study will in-
volve national and divisional analyses for residential
categories.

Since this study focuses on fertility within residen-
tial sectors, a discussion of how the review studies treat

the residence variable is appropriate. While studies were

100

chosen partly on the basis of their contribution to the
understanding of urban and rural fertility, there have
appeared very few distributive studies which compare urban
and rural fertility. There is a definite tendency among the
studies selected to emphasize the investigation of urban fer-
tility (Table l3:2,4,5,7,8,10,12,17,20,21,23,24,25,26,27,28,
29,31) and this is fairly representative of fertility studies
in general. If the studies which equally emphasize urban and
rural fertility are added (Table 13:1,3,13,14,16,18,19,30),
there are 26 of the selected 31 studies which contribute to
the understanding of urban fertility. Calculating in the
same way, there are 18 studies which provide implications
toward the understanding of rural fertility, although studies
focusing on the rural components alone are relatively few
(Table 13:6,9,11,15,22). It is unfortunate, however, that

so many of these review studies in employing residential
categories apply the aggregative approach rather than the
distributive. As a result, most of the 31 review studies
treat the residential populations as being homogeneous with
respect to the variables related with a fertility measure.
Relatively few studies provide insights as to the relation—
ships one might find assuming variation within each residen-
tial group and how they might differ in this respect. In
spite of this deficiency, it is hoped that the findings and
insights of the review studies will provide enough informa-
tion on the nature of comparative variations within residen—

tial groups to make proposition construction possible.

101

Summary of Findings, by Variable

With the assistance of Table 14 a systematic review
of the relevant findings is made more complete. Each of the
general independent variables is discussed under separate
sections. Upon a quick glance at Table 14 it can be seen
that the three primary measure of socio—economic status have
been most frequently employed in the 31 selected analyses of
factors affecting fertility. Education has been included in
21 of the 31 studies, occupation is second appearing in 20
studies, and income runs close behind, appearing in 18 of
the studies. The remaining three variables have appeared in
sporadic fashion. Both age distribution of women and female
employment have been employed in six of our 31 fertility
studies; measures of distance have appeared in only five of
the studies represented in this review.

It must be mentioned that equal weight will not be
granted to all studies reviewed. Studies which have inves—
tigated fertility differences within residence groups or
have study designs comparable to this study, e.g., causal
and distributive, will weigh heavily in the development of
propositions. These studies tend to be those most current,
such as, studies by Hashmi, Hughes, Goldberg, and Duncan, to
name a few.

Let us now consider independently the pattern of
association between fertility and the independent variables
indicated in Table 14: distance, occupation, education,

income, female employment, and age composition.

102

.mmuom .mm .ma manna ca uouau mm «Emu ous nousououom;

 

oau¢>ca a“
.m.z m .uuoum
wa no ca
uuouao auao
Baum . 9.5

m0 condunao

coaxumn a
caumaoza .o

 

Ammauomemo

m macov nodumwou
omum>Ca " aasmm

mmmau .mmooo vmoun

sumo ca camumwwu

mmum>ca "newsman

uomamao
.uaco canva3 so>o
.coaumaou unu0>sa

:aouwouoz .m

 

.mom a ..m.z N
.mmuo>ca noauau
m "enuow own
cvuma amen: x

rom.- 0» ov.u mama“
m.u. .uuomsa umoe
“muuo>Ca "00>OaQEm
+oa c0503 mo x

laws. on em.

omcmu u.uv :Oaumaou
o>auanom «.mme :a
+oa 0mm mnemuom NO x

sounaons .e

 

 

ummwu

90: has nan

uoouav unoao
hm .QOaunaou
uoouav awash
a .mom .80»

«.00 no uoau
an ousmunan

DAD! Q
Hflflgkm . n

 

ouuo>¢a mnouum
.umaoum .mnooo vaoun
m new “manncoauaa

 

 

 

 

 

 

 

 

 

 

 

o>aucousa Hon couooaou moacsun avaaauuou

so comma uoanaauu> unocaoaousa ou huaaauuou no amaaas0auuaou no oaauu finallflm

Iou uaoao o: .omnooo aaflaamm a
uauaoomu aam uo>o saouuouoz .N
canny can» um30a
.nmauoo auuauaao
Iauma .onuo>:a msouuo :aouuouoz
"aauum .amush can» can
Homsouuo .aoaunaou noxoauuu
ouHo>=a macho. «cankD unocmm .a
scauaaomaoo cud unoEaOamem oamaoh oaoona :Oaunuscu abaummuooo ouauunaa auvuonund
¢3Da>ou

.Qa OHAQB

.l()3

 

   

ceaum 0n 0mu0>ca
" aEmw mo 0E00
-aa Mecca mama amuos

0mu0>ca Lyon

 

muouonma Ehmw can
mumEumu new 0mu0>ca
pan mascum .msouo
aam macaw GOaumaou

 

 

 

 

 

 

 

 

 

 

 

 

hcma a ocma "mua3 0c "mvma a h¢ma {F nomax .Na
ouam.
.aaa> m anon nom1
1 aouusoo :05:
00cm». .uuomfia no:
IuomBa manamaamocm “muasmnua H030a
mo @mimu 0km 0:3. ..uuow uo3ba
vclma c0503 $_ :Oaumamn uuouaa. :mocsa .aa
M .m.z mumsuo 1
1 .0mu0>ca maco 0wa3 GOauMaou o:
"nonsmam “sman .uuou .mmamsou coccman
.aa00 :uon :0n3 um00 macaw «uuouac 1
m masm Ixo 0mm0>sa «mausou cnm3ou uu0xomun
IIIIPII 1
1 ICOaumaou o: upmcsmam .Dmss can» ummcouum mo» ca NGOMumwo
. wa0>0a ummm: um00x0 .0mu0>:a “ova: Ixu van .0mu0>:a 1 couma0£3
1 0mu0>§a "camnmsm. 0mu0>Ca "unmnmam cuwzbu mucoccmb 1 a uuuax .oa
11‘ ummao uom4.m00.
d _ .nuu abuu 0>OE
unmoauacmamw .cmamlco: .0mu0>:a 0mu0>¢a ".00 no mm aamu mmumu
Icon nm~lmm 05m” "mmma .0EOUCa Sham uumm mm new .9800 .uuom um030a umwmnom w
«elma coao3_fim muammo hum 0mmu0>< .mum .num guano: 1 mmauao .ou0021umasvc0mom .m
4‘ .m9: can» ucmuuo Ba 1 1
1 once .0900 n.0ma3 1
"mmmmmm .ma0>0a h
.aaou uOu HmmHo 00: 1
van 0mu0>ca "camnmsm 1
won an s0>0 .0mu0> . ummax a
Isa haum0a0 ”omas 1 nonmaons .m
3H
swan: new mucouuomaa ca H030a amazoa swan: «nonma
.um> ucmuuomaa unOE 09cm mzbnn .3»: Mom gnaw cmnu .uumm M0
uaooca «scam unmao unwaUImuouu “umoHu |3Oa nuofiumu ..mmsuoo
Imu0u0 ”cuaammm uo: 0mu0>sa "amunu summco: new 0mu0>
awash “mommmau humus “haco .uuac aamfin Isa «amusm «ceaumauu
:a aamfim nan .ouu0> .m.m can .HHOU pan noouac nonunao mo» :
Isa swans “NaaEmm ouu0>Ca sang: nouas van 0nu0>sa "cmnub uanax .n
scauaaomaou 0m¢ 0:03a0amam oanalm oaouca ceaumosom scaummnooo 00cmuman anvuonu=<

 

 

 

 

 

 

 

 

nossaucoonnea canoe

1(14

 

cuouum

ua0H0 on "amma 0200
a 5

 

   
 

 

Eumm now swan
a mcoaummsuoo Eumuco:

 

 

 

 

 

  
  

 

 

 

 

 

 

 

 

 

 

 

 

I am mam .0mu0>: 0mu0> macaw .uuou 30a ammo
umvma oaooca aaEmm Isa msouun "ouaz uxo coaumamu Hm0a0 oz aaanmuo .ma
0mu0>ca mm can mzm
1 "m 0mquca aaaEmu .mEa amsvm mo .05u0 can» :mnus ca 0wu0>ca
1 swans .muo .uwmo Eumm 0mu0>Ca w .maooo unmulmmouu who: .mmsuoo 00a>u0m
1 .0uOE xuo3 c0803 .mv u0>o c0503 "HmmH “m.©:mnm5£ can» a .u0a0 qu ceauamo
1 .m H0>O noun can ms: mo 0&OUCa .uuomfia OHOE n.0mm3 Ixm .muufiumw cmcu
1 Iaanu c053 100u0u osoa amuoe nomuw>c “0mu0>:a "panama: .uumm u0£mac .uonma caumaonz
1 nonma ca :0503 mumnm "mvma IHHEmm1 museum .mmu aam ea Eumw .mQSOHm 00:0ca w uwmax
1 now u030a .uumm mmmuca amcos amuoa1 0mu0>ca ummau "mmaz Imou aam ca 0mu0>ca .aaanmuu .ma
w J 0mu0>
1 Ina 00a£3|0>aumc
. . “0mu0>ca .mom amuoa camumcaou .ma
. ceaummmmuo sun»
1 ucmquQEa whoa n.uu0w
_ cu mmcmnu 8 acaumauu muwuonma Emma can»
0>auamoa no ouamm .uuuw ~03Oa uweumw
:a .mQSOum .mmu “museum .wmu aam ca
1 aam .momm aam .GOau,am0au0a0 uo coaumwoxm
_ Imamu 0mu0>¢a "mwaz :ua3 ceaumawu 0mu0>ca nomax .ma
_ toaaaxmcs can»
. .uuwu u030a .uonma
Emma .cOaaaxm .uuom
“030a Hoeumu "acoum
luoan oasocooo msba
w IaOu acaumaou 0mu0>ca Smmm .ma
1r muouonma
.mmmau macaw .muwac shun can» .uumw
mo macauomuucou mo uoaba uafihmw unmsuoo
manm :a scaumaou amusuasoaumm “acoau
0mu0>sa um0a0 303m Immoxm nua3 .mmma
h¢ma m oema "0ma3 .ovma .oama onu0>sa uuoummz .va
acaumaau oouo>ca Ȣ
00 conmmaaoo museum
cocaoaa .0231 «.2308 .umm How
.cva o: .GOaumaou nacho ucuuuaucoo a:
on nova) «0 00¢ .:0aana0u unoao 0:
now ooaaouusou museum .msouo onaz auxcan .ma
coauanoaaou om< ucoahoamau cacaom oaoosa scaumoscm soaumnuuuo ouanuuao auvuonuad

 

 

 

 

 

 

 

 

 

coucaucoou-ea canny

105

 

 

ucmuauacmam

Ir
0:0 0mu0>0a "0ma3

mucmumaa can

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“mucmumae can scan: swans msoc0ma©sa now u0mca00am
msoc0macca Ham 00a» a0>0a 0m0aa00 Ham 0:0
Imaou uoouac "panama: .0x0 .0mu0>:a u0ua3 amavooum .om
. ma£000a00a0u u0mcouun
1 mucmumaa .00u0>:a
1 0mu0> "0wa3 "vacuum mm 00:
_ Ina mucmumae ..m.z pan 00am m0uacmnus
m0uacmnu0 .0mu0>0a .0mu0>0a mucmumma
1 m0ad000 aam ”panama: .m0am000 aam “camnnsm mu0nwa00 .ma
. .aa00
m 0>auamoa van 00H0>Ga
. IILILII.
. ma0>0a u0mms .0x0 "ccmnmsm «a0>0a .HHOU ma0>0a u0mms
.0mu0>ca Hosanna: .0x0 0mu0>sa "0.0ua3 .0000X0 0mu0>ca 00amm0m .vm
. mmma 00:» Nmma 00:» 0mu0>
1 0HOE 0wu0>ca hmma Isa 0u05 hmma nu00a0
1 0mu0>0a umoao 100mm aam 0mu0>ca mwo no: mmo nm¢lm~ 0mm
_ smma u08.3 o>onm 1o¢-m~ uoouao .emuma uoouao .vaama own
_ uoouac uQ0UX0 0mu0> 0mm 0mu0>0a caumu 0mu0>sa.om a ovma
-aa mmma "panamam .uumw amps: u0ua3 oaumu .uu0m gonna nouax .mm
0mu0>ca 000 m
o>au mo anon :04 mo mums amazon: ”mm
..m.z ups .ma: "mm 1m «o .uuoou upon
00009 m we cam ..smam um0£ma£ ..mom amuOu
can 0>au 000 .mom ca 0mu0>0a .+mm 0mm
amuOu "osoosa aaamm new cmac0a ”00am: nonmam .NN
.mucmumaa
munmumaa 00am “.m.z Show 000 mouacmnun
00u0>0a .m00acmnu0 zuon. .00000 no 0800 nucmumaa Sham no
h.w.z 009 .0mu0> Isa :05» ucmuuomaa n0uacmnu0 now 00a»
Isa moansoo aam "000m 0HOE .0mu0>:a "ova: Imaou unnuauanmau oz muonuaou .aN
0wa3 .ououonma Shaw 00:0
00:0 unnuuomaa 000a .uuom u030a Hoaumu
00>a3 mcaxuo3 00 "vsmnupm .aoeouca “finance: .n> Shaw aa0n
ucmuuomfia 000E 000 .0mu0>:a "Naaamm .nmsc mo aouucouv 0nu0>ca vuoumoum Imamu a
mo 000 ”ouu0>sa unmUaMacmamco: 00m 0600 .m.m a .HHoo “umaoum Shannon .couma0zz
"0wa3 mo usoaaoamau. 0>auamom "cannnsm pan 0mu0>ca "ouaz macaw .uuac 0avuaa .005000uh .om
scauauomeoo out 000:50amafi camsfim osouaa coaumosvu :Oauamauuo anchorage

 

 

 

 

 

oocuuuao

 

cuscauGOUIuva canoe

 

 

{I‘ll

 

 

l()6

 

0m0auume 00 0mm

0 .mauuo ..0500 now
aouucoo 00:3 .a0u
.000 mcouu0 "+NN um
00:3 .uom .a~u¢a um
voauumfi 0ma3 0003

IPI
0000>0a "0000000

+N~ um ooauumfi 00:3
.000 .auuea 00mm um
voauuna 00m) 00m”

0000>0a ”000A0sm

00uuaao

.0m5000 .000 "+NN
0mm 0003 000%: “am
Ica 00mm 00 00auudl
00a: 0003 0000>0H

 

.05 :2: .233
“300N000 00w
0000a .0000ca00u
Sham 0a 000m0ouum
.00auma00 00u0>0m
"0ma3 no .0000 ha

0ua3 00
0000000000 00 000
050 0000000 0000>
I0a 0500 «mammmmm
.hz Mow 0000:00uum
“00noumxuan Show 000
.000 Shaw 000 00u0>
I0a 0000000000 "00am

 

.om

 

00a»

Ia000500 amuaumfi uOu
mcaaaouucoo uuouap
.0000 amuauma “0000>
I0a 03000 0000 .0000

amuou ..u0a 00a005 .

mo 00000a0 "Naaamm

 

 

0mmauumfi mo 00%
maumo 000020
.uuomfia .a0u 000
“.000> va uo aa
.0H» +oa coauuua
«0a u0 0009 :00
.00» mum c0auuma
"0mm 0.0ua3

 

.a0u 0000>0a
h00a00a00> Va 0a
sum “0 0a 0003 sum
«00> +oa v0auuafl
"00a00a00> 0a 0a
sum ”0 0a 0003 can
.0u> mum c0auuma
"Oua3 HO ufluwuu fig

0a 00
sum .00» +oa rayon
:uoa .00» mum "mam

tha000 .00> ea 0a «w
00 upon :00 .aa» +oa
no no 0003 vnm .0u>

 

mum 0a “0000>0a "0ma3

“anon :04 .aa» +oa
.ouon numa .ma» mum
"0a0ma000 .um> 0a 0a
"00000 0 0a .0.2 .00»
mum 00auuma "0000000

 

00000 0000
0a .uuomaa hu0> 000
“anon nu¢a an» +oa
1: 5 3E
..0u> mum ”000n000
“muon nuaa .nua +oa
.0a 0a 000A 000 .00»
mum 00a0000 «00a;

 

.mN

 

0““
.a00 000 “70000
a .3000a .00000.

..0000 umgman
.0msoum 0mm

“code a 00:00:

.uoouav "000300

a000aunu0000

.uuomaa

.a0u nun ".0000

000 03000a 00000
I000 mcavaong 0000»
I0a mcouua «combam
IE0 +ma 00in: R

00000000Ia .a0u
00 qua a.0500 00000
neon unavaoso coda

Inaou uoouav ..uu
8.85 x .1050

00a30

Iaum> aHMINb .uuom
Iaa .aou 000. “A0I00
I0a 00000000 mCaUaon
00>0. 000023 000.130
.vouoamloo .m00

+m~ 00am! 8 «00am!

000

Inuuomfla .a0u 0auuaa
«n.0000 000 .00a
00000000 unauaoao
.m.z .00a0 «uoauau
.bhoa 0.0nua % «onuob
10a .0uu000 w .uoum
R.»000auaaou canal

N058

000000

000000

000

030mmaax
.

00-00000
.uuomla
.aou sum.
«.0060 a 0I001
.aa 000 aouu.
I000 £00.30.
huao «0 000000.

luau 000000a0 al§000

F

.hN

 

00aua0omsoo om:

 

u00lh0amafl OaQIUh

 

OEOUGH

 

coauauaou

 

00a0005000

 

1

000000an lbwuvuozuud

 

 

voucauaoouuva manna

107

Distance.—-Studies which have related fertility to
distance from urban areas are few, and even fewer from met—
ropolitan centers. Thus, most of these studies place their
findings in the realm of "urban influence." Nevertheless,
these analyses of fertility have been highly descriptive,
for want of a theoretical framework and a better comprehen-
sion of just what distance does measure. Often "urban
dominance theory" is only an implicit theoretical framework
underlying such studies. Let us review what the empirical
studies of fertility declare overall to be the relationship
between fertility and distance.

In Table 14 there are two early studies which employ
data for no later than 1930. They have obtained fairly con-
sistent results. Brunner and Kolb (Table 14:3) investigated
fertility levels by tiers of counties from 18 large urban
centers (all but three having populations over 50,000) and
found a relatively clear positive association when consider—
ing total population and total rural population by county
tiers. The contrast of rural-farm and rural-nonfarm by
distance found some exceptions to the positive association
of fertility and distance, but the pattern persisted.
Thompson and Jackson (Table 14:6) focused on tiers of town-
ships rather than counties and considered several other
variables to denote economic conditions of rural communities.
Distance and a measure of the percent of the rural popula—

tion dwelling on farms were used as indices of the degree of

108

isolation of a rural community from the influence of cities.
Although exceptions appeared, the predominant relationship
was a positive one. Thompson and Jackson, interestingly,
noted the possibility that distance was merely reflecting
variations in general economic status of communities and,

in turn, economic conditions measured to some degree the
isolation of rural communities.

Rosenquist and Schafft (Table 14:9) considered rural—
farm parts of counties in Texas with respect to fertility.
There was a tendency toward a positive relation, but it was
not necessarily a clear, linear one. Duncan (Table 14:11)
considered the location of villages (indicative of the rural-
nonfarm population) in Pennsylvania with respect to metropol-
itan and urban centers in non—metropolitan counties. The
location of villages was positively related with respect to
fertility levels, but the effects of location were reduced
to negligible proportions when rent and village size were
held constant. As did Thompson and Jackson, Duncan's
results hint the possibility of location actually reflecting
differences in socio—economic status. Finally, Hashmi also
found a positive association among census tracts of the
Chicago metropolitan area, but this case does not pertain to
our immediate problem, since his results focus on the pat—
tern within a metropolitan center rather than between the

metropolitan center and the hinterland counties.

109

Two possibilities appear in terms of interpretation
of what distance measures. On the one hand, it may merely
be a "conglomerate" variable which subsumes several latent
factors highly correlated with distance. On the other hand,
the conventional procedure is to couch the explanation in
terms of metropolitan dominance, rural isolation, urban in—
fluence and the like. It is understood that influence wanes
with increasing distance from the centers of influence. Two
variations occur on this explanation and they take the form
of the "homogeneous area" vs. "nodal area" issue discussed
by Bogue and touched upon in the first chapter of this
thesis of non-random distribution of certain attributes in
space.8 The "homogeneous" region stresses the internal
homogeneity of delineated areas in contrast to comparable
regions. The metropolitan community, or "nodal" region,
stresses non—random distribution of attributes within the
area along a continuum of distance from the central city.
These two positions may be merged if homogeneity is under-
stood to refer to the spread of dominance within the region,
a notion directly opposed to the concept "urban influence"
which is viewed as diminishing or "trailing off" with in—

creasing distance. As a result, the entire metropolitan

 

8Donald J. Bogue, "Population Distribution," in P.
Hauser, and O. D. Duncan, The Study of ngulation (Chicago:
University of Chicago, 1959), pp. 393-394. Discussion of
this issue appears in this thesis on pp. 67-69.

110

region is considered dominated by the metropolitan center
with equal force. Concern should not be so much for how far
does the influence of the center extend but what is the
impact of the centers influence on the structure of the
region. Metropolitan regions may differ, then, in the exten—
siveness of dominance reflected by size of the metropolitan
center, but the non—random distribution of attributes within
the metropolitan region results from the prevading dominance
within both the urban and rural hinterlands. This last
ntion, in turn, absorbs the earlier comment that distance
reflects a number of latent factors. On this basis, it is
expected that fertility, as well as many other attributes of
rural and urban communities within the metropolitan regions,
will reflect a non-random pattern of distribution.

In view of the consistent positive relation between
fertility and distance apparent in the empirical studies

discussed above, a strong positive relationship_is also

 

expected in our analysis; i.e., with increasing distance

 

from a metropolitan center, an increase in fertility within

 

all residence groups of the hinterland. Furthermore, a mea—

 

sure to differentiate size of metropolitan center must be
applied to the distance measure. Finally, considering the
necessity of subordinating most attributes revealing non-
random distributive patterns in the metropolitan region to
dominance exerted by metropolitan centers, the distance mea—
sure as a "conglomerate" index must receive primary impor—

tance among all variables considered, again, for all

111

residence groups. On the other hand, as some of the empir—
ical studies indicate it may very well occur that the rela—
tive importance measure is reduced considerably when other
variables are held constant, such as, education, income, etc.
In such a case we would conclude that distance reflects dif-
ferences in socio—economic status and, therefore, has no
predictive value by itself. Hence, there remains the

dilemma as to the relative importance of distance in predict—
ing fertility levels. There must also be asserted, then, a
contradictory proposition, that distance may also reflect a
slight positive association with fertility and low relative
importance in accounting for fertility variation in both
urban and rural areas. This conclusion suggests the need

for further research.

Occupation.—-Occupational data have been available

 

through census much longer than any other criterion of socio-
economic status. This explains why the early studies of
differential fertility, for the most part, selected occupa—
tional measures to relate with fertility (Table 14:1,2).
Occupation was employed in these early studies as a measure
of "social class" or "social status." Since that time,
additional measures have been developed as indices of social
status (education, income, monthly rent, etc.) to complement
and at times replace an occupational measure.

Grabill, Kiser, and Whelpton have noted some inher-

ent weaknesses in the use of occupational group of the

 

1“

112

husband as an index of socio-economic status:9 (1) wives
do not describe their husbands' occupations with sufficient
accuracy; (2) unlike educational attainment (but like income
status), occupation is subject to change, and a previous
occupation may be in some instances more influential with
respect to fertility than the present; (3) whereas educa-
tional status and income are quantitative and continuous,
occupational groups are more qualitative and discrete; (4)
mxrupational group of husband, by definition cannot be
applied.to single women. In addition, the broad occupational
groups themselves contain a wide variety of specific occupa-
tions and a wide range of social and economic gradation. It
could be argued that, compared to the early studies, at
Fussent this variation within each of the broad occupational
9Toups has widened because of the continuing diversification
°f<xxrupationa1 classifications. For these reasons occupa—
tiohis probably a poor indicator of socio-economic status.
Whereas Sydenstricker and Notestein (Table 14:1)
fOUHd iflor 1910 a clear inverse relation for broad occupa-
tional qgroups. Notestein and Sallume (Table 14:2), by in—
vestigating the variation of specific occupations constitut-
ing the broad occupational groups of urban areas, found no
(”ear—(nit relationship and much overlapping of variance.

\\_—

‘9
The Fe .W. H. Grabill, C. V. Kiser, and P. K. Whelpton,
1958) rill—ity of American Women (New York: John Wiley,
' . 116.

113

The relationship of fertility to broad occupational groups
was persistently shown to be inverse (Table 14:1,2,4,5)
until data from the 1930's began revealing a trend toward a
direct relationship with fertility, at least among the upper
classes of urban populations (Table 14:7). Other studies
appeared which concurred with this finding, i.e., the estab-
lishment of an approximate "J“ curve among the broad occupa—
tional groups (Table 14:10,14,16,18,24). It is interesting,
however, that Kiser compared occupational data for 1952 and
1957 and discovered an indication of a reversal of this "J"
curve pattern toward a stronger inverse relation.10 These
data, however, do not indicate a clear inverse relation nor
are residential categories presented separately.

On the basis of the studies represented in Table 14,
ii must be concluded that there is no clear-cut pattern of
relationship between fertility and occupational group,
whether this be due to the occupational measure employed or
the essence of the relationship itself. Four of the 20
Studies; considering the relationship found a clear inverse
relationship, seven reported an inverse relationship with
‘fistortxions in the upper brackets, and nine indicated no

Slgnifszant or inconsistent patterns of relationship.

\—
l . . . . . . .
Stat 0C. V. Kiser, "Differential Fertility in the United
egg," in National Bureau of Economic Research, Demographic

 

an
BFEEE¥ESEIEEEiC Change in Developed Countries (Princeton:
eton University Press, 1960), p. 104.

114

Most studies of occupational groups and fertility

have concentrated on urban areas. It can be concluded that

this relationship in the rural components, especially the

rural-farm, is a much more stable one. Sydenstricker and

Notestein found a strong inverse relation among nonagricul—

tural occupations in the rural areas. It appears, however,

that the most persistent patterns within a residence group
is that of higher fertility rates among farm laborers and
lower for farmers, usually indicated only for the rural
mmnponent (Table 14:1,7,12,14,15,l6,18,20). Dinkel, who
analyzed 1940 and 1910 census data by regions, was the only

dissenter. He concluded that the rural occupational groups

(farm owners and farm laborers) maintained no consistent

order in their relative positions. In 1940 farm laboers

e‘Xceeded farm owners in fertility only in the Western region
“4th'the other regions indicating mixed results among ten
age grwaups.ll In contrast, Westoff found that over the
erioél 1910-1950 fertility had continued to increase among
farm laborers and foremen and had decreased among farmers

and farnn managers. Differential fertility by occupation and
ciontra<::ted but the contraction was due wholly to change with—
in the Inore exclusively urban occupations.12 It appears

.1_‘______‘______

l . . . . .

Uhit lRobert M- Dinkel, "Occupation and Fertility in the
19 eCi 81:3t63." American Sociological Review, XVII (April,

52) ’ 181.

l
2C. F. Westoff, "Differential Fertility in the U.S.:

19101
659- 1495(3." American Sociological Review, XIX (October, 1954),

k

115

that among the agricultural occupations the trend is in the
opposite direction than among urban occupations, i.e., in
reverse of the "J" curve direction. This suggests the
inverse relation among agricultural occupations may continue
for some time.

Two problems should be noted with respect to occupa-
tional group within residence categories which may affect
some of these relationships mentioned above. Occupational
groups by definition separate the urban and rural-farm popu—
lations (agricultural vs. nonagricultural), whereas educa—
tnon.and income do not directly do so. Rural people are
more heavily weighted in the lower educational and income
groups and, therefore, tend to enhance an inverse relation
between these variables and fertility. On this basis, a
Strong inverse relationship is not as probable along the
ocrcupational dimension. Moreover, a look at fertility dif-
fenau:ials within residential groups would tend to reduce
theSe eaffects, although Goldberg finds that even within the
urban Chumponent there is the rural migrant which tends to
dj-Stort: the real pattern of socio—economic differentials
Within the urban population.13

Iknother problem is the continuing decline of propor—
tions (Df? farmers and farm laborers reflected in the

\—

.1
(”bari‘ 3David Goldberg, "The Fertility of Two-Generation
lteS," Population Studies, XII (March, 1959), 214—222.

116

diminishing proportion of wives of farmers and farm laborers.
Grabill, Kiser, and Whelpton remark that

marked changes took place during 1940 and 1950 in the
distribution of women of childbearing age according to
major occupation group of the husband. Most conspic-
uous . . . has been the reduction in the proportion of
farmers and farm managers. In 1940, among native white
women of child-bearing age, married once and husbands
present, the proportion classified as wives of farmers
and farm managers extended from about 12 percent at ages
20 to 24 and 25 to 29 to 22.5 percent at ages 45 to 49.
In 1950, the range for total whites was from 8 percent
at ages 20 to 24 to 12.5 percent at ages 45 to 49. A
decline in the proportion of women whose husbands were
farm laborers is found for the United States as a whole.
Thus, among native white married women, the proportion
classified as wives of farm laborers extended from 13.2
percent for wives 15 to 19 years old to 1.6 percent for
wives 45 to 49 years old. In 1950, the comparable fig-
ures for white wives were 8.0 and 1.5 percent.

These same data reflect also the different age distributions
of wives of farmers and farm managers vis—a-vis farm laborers
and foremen, which consequently may influence fertility dif-
3fiHEnces between agricultural occupations.

The previous discussion in many cases points to
factorms which would tend to compound the relationship of
occupational group with fertility and, consequently, to
reducerthe relative importance of occupational group in
detiermining fertility levels as measured by a statistical
relationship. Moreover, several studies have commented on
the reliitive importance of occupational measures in deter-

mining fertility levels. Among urban occupations Notestein
\M—

.14 . .

QELaEEEE, W; H. Grabill, C. V. Kiser, and P. K. Whelpton,

pp. 118-19.

117

and Sallume (Table 14:2) found a wide range of variation in

fertility left unexplained by specific occupations. Kiser

found occupation less important than income in explaining

urban fertility differences (Table 14:7) and occupation less

important than educational attainment (Table 14:16). Freed—

man, on the other hand, concluded that occupation and income

were of equal importance in both the urban and farm catego—

ries (Table 14:20). Hashmi, applying the distributive

approach to Chicago, concluded that occupational group was

of numb.less importance than income and education in influ-

encing urban fertility levels.

In conclusion, depending upon the measure of occupa-

tion employed, it would seem that its relationship_with fer—

Eility could range from a slight inverse or a non-signifi—

EEEE to a moderately positive relationship. Furthermore, it

Wouhi be expected not to rank relatively high in determining
fbrtilrity levels in any residential category, although a

StNNK; inverse relationship is more likely to appear in the

rurEll—farm sector than in the urban. Furthermore, employing

Separatxe measures for two broad agricultural occupational
groups ‘Nould produce results of an inverse relation with
fertility for farm owners and farm managers and a direct
relaticni for farm foremen and laborers. Such measures being
more relevant to the rural—farm sector, would be of greater

rel - . .
atlve importance in the rural areas than urban. These

twO .
a“gr1—<1'ultura1 occupation groups constitute 68 percent of

...

118

the employed rural-farm labor force, 8 percent of the rural—

and only 1 percent of the urban labor

that

nonfarm labor force,

force. One should recognize the possibility, however,

employment in agricultural occupations is an index of off-
farm occupation employment and therefore may be a better
reflector of the extent of urbanization of a population

rather than the usual assumption that it reflects socio—

economic status. In view of this possibility, agricultural

employment may be of greater relative importance than sug-

gested by the empirical studies, although again more so for

rural than urban areas.

Education.—-The availability of educational attain-

ment data in the census is relatively recent compared with

occupational information. Only in 1940 was an illiteracy

Question replaced by educational attainment. Before 1940

edumational attainment information related with fertility

couhi only be obtained via a sample survey (Table 14:5,7,8.

10), lflowever, the popularity of educational attainment in

fertilrtty studies has increased and in many cases it has
been'qnlite successful as an explanatory variable. Among the
31 Selected empirical studies, 21 include a measure of educa-
tiQn ft”? either the wife, the husband, or the couple. In
Contrast to occupation these studies have tended to appear
in the more current studies.

fPhere are perhaps certain inherent characteristics

in t
he education measure which make it attractive and more

read '
J‘lY Correlated with fertility. In the first place, the

‘

9“

..u

 

h-

119

"highest grade of school completed" is a fairly stable
attribute after a certain age is reached. Unlike occupation,
then, fertility is more likely to take place while the woman
is of the same educational class. Second, and in the same
breath, education acts in many cases as a direct deterrent
to fertility in that marriage is generally postponed for the
sake of education. Of course, recent statistics on propor—
tions marrying by educational level indicate a sizeable in-
anease within the college level, thus slightly reducing the
detemrent effect of educational attainment in relation to
fertility. A third attribute is that educational attainment
is more frequently measured for the wife and, therefore,
more relevant to fertility levels measured also for the wife.
However, interest has been directed in more recent studies
t0 mnployment and income measures for the wife. Fourth,
edmxitional attainment is a quantitative measure, unlike
°C°m@ational group data as has been employed in conventional
Studies.
Three measures of educational attainment are rele-
Vant tc> explaining fertility levels. It appears from
Table ll4-that wife's educational level is more popular in
the eHHPirical studies selected for review. Eighteen of the
21 Stlldixes employing one or more education measures used a
measure for the wife. Only ten of these studies employed a
husband' S educational attainment measure (Table l4:4,8,10,18,
20’24'25 28 3o 31) and all but tw f
. , , o 0 these also employed a

"leas
tire for the wife, only three empirical studies used a

‘

 

 

.U

...

120

Combined measure for both husband and wife's educational
level (Table l4:8,9,10). Furthermore, in several studies it
was concluded from the evidence of the results that wife's
educational attainment level is much more influential than
the husband's in predicting fertility (Table 14:8,10,18,20,
25,30). Because of this, the educational level of the
couple always revealed the same relationship with fertility
as that between wife's education and fertility.

Overall, educational attainment, by whatever measure,
tencm»to show a persistent inverse relationship with fertil—
ity. For all three measures mentioned above, studies which
reveal a clear inverse relationship far outnumber those

which either found exceptions to the inverse relation, just

as the "J" curve for occupational data, or found a relation-
ship not statistically significant. In support of this
Observation it is to be noted that wife's educational attain-
ment.lewel maintained this strong inverse relationship with
fertilrity during the period 1940-50, in spite of the fact
that charing this time the proportional increases in fertil—

ltY were directly related to educational attainment.15

This
Fmtteril occurred in all residence categories, although more
pronounced in urban areas. Kiser later found educational

differentials in fertility widened in 1950—57 (Table 14:23) -

\fl

Eton 15C. V. Kiser, "Changes in Fertility by Socio—
Cnnigz Status during 1940—1950, " Milbank Memorial Fund

Qu
\~§E:E£i£_JZ. XXXIII (October, 1955), 417.

121

Both Kiser and Grabill (Table 14:16,l8) found a
clear inverse relationship for wife's educational level
within all age groups and residence groups. In an earlier
study, Kiser (Table 14:7) found the relationship stronger
in his rural sample. Rosenquist and Schafft with data of
approximately the same period also found a relatively sig—
nificant negative correlation for the rural-farm part of
counties in Texas (Table 14:9). Hughes (Table 14:22) like-
wise found in Tennessee that for the rural-farm part of
counties educational attainment (using a median as did
Rosenquist and Schafft) has a strong inverse association
with fertility (revealing the highest beta coefficient of
the variables in the equation).

Goldberg in two important studies (Table l4:21,25)
found that the usual inverse relationship of fertility to
Socio~economic variables is largely due to farm migrants in
urban areas. It is worth noting, however, that he finds
"onlyeeducation survives as a status variable capable of
differentiating levels of fertility among the two-generation
‘urbanites."16 Thus, among both the second-generation urban-
.ites and the farm migrants education possesses much predic-
'tj”e Power, whereas occupation and income do not. However,
'iIlthiS and a later study by Goldberg (Table 14:25), educa—

tllon indicated a stronger inverse relationship for the farm

1___‘____________
l . I
6DaVid Goldberg, op. c1t., p. 218.

122

migrants than for second—generation urbanites. Duncan
(Table 14:30) supported this residential difference in find-
ing that the inverse relationship of fertility and wife's
education is more pronounced for wives of farm residence and
farnlbackground. A study of Freedman and Slesinger (Table 14:
26) concurred with the initial finding of Goldberg that edu-
cation remains inversely associated with fertility in the
"indigenous" nonfarm group, whereas income indicated a posi—
tive relationship. Finally, Hashmi's study (Table 14:27)
focuses on a single urban area and concludes that education
and income together are the most important factors in urban
fertility, together accounting for 71 percent of fertility
variance. Education in his study indicates a strong nega—
tiVeCorrelation with fertility, plus producing the higher
betatzoefficient over income.

In conclusion, these empirical studies strongly
SuPport the expectation of a strong inverse association in

all residence groups, in fact, stronger than such socio-

 

economic status indicators as income and occupation. More-
over, the relative importance of this variable compared
anmng residence groups is difficult to determine, primarily
leecause the relative importance of educational attainment
Should be high in all categories of residence. Some empiri-
C:al StUdies suggest that educational attainment is more

.ElEEEEEEHE_in the rural sectors (Table l4:21,22,25,26,30) in

e"(pl-aiming variation in fertility levels.

 

   

0'

-v

on

 

 

 

 

 

 

 

123

Income.-—Before discussion of the findings on income
from the selected list of empirical studies, a review of the
inherent characteristics of income as a variable is appro—

priate. In contrast to education, and like occupation,

income is a changeable measure; but unlike occupation, it is

quantitative. Second, the usual measures of income consider

income only for the previous year. The sensitivity of such

a measure, it would seem, is dependent upon events which

have transpired in the immediate past. Education, in con-

trast, is a more cumulative index. It is, therefore, neces—
sary to be cautious of results when applying income measures
to the study of cumulative fertility. Kiser explained his
conclusion of the importance of income over occupation and

Education in his study simply because income and fertility

measures pertained to the previous year, whereas occupation

17

and education did not. Third, it would seem that there is

a tendency for income to increase with advancing age.
Similar to occupational group, a previous income level may
have more influence on number of children ever born that
CUrrent income. Finally, a measure of family income is con-
founded by working wives which tends to raise income level
while simultaneously diminishing the fertility level.

\‘V

(B . 17C. V. Kiser, Group Differences in Urban Fertility
altimore: williams and Wilkins, 1942), pp. 172—3.

124

Many indirect indices of income level have been

devised, e.g., relative income, rental value, place of

living, welfare, tax lists, etc., but our primary concern is

with family income per se. The popularity of income as a

measure of socio-economic status, like education, has in—

creased in current studies. Three variables in measurement

have been employed: family income, husband's (sometimes
head of household) income, and wife's income (seldom used

until recently). The relationships of these measures with

fertility and among themselves are much more complex than

for educational attainment measures. The confounding

effects of husbands' and wives' educational level on the

Couples' educational attainment level are not as great as

that for husbands' and wives' income levels and family in-

Come.
The relationship of income and fertility as eXpressed
by empirical studies in Table 14 employing an income measure,

is not clear. There are 18 of the 31 studies in the table

Whleh consider at least one measure of income level. Hus-

band's and family income have been used in about the same

nunIber of cases. In an early study (Table 14:5) family in—
Come for urban areas exhibited an inverse relationship with
fel‘tility, but Kiser (Table 14:7) very early found the ten—
dency for upper income classes to be directly related with

f . .
er1:::n_lity level. Succeeding research established a similar

pattern of relationship (Table l4:10,23,24,3l) primarily for

Urb
an areas, however.

 

125

Some provocative data on income and fertility
appeared in the Current Population Surveys of 1949 and 1952
(Table l4:18,l9). The 1949 Survey found a sharp inverse
relation of number of own children under five per 1,000
Imuried.women 15—49 to total money income of families. The
1952:3urvey, however, found no corresponding inverse rela—

tion between the number of own children under five per 1,000

married men 20-59 and money income of the man. In this same

survey children ever born per 1,000 women, age 45 or older,
married and husband present, was found to be inversely asso-
ciated with husband's income, but for women age 15—44, no
such relation was found. The difficulties of interpreta-
tion arise from the variety of measures used for fertility
and income level as well as the failure to control for high
fertility groups such as rural—farm and non—whites. These
data raised the question, however, of a possible transition
at least for urban areas toward a positive relationship of
fertility to income level. The rural elements appeared to
maintain the inverse pattern of relationship. In the same
publication (Table 14:18) data were presented indicating an
inverse relationship between farm operator family income in
‘19‘49 and fertility. Using 1950 county data for Tennessee,
Hughes (Table 14:22) found a similar inverse association for
the rural-farm parts of counties.

Hughes in the same source, however, established a

Po ' - -
sJ.t_‘1_ve relationship for income of the total population

 

   

 

4..

Il'

...

‘v

 

126

among the counties of Tennessee. This correlation was
strong with income second among five beta coefficients.
Other indications of a positive association are available,
employing relatively current data. The Growth of American
Families Study (Table 14:20) uncovered a positive relation
for husband's income, although it was not significant. The
authors, however, explain that whereas "the average number
of children born by 1955 was actually larger for the high-
income groups than for those with low incomes . . . this
differential is due to the influence of age--the higher in—
come couples are older and so the wives had an opportunity
to bear more children by the time of the interview."18 The
same study found a stronger inverse relation between family
M and fertility. The authors, in addition, pose the
fOJ-lowing explanation as to why the relationship of fertil—
itY to income is somewhat stronger when family income rather
than husband's income is used:
It is because the couples with low family income are
more likely to be those in which the wife does not work
and has relatively many births, while the couples with
high family incomes are more heavily weighted with those
that include working wives who have relatively few chil-
dren. Because the stronger relationship of fertility to
family income is due primarily to the influence of the
Wife's employment status and not to the family income

N, we cannot say that family income alone has much
.‘Lnfluence on fertility.

\~\——

18
Fam' Ronald Freedman, P. K. Whelpton, and A. A. Campbell,
McGLl Planning, SterilitL and Paulation Growth (New York:
raw'fllll. 1959), pp. 296—7.

19 _
334$, pp. 302-3.

127

Others have echoed this same explanation that wife's
income strongly enhances the inverse relationship between
family income and fertility (Table l4:7,l8,26,27) . The
subsequent studies associated with the Growth of American
Families Study, therefore, moved to relating husband's and
wife's income separately with fertility. Using husband's
income Goldberg (Table l4:21,25) found the relationship for
second-generation urbanites tended toward the positive,
whereas farm migrants indicated the traditional inverse rela—
tion. The investigation of wife's income and fertility is
relatively recent, largely because the working wife has
become such a common phenomenon today. Freedman and
Slesinger (Table 14:26) found a consistent negative relation
0f fertility to wife's income, being strong in the first
years of marriage, but a tendency toward the positive for
busband's income. Deborah Freedman (Table 14:28) discovered
a Positive relation only when employing husband's "relative"
income (the ratio of a man's actual income to the income
Customary in his socio-economic reference group). Kunz
(Table 14:31) reinforces this with the discovery that con-
trolling for education occupation, and age at marriage makes

this POSitive relationship between fertility and husband's
inCome even stronger. Surprisingly, Hashmi indicates a
positive relation for an urban area even when employing
m incOme (Table 14:27). Kitagawa (Table 14:29) employed

f - .
amlly income for an urban area but found it related

 

 

.
~

 

 

128

inversely with total fertility, but positively when marital

fertility was employed. This, apparently, was due to the

tendency for women in the high-status groups to marry later

than women in the lower—status groups.

In conclusion, on the basis of the above evidence,

it would not be surprising to discover a moderate positive

correlation between income and fertility in urban areas even

when family income is applied. Rural areas will retain the

more traditional inverse relationship. The income of women,

on the other hand, would tend to consistently correlate in—

versely with fertility in both residence groups. In terms

of relative importance of income variables, although the
eVidence has not been fully discussed, analyses of the
empirical studies suggest that income would rank below edu—
cational attainment but above an occupational measure. The
income of the wife, considering that little research has
been done in this area, will not carry much weight in an
anEllysis involving other major socio—economic variables.
HOWever, if family income is employed as the income measure,
Controlling for women's income as well as women employed in

the labor force and age composition should reduce the rela—

1Ve Jimportance of family income in accounting for fertility

v - . . . .
afiance. Furthermore, Since family income is perhaps a poor

In . .
eEnsure to apply to the rural-farm population, and Wife's

in ~ . .
m 13 perhaps more of an urban phenomenon, it is expected

 

129

that family income is relativeiy more important in the urban

sector than the rural—farm.
Female empioyment.--Interest in the effects of work-

ingxnomen on fertility levels is very recent. Although

Vfimlpton had employed this variable in an analysis of census

tract data in eight northern cities in the 1930's and had

found a strong inverse relation, as well as the fact that it

Mes one of the most important variables in explaining fertil-
ity differences, most research using this variable is recent.
zuxhough Table 14 denotes that only six of the empirical
studies (Table l4:4,18,20,27,28,30) have included a measure
fluffemale employment, the studies which have been conducted
are quite intensive in dealing with this factor.

.A11.six studies report a relatively strong inverse
relation of fertility to women employed in the labor force.
Whelpton (Table 14:4) was the first to point out the diffi-
Culty of specifying the direction of the cause-effect rela—

tiOnship regarding fertility and female employment. That
is: doxmmnen work because they prefer it to marriage and
Chi-ldbearing or because of economic pressure of the family?
A r‘elated question is whether women tend is whether women
tetufl to keep>the number of children low in order to work or
does; the fact that they are employed in the work force keep
fart:ility.low? It appears that a selective process may be
the rnore 1cxgica1 explanation, i.e., the labor force tends to

Se . . .
leaczt womeri of low fertility than Vice versa. Freedman

130

(Table 14:20) explains the relationship with reference to
fecundity: for subfecund women the small family leaves the
wife free to work, for the fecund, on the other hand, advan-
tages of employment may influence working wives to limit
their families.

Grabill finds that age of children, and not merely
whether children are present, is very important to labor
force participation of women (Table 14:18). This relation-
ship is more pronounced when fertility is measured as "own
children under five per married women," than "children ever
born." Thus, after children reach an age when close super—
vision of the mother is not required, there is less inter-
ference with participation in the labor force. This pattern
is distinctly different from an earlier pattern in which
women tended to drop entirely out of the labor force to
enter marriage and raise a family. A census publication
describes a life pattern of married women which may act as
a confounding factor in the relationship of fertility to
labor force participation of women:

In the last decade or so . . . a life pattern seems to
be developing among many married women in which they

work until the arrival of the first baby, temporarily
withdraw from the labor force while their children are

young, and then return to the labor force after Sheir
children are old enough to require little care.2

 

20U. S. Bureau of Census, U. S. Census of Ropulation:

1950, Vol. IV, Special Reports, Part 5, Chapter C,"Ferti1ity,"
p. 11.

 

131

Such a pattern would tend to reduce the importance of labor
force participation among women of long marriage duration.
This is precisely what Freedman found in a recent study
(Table 14:28). In the group of women married 5-9 years,
wife's labor force status was negative and relatively impor-
tant in predicting fertility; in the group married 10 or
more years, no significant relationship was revealed.

Another factor which may compound results rests in
whether marital status is considered in the measure of labor
force participation. Grabill (Table 14:18) finds that chil-
dren ever born per 1,000 ever-married women 15—49 is lower
for women in the labor force, regardless of marital status,
in all residence groups. If only ever-married women are
considered in the measurement of labor force participation,
omitting partially a segment of the female labor force which
may be subfecund, the relationship of fertility and labor
force participation would tend to be stronger. Thus, relat-
ing a measure of the participation of 3ii_women in the labor
force would tend to reduce the predictive power of this
independent variable in all residence categories.

A major share of the interest in female labor force
participation and its effects on fertility has been focused
on urban areas. This is understandable in view of the fact
that in 1950 the proportion of white women age 18 to 49
employed in the labor force was 37 percent for the urban

component of the nation, 26 percent for the rural—nonfarm,

132

and only 17 percent for the rural—farm population. Women
employed in the labor force is more largely an urban phenom-
enon. Hashmi (Table 14:27) found for Chicago census tracts
that employment status of women ranked only below income and
education in relative importance in determining fertility
levels. Furthermore, Grabill speculates that:
It seems possible that the employment opportunities for
women in rural—farm areas may interfere less with house—
hold activities than is the case in urban and rural—
nonfarm areas. It is also likely that by virtue of
higher fertility and by virtue of a greater tendency to
take parents and relatives into the household, the pres-
ence of children per se does not tend to tie the mother
to the home as much in rural-farm areas as in the other
cases.
If this is an accurate observation, it might be argued that
women in the labor force would not affect fertility as much
in rural areas as in urban areas. Thus, the stronger inverse
relation should appear within the urban sector and the rela-
tive importance of women in the labor force should be greater
in the urban population than the rural.
In conclusion, it is expected that female employment

will reveal a stronger negative relationship with fertiligy

in urban than rural areas. Female employment, because of

 

confounding factors, will have ggeateripredictive power in

urban areas than in rural areas and perhaps will be of less

 

importance than educational attainment and family income,

but above occupation and female income for urban areas.

 

21W. H. Grabill, C. V. Kiser, and P. K. Whelpton,
op. cit., p. 265.

133

Age composition.—-Age composition of women in the

 

child bearing years has been dealt with in past studies as

a secondary factor, generally in the form of standardizing
the fertility rate. This procedure, it is suspected, is
usually performed so automatically that the implications for
the results of the analysis are not seriously considered.
The fact that age is dealt with in this fashion implies that
age is possibly a major disturbing factor in the relation of
fertility to conventional socio-economic variables, although
in many cases it is impossible or difficult to determine the
effects of age.

Age not only bears heavily upon fertility because of
its biological foundation, but essentially it provides bench—
marks in the life cycle, and indirectly reflect such socio-
economic variables as income, female employment, occupation,
etc. Age is used simultaneously in qualifying, refining or
isolating dependent and independent measures used in fertil-
ity analysis. As examples, it is customary to limit the
measure of fertility only to the child—bearing ages of
women; occupation is represented only for employed adult
persons: educational attainment is limited to adults beyond
college age, etc. The charge could be made that age has
seldom been employed seriously as a primary variable in
fertility analysis, with the exception of a handful of
studies. Because of this fact, there seems to be very little
evidence with respect to the explicit relationship of fertil—

ity to age.

134

In Table 14, six studies of fertility are listed for
review that have employed some measure of age composition of
women in the childbearing age span. Even in these studies
the results on age either have been omitted, ignored, or
dealt with lightly. The Whelpton study of the 1930's
(Table 14:4) found fertility associated with the percentage
of white women 15—44 who were age 20—34 in an inverse direc—
tion for five of the eight cities studied. Three later
studies concluded that age of women had little to contribute
to the explanation of fertility differences (Table 14:9,11,
13). Rosenquist and Schafft found no significant relation—
ship among the rural—farm parts of counties in Texas when
employing a measure of the percentage of women 15—44 in the
age group 25-29. Duncan used the percentage of women 15—44
who were ages 25-34 ("the most fertile ages") in analyzing
village population in Pennsylvania, but found the age factor
of negligible importance. Dinkel controlled for age of wife
in each broad occupational group by region but concluded
that no additional information was yielded.

Two recent studies obtained more favorable results.
Freedman (Table 14:28) included wife's age in a multiple
regression analysis of fourteen variables. No direction of
association is reported, but for the group married 5—9 years
-Lts beta coefficient ranked fourth in relative importance;
fttr women married ten or more years, age of wife appeared
urljnmmmtant. Assuming that women of the younger ages in the

cblind-bearing period would tend to fall into the "married

135

5-9 years" group, the results suggest that proportions of
women in the younger ages are more important in determining
fertility levels. Hashmi (Table 14:27) recorded a very high
positive association between the proportion of age groups
and fertility, i.e., high proportions of older age groups
correlated with high fertility levels or, translated, high
proportions of younger age groups correlated with low fer—
tility levels (an inverse association). For the metropol—
itan area of Chicago the relative importance of age compo—
sition ranked below education, income, and female employment
and above occupation, distance, nativity, and ethnicity.

In conclusion, the review of empirical studies seems
to lead to a consideration of age as a relatively unimpor-
tant factor in fertility analysis. It could be argued that
age distribution of women of child—bearing age is probably
very similar within residence groups, though admitting
regional differences, largely due to a selective migration
factor which would tend to reduce proportions of young women
(largely single) in the rural-farm group and pad this age
group in the urban and rural-nonfarm groups. Considering
the well-documented fact that changes in fertility in the
last few decades have been due largely to earlier marriage
aindthe tendency to have children at younger ages, plus the

ability of women to conceive being strongest between the

136

ages 18 and 24,22 it appears that the greatest fluctuation

for cumulative fertility is in the early ages of the child—

 

bearing period. In addition, consider the fact that age is
not only an important factor per se, but also is highly
related to other socio-economic variables. Finally, distri-
bution of women by age could be expected to have a direct
limiting effect on fertility in the sense that children will
not be born when the supply of mothers of the peak reproduc-
tive ages is relatively low. Upon consideration of these
aspects, it seems reasonable to expect that age variation in
proportions of women in the childbearing ages would be rela-
tively important in all residence grogps. Furthermore, on
the basis of the two most recent studies reviewed above, 3
strong inverse relationship will occur between proportions
in the young childbearing ages (say ages 15—24) and cumula—
Eive fertiligy, but will not be strong at later ages where
fertility rates level off in the lager age groups of the
reproduction age span.
Resume of Empirical Propositions from
Intensive Review of Empirical Studies

As mentioned at the beginning of this chapter, one
function of a review of literature is the development and
menstruction of a set of meaningful propositions. We are

especially looking for urban-rural differences in the

\—

22Ibid., pp. 29—37.

,_._ .

 

 

 

137

association of independent variables and fertility. Several
propositions have been asserted in the review of findings of
the thirty-one empirical studies categorized by the six
broad independent variables considered in this analysis.
The resume of propositions will follow the same order as

presented above.

Distance
Although the studies reviewed above failed to treat

the relationship of fertility to distance from cities in any
way other than description, additional comments enable the
extraction of more substantial propositions. Even among the
studies considered, a direct relationship between fertility
and distance is quite consistent for all residence groups.
Since it is the intention of this study to consider an
ecological framework as a theoretical position, it is assumed
that distance is an overriding factor in all residence cate—
gories, and, therefore, a very important variable in predict-
ing fertility levels, especially when size of the influencing
metropolitan center is taken into account. It is possible,
'however, that controlling for several other variables, such
as, income, education, etc., will reduce the relative impor—
tance of ecological distance as a predictor of fertility
Variation. The following are propositions for the distance
me—‘asure:

l. Fertility is directly related to distance from metro—

politan centers (differentiated by size) within both
residence categories.

138

2. Distance from metropolitan centers (differentiated
by size) ranges from high relative importance to low
relative importance in determining fertility levels
within both residence groups.

Agricultural Occgpation

 

When considering measure of all occupations, it was
found that relationships obtained varied from a clear in—
verse, in some cases, curvilinear or nonsignificant, to a
slightly direct one. In general, it appears that measures
of the complete broad occupational group distribution were
relatively unimportant in accounting for variation in fer-
tility, especially in urban areas. An inverse relation per—
sisted in rural—farm population, especially among the two
primary agricultural occupations: farmers and farm laborers.
Although an inverse relation of fertility to proportions of
farmers and direct relation to proportions of farm laborers
should be persistent within rural areas, it is concluded
they will not be as relevant in urban situations, nor will
relative importance be very high. The following propositions
for the agricultural occupations are extracted from the
previous discussions on the occupation variable:

3. Fertility is inversely related to proportion of
farmers and farm managers in the labor force within
rural areas.

4. Fertility is directly related to proportion of farm
laborers and farm foremen in the labor force within
rural areas.

5. Fertility ranges from an inverse to a nonsignificant

relation to proportion of farmers and farm managers
in the labor force within the urban sector.

 

 

 

139

6. Fertility ranges from a direct to a nonsignificant
relation to proportion of farm laborers and farm
foremen in the labor force within the urban sector.

7. Agricultural occupations are of intermediate rela—
tive importance in accounting for rural—farm fertil—
ity variance.

8. Agricultural occupations are of low reletive impor-
tance in accounting for urban fertility variance.

Education

 

Wife's educational attainment level revealed a
persistently strong inverse relationship with fertility.
Because wife's educational level tended to be more important
than the husband's in accounting for fertility differences,
couple's educational attainment level follows the wife's
lead and appears, then, as a compromise level of association
between the husband's and wife's educational attainment, an
association still strongly inverse, but weaker than wife's
education. This relationship is inverse for both residence
groups and adult educational attainment is relatively impor—
tant in both residence groups. A list of extracted proposi-
tions follow for adult educational attainment level:

9. Fertility is inversely related to adult educational
attainment level within both residence groups.

10. Adult educational attainment is of high relative
-importance in determining fertility levels in all
residence groups, to a degree more within rural
areas, and less within urban areas.

'DA‘V
‘

...x. .

any--

on r

-o;

 

l
“‘1‘.‘

 

 

140

Income

Husband's income was shown to vary from a slight
inverse, sometime curvilinear or nonsignificant, to a moder—
ate direct association, more so in urban areas. Family in—
come revealed a more consistent inverse relation with fertil-
ity, although tendencies toward a moderate direct relation
is possible in urban areas. This difference is largely due
to working wives. Controlling for female employment and
female income would tend to reduce the importance of family
income especially in urban areas. In addition, female in—
come tended to exert a consistent but moderate inverse
influence on fertility in both residence groups. Its rela—
tive importance is slight, but perhaps greater in urban
areas contrasted with the rural. Propositions dealing with
family income and female income are considered below:

11. Fertility is inversely related to family income
within the rural population.

12. Fertility ranges from a slight inverse, perhaps non—
significant, to a moderate positive relation to
family income with the urban population.

13. Family income is of intermediate relative importance
in accounting for fertility variance within both
residence groups, of possibly higher importance
within urban areas.

14. Fertility is inversely related to female income
within the urban population.

15. Fertility ranges from a slightly inverse to a non-
significant relation to female income within rural
areas.

16. Female income is of intermediate relative importance
in determining urban fertility.

141

17. Female income is of low relative importance in
determining rural fertility.

Female Employment

A measure recognizing all women in the labor force,
regardless of marital status, is a weaker factor than ever—
married women in the labor force, but an inverse relation is
nevertheless expected. Female employment accounts for a
greater portion of fertility variations in urban areas than
the rural. Extracted propositions for proportions of women
employed in the labor force are presented as follows:

18. Fertility is inversely related to females employed
in the labor force in urban areas.

19. Fertility ranges from an inverse to a nonsignificant
relation to female employment within the rural popu-
lation.

20. Female employment is of intermediate relative impor—
tance in accounting for urban fertility variance.

21. Female employment is of low relative importance in
accounting for rural fertility variance.
Age Composition
Relative concentration of women of the child-bearing
ages in the early segment of this span is much more important
irnpredicting fertility levels than relative concentration in
tine later segments. A strong inverse association obtains for
Preportion of young women of childbearing age in both resi-
dence groups because of fluctuations in age of marriage in
this period. Proportions of intermediate aged women in the
childbearing age group varies from a possible slight inverse,

in some cases nonsignificant, to a slight direct relation

142

with fertility. Fluctuations in fertility level off sharply
among this group, but the tendency for urban women to delay
marriage suggests the possibility of this factor being
slightly more important within the urban population. Prop—
ositions for these two factors are suggested as follows:

22. Fertility is inversely related to proportion of
young women of childbearing age within both resi-
dence groups.

23. Proportion of young women of childbearing age is of
intermediate relative importance in accounting for
fertility variance within both residence groups.

24. Fertility ranges from a slight inverse, at times
nonsignificant, to a slight direct relation to
proportions of intermediate aged women of childbear—
ing age within the urban population.

25. Fertility ranges from a nonsignificant to a slightly
direct relation to proportion of intermediate aged
women of childbearing age within the rural popula-
tion.

26. Proportion of intermediate aged women of childbear—
ing age is of low relative importance in both resi—
dence groups, although slightly higher for the urban
population.

Conclusion

 

Table 15 provides a tabular summary of the expected
results of an analysis of residential fertility variations
(both expected direction of association and expected rela—
‘tive importance of the independent variable in accounting
ifor variation within residential fertility) based upon the
empirical propositions which have been constructed from the
intensive review of thirty—one selected empirical studies.

This table, it must be recalled, is not the product of

143

 

 

 

.mmmmcpcmumm CH pwmoaocm uoc mmOQu mm
Hoooo on mamxfla mm uoc mum nofln3 mmflcmcowumamu pmpommxm mmocu muocmo A V mommcucmummt
uomuao
A.m.zv Auomuflov moms mcwummnpaflno CH
canon 30d Ammum>ch 3oq .m.z cmEo3 pmmm mumaomfinmucH .m
mmmm mcflummn
omcflaumumocb mumeomaumucH mmum>cH mumopmaumucH mmum>cH Itaono CH cwEoB mcoow .m
.m.z
chHD mumnomauoucH mmum>cH 30g Ammnm>ch ucmfimoamfim mHmEmm .h
.m.z
chHD mumflomfiumucH mmum>cH 30A Ammum>ch mEoocH mamfimm .o
uomnno
roam 0n A.m.zv
chHD mumoooaumucH Ammum>ch muMflomEHmucH mmum>cH mEoocH haflfimm .m
Hmuom roam mmum>cH roam mmum>cH coaumuoom paste .e
.m.z
Hmuom 3oq Anomuoov muMHomEumucH uomuflo smEmHom o mumuonmq Eumm .m
ImOZ
Hmuom 304 Ammum>ch mumflpmaumucH mmum>cH mummmcmz Eumm o mHmEHmm .N
“.m.zv *A.m.zv AmNHm mnv Hmucmo cmufl
pmcHEumumocb scalrmfim common Boninmflm uomuflo Iaooouuwe Eonm mocmumao .H
mocmuHomEH mocmuuomEH coauomnfla mocmunomEH coauomnflm mmHQMflHm> ucmpcwmwocH
m>HumHmm m>numawm mwnmcowumamm m>HumHmm QHSmCOAumamm
“mamas nuns
mocmoflmmm canon Hmuom

 

 

muHHHuHmw mGHcHEHmumU CH mmaflmaum> ucmpcmmmocfi mo mocmuuomfifl m>flumamu pmuowm
mmHQMHHm> ucmflcwmmficfl Umuomamm 0U muflafluumw MO mmHSmGOflumamH Umuommxw
wvcwfiflmwu UGm mHQMHHm> ucmwcmmmwcfl ND mQOHUHmomOHQ HMUHHflmEm m0 NHMEEdw HmHSQmB

monoum wocmoflmmn can: cod

.umwnm>
No new
«WhomquU

J

‘I

‘q

An;

I4:

144

theoretical considerations but of the consistency of find—
ings from previous empirical studies. On the basis of what
other researchers have found regarding differential fertil-
ity patterns within residence categories, then, these are
the empirical relationships which one might expect to obtain
in a study similar to the present one.

We may employ this summary table to reconsider some
of the ideas presented in the previous chapter concerning
urban and rural differential fertility patterns. Recall it
was suggested that there might be discovered different pat-
terns of differential fertility in each of the residence
categories, thus, justifying the investigation of urban and
rural populations separately. Table 15 indicates that we
should expect some differences between residential catego-
ries, but these differences may not be extreme. This may be
due to the difficulty of summarizing the findings of such
diverse studies and/or the inability to differentiate finer
distinctions with the categories employed, i.e., the use of
"inverse" or "direct" to describe the direction of expected
association and the terms "low," "intermediate," and "high"
tO indicate relative importance. Generally the independent
Variables which are expected to be most important in deter—
minixug residential fertility reveal similar patterns in.both
reSindence groups, e.g., education, distance from metropolitan
Cent er, and proportion of young women in childbearing ages.
Differences in patterns of relationship and relative impor—

tance are greater for independent variables which are

 

..
-1

bu

 

.-

in

m

’1

145

expected to reveal inconsistent, fluctuating relationships
and are expected to have less influence in determining res-
idential fertility, e.g., agricultural occupation, female
income and female employment. Slight differences in direc-
tion of association with fertility are expected for six of
the nine independent variables: agricultural occupations,
family income, female income, female employment, and propor-
tions of intermediate aged women in childbearing ages. The
greater contrast in considering differential fertility pat—
terns by residence seems to be for the relative importance
of independent variables in accounting for fertility varia—
tion. This might suggest the possibility that, while the
direction of association between fertility and related inde—
pendent variables may actually remain the same within resi—
dence categories, the same independent variable may have a
different impact on fertility in rural areas than in urban
areas. This would still support the original hypothesis of
contrasting differential fertility patterns within residence
groups.

A distinction between rural—farm and rural—nonfarm
fertility has not been established in Table 15. There did
not appear enough evidence in the empirical studies reviewed
todifferentiate clearly enough the differential fertility
Pattxarns which one might expect in each of these populations.
Shmilar to the discussion of Chapter I, we have treated
"rurzflfl as one category. Though it was impossible to con—

Struct empirical propositions for these two subcomponents

146

of the rural population, i.e., farm and nonfarm, it is
expected that the rural—nonfarm fertility patterns will
fall somewhere between the urban and rural-farm patterns,
perhaps more approximating the rural-farm than the urban.
In the major analysis of this thesis, all three residence
categories will be considered. Hence, in terms of expected
relationships, we would maintain that the patterns for the
rural population as described in Table 15 would apply
equally to the rural-farm and rural-nonfarm populations.

But the most important function this summary table
of empirical propositions will perform in this study is as a
guideline for direction in theoretical considerations. One
could simply test these empirical propositions as presented,
but what gain would this be considering the amoung of empir-
ical descriptive documentation that has already been com-
piled? What theoretical contribution would this make to the

explanation of differential fertility patterns? The inten-

tion in the next chapter is to review some theoretical frame
works relevant to the explanation of residential differen-
tial fertility, but then to use these empirical propositions
t0 determine which theoretical framework has the greater
Prdbability of successfully predicting residential fertility
Pattuerns. This, it seems to me, is the proper use of past
reseearch findings. This is the method by which a science
Imnmes most efficiently from a descriptive to an explanatory

leve: 1 .

 

 

u...

5,.

-...

 

v-

I
H,

 

CHAPTER III

THEORETICAL FRAMEWORK FOR STUDY OF

DIFFERENTIAL FERTILITY

In view of the fact that this study is primarily
demographic in nature, this particular chapter dealing with
a theoretical framework for the study of differential fer—
tility has proven the most difficult to construct. One of
the chief intentions of this thesis is to meet "head—on"
some of the many criticisms which have been tossed at dif—
ferential fertility investigations based on census data and
to correct in the design of this study for the glaring defi—
ciencies of past studies. Mention has been made in Chapter I
of some of these criticisms and corrections have been pro—
posed. However, those mentioned previously are primarily
methodological. To be comprehensive and consistent, atten—
tion must now be focused toward solving theoretical criti-
cisms and deficiencies. Indeed, it is a temptation to sim-
Ply proceed to test the empirical propositions abstracted
from relevant literature, a procedure which is normative for
research in this area. However, it is felt working from a
theoretical framework, however formidable and unwelcome the
task, is necessary if this area of demographic research is

tetuantinue to make advances. Vance makes a more dramatic

147

148

emphasis when he says, "Empirical operations without basic
theory, no matter how carefully safeguarded, are now proved
dangerous."

The procedure for this chapter is, first, to con—
sider the status of differential fertility theory (and popu-
lation theory in general); second, to suggest critical guide—
lines for theoretical design which will correct for past
weaknesses and deficiencies in the study of differential
fertility; and, third, to construct a theoretical framework
which can be used to provide adequate explanation of the
patterns found in the empirical data of this study as well
as can be considered a solid contribution to demographic
theory, and, more specifically, differential fertility
theory.

Status of Differential FertilityiTheogy

It is precisely the status of differential fertility
theory, and for that matter population theory in general,
that makes the inclusion of a chapter on theoretical frame—
Mmrk in a demographic study so absurd. For, according to
the experts, the keystone characteristic of this field and,
hence, the most devastating criticism, is the lack of theory
in demography and the paucity of theoretical activity pro—

duCed by demographers.

lRupert Vance, "Is Theory for Demographers?" Social
Ehorces, XXXI (1952), 13.

149

It was this situation which lead Vance to entitle
his disturbing article "Is Theory for Demographers?" His
answer to this question, of course, is affirmative, but his
description of demographic theory it as a "wasteland."2
Others have made exactly the same claim. Hauser asserts,
"there is still too much of a tendency among population stu-
dents to produce discrete, descriptive studies with little
or no attention to theoretical framework as a basis for
their research orientation or for the formulation of their
conclusions."3 Moore also makes note of this fact in com-
paring sociology and demography.
If a standard complaint about sociology is that is has
"too much? theory, a standard complaint about demography
is that is has "too little." . . . What is generally
meant in the case of demography is that a pervasive pre—
occupation with refinement of measurement and with is
hoc explanations for observations leads to an avoidance
of the fundamental question, what do we want to know?

Of special interest to this study is that in this same chap—

ter Moore cites "the preoccupation of many demographers with

the analysis of census data" as a partial foundation for the

exaggerated charge that demography has "no" or "too little"

theory, and in the same breath he asserts "the absence of

21bid.

3Philip Hauser, "Present Status and Prospects of
Research in Population," American Sociological Review, XIII
(August, 1948), 377.

4Wilbert E. Moore, "Sociology and Demography," in
E>, Hauser and O. D. Duncan, The Study of Population (Chicago:
The University of Chicago Press, 1959), p. 845.

‘
Aa‘

....

I.—

o.»

I-~.

|i~.

 

 

 

150

any clear—cut theory of fertility differentials."5 Though
all sorts of answers have been presented in response to the
charge that there is a dearth of demographic theory, one

fact is sure, that there is a predominant tendency for demog—
raphers to prefer a raw empirical approach vis—a—vis a Eheof
retically sophisticated approach in demographic research.

The result is, as Davis states, "the tendency to initiate
research either with no explicit hypotheses at all or with

hypotheses pulled out of a hat."6 Hence, regardless of the

excuses put forth to explain the situation, it is true that

differential fertility theory, and demographic theory in

general, is "poverty stricken."

Critical Guidelines for a Theoretical Design
for the Study of Differential Fertiligy

In response to this general charge against demo-
graphic theory, however, several demographers have proposed
constructive points which could strengthen the theoretical

condition of demography if effectively incorporated into

basic research. There are a select number of such points

Vflhich I wish to call upon for the purpose of providing "crit-

irzal guidelines" for a theoretical design. There are four

altogether .

5Ibid., p. 849.

6Kingsley Davis, "The Sociology of Demographic
Behavior," in R. K. Merton, L. Broom, and L. Cottrell,
Socnxology Today (New YOrk: Basic Books, 1959), p. 323.

 

..7\..

151

The first guideline flows from a criticism leveled
at demographic studies that the "explanatory" function of
theory has been neglected. After reviewing several examples
of what demographers consider "theory," Hauser and Duncan
make this their first "critical observation.”

. . . there is relatively little emphasis in the state—
ments of demographers on the explanatory_snd predictive
functions of theogy. At least in the opinion of many
writers on scientific method, what distinguishes proposi-
tions of a "theoretical" character from mere empirical
generalizations is that the former state considerations
as a consequence of which certain empirical regularities
are expected to obtain and indicate conditions under
which such regularities will hold, i.e., theories are
supposed to "explain" and "predict" the facts at the
command of a discipline. . . . If this crucial feature
of scientific theory is neglected, it becomes possible
to accept as "theories" various bodies of discourse
which incorporate several of the ingredients of theory
(such as concepts, definitions, and deductions) but
which fail tp_perform theipringipal function of theopy.

Later in the same chapter Hauser and Duncan seem to provide
a plausible explanation as to why demographers have neglected
this function of theory. In a critical comment on "psycho—
social theories of fertility" (primary reference is to the
Indianapolis Study) they make the statement that:
it may be, too, that they have hit on an issue of signal
importance for population studies in general in stating
an antithesis between testing a particular theory and
the_prediction of a concrete event. If, like most demog-
raphers, they prefer to attempt the prediction or expla—

nation of concrete events, they seem prepared to sacri-
fice some of the elegance of an integrated theory.

 

7Philip Hauser and Otis Dudley Duncan, The Study of
Population (Chicago: University of Chicago Press, 1957),
pp. 84-5 (italics mine).

8Ibid., p. 97.

 

152

Hence, theory development has been sacrificed for the narrow
conception that the primary objective of any scientific
research is the mere prediction of a concrete event. In
View of such an objective, of course, the most appropriate
and efficient research design is the "dragnet" approach (a
continuous accumulation of independent variables until the
dependent variable is satisfactorily "explained" to the
neglect of any understanding of how all the employed vari—
ables fit together in an integrated, logical fashion). All
too often it is this approach which has been found popular
in past differential fertility studies. This must be
altered.

To correct for past deficiencies, therefore, the
first critical guideline requires that the "explanatory"
function of theory be the foremost consideration in theory
construction and that the nature of the problem contained in
this thesis be discerned as "theory testing" rather than
merely "predicting a concrete event." To follow this guide-
line, of course, is not going to be without its casualties.
Vance skillfully describes the risks involved, but also the
challenge:

If there is room in demography for the timid souls,
is there also room for the bold and audacious? In
science as in poker, we realize we can play it one of
two ways. We can play it close to the vest, that is,
maximize description and minimize synthesis or we can
play it for maximum gains of human knowledge. . . .
Accordingly the closer one sticks to his data, the less
vulnerable are his generalizations and ofttimes the less

important. A loose thoughtsystem sacrifices accuracy
for the sake of generalization.

153

In science when one plays for double or nothing, he
runs the risk of evolving a system of high generaliza-
tions and low validity. Obviously this represents high
vulnerability and we are all cautious enough to dread
the results. But we should remember there are two forms
of maximum error: the first is a system that misses con—
tact with the known facts at every point of observation.
The second is no system at all. This is maximum error,
for it equates with total ignorance. As a matter of
fact, I am willing to make the claim that he who devel—
ops a theory capable of being proved invalid makes a
contribution. In statistics the disproof of any hypoth—
esis is accepted as a way station on the road to knowl—
edge. Demographers should become brave enough to so
state their hypotheses that they are capable of dis—
proof.

The second critical guideline follows from the crit—
icism that demographers have maintained a limited conception
of what constitutes the essential ingredients of theory and
that demographers have not made full use of all these levels
of ingredients. This criticism is not difficult to substan-
tiate. Hauser and Duncan suggest this in making the follow-
ing critical observation of demographic theory:

It is not evident in the cited writings of demog—
raphers that "theories" may be stated at widely varying
levels of generality and with greater or lesser scope of
applicability. It is possible that some of the concern
over lack of theory in population studies reflects a
failure to recognize that the functions of theory can be
performed at different levels of specificity and that
each level is potentially significant for some purposes.

As one might expect there exists the eternal ques-

tion of what are the essential ingredients of theory?

Though disagreement would be forthcoming upon whatever one

 

9Rupert B. Vance, op. cit., p. 12.

10Philip Hauser and O. D. Duncan, op. cit., p. 85.

10

 

 

154

proposed as the "essential” ingredients, there is sufficient
consensus concerning this question which will permit the
assertion of a significant observation about demographic
theory. For example, Gutman notes that while there is con-
fusion in the meaning of the word "theory," in practice
"generalizations on several different levels of abstraction
with varying degrees of comprehensiveness are labeled popu—
lation theory."ll Though this list is not exhaustive, some
of the ingredients of theory suggested by Gutman are "con-
cepts,"12 "general orientations toward substantive mate-

rials,"13 "empirical generalizations,"14 and "societal

 

llRobert Gutman, "In Defense of Population Theory,"
American Sociological Review, XXV (June, 1960), 329.

12Gutman writes: "The term 'theory' has been used as
a synonym for a great number of concepts used to organize
population data. The task of analyzing the assumptions and
implications of these concepts, and refining them, has been
considered a kind of theoretical activity." Ibid., p. 329.

l3Gutman uses Robert Merton's phrase to describe
this ingredient: "Such orientations involve broad postulates
which indicate types of variables which are somehow to be
taken into account rather than specifying determinate rela—
tionships between particular variables." Ibid., p. 331.

14Of these Gutman writes: "Several theoretical key-
stones of contemporary population study are really empirical
generalizations; that is to say, they describe 'a set of
uniform conjunctions of traits repeatedly observed to exist,
without any understanding of phy the conjunction occurs;
without a theory which states its rationale.I Transition
theory and goneralizations about the inverse relationship_
between fertility and socio-economic status probabiy belong_
in this category) even though demographers do have positive
notions about why the demographic transition has taken place
and the reasons which explain the higher fertility among the
poor. But they are notions rather than carefully conceived
and well-reasoned explanations." Ibid, p. 331 (italics mine).

155

laws."15 Hauser and Duncan earlier attempted to synthesize
selected demographic materials and concluded that the ingre—
dients of theory according to demographers include:
(a) concepts defined within a frame of reference;
(b) empirical propositions, generalizations, or laws
(which vary in generality and “ credibility); (c) prop-
ositions, hypotheses, or theorems deduced or constructed
from other elements of theory, with these taking the
form of (d) "necessary relations," "models" incorporat-
ing necessary and/or empirical relations, or "purely
hypothetical constructions."

Although Gutman makes a strong case that demograph-
ers have contributed more in the way of theory construction
at the several levels of theory ingredients than what others
are willing to credit them, he does not suggest the quite
realistic possibility that demographers' contribution to
theory—building has been to a large extent concentrated in
the area of "empirical generalizations" (to use Gutman's
phrase). Hence, while demographers should be commended for
not entirely neglecting theory development in their disci—
pline, it must be emphasized that any mass production of
empirical propositions alone is not a sound approach to
theory development. In such a case there is still too much

wanting. Attempts to clearly spell out the variety of

ingredients which constitute theory are intended to influence

 

15According to Gutman "societal laws differ from
empirical generalizations because they are derived from
statements on a higher level of abstraction, and are the
product of deduction rather than induction." Ibid., p. 332.

l6Philip Hauser and O. D. Duncan, op. cit., p. 84.

156

the initiation of theoretical activity at many levels, not
to encourage specialization tendencies in selected direc-

tions. In practice, then, demographers have not met this

theoretical need.

As a result, and Gutman himself suggests this, the
theoretical contributions of differential fertility studies
employing socio-economic status variables fall mostly under
the rubric of "empirical generalizations." But there is
more to "theory" than just "empirical generalizations."
Dinkel clearly reveals the inadequacies of such an approach
in his criticism of a fertility study which he described as
"another inheritor of the Indianapolis Study legacy."
Dinkel writes:

Even if correlations of high order were found and
theoretical inconsistencies between supported and unsup-
ported hypotheses were ironed out, the authors still
would have the task of explaining the processes by which
operation of the model yielded the correlations. Corre-
lations need to be buttressed by_qgalitative materials
and logical connections that indicate the processes
through which the independent variables are associated
with fertilipy.

Physical scientists are more accustomed than social
scientists to so finishing the job of research and not
letting it stop with the obtaining of statistical asso—
ciation between the variables. In the physical sciences,
of course, possession of more closely—reasoned theoreti—
cal models that are supported by much past research of
an additive character makes more possible the crucial
experiment from which results can be taken directly and
fed back into the analytical structure; in other words,
the explanation of quantitative associations found in
the test are at hand before the test is made. In the
social sciences, on the other hand, it is necessary to
backtrack from the correlations to other research and to

 

 

 

 

 

~e

157

the conceptual framework in order to juggle the_pieces
together in a new whole.17

 

It is strongly felt that criticism such as this must
not be written off as only an "interesting comment" which
might be wisely considered by the designers of that partic—
ular study, but it can be legitimately and accurately ex—
tended to all demographic investigations which seem to
proudly hail as their ultimate objective the empirical sub—
stantiation of a battery of disconnected hypotheses, implic-
itly or explicitly stated, which leads to the careless stock-
piling of irrelevant empirical propositions. On the other
hand, it must be noted that deploring this tendency in demo—
graphic studies does not in any fashion give full blessing
to the type of "theorizing" which produces abstract proposi-
tions which are finally systematically and logically related
but which have no empirical referents. We must concur with
Davis when, in condemning this trend in the social sciences,
he acclaims:

In social science this term (theory), instead of meaning
the widest body of rigorous reasoning about a set of
observed relationships, has come to mean a long stretch
of purely verbal analysis. If a publication contains
any empirical evidence, particularly of a statistical
kind, it is not theory; but if it contains only verbal

generalizations, no matter how loosely connected, it is
theory. 8

 

17Robert M. Dinkel, "Fertility in Mid-Century
America," Eugenics Quarterly, III (March, 1956), 26 (italics
mine).

l8Kingsley Davis, "The Sociology of Demographic
Behavior," op. cit., p. 313.

-n
.5

 

"1

.n

A».

-..

...“

a...‘

o

‘I-‘

“l‘ ~

5‘.‘

u
‘-

 

 

 

158

There is an all too obvious solution by which the
making of obeisance to either undesirable form of theoretical
activity mentioned above can be avoided, and this introduces
the statement of the second critical guideline for the theo-
retical design of this study. A reasonably sound approach
to theory-building is the unending attempt to integrate all
levels of theory ingredients into the same theoretical frame—
work, especially empirically produced and theoretically
deduced propositions. Hence, theorizing must take the form
of bridging the gap between empirical propositions and theo—
retical hypotheses, of employing to the fullest extent pos—
sible all empirical propositions which have been established
and all theoretical frameworks which are relevant. obviously
this guideline is not original, but it is felt the explicit
statement of this research ideal will go far in correcting
for mistakes made in past differential fertility studies.

The third critical guideline for the theoretical
design of this study takes its cue from a plea by a number
of demographers for the development of "theories of the
middle range." In Robert Merton's own words, such theories
are those "intermediate to the minor working hypotheses
evolved in abundance during the day by day routines of
research, and the all—inclusive speculations comprising a
master conceptual scheme from which it is hoped to derive a

very large number of empirically observed uniformities of

159

social behavior."19 In the eyes of these demographers
"theories of the middle range" are viewed as the best and
only answer to resolving the dearth—of—demographic—theory
problem. Furthermore, it is speculated that in no other
area of the social sciences than demography is the theoret—
ical void at this level so noticeable and obstructive to the
advancement of the discipline.

Hillery perceives a critical lacking of middle range
theory in both sociology and demography, but foresees this
type of theory as a significant key to the integration of

both fields.20

Like Hillery, Vance also makes a firm appeal
for making middle range theorizing fashionable. Vance's
hope expressed in his article was that a whole new era of
demographic theory would now be possible because of this new
directive. This means that demographic theory would now be
constructed at a level that would permit a procedure of test—
ing, disproving and refining. Now it would be possible to
correct for another major deficiency of past population
theories (to put it in the words of Hauser and Duncan):

. . . that they are stated so generally or abstractly

that they fail to "make contact" with the facts or with

regularities which have thus far been established empir-
ically or, what comes down to the same thing, that their

 

19Robert K. Merton, Social Theory and Social Struc-
ture (Glencoe, Illinois: The Free Press, 1957), p. 9.

 

20George Hillery, Jr., "Toward a Conceptualization
of Demography," Social Forces, XXXVII (October, 1958), 49-51.

 

0!

.s

\u

I'.
g

i.

9.,

'\

‘A‘.

'5‘:

N-~

a
ll'

4.

Ill

 

 

160

predictions are of such a general character that they

are not readily refuted or confirmed by evidence now

available or likely soon to become available.21
This is, indeed, a quite opposite criticism posited against
demographic theory when the discussion of the second criti—
cal guideline presented above is considered. Both criti-
cisms are valid, nevertheless, as Vance indicates in his
article. Demographic theory has been guilty of "theorizing"
too close to empirical data as well as "theorizing" out of
contact with empirical data. Middle range theory is pre—
cisely what is needed to bridge the gap.

Though several years have passed since Vance wrote
his article, the fact that his prediction that "when all the
hypotheses of the Indianapolis Study are finally fused, popu-
lation will have a healthy young theory of the middle range"22
did not come true is perhaps indicative of the snail's pace
at which demography has attempted to rise to Vance's request
for the development of middle range theory. Nevertheless
there are "sproutings" of middle range theory which are now
developing in demographic fertility analysis and one shall
have to consider these in any attempt to bring theory into
contact with empirical data.

Finally the fourth critical guideline which will

influence the theoretical design of this thesis is easier to

 

21Philip Hauser and O. D. Duncan, op. cit., p. 85.

22Rupert Vance, op. cit., p. 90.

161

spell out than to operationalize. Mention was made above
concerning the recent emergence of some middle range theo—
ries dealing with differential fertility. In one sense

this is exactly what Dr. Vance had ordered, but from another
point of View this places the researcher in direct confronta-
tion with the sticky problem of choosing which middle range
theoretical framework to employ in fertility research. The
fact that none of these theories are what one might call
"full—grown" theories, since all of them contain the usual
gaps and lack repeated empirical substantiation, adds to the
dilemmatic nature of the choice. This "pieces and patches"
condition of fertility theory is perhaps the underlying rea—
son why, quite subsequent to the classic Indianapolis Study,
investigators have repeatedly rejected the scheme of select-
ing a particular theory to direct fertility analysis.
Mishler and Westoff, principal investigators for the Prince-
ton Study, reasoned that developing a particular theory is
unwise because there are "insufficient grounds for selecting
any particular 'theory' which would automatically restrict
the types and ranges of data gathered."23 The question can
be raised, however, as to whether this was the only and

wisest choice in the light of the status of fertility theory.

 

23Elliot Mishler and Charles Westoff, "A Proposal
for Research on Social Psychological Factors Affecting
Fertility: Concepts and Hypotheses," in Current Research in
Human Fertility (New YOrk: Milbank Memorial Fund, 1955),
p. 128.

 

 

 

 

 

 

 

 

 

162

Their choice was one of “no theory" or the hope that a uni-
fied theory would evolve from their efforts, as Hauser and
Duncan describe this procedure:
What the investigators do propose to use is a "concep—
tual framework or 'model' which (permits) the descrip—
tion of the major elements of the concrete situation
within which the fertility event takes place." As
expounded, this "conceptual framework" appears to amount
to an a priori classification of "dependent variables"
and "independent variables," whose use "entails a cer-
tain risk that the hypotheses which are formulated will
form together less of a unified whole than would be the
case if the study were developed in terms of a single
body of theory." 4
Ignoring existing theory is herewith rejected as an
acceptable procedure for the present study. To make use of
existing theory, however, considering the current status of
fertility theory, it is felt that some rather untried pro-
cedure must be adopted. Casting about for instances of such
a procedure, I find that Westie25 most closely approaches
the model in mind. Although "Westie's procedure" will not
be replicated, the "spirit" of what is called for in this
fourth critical guideline is captured in his introductory
comments to his suggested procedure. He proposes the util-
ization of his procedure in areas of investigation where

there exists "a high degree of theoretical incoherence,"

where "knowledge consists of numerous contradictory theories

 

24Philip Hauser and O. D. Duncan, op. cit., p. 97.

25Frank R. Westie, "Toward Closer Relations Between
Theory and Research: A Procedure and an Example," American
Sociological Review, XXII (April, 1957), 149-54.

 

163

and fragments of theories that have been constructed to
explain 'empirical relationships,‘ which may or may not
exist."26 His procedure is described as an alternative to
procedures the researcher more customarily follows under
"incoherent" theoretical conditions, such as:

1. He resorts to a rigid empiricism in which the
"facts" (meaning the empirical findings) are seen to
speak for themselves. . . .

2. He selects from among the many contradictory
propositions already held in the field a particular
proposition or set of propositions which are relevant
to the problem at hand and which appear to make sense
in terms of what the investigator already knows about
the aspect of society under investigation. 27

3. He creates a new set of propositions of his own.

Briefly his procedure proposes the utilization of all the
theoretical propositions in the area of investigation as
they exist, with all their contradictions and inadequacies.
It involves "listing a comprehensive range of presupposed
empirical relationships . . . which might possibly turn up
in the research at hand and explicitly listing a range of
interpretations . . . for each possible empirical finding."
The relationships that are supported by empirical investiga-
tion are retained and the correct theoretical interpreta—
tions are selected from the array of contradictory though

"plausible" interpretations attached to the empirical rela-

tionships that have survived the research test.

 

26Ibid., p. 149.

27Ibid.

164

It is not my intention to follow this procedure
wholly, but Westie's suggestion that we begin testing con—
tradictory theories and hypotheses against each other is
promising for differential fertility analysis, given its
current status and deficiencies. To meet this critical
guideline I intend to consider two theoretical frameworks
which have been employed in differential fertility analysis
but which lead to contrasting hypotheses. These two theoret—
ical frameworks are "urban dominance" theory (presently in-
corporated in the logical extensions of demographic transi—
tion theory) and "metropolitan dominance" theory. Empirical
propositions have already been established by which to deter-
mine which of these frameworks leads to the more valid con—
clusions. In this manner Westie's suggestions (making up
the fourth critical guideline) will be implemented. However,
this procedure will also, I claim, face up to the other
three critical guidelines established above. I consider
both theories to be of the middle range variety. Further-
more, the design of this study will go far in bridging the
gap between empirical propositions and theoretical hypoth—
eses. All levels of theory ingredients will be employed.
.Finally, the explanatory function of theory is especially
emphasized in this manner over against a narrow concern for
Preediction of concrete events.

Let us now turn to implementing these guidelines.

Beifore contrasting the two theoretical frameworks and

165

generating hypotheses, however, an introduction to the
implications of a distributive design for the theoretical

framework is necessary.
Theoretical Framework

It has been stated in a previous discussion of the
research design of this study that the distributive approach
is to be employed to investigate comparatively variation in
rural and urban fertility. This approach necessitates the
study of a given population in terms of spatial patterns
among designated areal subdivisions of its territory. Anal-
ysis involves the discovery of population characteristics of
these areal units which covary with population characteris—
tics of these same units. Explanation rather than descrip-
tion should be the objective of the distributive approach,
but the discovery of a statistical relationship or the uncov—
ering of factors that are nonrandomly related to population
events cannot be interpreted as being causal. The distribu-
tive approach is only a particular type of research design,
not a theoretical framework. Cause must be inferred from
theoretical considerations. Hence, at this juncture a theo—
retical framework should be selected to complement the dis-
tr‘:i‘butive approach as a basic research design of this study.

From the point of view of this study, an ecological
fréamework is best suited to a situation wherein the distribu-

titre approach is employed and provides the best means for

 

 

U.

 

‘F.
p

 

'\

. u.

-"

‘v

”o

 

("

I
1,,

h‘~

v“.

 

.t‘

166

ordering demographic data intelligibly. Noting that human
ecology focuses on the impact of man's environment on human
behavior, Bogue himself infers an ecological framework when
he states that "population distribution studies are capable
of contributing a great deal of indirect information about
environmental factors that underlie population events."28
The unit of analysis for human ecologists, however, is not
merely a population aggregate, but a human population more
or less circumscribed territorially. Though the community
is not the only ecological unit of analysis that fits this
description, it certainly has been a most frequently chosen
unit of analysis in ecological research. Hawley himself, a
leading contributor to ecological theory, argued that a
community is "the basic unit of ecological investigation,29
though others have questioned the expediency of this asser—
tion.3O Nevertheless, from an ecological point of View the
community definitely maintains a territorial dimension, as
Suggested by Duncan and Reiss' definition of community:

"The territorially oriented complex of human relationships
through which a more or less ecological population meets its

28Donald J. Bogue, "Population Distribution," in

P. Hauser and O. D. Duncan, op. cit., p. 393.

29Amos Hawley, Human Ecology (New York: The Ronald
Press, 1950), p. 180.

30O. D. Duncan, "Human Ecology and Population
Studies," in P. Hauser and O. D. Duncan, op. cit., p. 684.

b.-

 

'

I.”

...n

5') I

V‘a.
v .
‘-

,—

'V‘v.
.-
‘I‘. .

a.

b..-

u. p.
I

“v..

 

 

167

sustenance and residence requirements."31 The unit of
analysis for this study, as reported previously, is the
residential part of a county and is to be considered an
approximation of the ecological definition of community.

Given the community as the unit of analysis, the
next logical consideration is inter—community variation.
That communities vary one from another, even in the same
geographical region, is common sense, but why communities
vary must be answered by a theoretical framework, specifi-
cally ecological. When ecologists speak of community varia-
tion, they speak in terms of structural differentiation,
variations in social structure or social organization. The
social organization or structure of a community is viewed
by ecologists as "a collective adaptation on the part of a
population to its total environment (including other orga-
nized populations, as well as physical features), an adapta-
tion that is strongly modified by the technological equip-
ment in use and by certain 'purely' demographic attributes
Of the population itself, notably its size, rate of growth,
and biological (age—sex) composition."32 In this View of
<20mmunity social organization one can readily identify the

four main referential concepts which human ecology embraces:

31O. D. Duncan and A. J. Reiss, Jr., Social Charac-
teristics of Urban and Rural Communities, 1950 (New York:
John Wiley, 1956), p. xiii.

32L€0 Schnore, "Social Morphology and Human Ecology,"
Aperican Journal of SociologY, LXIII (May, 1958), 629.

168

population, environment, technology, and (social) organiza—
tion. These are referred to collectively as the "ecological
complex" or the "ecosystem."33 To understand variation in
community social organization or social structure, then, one
must perceive it as a product or outcome of the unique demo—
graphic, technological, and environmental pressures which
the community confronts, though one must not discount the
interdependent relationship of the elements of the ecolog—
ical complex and the reciprocal effects which social organi—
zation may bring to bear on the other elements. Community
variation is not random or unexplicable, therefore, but
reflects the operation of unique combinations of local
conditions.

The dependent variable of this study, however, is
not social organization but a population event, namely, fer—
tility. In the same manner that we may consider social orga—
nization as a collective adaptation to unique demographic,
technological, and environmental pressures which a community
Confronts, we may also view fertility, being a component
Of population growth, as a collective adaptation of a com-

ITl‘unity to the influences of its peculiar organizational,

33O. D. Duncan, "Human Ecology and Population
S‘tudies," op. cit., pp. 681-84; Leo Schnore, op. cit.,
I). 629; O. D. Duncan and Leo Schnore, "Cultural, Behavioral
Eand Ecological Perspectives in the Study of Social Organiza-
tion," American Journal of Sociology, XLV (September, 1959),
135-6; and O. D. Duncan, "From Social System to Ecosystem,"
§£§iological Inguigy, XXXI (Spring, 1961), 140-9.

 

169

technological, and environmental (including other organized
populations) conditions. Duncan himself suggests that one
of the distinctively ecological contributions to the study
of the vital processes (fertility and mortality) is the
study of "vital rates as indexes of the adjustment of a
population to its environment and investigation of the
impact of variations in community structure and function
on the vital processes."34
It is hypothesized, then, on the basis of an ecolog—
ical framework, that community social structure or organiza—
tion will play a significantly determinative role in fertil-
ity variation among communities. Discussions on the distrib-

utive design, human ecology and demography point to popula-

tion composition as an excellent source for operationalizing

 

indices of community social organization. Bogue asserts
that the "study of how differential population composition
leads to differential population behavior (in our case,
fertility behavior) is one of two major aspects of distri-
butional analysis."35 Duncan states that "the principal
interests of the human ecologist in the study of population

<30mposition are the exploitation of data on composition as

34O. D. Duncan, "Human Ecology and Population
Studies," op. cit., p. 698.

5
Donald J. Bogue, "Population Distribution,"

.Op. cit., p. 384.

170

indicators of ecological organization."36 Likewise,
Schnore maintains that "a compositional View of population

inevitably provides a proximate description of social

structure. . . ."37 Hence population composition indices

operationalized on an areal basis at the community level
reflect community social structure and may be expected to
significantly influence fertility behavior. In commenting
on aggregate social data for areal units as indices of
social structure, Feldman and Tilly argue that

. . . the residential area is one important context
within which personal behavior takes place. Thus, the
fact of living in an area with certain characteristics
in income, education, race, and so on (in other words,
various aspects of population composition), is sociolog-
ically relevant, whether or not the personal traits of
residents are similar to the averages of the areal units
in which they reside.

Hence residential area is a significant context in which
human behavior takes place. The importance of spatial rela-
tionships in ecological analysis is expressed by Duncan in

three points:

First, territoriality is a major factor giving unit char-
acter to populations. Second, space is simultaneously a

36O. D. Duncan, "Human Ecology and Population

S”Cudies," op. cit., p. 693.

37Leo F. Schnore, "Social Mobility in Demographic
IPerspective," American Sociological Review, XXVI, No. 3
(June, 1961), 411.

38Arnold S. Feldman and Charles Tilly, "The Interac-
tion of Social and Physical Space," American Sociological
'Review, XXV, No. 6 (December, 1960), 879. Statement in
‘Darentheses mine. For related comments on ecological corre—
lation see pp. 69-72 of this thesis.

 

 

 

 

.A‘I

‘3':

mm.

.
.'A4

'uv.

-
~V-u

~..

 

(I)

 

 

 

171

requisite for the activities of any organizational unit
and an obstacle which must be overcome in establishing
interunit relationships. Finally, space--like time--
furnishes a convenient and invariant set of reference
points for observation, and observed spatio-temporal
regularities and rhythms furnish convenient indicators
of structural relationships.

The conclusion from an ecological perspective is
that fertility behavior, operationalized on a residential
area basis, is a function of residential community social
structure, operationalized in terms of various indices of
population composition, as well as technological, environ-
mental, and other demographic variables. For this study
these independent variables include such indices as distance
from a metropolitan center, employment in agricultural occu—
pations, education, family income, female income and employ—
ment, and the age structure of women in the reproductive age

period.4O

 

39O. D. Duncan and Leo F. Schnore, "Cultural,
Behavioral, and Ecological Perspectives in the Study of
Social Organization," op. cit., p. 136.

40It is difficult to classify these independent
Variables in terms of the four broad categories of the eco-
logical complex, i.e., population, environmental, technolog-
ical, and organizational, since several of these could fall
-into more than one category. To facilitate the discussion
<?f these independent variables, however, we may group them
iln.the following manner: urbanization (distance from a
Imetropolitan center and employment in agricultural occupa-
tions, which is a reflection of off-farm employment), socio-
Sagonomic status (education and family income), wife's oppor—
.Egmities alternative to child—bearing (female income and
female employment), and demographic age structure (the dis—
tribution of women in age groups of the reproductive age
period).

 

 

 

 

 

(

1
q

 

1

u
‘

 

172

Since the primary concern of this study is a compar—
ison of urban and rural fertility patterns, we must extend
our ecological framework so that hypotheses can be generated
dealing with the urban-rural comparison. Up to this point
we may only assert that the general hypothesis of the covar—
iance of community fertility behavior and community social
structure applies both to urban and rural areas. Our eco-
logical framework has not been developed far enough to estab-
lish hypotheses concerning the direction and nature of the
relationships between fertility and community social struc—
ture or the comparative fertility patterns for urban and
rural communities. Let us proceed to investigate ecological
theory in more detail in order to establish such hypotheses.

At this juncture, however, there are two possible
directions we could take in the ecological comparison of
urban and rural community social structure and their influ—
ence on fertility behavior. These two theoretical direc-
tions may be designated as "urban dominance" or "urbaniza—
tion" theory and "metropolitan dominance" theory. Both of
these theories may be tested by a distributive research
design, i.e., by an investigation of the association of the
5Spatial distribution of fertility levels and community social
EStructure. But each theory leads to a different, in fact,
<lontradictory, set of hypotheses concerning the comparison
of urban and rural fertility patterns. Urban dominance

theory will be rejected in favor of metropolitan dominance

173

theory as the most suitable theoretical framework for this
study, but let us consider these two theories for a brief
duration in order to fully explain the rejection of the one
in favor of the other.

There seems to be in the literature of ecological
and demographic research on urban and rural communities a
subtle confusion as to the definition and differential
effects of urbanization and metropolitanization, urban
dominance and metropolitan dominance. Bogue, in his classic
work on the metropolitan community, is quite insistent on
distinguishing urbanism and metropolitanism, urban dominance
and metropolitan dominance. Very simply Bogue states, "by
urban we mean 'pertaining to all cities,‘ by metropolitan we
mean 'pertaining to the metropolis.”41 Bogue also points
out that his study "is an attempt to explore some aspects of
the hypothesis that great cities, or metropolises, dominate
the social and economic organization of technologically
advanced societies."42 Metropolitan dominance theory, it is
mentioned, is "an extension and attempted refinement of the
.more general theory of urban dominance, that is, that cities
in general are the foci about which life of modern nations

is organized."43 Likewise Vance and Smith differentiate

41Donald J. Bogue, The Structure of the Metropolitan
Community (Ann Arbor: Horace H. Rackham School of Graduate
Studies, University of Michigan, 1950), p. 6.

42Ibid., p. 3.

43Ibid.

 

 

r

 

174

between urbanization and metropolitanism. They point out
that metropolitanism is one aspect of the more general and
broader process of urbanization. Urbanization is "the con—
centration of population and human activities at focal
points in space. This concentration proceeds in two ways
simultaneously: the multiplying of the points of concentra—
tion and the accompanying increase in the size of the indi—
vidual concentrations."44 Hence the emergence of exception—
ally large cities is the process of metropolitanism. Let us
consider first urban dominance theory with respect to the
effect of urbanization on community social structure varia-
tion and its consequences for fertility behavior for both

urban and rural areas.

Urban Dominance Theory

 

Earlier in this thesis we made reference to a cardi-
nal proposition of urban dominance theory (though it was not
called by this label): that urban and rural differences are
disappearing as urban centers extend their influence further

into rural areas. Comments were cited from Robin Williams45

44Rupert B. Vance and Sara Smith, "Metropolitan Dom—

inance and Integration," in R. B. Vance and N. J. Demerath
(eds.), The Urban South (Chapel Hill: University of North
<Iarolina Press, 1954), p. 114. Original statement in Hope
'Tisdale, "The Process of Urbanization," Social Forces, XX
(1942), 311.

45Robin M. Williams, Jr ., "American Society in
Transition: Trends and Emerging Development in Social and
Cultural Systems," in J. H. Copp, Our Changing Rural Society:
Eerspectives and Trends (Ames: Iowas State University Press,
1964), 23-4.

 

 

 

175

and Comhaire and Cahnman.46 These last authors conclude

that American society is "urban through and through."

Though others have eschewed the positing of such a bold
pronouncement, implicitly or explicitly they assume an urban
dominance framework and, on this basis, suggest that a near
future outcome will be the vanishing of urban—rural differ—
ences. Ryder, a demographer, writes of urbanization as part
of an overall tendency toward homogeneity in American
Society.

. . . We are becoming more and more like oneanother all
the time, more and more homogeneous as a society, more
and more similar in big and little ways. . . .

The contrast between past and present is equally
striking in terms of where we live. It used to be that
the population was sharply split between those who lived
in sparsely settled rural areas and those in dense city
concentrations. But now the farm population has been
dwindling for some time, and will continue to do so;
more and more farmers also have city jobs; and the urban
ways of living have been exported out to the farmer.
Meanwhile the urbanites are fleeing the city. The cen—
ters of our big metropolitan areas have pretty well
stopped growing, and the city-dweller is transformed
into a suburban commuter, with a setting as rural as he
can afford. One other way in which the country and the
city are merging is that we don't actually live in one
place any more. We are the most mobile people ever to
have lived on this planet, and our constant peregrina—
tions are profound erasers of regional differences and
provincial attitudes.4

46Jean Comhaire and'Werner J. Cahnman, How Cities

Grow (Madison, New Jersey: The Floram Park Press, 1959),
pa 1.

 

47N. B. Ryder, "Variability and Convergence in the
American Population," Phi Delta Kappan, XLI (June, 1960),
381-2.

 

 

176

Hence regional differences, perhaps due to varying rates of
urbanization, will also disappear. From this one concludes
that the more urbanized a region, the more the rural popula-
tion will approximate the urban.

Many studies of differential fertility have con—
cluded the disappearance of urban and rural fertility dif—
ferences on the basis of urban dominance theory. The theo—
retical framework frequently employed in such fertility
studies goes by the label "the demographic transition the-
ory, " though one readily sees that this is a variant on

urban dominance theory with a historical perspective on
Charlging fertility levels between different groups in soci—
ety - The pervading proposition underlying demographic
trans ition theory is the inhibiting effect of urbanization

on fertility.48

Generally the theory hypothesizes an in—
verSe relationship between fertility and urbanization.
Acc-‘-c>:l:‘dingly populations manifest characteristic types of
population growth dependent upon their stage of urbanization.
Hence populations which are highly urbanized have low fertil—
ity and mortality rates; populations beginning the process
of urbanization portray rapidly declining mortality rates
\

Am 48A. J. Jaffe, "Urbanization and Fertility, "

fiiaan Journal of Sociology, .XLVIII (July, ls4z), 48-60;
d Warren C. Robinson, "Urbanization and Fertility: The

an‘Western Experience," Milbank Memorial Fund Quarterly,
31. No. 3. (July, 1963), 291-308.

 

177

but relatively high fertility rates; populations which have
not begun the transition reveal both high fertility and high

mortality levels. Transition theory is traditionally

extended to explain the phenomenon of the inverse fertility
differentials found among the socio—economic and residential

groups of populations undergoing and having undergone the

transitional phase. Grabill, Kiser and Whelpton succinctly

describe this process as follows:

The phenomenon of differential fertility according to
occupational or socio-economic status has sometimes been/
described as a transitional phase of declining fertilityfi v.4"
The theory is that the decline begins in the so-called ’
"upper" occupational classes in urban areas. Later, the
declines affect the so-called "middle" classes and
finally the so—called "lower" occupational classes.
the meantime the declines spread outward to the rural
areas and presumably the process runs the same type of

c ourse there .

Abu—Lughod has proposed demographic transition the—

In

City as an analytical model which can be used to predict the

\__

 

" 49The theory was first stated by Warren S. Thompson,
9ggptllation," American Journal of Sociology, XXXI (May, 1929),
L 9~75; reformulated by Frank W. Notestein, "Population--The
U21}? View," in T. w. Schultz, Food for the World (Chicago:
Selversity of Chicago Press, 1945) . For other statements
The : Kingsley Davis, "The World Demographic Transition,"
\H e Annals, CCXXXVII (January, 1945), 1-11; Kingsley Davis,
fin Society (New York: Macmillan, 1949), pp. 603-8; Dennis
Doohg, Population (New York: Random House, 1956); and
Thnald O. Cowgill, "Transition Theory as General Population
eory," Social Forces, XLI (March, 1963), 270-74.

50W. H. Grabill, C. V. Kiser, and P. K. Whelpton,

{fiFertility of American Women (New York: John Wiley,
58), p. 180. See also D. O. Cowgill, op. cit., proposi-

t
J~<>ns 6 and 7, p. 273.

 

 

178

presence (or absence) of urban—rural differentials at

various stages of the transition. In her introductory

statement she says:

Within Western industrialized countries certain dif—
ferences have been noted in the demographic structures
of urban and rural areas. Early in the present century,
when these were first being probed, the observed varia-
tions were emphasized and urban areas were considered
to differ sui generis from rural ones. More recently,
as urban culture has spread out from the metropolitan
centers to encompass more and more of the rural hinter-
lands, and as rural values, patterns, tastes, and stan-

dards tend increasingly to approximate those set in
cities, many of the differentials hitherto considered
inviolate have begun to blur. The wide spread between
1irban and rural fertility rates commonly observed sev-
eeral decades ago in the United States has been narrowing
Iorecipitously, and further diminution is anticipated.

True 1:raditional View of transition theory contains only
three stages. Abu-Lughod proposes four phases: pre-indus—

tITiiaZL, semi—industrial (early transition), industrialized

(tzriarisition proper), and post-industrial society. Though

SE16! (does not state it explicitly, she implies that Western

colllllfl‘tries, including the United States, are moving into a

rnaVV 'transition period. the "postindustrial society." During

1:}“3 transition proper period, urbanization first tends to
e1"laggerate urban—rural differentials by increasing member-

ssrlilp in classes likely to be experiencing fertility decline

eefigggg

E‘ SlJanet Abu—Lughod, "Urban—Rural Differences as a
aunction of the Demographic Transition: Egyptian Data and
11 lknalytical Model," American Journal of Sociology, LXIX

(March, 1964) , 476-90.

521bid., p. 476.

179

and reducing membership in classes least affected by the new
fertility pattern. Hence the change in the percentage of
tihe population engaged in non—agricultural pursuits results

iri "an increase in the number of persons 'exposed' to condi—

53

tjxons and values favoring lowered fertility." One signif—

icant change that occurs in industrial society affects the
Italiationship between cities and their hinterlands. "With
iruiiistrialization and more particularly with the prolifera—
tiJDII of the transport and communication networks prerequi—
Siilee ‘to industrial growth comes a radical expansion in the
(zitzir':s sphere of influence that is manifested physically in
true cievelopment of a transitional suburban ring and socio-
1-Ogi<:ally in the increased capacity of the city to affect

ecCDrlcnmic conditions, aspirations, and ways of life in an

54

ever widening hinterland." Hence as city influence ex—

pands, urban-rural differences in fertility contract. The
everitual outcome is post—industrial society. In speaking
Of this phase Abu—Lughod says:

The basic characteristic of a postindustrial society
is that urbanization and industrialization have become
so pervasive that their presence or absence within any
given geographic subarea is culturally "irrelevant."
Just as preindustrial cites absorbed their dominant
ethos from agrarianism, so postindustrial rural enclaves
derive theirs from urbanism. In neither instance can
simple criteria such as size of community, density of
settlement, etc" serve as reliable indices to values
or "ways of life." And just as preindustrial societies

‘-\l\_

53Ibid., p. 488.

54Ibid., pp. 488—9.

180

maintained a relatively stable demographic balance in
the absence of major urban—rural differences, the post-
industrial society appears to develop its own equilib-
rium despite urban—rural uniformities.

 

One of the essential elements of transitional theory
is the differential diffusion of contraceptive knowledge and
use from urban to rural areas. Some historical studies of
the long—term fertility decline have criticized this differ-
ent ial diffusion hypothesis. They have found that there is

little evidence of a delay in the reduction of rural fertil-
ity compared with urban fertility. Bash provides historical
evidence for the United States that indicates the secular
de-‘c-'L:'Lne in birth rates was a simultaneous process for both
urban and rural areas, though rural rates always appear

56

higl’ler than urban. Bash tries to explain this inconsis-

tency in transition theory by disregarding social structural
features of urban and rural populations. He suggests the
need to think of a "dominant value orientation" or some
peI‘Vading cultural factors of American society which influ—
erice—ed simultaneously the secular decline of both urban and

rural fertility. Bash suggests that:

\

55Ibid., p. 489.

56Wendell Bash, "Differential Fertility in Madison

Ouhty, New York, 1865, " Milbank Memorial Fund Quarterly,
3111 (April, 1955), 161—86; Wendell Bash, "Changing Birth
Ma tes in Developing America: New York State, 1840-1875,"
S;\lbank Memorial Fund Quarteriy, XLI (April, 1963), 161-:82.
Ste also Bernard Okun, "Trends in Birth Rates in the United
Hi etes since 1870, " The Johns Hopkins University Studies in
reޤytorical and Political Science, Series 76, No. 1,1958;

nd W. H. Grabill, C. V. Kiser, and P. K. Whelpton, op. cit.,
913- 16-19.

 

r (‘1

p-

181

. . . we should study more the culture within which a

particular kind of urbanism emerged. Then we might find
that changes in urban and rural birth rates, rather than
being responses to different cultural values, are differ-

ent responses to the same ones.
Efireedman, in commenting on this same finding, supplies

arnother explanation based on a more structural rather than

cultural analysis .

The higher fertility of the rural population—-and
especially the farm sector--has been well documented for
a long time. Recent analyses have added the important
conclusion that the long—run secular decline occurred
simultaneously in both the rural and urban sectors and
‘vas not primarily a direct consequence of the transfer
(of population between the sectors. Changes in the rural
asector, although undoubtedly linked to changes in the
lirban sector, accounted for a large part of the long—run
(decline. Probabiy changes in the rural sector were pro-
gguced hy its involvement in a specialized market economy
ggentered in the city, It suggests that the farmer need
riot go to the city to become urbanized. In various ways

1:he city can come to him.

It is difficult to accept Bash's proposal of a dif—
fereeritial response to the same cultural values, but the
rRDtlixon of a differential response on the part of urban and
ruill‘alpopulations is worth pursuing. It is puzzling as to
VVrVY' the urban-rural fertility differential could not be per-

<2eeifiaed as a differential response to structural factors

\

 

57Wendell Bash, "Changing Birth Rates in Developing
flueairica: New YOrk, 1840—1875," op. cit., p. 181.

3’1. 58Ronald Freedman, "American Studies of Family

I: ethning and Fertility: A Review of Major Trends and
Eseyues," in C. V. Kiser, Research in Family Planning_(Prince-

ju:?I1: Princeton University Press, 1962), p. 213 (italics
llle).

182

rather than cultural values. However, as it has been

pointed out above, the more frequent interpretation is to

perceive the process of urbanization as altering rural

social structure similar to urban. When social structure

becomes similar for both urban and rural areas, fertility
levels will converge. As Freedman has suggested, the rural

sector can become urbanized without becoming part of the

city because of " its involvement in a specialized market

economy centered in the city." Though this conclusion is

in line with urban dominance theory, it is definitely incon-

gruous with metropolitan dominance theory. Metropolitan

dominance theory emphasizes the notion of "specialization"

or differentiation. Hence the social structures within the

area of influence of a metropolitan center will tend toward
s'peeCIialization, not uniformity. In contrast urban dominance

t'he<:>ry hypothesizes that urbanization will affect rural

Social structure similar to urban social structure. Assum-
ing that fertility behavior is an adaptation to community
SOC ial structure, one concludes from urban dominance theory
that urban and rural fertility should eventually converge.
Th3 desire of this thesis, however, is to suggest the possi—
ility that urban and rural fertility are different re—
Spohses to the same process, metropolitanization, not urban-
i2ation. Assuming again that fertility behavior is an

adeptation to community social structure, we must conclude

from metropolitan dominance theory that urban and rural

183

fertility may never converge, because of the specialization
effect that metropolitan centers have on the social struc-

1:ures of urban and rural communities in their sphere of

iJifluence. Abu—Lughod hints at this as a possible develop-

nuant in postindustrial society which is characterized by
"Iralatively stable birth rates comparable in so-called urban
aIICI so—called rural areas, with significant intraregional

VEirfiiations due_primarily to ecological specialization within

tliea Inetropolitan complex."59 However she fails to develop

this idea further.

We must move away from the theoretical bias of urban

dCHntiriance theory. Freedman himself has suggested that demog—

raphers and sociologists alike have shared this bias.

further states :

He

this was the View that urbanization with its accompany-
:ing specialization and high rate of mobility inevitably
Vvould lead to a growth of secularism and rationality, to
'the declining influence of such traditional forces as
:religious faith, to a shattering of traditional family
ties and other primary group influences, to a growth of
individualism, and to the attachment of the individual
to larger, impersonal, and rational organizations. . . .
The dominant view among both demographers and sociolo—
gists was that as all of the population becomes closely
involved in an urban society, family planning would
become universal and the size of families planned would
continue to decline. . . . A continuing revision of the
older View of urban society since the war gives more
weight to the persistence of religious and other tradi-
tional allegiances. There is growing emphasis on the
persistence and even resurgence of the family and other
primary groups as the channels through which the larger

bureaucratic organizations reach the individual, in

59Janet Abu-Lughod, op. cit., p. 489 (italics mine).

184

larger measure. Urbanization and industrialization are
seen as leading to the reorganization of society in new
forms rather than to inevitable disorganization and mass
anomy.
Urban dominance theory suggests the emergence of a mass
society in which all traditional differences have disap—
pearecL. This is highly unlikely. As Freedman suggests,
societqris undergoing a reorganization and new patterns of
differrnmdation will emerge. In fact, it is possible that
mums former differences will persist if maintained by the
differtantiating effects of metropolitan centers on their
hinterfiland. It is proposed that the urban-rural fertility
differermial is one of these traditional differences that
will persist.
we must avoid inaccurate interpretations of ecolog—
ical<mita produced by this urban "bias" of urban dominance
theory; One pitfall in ecological analysis that continually
crops up) is how to interpret the "gradient." The gradient
PrinCiplra is based upon the work of scholars who have empha—
sized thee concentric zone and gradient effect of city influ-
ence on rairal areas. The gradient principle is consistent
with the concept of urban dominance.
This principle holds that at any given moment of time
urban characteristics are distributed in the satellite
rural area so as to form gradients of decreasing inci-
dence with distance from the city. In a dynamic sense
the influences of the expanding urban center can be
thoughtof as extending into the rural area in a gradient

Which declines in accord with diminishing communication
and transportation facilities. Stated in brief form the

x

6
0Ronald Freedman, op. cit., p. 215.

R
V
the;

o

 

 

 

+-

.3239
Thou
the
p02"

3 ti

‘
vuii
«MIL-‘1
(I.
yfi'b

ance fr
‘23“

I

.
“nu
.E

.C e
O;
., ,
~I at-
\
n...»
. .
gnbv
..
‘8‘
Y"?

 

C. T. .l .c n. .. Z T r

C. .. a : . .1 O 1 ...c a:

2. «J at P“ .C Do I. C m u 6 E D; ad
a: C NV \d h u A r O
.1 a u u .. C A c 9 C A \I s .
.fih 0 Q» - .V e Vi r G» c 2v

“an ..wH 11¢ ~H u u u “U. h 1 him

. i a . i :u nu. 3 ~11 :u
a,“ a... ..I L ... J ...: . .
nil .Tu v. i. W o H.- u - “hunk: NW. N

185

gradient principle reads: the extent of urban-influ-
enced changes in rural areas varies inversely with
distance to the nearest city and directiy with the size
of that city.eI

 

 

Though the gradient (essentially, then, a measure of

distance from a city) invariantly reveals a difference

between the urban and rural populations considered,62 the

 

underlying assumption as to what the gradient reflects is
generally the extent to which the characteristics of the
rural population resemble those of the urban population.
Given time and progressive urbanization, these differences
will disappear as the rural population approximates the
urban. Duncan, after noting that the gradient of urban
influence is more pronounced for the rural—farm than for the
rural-nonfarm and that it is steeper in less urbanized geo-
graphic divisions, concludes that:
Both these observations suggest that the effects of
urbanization on the rural population will appear less

pronounced, the more urbanized the national economy
becomes. The distinction between rural and urban,

 

6J‘Walter T. Martin, "Ecological Change in Satellite
Rural Areas," American Sociological Review, XXII (April,
1957). Reprinted in G. Theodorson, Studies in Human Ecology
(Evanston: Row, Peterson, 1961), p. 610.

62See for example such studies as Warren S. Thompson
and Nelle E. Jackson, "Fertility in Rural Areas in Relation
to Their Distance from Cities, 1930," Rural Sociology, V
(June, 1940), 143-62; E. T. Hiller, "Extensions of Urban
Characteristics into Rural Areas," Rural Sociology; VI (Sep—
tember, 1941), 242-57; Edmund de S. Brunner and J. H. Kolb,
Rural Social Trends (New York: McGraw—Hill, 1933), Ch. V;
and O. D. Duncan, "Gradients of Urban Influence on the Rural
Population," The Midwest Sociologist, XVIII (Winter, 1956),
27-30.

 

 

 

 

‘.3 i
-
‘ .
’r ha
5‘
«.-lc:
.
rczn»;
Own»..\..
A»
L c
Q
~:.: {:va“
.aa-‘1 V'»-‘
*.r~fll- =.._V_‘
-‘.~‘& ”uh-Iv
1" a
(. \“Rh K‘Y
v ~V¥.. '5‘
R
1 we‘“ -
«A
:‘u- ‘a‘. .
“K,
L
‘fl—a
‘ev..:. "(a

‘
."P:l N
.«M§-‘* r“ 2..
V M
. §
\‘(v 1 -
“Jioxlh 2*
‘0
u..
u.
“‘"I Pr! A
§ .kr
4“ s s A“
“5“. u , _
I -4“:

 

186

though important, is blurred. Galpin, among others, saw
this, when he proposed the neologism, "rurumicommunity."
The cross—sectional analysis here, together with a vari-
ety of supporting trend data, presents a prospect of
still further blurring. Sorokin and Zimmerman, twenty—
five years ago, concluded that in highly urbanized coun-
tries the magnitude of rural-urban differences had
reached a peak and was due to diminish.63

Others have made this same conclusion from gradient
data, even when using a metropolitan dominance framework,
which suggests, of course, a lack of awareness on the part
of such writers of the fact that urban dominance and metro-
politan dominance theories lead to conflicting interpreta-

tions. Consider, for example, an article written several

 

63O. D. Duncan, "Gradients of Urban Influence on the
Rural Population," op. cit., p. 30. Duncan's reference to
Sorokin and Zimmerman is from P. A. Sorokin and C. C. Zimmer-
man, Principles of Rural-Urban Sociology (New York: Henry
Holt, 1929), Ch. 27.

It is interesting that while Duncan calls upon
Sorokin to support the urban dominance hypothesis of blur-
ring urban-rural differences, he may also be called upon to
support the metropolitan dominance hypothesis of continued
differentiation. Sorokin said two years earlier: "urbani—
zation" . . . means only an approach of (phe rural world's)
characteristics to the characteristics of the urban world
but does not mean a complete obliteration of all differences
between them. In order that such an obliteration become pos-
sible it is necessary for the urban and rural areas to have
the same density of population, the same size of community,
the same homogeneity or heterogeneity of population, the
same occupational milieu and nature of the occupation, and
the same intensity and character of the interaction system.
Sorokin's statement from Pitrim A. Sorokin, Social Mobility
(New York: Harper, 1927), p. 624. Italics in original.
Quotation taken from secondary source: Leo F. Schnore, "The
Rural-Urban Variable: An Urbanite's Perspective," Rural
§ociology, XXXI; No. 2 (June, 1966), 132.

 

 

 

. . . w.
J r“ / .,. C 3 3 I . . .
m. m: .l n». . i .1 2. 3 ... C .
no. u“ t :1 “...: ....i sc.r.l~
r. .1 S .l E .3 n“ u. 3. E .r.
.1. ... n-.. ... .w. ~o. we r. vi v. n. C.
A: a: a: H o. v. V; «J .: Ac 2. ... .3
N .3. “a .C 3. A» r S r. 1; S :1
Va« .h... (>5 ...
A: we. a“ .3
1r“ av ... 3. .s
s c» .a« .~h ~u~

.... u“
we .3 ..M Y‘
e ...u Ci .2
.. .7. .C
T. Z. S m.
o. .1 it ..u.
C .n.. .l r...

P.,‘
‘5

.-. E T. t

3 .n... E

3 t C E
L~ v.

T .
n J O .3. (\ rC
t A:
e O C T. «L
..I‘ kk has .1 Q»
”H4 «Fu .1 V‘ .‘\
c. L. yr to. at
C. ... \u. 1. .J
\-5 .'\ 1.! be. 9“

187

years ago by Firey, Loomis, and Beegle.64 The authors use
'the metropolitan dominance concepts of "field" and "center"
to distinguish rural and urban and also speak of the role of
highways in bringing rural communities into an interdepen-
dent unity, a metropolitan community, in other words. They
state:

Briefly, highways are binding the field areas into
organic, functioning unities and subunities which sur-
round and tie in with centers and subcenters. Villages
serve as centers for little towns; towns are centers

for large fields; and cities function as centers for the
largest fields. Each field, with its center, is succes—
sively subsumed into the next large one, in hierarchical
fashion.

Thus there emerges a functional pyramid of field-
center "organisms," all bound together by a network of
highways. The height of the pyramid and the degree of
its functional unity is directly contingent upon the
number and layout of its sustaining highways.

The authors use a metropolitan dominance framework and
describe clearly the structure of a metropolitan community,
but unfortunately conclude the convergence of urban and
rural areas rather than differentiation. They continue:
Such a pyramid implies interaction between rural and
urban people, a reduction of their differences, a fusion
of their interests. . . . In between these centers and

subcenters, as well as out beyond them, in the areas
more truly rural, are farm families, whose new proximity

 

 

64Walter Firey, Charles P. Loomis, and J. Allan

Beegle, "The Fusion of Urban and Rural," in Jean Labatat and
Wheaton J. Lane, Highways in Our National Life: A Symposium
(Princeton: Princeton University Press, 1950), pp. 154-63.
Reprinted in P. K. Hatt and A. J. Reiss, Jr., Cities and
Society (2nd ed.; Glencoe: The Free Press, 1957), pp. 214-22.

65Ibid., pp. 215—17.

#

 

 

188

to the city, made possible by the highway, renders them
a little less rural and a little more urban than they
had been before.66

After noting the possibility of delineating with remarkable
precision the gradients which urban cultural patterns take
on in the rural areas contiguous to a city, they conclude by
saying:

Making due allowance for some exceptions and for some
degree of variability in the centers, the general prin-
ciple still seems to hold that rural areas, in direct
proportion to their proximity to urban centers, are
becoming culturally urbanized. Since proximity is con-
tingent upon time-cost accessibility between country and
city, itself a function of highways, the causative agent
in this urbanization of rural culture must be evident.
It is the highway that has brought city values, ideals,
and standards to the country dweller. Notions about
life objectives, about loyalties, about modes of living,
about consumption tastes, about well-being--all of these
are becoming more alike as between country and city.
While all this perforce means the loss of quaint, rustic
ruralisms, it means, too, the fuller integration of the
American people around basic and historic ideals of the
nation. More truly than ever before a homogeneous,
internally consistent, and universally accepted value
system, shared alike by urbanite and ruralite, is coming
to characterize American society. The role of the high-
way in effecting this cultural rapproachement between
country and city has been decisive.

It seems that a too hasty conclusion has been posited by

these writers: that increased integration of the rural and

 

urban populations necessitates homogeneity. But the notion
of the metropolitan community contains the integration of

heterogeneous elements into an interdependent whole. Indeed,

 

there are dynamic changes taking place between urban and

 

6§£E£Q-, p. 217 (italics mine).

67Ibid., p. 222.

.1.
‘1

7.

Av Tr.

..t.

..t.

.. I

.aa

A: ‘L

a:

C. i.e.lmi 3 L...

r\ it I

«C ~Ho

.C t

i

..f

:u

T.

 

 

189

rural parts of the metropolitan community, but this is not
necessarily a move toward the disappearence of urban and
rural differences. A reorganization of society by metropol—
itanism suggests the emergence or persistence of differences
in functional interdependence. Quinn, in his classic text—

book on human ecology, points out that homogeneity and func—

 

tional integration are two distinct bases for the classifica-

 

tion of substantive areas. In distinguishing these he says:

. . . a homogeneous area is characterized by similaripy
of differentiating attributes throughout its extent.

For example, a region may be characterized by a hot, dry
type of climate, by rugged topography, by a distinctive
type of population, or by a characteristic culture.
Whatever attributes have been selected for purposes of
delimitation, these attributes must show sufficient
similarity and importance throughout its entire extent
to give character to the area, and at the same time
sufficient difference from adjacent territory to mark it
off as distinct. In contrast, the integrated area
typically includes contrasting_parts organized into a
larger areal unit. For example, the diverse sub-areas
of a metropolitan region, including the agricultural
hinterland with its town, village, and farm communities,
and the metropolis with its business center, factory
districts, residential areas, and satellite suburbs,
together constitute a functioning area unity. The chief
attribute of such a metropolitan region is that of inte—
gration itself. The contrasting types of sub—areas are
welded together in a complex where larger unity itself
makes the metropolitan area distinguishable from adja-
cent territory.

 

 

 

 

Finally, Bogue seems to suggest that urbanization, characters
ized by the emergence of a metropolitan economy, will not
have the effect of diminishing geographic differences, but

will actually enhance them. He asserts:

 

68James A. Quinn, Human Ecology_(New York: Prentice-
Hall, 1950), p. 39 (italics mine).

 

.» Ce
5 at
L” .1
w t

A:
w».
n\~

u.
«N.
C
“V.

. v

“.V.
O
nu.

L'

§

”57'

b

..\-~¢

INA-“bus .
\’ Uv-l-l‘qr l ‘-
0 5

val-‘6."
‘&\-\$

-
a.
‘.

., 1. .
“‘Van n»
\A\.

‘-
\-

Q31!
”snce

ere:

.-
‘
es

N .
u

190

. . . the notion that progressive urbanization and
industrialization will cause these distinctive clusters
to disappear cannot be supported either by theory or by
observation. In fact, the contrary appears to be true.
A market economy seems to cause places to seize upon
whatever unique sites, location or physical characteris-
tics they have that may provide the basis for profitable
specialization. . . .

Thus, instead of minimizing geogrophic differences,
the modern metropolitan economy may emphasize them. In
this process, interregional differences that once were
large may disappear, while new differences may appear.
Instead of becoming a homogeneous mass, industrialized
populations tend to become a patchwork of specialized
populations tied together by a geographic as well as
intracommunity and intercommunity division of labor.6

 

In conclusion to our discussion of urban dominance
theory, we must propose that urban—rural differences in
community social structure will persist. Assuming the eco—
logical principle established previously, that fertility is
a function of community social structure, we must also
insist that urban-rural fertility differences will persist.
Urban dominance theory does not lead to this conclusion,
since it must posit the eventual blurring of urban—rural
differences. We must conclude also that fertility studies
which assume an urban dominance theoretical framework do not
have a sensitive research design essential to understanding
possible differences which may exist in urban—rural fertil-
ity behavior. To understand why urban-rural differences
will persist, we must move to a consideration of metropoli—

tan dominance theory. Though much has already been said

 

69Donald J. Bogue, "Population Distribution," op,
cit., p. 396 (italics mine).

191

concerning metropolitan dominance theory in contrast to
urban dominance theory, let us proceed to systematically
consider the elements of the former as the preferred theo-
retical framework for the analysis of urban—rural fertility

differences.

Metropolitan Dominance Theory

 

The metropolitan community (or region) is becoming
an increasingly important form of organization in modern
industrial society. We have established previously the
expectation of an orderly spatial distribution of population
characteristics within given geographical areas. An essen—
tial assumption in metropolitan dominance theory is that the
metropolitan center is a primary organizing agent which pro—
duces the spatial distribution patterns of community social
structure within the metropolitan region or the sphere of
influence of the metropolitan center. The economy of the
metropolitan community is viewed as "the characteristic and
dominant type of modern social and economic organization."70
It is assumed, then, that American society is a "metropoli—
tanized" society. Our present day society operates in terms
of, and is conditioned by, the metropolis. The metropolitan
economy is the modern form of social organization by which

man makes effective use of his advanced technology. The

 

70Donald J. Bogue, The Structure of the Metropolitan
CommuniEy, op. cit., p. 8.

 

-..
....

WV-

«p
t.
.3

pl.

Er.

In 1 ‘
C Li

.Ql‘

“1..

«G ...H Mix
c AU “U4.

192

pervasive influence of metropolitan centers has spread to
such an extent that "the entire area of the United States

may be broken down into a series of areas, each of which is

dominated by a metropolis."71

If we are to understand how the metropolitan center
orders the Spatial distribution of community social struc-
tures, we must understand first the nature of the metropol—
itan community. The general spatial structure of the metro-
politan community was described by R. D. McKenzie some time
ago:

The metropolitan region thus considered is primarily
a functional entity. Geographically it extends as far
as the city exerts a dominant influence. It is essen-
tially an extended pattern of local communal life based
upon motor transportation. Structurally, this new
metropolitan regionalism is axiate in form. The basic
elements of its patterns are centers, routes and rims.
The metropolitan region represents a constellation of
centers, the interrelations of which are characterized
by dominance and subordination. Every region is orga-
nized around a central city or focal point of dominance
in which are located the institutions and services that
cater to the region as a whole and integrate it with
other regions. The business subcenters are rarely com-
plete in their institutional or service structure. They
depend upon the main center for the more specialized and
integrating functions.

As McKenzie suggests, the rise of the metropolitan
community was made possible by the emergence of rapid trans—

portation, especially motor transportation. Prior to the

 

71Ibid., p. 13.

72R. D. McKenzie, The Metropolitan Community (New
York: McGraw—Hill, 1933), p. 70.

q.

C A .2 l C c.“ C . S O N.
G o o . .1 . , r. 0 ML e .s a S r;
w. L. 2. 5 ~. 7:. u . 1(a fig Cu 7. « LI r e .3 Qt
... c" :u . O . . ...». o. t E :1 a-..
.2 f .n. J 5 a no 3 C. wt. mt. Min 4w. one. :5. Wu t .1 a a
la a . a o‘ H :3 .l c 7 .A ‘Lv 1* ~,Q .1 S in
“w 3.. it at «we. .... C MW M» .5 a C u. 0 S F. Au -1 t
a c _ .. o. u . cc. 3 .1 . .
. .. ,. .... H s c M... m f.

 

193

development of modern highway systems, in the waterway and
railroad eras, the city was in very large degree autonomous
of its own rural hinterland. "But the highway has changed
all of this. Because of the peculiar superiority of the
automobile as a short—distance, small load carrier and as

a 'free—agent' whose course and destination need not be
confined by water ways or railroad tracks, the advent of
highway transportation has meant, for the first time,
intimate contact between a city and its hinterland."73
Hence motor transportation has opened up the hinterland in
such a way as to increase the interdependence between the
central city and surrounding hinterland populations. Inte—
gration or interdependence, then, is the essence of the

metropolitan community, i.e., the functional interdopendence

 

of metropolitan center and hinterland. As McKenzie states:

 

The super community therefore absorbs varying num-
bers of separate local communities into its economic
and cultural organization. In this pattern a dominant
city--that is, dominant relative to surrounding settle—
ment, functions as the integrating unit. . . . In
other words, there is developing within the United
States . . . a pattern of settlement which may be desig—
nated as city regionalism. This new city regionalism
differs from the regionalism of former times in that it
is a product of contact and division of labor rather
than a mere geographic isolation.74

 

73Walter Firey, Charles P. Loomis, and J. Allan
Beegle, op. cit., p. 215.

74R. D. McKenzie, op. cit., p. 313.

I . . .. J ....
v. c. e C. I 3 ... l.
.o .
.... e a u S R C ..C “O
. .. ...U .. . m“ T. n u 0..
. s l a: .
woe ~ 1 C .... «w .r..
1. .1 F
. m

 

 

194

But the notion of integration or interdependence of
elements with the metropolitan community implies "a division

of labor," i.e., a differentiation or specialization of func—

 

tion by place. "The modern metropolitan community, unlike

 

the pre-motor city, obtains its unity through territorial

differentiation of specialized functions rather than through

mass participation in centrally located institutions."75

Bogue, in describing the function of cities in the metropol—
itan community, differentiates between "metropolitan centers"
and "hinterland cities." Commenting on the function of the
metropolitan center he says:

The metropolis is usually the largest and most complex
(the farthest removed from the "average" city) of all
the cities in the territory. Because it is able to
assemble cheaply a varied array of raw materials and
products from all parts of the world; because a larger
number of specialized components and skills are required
in the production of the goods required to sustain human
beings at their present level of living; because up to a
certain point machine production increases in efficiency
with an increased scale of operations; and because cer-
tain mutual benefits appear to accrue to business enter—
prises from their location in proximity to each other,
the large city is able to produce and distribute more
varied goods and services than is a smaller city. The
more specialized the goods, and the more the goods are
amenable to mass production, the greater these industrial
and commercial advantages of large cities seem to become.
From these facts it has been concluded that the metrop-
olis, or modern large and complex city, exercises an
organizing and integrating influence on the social and
economic life of a broad egpanse of territogy far beyond
the civil boundaries, and thereby dominates all other
communities within this area.76

 

 

 

75Ibid., p. 71.

76Donald J. Bogue, op. cit., pp. 6-7 (italics mine).

~

53*-
y»--”
use“,
”the"
r.«*“

id A

... a.“

«L .

.mv la
.

... “J

a. .5

«i

‘iej

F‘rke!

..

195

The metropolitan community, then, is really a network of
smaller communities, rural and urban, distributed in a
definite pattern around a dominant city, and bound together
in a territorial division of labor through a dependence upon
the activities of the dominant city. It is the metropolitan
center, then, which is the organizing agent of inter—commu—
nity differentiation. Communities lying in this region
about the metropolitan city, drawn into a division of labor
with this center, exchange for specialized goods and ser-
vices of the metropolis such other products as can most
effectively be produced from the resources in their immedi-
ate locality. All subordinate communities become dependent
upon metropolitan markets, including farm operators who
regulate their activities to produce those products which
will yield them the greatest return in the metropolitan
market.

But the metropolitan center is not just an economic
center; nor does it influence only the economic activities
of its hinterland communities. With the exchange of mate-
rial goods there is also an exchange of ideas and human
values. "The metropolis appears to have become the focal
point not only of our material activities, but of much of
our moral and intellectual life as well."77 Bogue, in his

attempt to demonstrate the pervading dominance of the

 

77Donald J. Bogue, op. cit., p. 6.

metropolis,
activities ,
:raée, serr
hwe'xcer, fr

| I
“ "LQYV‘ - ’-
VaLV-VAA... Anl»

less clear]

L, ‘3 tL

Ffi~l ~
.JAblaer ‘1‘:
5“

N
511’s Ha
\. 5‘
‘K WY
q‘uifi:
t':+
.. ‘

Ik‘ h I
.mEr 1P 1
A f‘
0.1“-
"is l‘r-y‘
O‘jn CFC:
¢.\‘
:flr‘;
nu“...
.dl Stru
1‘32}
«lLan er,
\.

196

metropolis, concentrated on what he termed "human sustenance
activities," viz., the functions of retail trade, wholesale
trade, services, and manufacturing. Bogue does not deny,
however, that dominance of an equally intense or identical
pattern might operate in other human activities which are
less clearly related to human sustenance. In fact he defi~
nitely feels that ”many other conditions of life undoubtedly
are subject to control or modification by the central city.
The complete structure of the metropolitan community may
include the functions of finance, government, education,
religion, and innumerable other aspects of the institutional
composition of the individual hinterland community."78
Though the notion of metropolitan dominance in its original
conception contains many economic overtones, it is clear
that the theoretical framework can and must be employed to
consider the organizing effect the metropolitan center has
on all activities of hinterland communities. The all-inclu—
sive nature of metropolitan dominance permits the assumption
that almost any community activity is influenced by the
metropolis. These comments suggest the need to consider the
manner in which fertility behavior is influenced indirectly
through the organizing effect the central city has on the
social structure of communities which lie within the metro—

politan region itself.

 

781bid., p. 61.

_

l"+.h!.v~ ' ‘n
“11...“. C..

. ~ a;
C
0. s45 51.

,
.rzctlons .
“'0

197

But metropolitan dominance cannot be understood only
in terms of the function of the metropolitan center. The

interdependence of central city and hinterland must also be

 

recognized. The high density of population in the central
city precludes the possibility that food stuffs for the
population or raw materials for industry can be provided
within the central city. It is impossible for the city to
be self-sufficient, hence, the need to consider the function
of the hinterland. Hawley assumes this distinction when he
describes the community as "comprised of two generalized
unit parts, the center and the adjoining outlying area. In
the one are performed the processing and service functions,
and in the other are carried on the raw-material producing

functions. The two develop together, each presupposing the

other."79 Bogue in his investigation of the metropolitan

community also emphasized the interdependent relationship of

the central city and its hinterland:

The one situational factor which Gras80 holds to be
absolutely essential for the development of a city
aspirant to metropolitan status is the possession of a
hinterland, "a tributary adjacent territory, rich in
natural resources accompanied by a productive population
and accessible by means of transportation." We are
warned by Gras not to overemphasize either the metrop-
olis or the hinterland in considering the metropolitan
organization.

It is true that in studying this organization we are
inclined to emphasize the great metropolitan center, but

 

 

79Amos Hawley, op. cit., p. 245.

80N. S. B. Gras, An Introduction to Economic History
(New York: Harper, 1922), quotations from pp. 185—87.

 

t.
.3
v.
n J

72.5;

’ ,

J5

‘0

Ac

5 a

‘3 LL

3. xfiu .

-;N II.

at

Q. d
a:

Win

... c

v‘

o: it
A.» «\~

.2
«C

3 .1 c
by; .l

. . C
... S\ O
.5 (... C. 3.
2..
.VA
to
A:

198

to forget the large dependent district would be fatal to
a correct understanding of the subject. Perhaps, indeed,
it is somewhat incorrect to speak of the area as depen-
dent upon the center, for though that is true, the center
is also dependent upon the outlying area with its towns,
villages, and scattered homesteads. Interdependence of
parts is really the key to the whole thing.81

 

Though this caveat was voiced rather strongly by
Gras and Bogue, there seems to be some evidence that it has
not been adhered to rigorously by others. Grigg claims:

The proponents of this position (the interdependency
of central city and hinterland) emphasized one aspect of
this principle of dominance—~the function of the center-—
and have ignored the second--the hinterland. They have
taken Gras's statement of the relationship of the two
and have failed to consider the broader implications of
his writings. Gras points out the necesSity of a hinter—
land for the existence of a metropolitan center, but at
the same time he insists that you cannot reify one at the
expense of the other. Indeed, implied in the writing is
the injunction that extent and degree of functional inte—
gration imposed on the hinterland by the metropolitan
center is an empirical question wanting to be demon—
strated rather than an ad hoc assumption to be treated
as a reality. Thus, the ecologist finds himself in the
position of stating dominance exists because metropol—
itan centers exist ipso facto; and the concept of the
hinterland is dragged along behind, not because they
(pig) exist, but because it is obvious Ehat each metro—
politan center must have a hinterland.8

 

Grigg calls for more intensive consideration of and possible
refinement of the concept of hinterland in metropolitan
dominance theory. It is essential that the hinterland be
given due consideration not only because of its interdepen-

dence with a metropolitan center, but also because the

81Donald J. Bogue, op. cit., pp. 7—8.

82Charles M. Grigg, "A Proposed Model for Measuring
the Ecological Process of Dominance," Social Forces, XXXVI,
No. 1 (December, 1957), 628.

 

Fm "ance
“var-o.

:zanizati

have some

» ' ‘
+1fl!‘ 1 - fi
~‘V00 -4; L..
I“ " 1
n" " .V‘n
“AaH JCUR‘
. Q
7 n
2.1‘Vnnfi
s».:-: d—e

. .
“‘u‘ ‘ -
..- er ‘v-p‘

Ct. -‘ ‘.U

(1.)
(...)

199

dominance of a metropolitan center is reflected in the
organization of the hinterland and the hinterland itself may
have some bearing on its own internal pattern of organization.
We shall consider this point later.

At this juncture we may assert that the theory of
metropolitan dominance states that the metropolitan center
more and more controls the conditions of life of the popula-
tion in the areas surrounding the central city. The hinter-
land populations, as a result, are spatially organized with
reference to the metropolitan center. We should next con—

sider the concept of dominance, the means by which the

 

hinterland is controlled.

Bogue considered the concept of dominance to a great
extent and suggested that "dominance, in its ecological
meaning, is a special kind of control over a community of
interfunctioning units."83 Simply put, the metropolitan
center establishes and controls the conditions of life which
set limits to the activities of the other communities in
its sphere of influence. The net effect is a multiple—
community complex, a constellation of communities, which may
be termed the metropolitan community in deference to its
dominant central city. The social structure of each of these
communities will be an adaptation to the conditions estab—

lished by the metropolitan center.

 

83Donald J. Bogue, op. cit., p. 10.

9

a

nu»
Arab

A—uqnc“
VI «~‘\A¢O

C

. - ~ A C
a a“ ”(do
1" h. ~

:~
0
K

I s
I
‘4

‘3

Wu

3..
.1
+5

3
...H
h...

I

Cy

v.
\Q

I

.ma

rm...

. ‘

r
s:
7V
Y

n
E
C

:

200

This is the nature of dominance, but how does one
operationalize it? Bogue suggests that if a definite non-
random distribution of attributes can be demonstrated, pat—
terned with respect to the metropolis, it can be inferred
that the distribution of these attributes is "controlled"
to some extent by factors associated with the metropolis.84
But dominance is not to be considered a fixed attribute,
but a variable. The amount of control exercised over the
communities of the metropolitan region will vary with the
size of the dominating center and the accessibilioy_(dis-
tance) of the hinterland community to the dominating center.

Accessibility is assumed to be a variable which
covaries with dominance. The task of delimiting areas of
like degree of dominance is a problem of delimiting areas
of like degree of accessibility. The metropolis is a
metropolis because of its superior ability to serve and be
served by the hinterland.

In terms of time, cost, and expenditure of energy the
entire area can enter more easily into a division of
labor with a city located at a highly accessible point.
Exchange and interaction with a city located at the
most inaccessible point could be achieved only at a
maximum expenditure of time, cost, and energy. Since
time, cost, and energy are all elements of life which
must be conserved in order to ensure most economical
survival, it can be reasoned that, since these elements
vary with accessibility, the following assumption can be
made: varying dogrees of accessibilipy must represent

varying degrees of interaction with the metropolitan
center.85

 

84Ibid., pp. 14—15.

85Ibid., p. 21.

L
L

Y:‘“r\
Q F‘s-l

an
USO
to

'1
V

\ iIC)

.

C

H

1:”
duo.“
pant
A-Ulob
L
L

. 337.13

.1

.a
C

A‘bv.
3"
‘5..(

Q...
“I.

.r-‘w‘Q"
D

00
o

‘
fin
u..

0

tut
h
at

Zatio

1'
§

larger

201

Because of the importance of motor vehicle transportation in
the development and establishment of dominant metropolitan
centers, distance to be traveled may be expected to covary
with accessibility. The distance to be traveled limits the
opportunity to transport goods, services, and persons from
hinterland communities to the metropolitan center. "One
permanent requirement for changing the location of any
object is the necessity of overcoming distance."86 Hence,
it is understood that varying distance of a hinterland com—
munity from a metropolitan center reflects varying degrees
of dominance.

Size, as an attribute of the dominating center, is
also an indicator of dominance. It is expected that the
larger the metropolitan center, the greater will be its
organizing influence on the hinterland, the closer the
integration of the central city and its hinterland, and the
larger will be the hinterland area which can be effectively
influenced by the metropolitan center. Increases in the
size of a population are related to the degree of special—
ization attainable by that population. Hawley points out
that "population size imposes limits on both the extent of
specialization and the number of different activities that
may be carried on simultaneously. . . . In a small popula-

tion the degree of specialization of activity is necessarily

 

86Ibid.

slight. On- th

creases the ex

11:35 this at

I n
~ .
”171“.” an” ‘ V‘
us‘..;,u div 1‘.
.4
AA
u n N!
- .
5 e a up
‘1 ‘3? Ira
oesn‘h‘,“u..u.

202

slight. On the other hand, every increment in size in—
creases the extent to which specialization may be devel—
oped."87 Bogue, in describing the metropolitan center, the
largest city in the metropolitan region, emphasizes its
greater capacity for specialization of activities. He then
links this attribute of the metropolitan center to the orga—
nizing and integrative influence it exercises over the
hinterland.88 Hence the larger the metropolitan center,

the greater its specialization powers, and the greater its
integrative effect on the hinterland. Therefore in opera-
tionalizing the concept of metropolitan dominance both dis-
tance from the metropolitan center and size of the metropol—
itan center should be taken into consideration.

Now at this point we may briefly summarize by saying
that the metropolitan center and its hinterland are func-
tionally interdependent. But the metropolitan community is
a network of communities, including both urban and rural
populations, which are territorially differentiated by the
dominant influence of the metropolitan center. The degree
of dominance of the metropolitan center over its hinterland
communities is a function of size and distance. We have
also established the fact that the metropolitan center has

a controlling influence, not only on the sustenance

 

87Amos Hawley, op. cit., p. 122.

88Donald J. Bogue, op. cit., pp. 5—6.

activities of
social struct:
or: the whole ;
reconsider ths
Warlae '

Aauv ‘,u.

P I
OI COIITCJIHU’

V
A

"V‘u
30». (.e 85:011.:

.5 y -
.63.: ‘ {wry-n _
ahu L U‘ a
\

 

203

activities of hinterland communities, but on the entire
social structure of such communities, and perhaps indirectly
on the whole gamut of community activities. We must now
reconsider the effect of the metropolitan center on its
hinterland. What patterns of territorial differentiation
of community social structural variables might we expect to
accrue among hinterland communities?

The most common type of analysis of the influence of
metropolitan centers on hinterland communities has been in

terms of gradients (concentric zones) which extend out from

 

the metropolitan center. In such studies it is usually
hypothesized that the hinterland will be spatially organized
with reference to the metropolis and that this organization
will manifest itself in a series of gradients in the charac-
teristics of the population along the dimension, distance
from the metropolis. The fact that such gradients exist has
been documented by numerous studies emphasizing a wide vari-

ety of indices of population characteristics.89 Unfortunately

 

89A selected list of such studies might include the
following: Theodore R. Anderson and Jane Collier, "Metropol—
itan Dominance and the Rural Hinterland," Rural Sociology,
XXI (June, 1956), 152-57; Edmund deS. Brunner and J. H. Kolb,
Rural Social Trends (New YOrk: McGraw-Hill, 1933), Ch. V;
Otis D. Duncan and Albert J. Reiss, Jr., Social Characteris-
tics of Urban and Rural Communities, 1950 (New YOrk: John
Wiley, 1956), Ch. XIII; O. D. Duncan, "Gradients of Urban
Influence on the Rural Population," The Midwest Sociologist,
XVIII, No. 1 (Winter, 1956), 27-30; 0. D. Duncan, "Note on
Farm Tenancy and Urbanization," Journal of Farm Economics,
ZXXXVIII, No. 4 (November, 1956), 1043-47; John Stoeckel and
J. Allan Beegle, "The Relationship between the Rural-Farm
Age Structure and Distance from a Metropolitan Area," Rural

 

I‘-

l

ltc

.

I033

O
1.:
u‘v "

"I.
A.

0m 1

.0

o

2.—

.75
I l

V

‘

OI COLE] m

p
o
d

ices n

q

813

ray-‘0.
5“ \—

*Ce 83'

5h

..
3.

nul

OI DO

:4“

ch

12:”:

311Gb S.

 

204

the test of the gradient pattern in the hinterland of a
metropolitan center is a test of only p§£E_of the metropol-
itan dominance theory. In fact, the gradient test is a test
for both metropolitan dominance and urban dominance theory,
and does not differentiate between the two. Gradient
studies usually conclude that the hinterland population
manifests a pattern of decreasing (or increasing) incidence
of population characteristics (depending upon which vari—
ables are employed), as one moves away from the central city.
Empirical findings of gradient studies have been used to
determine, for example, the territorial extent of the influ-
ence exerted by a central city, or the average of a number
of central cities; the difference between rural populations
located adjacent to cities and those removed from cities,
with the amount of difference indicating how much the
strength of urban influence is conditioned by distance; how
the gradient of one particular characteristic may be differ-
ent from that of another, with such differences suggesting
the spheres where urban influence is most pronounced; or how

much size conditions the amount of urban influence exerted

 

Sociology, XXXI, No. 3 (September, 1966), 346-54; Harold F.
(Goldsmith and James H. Copp, "Metropolitan Dominance and
(Agriculture," Rural Sociology, XXIX, No. 4 (December, 1964),
385-95; E. T. Hiller, "Extension of Urban Characteristics
into Rural Areas," Rural Sociology, VI (September, 1941),
242-57; and Warren S. Thompson and Nelle E. Jackson, "Fer-
tility in Rural Areas in Relation to Their Distance from
Cities, 1930," Rural Sociology, V (June, 1940), 143-62.

 

 

 

surrou:

01‘.

ath

I

etrop:

sin

.1. o.
a h .nu
.M +.
.. at
an at
flu a u
.-. .2
oL . n

‘ -
r5
5 a- s C

I.

.-
‘t-
.\r

V: d .0 \
my ad
w“ n1.
A:

. . .5
AL .
...ifi

Z; a»
A—v .
odd C

15 Co 5/

I Q» «luv

. in) -~« 6
a: 7 in;

205

on surrounding rural territory when gradients are examined
in relation to the size of cities.

These studies tell as much of the gradient_principle
of metropolitan dominance theory (and urban dominance theory
as well) but little in terms of the principle of differentia-
utign, which is also an essential part of metropolitan domi-
nance theory (but not of urban dominance theory). An excel—

90 draws especial attention to this

lent article by Martin
deficiency in metropolitan dominance research. This arti-
cle is not an empirical study but an attempt to summarize
and synthesize what research has been done dealing with
ecological changes taking place in the rural sectors of
satellite areas (hence, he does not consider the entire
range of hinterland components of a metropolitan community).
Martin uses the two broad principles of gradient and differ-
entiation as "organizing devices" for ordering the results
of recent studies. With respect to the gradient principle,
"the extent of urban influenced changes in rural areas
varies inversely with distance to the nearest city and

91

directly with the size of that city," he finds a wealth

 

90Walter T. Martin, "Ecological Change in Sattelite

Rural Areas," American Sociological Review, XXII (April,
1957), 173—83. Also reprinted in George A. Theodorson,
Studies in Human Ecology (Evanston: Row, Peterson, 1961),
pp. 607-16.

91Ibid., p. 610.

\
AC,
(1.;5

p.
v!

V‘
s
n
.1
‘
V‘,
to
2‘58.‘
‘1»;
..

c0
accivi‘
i
0

.cc

.
‘

de‘

.
.
QQ.
‘\
5'0va
fi“

{no ’10

of extirl
5.5% “y

r. ,‘
r. nu \Q ..u

5m ...u. 1 CU. MG
e S

.0. r it T. +.t 3 0. 9...!“
. 3‘ 2w «.ma ... LL no. C Q» nu MYCI .|« No &L
in ...i .... ...... m a e e e .1 i e c r. ....
I ~ .s . 3N: l 2 s .. . i
at .l .. e enure‘il‘a
.1 ..
YA pan
.9» a»

206

of empirical studies which document gradient patterns for
the deconcentration of industry, population, and business
activities, occupational composition, rural land values,
nature of the farming enterprise, income, age and sex com-
position, fertility, educational achievement, and partici-
pation in urban activities. The second principle, differ-
entiation, is to be considered complementary, rather than
contradictory, of the gradient principle, although "the two
are partially independent in the sense that demonstrated
tenability of the gradient principle provides no basis for
evaluating the differentiation principle; on the other hand,

acceptance of the second indicates that the first holds

true. . . ."92 What Martin says of the principle of differ-

entiation is crucial to this thesis:

The second principle (differentiation) . . . holds
that these influences are extended selectively rather
than diffusing uniformily, and that the over all effect
is to transform the homogeneity of the rural territory
into an urban-like heterogeneity with specialization of
labor, differentiation of subareas, and functional inter-
dependency of parts. In spite of the almost complete
lack of research concerned with this principle, it would
seem to have as much or greater implication for changes
occurring in rural areas than does the gradient prin—
ciple, which has been dealt with so frequently. The
differentiation principle concerns the dynamics of the
relationship between the rural and urban sectors of the
economy, and the increasing integration of rural areas
into the great regional urban complexes. While this
idea has been stressed by McKenzie and others, there is
a surprising lack of empirical research.

 

..J r\ 1. 4| .-. v. ¢\ .\ .L .. a. 7.. x:
.l r LL .1 at 3; AV .c ... .3 nu C» «a O .hd vi +» a»
O L . ow .W O nu C. C n . an at a at r C L t h... .C
D. A: n .nu .1 My n, ..n L» u“ 04]“ \C ~ u r 3» .1s Zn
.» an .. . g i. . 1| 4 w o F a: .. LI. .
A: m c A: r; .3 . t a - a1. CL LL ...y .1.“ . W“ 1an T ML.
«r- .... .nd V- O at «C .t h» “A an n ‘L No F. L t :4 hi. .3
O b b e no . t T“ .1¢ A.» .hd .l -l «\U C 1‘ Ad s i 0 IBM

T.

y
k

‘1.
l!“-
:50

liCor

i

207

The point Martin makes, then, is that an important element

of metropolitan dominance, i.e., the differentiating effect
the metropolitan center exerts on the hinterland communities,
even within the same distance zone from the central city,

for both urban and rural communities, is in need of empirical
testing. The crucial issue to consider at this stage of
metropolitan dominance research, then, is not that metropol—
itan centers exert a controlling influence over hinterland
communities or that its influence wanes as distance increases,

but that the metropolitan center exerts a differentiating

 

influence, transforming both urban and rural communities
into a functional interdependency of specialized parts.
Studies using the gradient principle tend more toward a
comparative study of the metropolitan center and its hinter-
land communities. The differentiation principle, on the
other hand, requires a different dimension of comparison.

If metropolitan centers do exert a differentiating influence
on hinterland communities, then an important comparison to
be made is between different classes of hinterland commu—
nities. Martin indicates that he is aware of only one study
dealing with this problem. This study, by Kish, classified
incorporated places by distance zones from the central city
and demonstrated conclusively that for a variety of varie
ables the cities of the distant zones made up a relatively

homogeneous universe while those in the inner zones were

«.3
n3
ct
ck

C‘Yfi“

ystvv
03“ us

.nyy. u:

‘Hba‘by
m

1‘1‘AFR
sO‘Y'

more c
h

C Co.

F.
ll

\q
wlt

.
hr-

#
LC

“ 3 1e

g,

208
highly differentiated on the same counts.94 But Kish's
study dealt only with incorporated places and excluded rural
communities. Furthermore the incorporated places were those
found only within the immediate metropolitan ring to the
exclusion of those lying in the more outlying areas. In
concluding his article Martin states that:
. . . it seems highly probable that the rural sectors of
the satellite areas, like the urban sectors, more and
more consist of well differentiated subareas as the in-
fluence of the central city is extended increasingly
throughout the larger metropolitan area. The changing
patterns in population density, age and sex composition,
occupational composition, and land values, to name a few
characteristics, are societal adjustments in the satel-
lite areas to the evolving spatial organization of the
metropolitan community.

Martin's comment clearly suggests that further com—
parisons of the differentiating effects of metropolitan
centers must concentrate on the differences that accrue
between urban and rural hinterland communities. Hence we
should expect that metropolitan centers exert a different
impact on urban hinterland communities than on rural hinter—
land communities. On the basis of these comments we should
be able to formulate some hypotheses concerning the differen-

tiating influence which metropolitan centers have on their

respective hinterlands. Recall that earlier in this chapter

 

94Leslie Kish, "Differentiation in Metropolitan
.Areas," American Sociolggical Review, XIX (August, 1954),
388-98. '

95Walter T. Martin, op. cit., pp. 615-16.

«G
A:

C A“ I
U..b.'.

3»

..l
T.

‘V- 9

PA- A
UVVA.

ual-T.

...Sr b -

.l
.c

L .

«G

h.§_"
‘~“\.i

~
“9"
c.“

a

209

we accepted a major assumption of ecological analysis, that
community social structure indicators will manifest a defi-
nite nonrandom distribution pattern in social space and that
fertility behavior, operationalized on a residential area
level, will be a function of community social structure in
both urban and rural populations. With the introduction of
metropolitan dominance theory, we now assume that this non—
random distribution of community social structure indicators
is due to the pervading influence of metropolitan centers in
their respective hinterlands. Metropolitan centers, then,
are the determining agents of the social and economic orga—
nizational makeup of both urban and rural hinterland commu-
nities. It is assumed that, given the differentiating
effect of metropolitan centers, community social structure
in the urban hinterland will manifest different patterns of
specializing than the rural hinterland. Accepting distance
and size as measures of the dominance of metropolitan cen-
ters, and the differentiating effect of metropolitan centers
on their hinterlands, it is hypothesized that:

1. Community social structure is a function of distance
and size of a dominating metropolitan center.

a. Distance and size of a dominating metropolitan
center will manifest a different impact on com-
munity social structures in the urban hinterland
than in the rural hinterland.

2. Fertility behavior is a function of community social
structure and distance and size of a dominating
metropolitan center.

a. Community social structure and distance and size
of a dominating metropolitan center will mani-
fest a different impact on fertility behavior in

the urban hinterland than in the rural hinterland.

‘.
n“
.329

metrosoxl
6

pl-
.4.-C*
.
n: rura
.ys.‘.|.L.

envy“. w

q
9
H

.
~~.‘Ar

>1.

:.
Wu

C»

v.

U

1' to L‘.

Ml CO”'
uterla:
QM

yg '

Vs
.Lb
‘
[1'
Oo§y
A
'5‘
*U\.
i h
N

“VJ.
(A

.

.
"vn.
O

210

To this point we have discussed various aspects of
metropolitan dominance theory. Our hypotheses emphasize the
differentiating effect of metropolitan dominance on urban
and rural hinterlands. If our discussion of this theory is
terminated at this point, we shall be in danger of violating
the caveat expressed earlier, that of overemphasizing the
role of the metropolitan center and overlooking the impor—
tance of the hinterland. Let us reconsider the place of the
hinterland in metropolitan dominance theory.

If we return to Bogue's study of the structure of
the metropolitan community, we will capture an important
point often overlooked by other studies of metropolitan dom-
inance. Many such studies grant the attribute of dominance
only to the metropolitan center. Bogue, emphasizing the
point that the metropolitan community is a "community of
local communities," constituting a central city and several
hinterland communities, recognized that "all local communi—
ties in the metropolitan community are considered to possess
some degree of dominance over some portion of the hinter-

land."96

For this reason he found it necessary to adopt as
his unit of analysis the individual local community, since
each community exercises some influence within the hinter-

land. In this connection he stated:

The hinterland contains a great variety of communi-
ties, ranging from cities of more than 100,000 inhab-
itants to small villages and local farming communities.

 

96Donald J. Bogue, op. cit., p. 30.

.1 fly «.aa Pu «a Po 1_ . A: .3 Y. Q»
_ u T. 3. .l o. ~U

«v ad a: C . a ..

v . .l . t . i. E :u

a. . . . .D I Y. Y.

raca

J»
E

C
‘“348

9'

‘W h

‘5‘.-
i,P\.-y.

211

For this reason it cannot easily be assumed that the
conditions set by the metropolis force one pattern of
adjustment in all areas about the metropolis. Nor can

it be assumed that all hinterland communities are
oriented solely toward the metropolis. . . . This
general observation makes it evident that some system

of classifying the hinterland communities must be
adopted, and that this classification must be in terms

of amount of influence exercised by the principal hinter—
land communities.9/

 

 

From this Bogue proceeded to propose a fourfold classifica-
tion of hinterland communities corresponding to different
steps in the dominance continuum:
Metropolitan centers . . . . . . Dominants
Hinterland cities (urban) . . . Subdominants
Rural—nonfarm populations
(rural-nonfarm). . . . . . . . Influents
Rural-farm populations
(rural-farm) . . . . . . . . . Subinfluents
It is understood that "decreasing values along this scale
refer to two types of change in dominance: (l) decreasing
range or area of dominance, and (2) decreasing number of
functions over which dominance is exercised."98 Although
Bogue considered in his analysis only the first two levels
of dominance to the neglect of the rural-nonfarm and rural-
farm populations, it is essential for a comprehensive View
of metropolitan dominance to take into consideration the

potential influence of the metropolitan center on all three

hinterland areas: urban, rural-nonfarm and rural—farm

 

97Ibid., p. 18.

98Ibid.

hinterla:

C)

.
tie metro;
‘HCYQ he
““h" Cu

QC
J-
vlslalizes
poiltan CC
:ug
" u

C i 3 x,
‘ioglcal Pr
Dir, 1957)
tnree‘folr‘i

212

hinterland communities.99 The present study is considering
all three types of hinterlands as well as the local commu—
nity as its unit of analysis.100

Getting back to our problem, what are the implica-
tions of the observation that all local communities in the
metropolitan community exert some degree, great or small, of
dominance? If the metropolis cannot be expected to set the
conditions for adjustment in all areas of the hinterland and
if some hinterland communities are not solely oriented to
the metropolitan center, what other source of influence is
there operating in the hinterland to account for the non—
random distribution of structural attributes in space?
Furthermore, if the power of the metropolitan center to

control the conditions of life, to which hinterland commu-

nities adjust, wanes with increasing distance, what other

 

99Bogue's is the only study that I am aware of that
visualizes the metropolitan community in four parts: metro-
politan center, urban hinterland, rural—nonfarm hinterland,
and rural-farm hinterland. Other metropolitan dominance
studies employ a three-fold classification: metropolitan
center, urban hinterland, and rural hinterland. See, for
example, Lewis Jones, "The Hinterland Reconsidered," Amer-
igan Sociological Review, XX (February, 1955), 40-44; and
Charles M. Grigg, "A Proposed Model for Measuring the Eco-
logical Process of Dominance," Social Forces, XXXVI (Decem—
ber, 1957), 128-31. Some urban dominance studies use a
three—fold classification also, but slightly different to
fit the theory: urban centers (by metropolitan and non-
metropolitan status), rural—nonfarm and rural-farm. See,
for example, 0. D. Duncan, "Gradients for Urban Influence
on the Rural Population," The Midwest Sociologist, XVIII
(Winter, 1956) , 27—30.

100See Chapter I, p. 62, for the specific definitions
Of urban, rural-nonfarm, and rural-farm communities. Census
definitions are employed on a county basis.

 

 

 

source of

land area:

he answers

tfl A"
\J C '
‘1 CON

 

 

213

source of influence might there be in the outlying hinter-
land areas of the metropolitan region? These questions must
be answered in order to maintain the basic assumption in
ecological analysis, that there exists a non-random distri-
bution of structural attributes in space. It is suggested
that though metropolitan dominance is an important determi-
nant of this non-random distribution of structural attri-
butes, it is not capable of accounting for all the variation
of structural attributes in hinterland communities.

Grigg has written two articles which have something
to say concerning the problem under consideration.101 In
the later article Grigg and Vance attempt a synthesis of
ecological (actually metropolitan dominance) studies and
regional studies. Both approaches can be examined from the
point of view of structure, process and content. The
authors conclude, however, that structure is the only basis
for synthesis.

This leaves structure--the basic element in area

study and the forte of both regionalist and ecologist-—
as a basis for synthesis. Somewhere in between intra—

metropolitan ecology devoted to its mosaic of natural
areas and the homogeneous region is found the analysis

101Charles M. Grigg, "A Proposed Model for Measuring
the Ecological Process of Dominance," Social Forces, XXXVI
(December, 1957), 128-31; Charles M. Grigg and Rupert Vance,
"Regionalism and Ecology: A Synthesis?" Florida State
University Research Reports in Social Science, III, No. 2
(August, 1960), l—ll.

214

of inter-community ecology. The regionalist can partic-

ipate in this because he sees the region developing as

a constellation of communities. The ecologist sees this

as the study of inter-metropolitan dominance and integra—

tion-—what R. D. McKenzie called the "new city regional—
ism."102

This "analysis of inter-community ecology" requires a con—
sideration of the structure of the homogeneous and metropol-
itan region as a "constellation of communities," each with
its own orbit of influence. After applying two models, one
of homogeneous subregions and the other of metropolitan dom—
inance, to rates of population change in the South, the

authors conclude that

. . . in an agricultural society the homogeneous sub—
region is the most appropriate spatial model to use.
However, (with) the development of cities, the spatial
model has to be modified to allow for the effect of
large metropolitan centers on the region. The most
appropriate model then is one which attempts to express
the relationship between the center and its hinterland.

This relationship can best be expressed in some measure
of distance.

This conclusion suggests that there is a dynamic change tak-
ing place in hinterland regions. Metropolitan centers are
becoming more and more influential in ordering their hinter-
land in some consistent pattern along a continuum of dis—
tance. This conclusion also suggests, however, that in out-
lying hinterland areas where metropolitan dominance is not
as influential, i.e., in the more rural areas of the hinter—
land, regional environmental factors may be operative in
effecting inter—community ecological patterns.

102Ibid., p. 4. 103Ibid., p. 10.

 

 

215

The purpose of Grigg's earlier paper, in connection
with the above comments, was to question the pervasiveness
of metropolitan dominance in the United States.104 Grigg
raises the question as to whether it can be shown that other
environmental factors rather than dominance can explain the
nonrandom distribution of attributes in space. Grigg con-
structs a research design which attempts to answer this

question.

As any test of the hypotheses of metropolitan dom-
inance would have to postulate other sources of non-
random distribution, we will establish two additional
sources other than the influence of the metropolitan
center. These will be called rural patterns and urban
patterns. Just as metropolitan dominance results in
metropolitan structure, the rural and urban result in
typical structures on non-random distributions.105

What Girgg's model assumes, then, is that all three elements
of the metropolitan community (the metropolitan center,
urban hinterland, and rural hinterland) may influence the
nonrandom distribution of attributes in space. His finding
that urban centers are chiefly responsible for nonrandom
patterns of population change in the South suggests a type
of dynamic model of urbanization. Remote agricultural areas
may first be influenced by their natural environment. As
urbanization continues, these areas come under the influence
of urban centers, but as urbanization takes on the form of
metropolitanism, metropolitan centers become the organizing

104Charles M. Grigg, op. cit., p. 128.

105Ibid., p. 129.

 

It..‘.l\“k t‘

216

agents of the hinterland. The point to be made, however,

is that hinterland communities themselves may have natural
environmental conditions which on a broad scale determine
the nonrandom distribution of structural attributes in out-
lying areas. The more distant the community from a metro—
politan center and the less influence exerted on the hinter-
land community, the greater the possibility that environmen—
tal factors determine the nonrandom distribution of commu—
nity structural characteristics. It is assumed that, given
the tendency for urban populations to concentrate in close
proximity to metropolitan centers and the tendency for the
rural-farm population to be dispersed in the outlying areas,106
rural hinterland communities will show the greater tendency
to be influenced by local environmental factors than urban
hinterland communities. These ideas are not new with Grigg,
however, Anderson and Collier, in an article published
earlier than Grigg's, employed a research design to test
both metropolitan dominance and urban dominance theory.

They concluded that:

These findings tend to cast doubt on the hypothesis
of metropolitan (as opposed to urban) dominance, and
even cast some doubt on the notion of urban dominance
over the rural hinterland. That is, when variations in
the gradients were found (in connection with size of
farm), the statistically explanatory factor which ex-
plained the variation proved to be a characteristic of
the rural areas (type of farming) rather than a charac-

teristic of urban areas. This finding seems to indicate
that the gradients result from a variety of forces, only

 

 

 

106Donald J. Bogue, op. cit., p. 35.

217

some of which can be said to be concentrated in the
metrgpolis and hence contribute to the concept of met-
rgpolitan dominanceIlU/

Hence it is possible that internal characteristics of the
rural areas themselves may be an important factor in account—
ing for inter-community variation. Bogue said somewhat the
same thing at even a much earlier time:

Within this multiple-community complex, which may be
called the metropolitan community in deference to its
dominant species, the individual local community must
occupy a subordinate position. The activities of the
local community are a function not only of its immediate
localityy but also of the relative ecological position
with respect to the dominant metrgpolis.IU87

 

Thus Bogue recognized that the nonrandom distribution of
community structural attributes is a function of BEER
metropolitan dominance and conditions of the local community.
This implies, then, that if the influence of the metropoli-
tan center, in other words, if distance and size of the met-
ropolitan center could be controlled, we should be able to
measure the relative influence of the local hinterland com—
munity, urban or rural, in determining its own activity and
structural patterns.

Now recalling our earlier hypothesis that community
social structure is a function of distance and size of a

dominating metropolitan center and that fertility behavior

 

107Theodore R. Anderson and Jane Collier, "Metropol-

itan Dominance and the Rural Hinterland," Rural Sociology,
XXI (June, 1956), 157.

108Ibid., p. 13.

 

218

is a function of both of these, it stands to reason that, if
the influence of a dominating metropolitan center(ixa, its size
and distance) can be controlled or held constant, then we
should be able to measure the influence of local community
conditions alone on fertility behavior. On the basis of

this reasoning we may propose a third hypothesis:

3. Community fertility behavior in both urban and rural
hinterlands is not only a function of the size and
distance of a dominating metropolitan center, but
also a function of conditions of its own immediate
locality, since all local communities in the metro-
politan region possess some degree of dominance over
some portion of the hinterland.

Since urban hinterland communities tend to concentrate in
closer proximity to metropolitan centers than rural hinter-
land communities, we should expect metropolitan centers to
have a greater influence in ordering community social struc-
tural patterns in the urban hinterland than in the rural
hinterland. A fourth hypothesis may be stated:

4. Community fertility behavior is more a function of
distance and size of a dominating metropolitan cen-
ter in the urban hinterland, but more a function of
local community social structure in the rural hinter-
land, when distance and size of the dominating metro-
politan center are controlled.

There is one final aspect to consider with respect
to the dominance of metropolitan centers on their urban and
rural hinterlands. We have established the fact that metro-
politan centers vary with respect to the amount of dominance
and extent of the area of dominance. It is also a fact that

a considerable amount of difference exists among various

geographic divisions of the nation with respect to the level

<1

219

of development of the metropolitanization process. Although
there are several ways by which one could demonstrate the
intensity of metropolitanization within the geographic divi-
sions, considering the distribution of population by metro—
politan status does allow us an approximate measure of these
differences. This information is provided in Table 16.
Although the nation as a whole reflects almost two-thirds of
its population residing in standard metropolitan statistical
areas, the geographic divisions manifest a range from a high
of 82 percent for the Middle Atlantic to a low of 36 percent
for the East South Central. The geographic divisions of
New England, Middle Atlantic, Pacific, and East North Cen-
tral indicate metropolitanization levels greater than the
nation; West North Central, Mountain, and all three Southern
divisions lower than the nation.

The question now arises, how can metropolitan dom-
inance theory cope with interdivisional variation with
respect to the metropolitanization process? Several comments
in the previous discussion of metropolitan dominance theory
have described metropolitan dominance as an emerging pattern
of organization for communities in the nation. The data of
Table 16 merely indicate the possibility that in some geo-
graphic divisions compared with others metropolitan centers
are perhaps more influential "organizing agents" of their
hinterlands. If this is true, we would expect to find that

the influence of metropolitan centers, measured in terms of

. CC:

.1):

ianau

1: q

220

.ma magma

.Aaoma

.moflmmo mcfluzflum ucmfich>ow ".U.Q

 

.aovmcflnmmzv ealflavom uuommm Hmcflm .mumEESm mmumum woufisb .musmuanmch

 

 

 

 

 

mo Hwnfisz .omma "cowumasmmm mo mamcmo .m .D .mSmcmo may mo samusm .m.D "mousom
m m.om mae.v m.me mmn.oa mmH.Hm unmaumm
a «.am hom.m m.me mem.m mmm.o :Hmucsoz
m m.o¢ mmm.n m.mm moo.m Hmm.oa Hmuucmo nusom ummz
m o.¢o oom.n 0.0m «em.¢ omo.ma Hmnpcmo nusom ummm
o m.me Hmm.ma N.om Heo.ma mem.mm oflucmana nusom
m n.9m mme.m m.m¢ oom.o emm.ma Hmuusmo npuoz ummz
a m.mm Hmm.HH H.ne emm.¢m mmm.mm Hmuucmo ruuoz pmmm
H m.mH mam.o m.am amm.nm moa.¢m onucmaua mavens
m e.mm oHH.m m.on mom.a mom.oa ecmamam zmz
. o.em mm¢.oo o.mo mmm.mHH mmm.mha mmumpm wanna:
QOHDMNHQmu ucmoumm Hmnfidz unmoumm Hmnfisz coauwasmom aOHmH>HQ
naaomoupmz Hmuoa wan mmumum emuaco
an xcmm m.¢m2m mwnmuso m.¢m2m meHmcH

 

 

Am.ooo CH aw>flm mumnfiscv coma "msumum Gmuwaom
IouuoE >9 .mconH>Ho tam mmumum ompwcb may Ga coaumasmom mo coflusnwuumfln .ma mamas

221

distance and size of the dominating metropolitan center,

is more important in accounting for the spatial distribu-
tion of community social structural attributes, and as a
result community fertility behavior, in both urban and rural
hinterlands, in the more metropolitan geographic divisions
compared with the less metropolitan divisions. On this
basis we may formulate a fifth hypothesis:

5. In the more metropolitan geographic divisions com—
pared with the less metropolitan geographic divi-
sions, size and distance of a dominating metropol—
itan center are more important in accounting for
variation in community social structure and fertil—
ity behavior in both urban and rural hinterland
communities.

But there remains more to be said concerning the
effects of emerging metropolitanization as the new mode of
organization for hinterland communities. On the basis of a
dynamic view of metropolitan dominance, one could predict
that eventually all hinterland communities will come under
the organizing influence of metropolitan centers, although
this is not yet the case in the United States. As expressed
above, metropolitan centers still compete with the effects
of local community conditions, especially in the more remote
hinterland areas of the metropolitan region. Hence, in the
less metropolitanized geographic divisions of the nation, we
should expect to find that local community characteristics
which are the result of natural environmental processes are

perhaps more influential in determining hinterland inter-

community variation than metropolitan centers. As a result,

222

fertility in these communities will be more a product of
local community conditions than metropolitan influence.
This suggests a sixth hypothesis for testing:

6. In the more metropolitanized geographic divisions,
size and distance of a dominating metropolitan cen—
ter will be more important in accounting for varia—
tion in community fertility behavior, in both urban
and rural hinterlands, than local community social
structure, when controlling for the influence of
metropolitan centers; in less metropolitanized geo-
graphic divisions, local community social structure
will be more important in accounting for variation
in community fertility behavior than size and dis—
tance of a dominating metropolitan center.
Considering the implications of inter—divisional

variation for metropolitan dominance theory greatly expands
the number of testable hypotheses and also enables us to
consider metropolitan dominance as a dynamic process. We
have hypothesized that as metropolitan centers continue to
emerge, hinterland community variation will continue to fall
more and more under the organizing influence of metropolitan
centers. An inter-dependence of metropolitan center and
hinterland communities will become the dominant characteris-
tic of community organization. But in this dynamic view we
must not commit the error of urban dominance theory. We
must not think of urban and rural differences as diminishing
while metropolitanization increases. An important point of

metropolitan dominance theory predicts that differences will

increase among hinterland communities under the dominance of

 

the same metropolitan center. Bogue found that in consider-

ing inter-regional variation along with the effects of

223

metropolitan dominance on the hinterland that "differences
associated with dominance and subdominance within each
region are greater than the average differences between

. "109
regions.

This suggests that as metropolitan centers
become more and more the organizing agent of hinterland com-
munities in the nation, that the patterns they effect in the

hinterlands will appear to be a universal pattern. Divi-

 

sional differences will become less pronounced and differ—
ences between hinterland community types (urban and rural)
will become more pronounced. Hence, given the fact that
fertility is a function of community social structure, as
community social structure becomes more and more influenced
'by metropolitan centers, the impact that community social
structure has on fertility behavior in urban hinterlands,
for example, will be more and more similar in all geographic
divisions than its impact on fertility in rural hinterlands
within the same geographic division. Metropolitan dominance

theory emphasizes the differentiating effect of metropolitan

 

centers on urban and rural hinterland community social
structure, and, as a result, on urban and rural fertility
behavior. A comparison of inter-divisional differences of
the impact of community social structure on fertility behav-
ior within the same class of hinterland communities (urban
or rural) should reveal fewer differences than an intra-

divisional comparison of the impact of community social

 

109Ibid., p. 61.

224

structure on fertility behavior between different classes of

hinterland communities (urban vs. rural). On the basis of

these comments a final hypothesis may be formulated:

7.

The impact of community social structure and metro—
politan dominance on fertility behavior will mani—
fest fewer differences when comparing the same type
of hinterland communities (urban or rural) on an
inter-divisional basis than when comparing different
types of hinterland communities (urban vs. rural) on
an intra—divisional basis.

Summary: Hypotheses Derived from
Metropolitan Dominance Theory

 

 

In this chapter urban dominance theory and metropol-

itan dominance theory have been contrasted. It was decided

that metropolitan dominance theory was the preferred theo-

retical framework from which to develop hypotheses pertain—

ing to the explanation of urban-rural fertility differential

patterns found in the United States. Seven basic hypotheses

were derived and are summarized below.

1.

Community social structure is a function of distance
and size of a dominating metropolitan center.

a. Distance and size of a dominating metropolitan
center will manifest a different impact on com-
munity social structure in the urban hinterland
than in the rural hinterland.

Fertility behavior is a function of community social
structure and distance and size of a dominating met-
ropolitan center.

a. Community social structure and distance and size
of a dominating metropolitan center will manifest
a different impact on fertility behavior in the
urban hinterland than in the rural hinterland.

Community fertility behavior in both urban and rural
hinterlands is not only a function of the size and
distance of a dominating metropolitan center, but

225

also a function of conditions of its own immediate
locality, since all local communities in the metro-
politan region possess some degree of dominance over
some portion of the hinterland.

4. Community fertility behavior is more a function of
distance and size of a dominating metropolitan cen-
ter in the urban hinterland, but more a function of
local community social structure in the rural hinter-
land, when distance and size of the dominating metro—
politan center are controlled.

5. In the more metropolitan geographic divisions com—
pared with the less metropolitan geographic divi-
sions, the size and distance of a dominating metro-
politan center are more important in accounting for
variation in community social structure and fertil-
ity behavior in both urban and rural hinterland
communities.

6. In the more metropolitan geographic divisions, size
and distance of a dominating metropolitan center will
be more important in accounting for variation in
community fertility behavior, in both urban and rural
hinterlands, than local community social structure,
when controlling for the influence of metropolitan
centers; in less metropolitan geographic divisions,
local community social structure will be more impor-
tant in accounting for variation in community fertil—
ity behavior than size and distance of a dominating
metropolitan center.

7. The impact of community social structure and metro—
politan dominance on fertility behavior will mani—
fest fewer differences when comparing the same type
of hinterland communities (urban or rural) on an
inter—divisional basis than when comparing different
types of hinterland communities (urban vs. rural) on
an intra—divisional basis.

To test these hypotheses requires an altogether dif-
ferent research procedure than employed in most empirical
studies of the urban—rural fertility differential. The
review of fertility studies contained in Chapter II of this

thesis makes this especially clear. Though the empirical

propositions established in the previous chapter hint at the

226

possibility of different fertility patterns in urban and
rural communities of the United States, these propositions
prove to be wholly inadequate to provide further direction
in the testing of our theoretical hypotheses derived from
metropolitan dominance theory. The theoretical framework

of metropolitan dominance, then, has proved fruitful by
generating new hypotheses and by extending the study of the
urban—rural fertility differential into new areas. All four
of the critical guidelines for theory design in the study of
differential fertility have been answered in this chapter.
Now we must consider the methodological procedures by which
to test the theoretical hypotheses as well as the theory of

metropolitan dominance itself.

CHAPTER IV
METHODOLOG ICAL PROCEDURES

Seven hypotheses have emerged in our consideration
of metropolitan dominance theory, each contributing to a
furthering of our understanding of the expected patterns of
community fertility behavior and what factors may be opera-
tive in determining those patterns. We are now confronted
with the task of indicating how these hypotheses are to be
tested. First, there appear to be several concepts used in
these hypotheses which need to be translated into forms
which will lend themselves to statistical testing. In other
words, what empirical indices can be employed to reflect
community social structure, metropolitan dominance, and
community fertility behavior? Our first task, then, is to
specify operational definitions for these concepts. Second,
after the empirical variables are specified, we should
review the statistical techniques which will be applied to
the data to test our theoretical hypotheses.

Conceptual Framework: Specification
of Variables

Two broad concepts have been suggested in our theo-

retical considerations as possible independent variables

which are expected to account for variation in fertility

227

 

228

behavior: community social structure and metropolitan
dominance. The sole dependent variable of the analysis is
community fertility behavior. Let us consider how these

concepts might be reduced to the form of empirical variables.

Community Social Structure

A countless number of indicators have been employed
in community studies for the purpose of representing commu—
nity social structure. As in any study of this nature,
variables must be carefully selected with fairly specific
reasons in mind as the basis for the selection. For the
purposes of this study indicators should be chosen on the
basis of whether they can be expected to reflect significant
internal characteristics of the organization of a community
and whether they can be expected to be of special importance
in explaining fertility behavior. In the preceding chapter
it was suggested that the broad concept of community social
structure should be broken down into four basic categories:
urbanization, socio-economic status, wife's opportunities
alternative to child-bearing, and demographic structure. It
is felt that these categories, though greatly narrowing the
range of possible indicators of community social structure,
will greatly facilitate the search for empirical measures
which should relate significantly to fertility behavior. To
reflect the extent of urbanization within a community we
should seek some measure of the amount of employment of

males in agricultural occupations. Such a measure may be

229

considered a function of the extent of employment in non-
agricultural occupations and in this way be considered a
possible indicator of the degree of urbanization of a com-
munity within a modern society.1 Socio—economic status
should be reflected in some measure of educational attain-
ment and income level for a community. Wife's opportunities
alternative to child-bearing within a given community should
be reflected in the amount of employment opportunities for
women and the amount of income a woman could obtain from
employment. The demographic structure of a community should
be represented by some measure of the distribution of women
within the reproductive age span. For each of these four
categories of community social structure, then, two empir-
ical variables were employed and can be specifically defined
as follows:

Employment in Agricultural Occupations (Urbanization)

1. Percent of the male labor force who are farmers
and farm managers.

2. Percent of the male labor force who are farm
laborers and farm foremen.

 

Socio—Economic Status
3. Median years of school completed by males and
females, age 25 and over.
4. Median family income.

 

 

lJanet Abu-Lughod, "Urban-Rural Differences as a
Function of the Demographic Transition: Egyptian Data and
an Analytical Model," American Journal of Sociology, LXIX
(March, 1964), 488; Hope Tisdale, "The Process of Urbaniza—
tion," Social Forces, XX (March, 1942), 311—16; and P. K.
Hatt, N. L. Farr, and E. Weinstein, "Types of Population
Balance," American Sociological Review, XX (February, 1950),
14-21.

 

 

 

230

Wife's Alternative Opportunities
5. Median female personal income.
6. Percent of females, age 14 and over, who are
employed.

 

Demographic Age Structure
7. Percent of ever—married females, age 15—44, who
are age 15—24.
8. Percent of ever-married females, age 15—44, who
are age 25—34.

It should be noted that, because of the employment
of census data in this study, the selection of variables to
represent community social structure was limited by the
types of information available on census tapes. It is felt,
however, that, in spite of this limitation, the selection of
adequate variables for the testing of our hypotheses is not
hampered severely. It should also be pointed out that each
of these variables (with the exception of female employment
and income) were computed separately for the urban, rural-
nonfarm, and rural-farm parts of every county in the nation.
Furthermore, because of the complexities and confusion which
could result from considering a fertility analysis for
whites and nonwhites combined, it was decided to limit the
present study only to the white population of the nation.
Hence, each of the community social structure variables

specified above apply only to the white population of the

residential parts of counties.

231

Metropolitan Dominance

Metropolitan dominance is another independent vari-
able to be included in accounting for variation in fertility
behavior. Because of the theoretical framework employed in
this study, the development of a measure of influence ex-
erted by metropolitan centers on hinterland communities
becomes very crucial. There are several possible measures
of metropolitan dominance that could be used, but our theo-
retical consideration has suggested that the influence of a
dominating metropolitan center on a hinterland community is
a function of the distance that community is from the metro-
politan center and the size of that center. The construc—
tion of such a variable requires the development of some
classification scheme by which every county in the nation
can be given some numerical designation which would approx-
imate the amount of influence exerted on that county popula-
tion by a dominating metropolitan center, dependent upon the
distance from the metropolitan center and the size of the
center combined. We may describe the procedure employed in
operationalizing metropolitan dominance in the following
manner.

It was decided to use the 1960 Census definition of
Standard Metropolitan Statistical Areas to designate metro-
politan centers. By definition

. . . SMSA (Standard Metropolitan Statistical Area) is
a county or group of contiguous counties which contain

at least one city of 50,000 inhabitants or more or
"twin cities" with a combined population of at least

232

50,000. In addition to the county, or counties, con-
taining such a city or cities, contiguous counties are
included in an SMSA if, according to certain criteria,
they are essentially integrated with the central city.

2
Each SMSA designated by the 1960 Census was located on a map
of the United States which outlined county boundaries. Each
SMSA county was assigned a numerical value that was a linear
function of the size of its popultaion. This value in-
creased by l for each 100,000 population up to a population
size of 2 million.3 Thus each SMSA county was assigned a
value within a range from 1 to 20. On this basis an SMSA
county with a population of 2 million or more was assigned

a value of 20, an SMSA of 1 million a value of 10, and an
SMSA of 100,000 a value of 1. Any SMSA county with a popu-

lation between 50,000 and 100,000 was also given a value of

1. Only whole integer numbers were employed to represent

 

2U. S. Bureau of the Census, U.S. Census of Populge
tion: 1960, Number of Inhabitants, United States Summary,
Final Report PC(l)—lA (Washington, D.C.: Government Print—
ing Office, 1961), p. xxiv.

3The establishment of a maximum size value of 20 for
all SMSA's with populations of 2 million or more is based on
the speculation that at some point increases in population
are merely duplications of technological functions and condi-
tions that exist in areas having maximum population density.
Thus, it was arbitrarily decided that any SMSA of 2 million
or more population would possess similar technological bases
and would exert the same influence over the same size hinter-
land area. See Dale Hathaway, J. Allan Beegle, and Keith
Bryant, Rural America, Census Monograph, forthcoming, pp. 15-
16. Furthermore, it might be stated that of the 212 SMSA's
designated by the 1960 Census, only 10 have a population
size of over 2 million: Boston, Chicago, Detroit, Los
Angeles, New York, Philadelphia, Pittsburgh, St. Louis,
San Francisco-Oakland, and Washington.

 

233

the size of an SMSA. Hence an SMSA of 150,000 or more was
assigned the next highest integer of 2. All counties of a
given SMSA were treated as a unit and therefore were all
given the same size value.

After SMSA counties were located and assigned
values based on size, the next task was to assign values
to all non—SMSA counties in the nation based on distance
from a dominating metropolitan center and the size of that
center. After determining a central point for each SMSA
(the center of the SMSA county containing the central city),
concentric circles were drawn radiating out from each SMSA.
The first circle was given a radius of 50 miles, the second
circle a radius of 100 miles, the third 150 miles, etc.
This created a series of distance bands around each SMSA,
each band being 50 miles in width. A county was not con-
sidered lying within a given distance band unless completely
covered by the most distant boundary of the distance band.
Because we assume that metropolitan dominance is related to
both distance and size, each county was assigned a numerical
value which was a combined function of the size of the influ—
encing SMSA and of the distance from the SMSA. Each band,
and consequently all counties covered by that band, was
assigned a value that declined as a function of the number
of 50 mile distance bands from the influencing SMSA. This
decline was by a numerical value of 2. As a result all non-
SMSA counties within the first 50 mile band from an SMSA

were assigned a value 2 less than that assigned to the SMSA

234

county or counties. Non-SMSA counties between 50 to 100
miles of an SMSA were assigned a value 2 less than the value
assigned to counties within 50 miles of the SMSA. Hence, in
the case of an SMSA of 2 million population, the SMSA coun—
ties received a value of 20, non-SMSA counties in the first
50 mile band a value of 18, the second band 16, etc. This
procedure was followed until the value zero was reached or

a competing SMSA determined the values for such counties.

An implication of this scheme is that influence of a metro—
politan center becomes negligible when the zero value is
reached. Generally the largerthe SMSA in population size,
the larger the hinterland area: the smaller the SMSA, the
smaller the hinterland area influenced by the metropolitan
center. Because of the numerical scheme employed, we assume
that a metropolitan center of 2 million or more can extend
its influence up to a maximum of 450 miles, though with
each distance band its influence is represented as gradually
diminishing. SMSA's of smaller size would extend their
influence less than this, in proportion to their population
size.

With this scheme it was possible that a county would
receive different values from different SMSA's. It was
decided that the final value assigned to any county should
always be the highest possible value obtainable through the
methodological procedures employed above. Hence, where sev—

eral bands overlapped the same non-SMSA county, the county

235

was assigned the highest possible value and allocated to the
hinterland of the metropolitan center which determined this
value.

A further implication of a scheme which recognizes
size of a dominating metropolitan center is the possibility
that larger SMSA's may influence not only non—SMSA counties
within their hinterlands, but also other smaller SMSA's with—
in their hinterlands. To distinguish these smaller SMSA's
from non-SMSA counties lying within the same distance band
from a larger SMSA, it was decided that the value assigned
these smaller SMSA's should be the value received from the
distance band in which it was located plp§_the size value,
determined on the basis of a numerical value of l for every
100,000 population in the smaller SMSA. Hence it was pos—
sible that an SMSA of 200,000 (with a size value of 2) lying
within a distance band 400 miles from an SMSA of 20 million
(receiving a distance value of 4 from that SMSA) could
receive a total size-distance score of 6 as the highest pos-
sible value it could receive using the methods described
above. It had to be assumed, however, that no SMSA county
located within the sphere of influence of a larger SMSA could
receive a total value (the sum of the size value and the dis—
tance value) larger than that assigned to the larger SMSA.
.As an example of this rule, suppose a county is designated
as an SMSA county with a population of 1,100,000. Its size

value is therefore 11. The counties included in the first

236

50 mile band from the SMSA are assigned the value of 9, the
second band 7, etc. Suppose that an SMSA county with a pop-
ulation of 600,000 is located between 50 and 100 miles from
the first SMSA. Using our addition rule, it should receive
a distance value of 7 and a size value of 6, or a total
value of 13. Since the influencing SMSA could obtain only
a value of 11, the value of the smaller SMSA is limited to

a value of 11 also. However, this occurred in few cases.

By this procedure all counties in the United States
were assigned a single numerical value from a possible range
of zero to twenty. Highest values reflect the greatest
influence from metropolitan centers based on proximity and
size, smallest values reflect least influence from metropol-
itan centers. All counties assigned the same numerical
value are assumed to be equally affected by the organizing.
influence of a metropolitan center. Since our analysis
required the breakdown of county populations into urban,
rural-nonfarm, and rural-farm components, it was decided
that each residential component of the same county should
receive the same value representing metropolitan influence
as assigned the whole county. It should be noted that these
values assigned to the residential parts of counties were
inserted on the census tapes. All other variables used in
the analysis were able to be computed directly from the data

included on the census tapes.

237

Community_Fertility_Behavior

 

The dependent variable of this study is community
fertility behavior. Numerous measures of fertility behavior
have been developed through the years by demographers.4
Though no single measure of fertility can be considered the
best overall measure, the choice of a fertility measure
depends upon many factors, such as, purpose of the investi-
gation and the source and nature of available data. The
measure of fertility used in this study is a measure of
cumulative fertility which is available on the 1960 Census
tapes, the single source of data employed in this study.
More specifically the dependent variable can be defined as:
the number of white children ever-born to ever—married white
females, age 15-44, per 1,000 ever married white females,
age 15-44. Similar to all variables in this analysis, only

the white population of the nation is considered. Also this

 

4See discussions of various measures in George W.
Barclay, Techniques of Population Analysis (New York: John
Wiley, 1958), Ch. 6; John Hajnal, The Study of Fertility and
Reproduction: A Survey of Thirty Years," in Thirty Years of
Research in Human Fertility: Retrospect and Prospect (New
YOrk: Milbank Memorial Fund, 1959), pp. ll-37; N. D. Ryder,
"Fertility," in P. M. Hauser and O. D. Duncan, The Study of
Population (Chicago: The University of Chicago Press, 1959),
pp. 400-36; Donald J. Bogue and James A. Palmore, "Some
Empirical and Analytic Relations among Demographic Fertility
Measures, with Regression Models for Fertility Estimation,"
Demography, I, No. l (1964), 316-38; and Ronald Freedman,
"The Sociology of Human Fertility," Current Sociology, X/XI,
No. 2 (1961-62), (Oxford, England: Basil Blackwell, 1963),
pp. 43-46.

 

 

 

 

 

 

 

. t..:.~\~,..
.C

238

measure is computed for the three residential components of
all counties in the nation.

The data on children ever born in the 1960 Census
arwe based upon a 25 percent sample of the population and are
deazrived from answers to the following question on the House—
hc>LLxd Questionnaire: "If this is a woman who has ever been

Huaz:tried—-How many babies has she ever had, not counting
srt;i.illbirths? Do ppp count her stepchildren or adopted chil-
<il:ea:11."5 The number of children ever born includes children
kDCDJZTri to the woman before her present marriage, children no
longer living, and children away from home, as well as chil—

dren borne by the women who were still living in the home.
Although the question on children ever born was asked only
of Women reported as having been married, undoubtedly the
ntllllikber of children reported includes some illegitimate
1>Jj~3=7i:hs. It is likely that many of the unwed mothers living
M’jL3t3511 an illegitimate child reported themselves as having
13th married. Stepchildren, adopted children, and still—
ijL“1:‘i:hs, nevertheless, were not counted in this measure of
betility.

The measure of fertility used in this study, then,

I‘Qb . . n . o
J:esents the cumulative fertility of married women until

llypppp‘ppp
1:;i‘ 5U. S. Bureau of the Census, U.S. Census of Pppula-
EET§3§EEEEi: 1960, Subject Reports, Women by Number of Children
Irleéfsil: Born, Final Report PC(2)-3A (Washington, D.C.: Govern-
]:lt: Printing Office, 1964), p. x.

 

 

239

the time of the census, i.e., it represents their reproduc-

tive histories. One might also consider the measure as the
"average number of children" per married woman or "average

size of family." It should be noted, however, that this is

not a measure of size of "completed family," since the num-
ber of children ever born is related to married women still

within the reproductive period, ages 15—44. On the other

hand, our measure of fertility cannot be likened to a mea—

sure of current fertility because past births are considered

as well as current births, regardless of the time period in

which they occurred. As do all measures of fertility, chil-
dren ever born has its advantages and disadvantages. Unlike
some measures of fertility, children evern born enumerated
in the census are related to the actual group of women who
Produced them. In this sense, the children ever born mea-
Sure is similar to cohort fertility. Since it is based on
c1"‘:i~1l.dren born over a relatively long, unspecified period of
time, the cumulative fertility rate will probably not be
iI‘ldslmenced greatly by temporal or short—time events, which
can affect current fertility measures quite radically.
Fun: thermore, mortality should not affect this measure, since
a1 1 children ever born are to be recorded, although the mea-
sure may suffer from inaccurate reporting because of the
6% bendency upon retrospection by the respondent. Birth

Q
VQrcounts and undercounts undoubtedly occur, though Grabill,

K -
a‘S~er and Whelpton, in commenting on the children ever born

240

measure, assert that while overcounts of children ever born
inay occur almost as frequently as undercounts, "it is prob-
aIDle that the bulk of the reports on children ever born are
(:cnnplete and accurate, at least for whites."6 The fact that
Inaalrital status is incorporated in our measure of fertility
gireaatly refines this measure, although perhaps accounting
.ftaczr age at marriage and marital duration would have contrib-
iitzesad even more to the development of a better measure of
:Eeazzrizility. It is very possible, however, that different
Cij_£3wtributions of married women within different age groups
C313' 1:he reproductive age span of 15-44 could greatly affect
tllilEi fertility level of a given population. That is, a popu-
Jréth::icn1Mdth a disproportionately large number of married
women in the ages, say, of 15-24 could produce a rather low
‘J‘€3“\7<al of fertility by this fact alone. This will be partly
QQI‘l’crolled, however, by the inclusion of age distribution
Theasures as independent variables in the analysis.

Finally, because the number of children ever born
ITEE’ICI-resents a measure of cumulative fertility, it has a some—
lefilfslt restricted value for an analysis of the comparison of
L1‘::t:.1?->an and rural fertility. Goldstein and Mayer, in consider—
J‘Irl5ér tide particular problem, state that "the current place
(32:3 residence of the mother does not necessarily indicate the

\

13. 6W. H. Grabill, C. V. Kiser, and P. K. Whelpton, The
'nggéisitilipypof American Women (New YOrk: John Wiley, 1958),
‘ 402.

241

residence at the time of birth of children." This resur-

:rreax:ts the problem considered previously in this thesis con—
cerning the effect of migration on fertility and the diffi-

culty of relating community social structure characteristics

of current residence to fertility events which may have
However, Goldstein

occurred previous to current residence.
in considering whether suburban fertility mea-

and Mayer,
S ured as children ever born is a result of current residence
conclude that suburban

or a selective migration process,
fertility is more a function of current residence than a
Though such deficiencies may

Q onsequence of migration.
‘VV'<EE JLl be inherent in the measure of fertility employed in
BY

it: 1715.5 study, we must accept our measure for what it is.
EE“:=<::epting children ever born as an adequate measure of com-
It‘"‘-—‘l-I'1:'L.ty'fertility behavior in comparing urban and rural com—
.ITI‘JlIEities, we assume that community social structure and
'It‘GErtzropolitan dominance do operate within the nation to

(3432t1ermine to some extent the average size of family within

‘:3klee various residential communities.

ggifiiilels of Comparison
It is perhaps appropriate at this point to describe

t113£a levels of comparison required in our research design to

 

l‘_p‘__-
7Sidney Goldstein and Kurt B. Mayer, "Residence and
fferences in Fertility," Milbank Memorial Fund

Etatus Di
uEarterl , LXIII, No. 3 (July, 1965), 295-6.

81bid., p. 297.

.

242

't:easst the theoretical hypotheses. Several references have
aailxready been made to the fact that comparisons are to be
Itleacfle only for the white population of the nation. This
<ileaczision was made because of the greater inaccuracies in
Izleazasuring fertility for nonwhites in the nation and because
<:>:Ef the complexity that diverse patterns of racial composi-
‘t::i_<3n of communities within different parts of the nation may
inflict upon the analysis. Analysis of nonwhite fertility
employing the same research design as this study may well be
C onsidered a greatly needed but necessarily separate study.
An important level of comparison for this study is
REZVITLEE comparison of urban and rural fertility patterns. All
variables, independent and dependent, are computed on the
71:>iElssis of urban, rural-nonfarm and rural-farm parts of coun-
1t:15_<es. Our unit of analysis for this study, then, is the
residential part of a county to be considered equivalent to
Ea- <2ommunity. Because of a metropolitan dominance theoret—
:i~<2ial framework, we shall consider all urban parts of coun-
t:iimes within the hinterland of a given metropolitan center
as its "urban hinterland." Likewise, all rural-nonfarm
“EDEtrts of counties make up the "rural-nonfarm hinterland" and
tlfme rural-farm parts of counties the "rural-farm hinterland."
I3<3<2ause it is expected that metropolitan centers will exert
E3 (different influence on rural-nonfarm than rural-farm com-
DruInities, it is felt necessary to divide the "rural hinter-

JLEiIid" into rural-nonfarm and rural-farm.

243

Though our analysis will consider the nation as a
whole comprising a constellation of metropolitan communities,
and the comparison of different patterns of how community
social structure and metropolitan dominance influence fer-
t ility behavior will be carried out for the nation as a
whole, theoretical considerations point out the necessity of
investigating these patterns on a "regional" (actually divi—
S ional) basis, i.e., by large homogeneous subareas of the

United States. It is suspected that metropolitanism is not
a universal pattern spread over the entire nation but that
d ifferences in the proportion of divisional population resid-
ing in metropolitan areas indicate a differential rate of
I‘11e‘tropolitanism. The procedure of applying our problem to
e ach of the nine geographic divisions in the nation does not
necessarily prove that there are differences among divisions,
but operating at a divisional level is a precaution which we
Q an afford to take in order to more clearly observe the
Q ffects of metropolitan dominance and community social struc—
ture on fertility behavior. If our hypotheses dealing with
pCDssible variation of fertility patterns among the divisions
33"-"e not supported, then we will gain some support of the
JCICD‘tzion that metropolitanism is a universal mode of organiza—
tion throughout the nation. If variation among the geo—
g'I‘aphic divisions of the nation is supported, however, we
InI-lst conclude that perhaps the differential rate of metro-

politanism is a process worthy to be considered in future

244

ssrtrudies and that the peculiar environmental conditions of
the divisions themselves may also have some influence on

creating variation in fertility patterns.
then,

A basic level of comparison in this study,
This level

pertains to possible variations among divisions.

of comparison is especially important with respect to the

manner in which the influence of metropolitan centers on
Because

their residential hinterland areas is considered.
Of the time and labor that would be involved in attempting
apply our research design to each individual metropolitan

to
it was decided to consider the com—

<== (:Irmnunity in the nation,
b ined effects of all metropolitan centers on their hinter—

:l—iEllndS, first for the nation and then for each division.
RIP7kihis means that for any given division we will be consider-
jLJIixg the influence of the average metropolitan center on the
‘E3”\7<erage urban, rural-nonfarm or rural-farm hinterland.
]E3<:nge, who followed this same procedure in his analysis of
1t-l‘lestructure of the metropolitan community, points out that
anrl investigation of individual metropolitan communities pro—

‘U'j—des a study of deviations of individual metropolitan com—
IT111r1ities from a general pattern. The approach of this
Estrudy, then, is to consider what that general pattern is
Einflating metropolitan communities with respect to fertility

 

\
Donald J. Bogue, The Structure of the Metropolitan

ig¥52gmggigx (Ann Arbor: Horace H. Rackham School of Graduate
1:14dies, University of Michigan, 1950), p. 29.

245

pattern comparisons between rural and urban hinterlands.
Even though divisional variations are permitted in our
analysis, the research design of this study still allows us
to consider this general pattern at the divisional level.
Statistical Framework: Specification
of Statistical Techniques
Already in the first chapter of this thesis the need
for the application of more rigorous statistical techniques
to the study of differential fertility based on census data
was pointed out. It was further stated that the traditional
"‘ aggregative" approach used in many past studies of differ—
e ntial fertility does not lend itself as well to the more
b igh—powered statistical techniques currently available as
I"t‘|-‘l~.‘l.<:h as the "distributive" approach. The "aggregative"
a;E>proach suffers especially from the limited number of inde—
pemdent variables that can be simultaneously incorporated
into an analysis of fertility behavior. The problem of this
1:l‘iesis, then, is how to relate simultaneously to fertility
t~11e several variables already presented above, 172E” metro—
politan dominance and the various indicators of community
SOcial structure, all of which are assumed to have some
influence on our dependent variable, community fertility
behavior? We not only need some statistical measure of the
degree of association of the independent variables to fertil—

ity behavior, but we need to know how important these vari-

aJDles are in accounting for variation of fertility behavior

246

in comparison with each other. In addition, we should also

seek an answer to whether these independent variables have
different effects on fertility in urban and rural hinterland

corrlrnunities as well as within the different geographic divi-

8 ions. It seems that the most appropriate approach to our

problem is a distributive approach and it has been shown by
Bog ue and Harris that the distributive approach lends itself

Very well to a multiple regression analysis.lo

wt iple Regression Model

Our analysis focuses on two levels: the nation

C Q cnterminous United States) and the nine geographic divi—

3 lens. For the nation and each division a multiple regres-

e 2.!ch equation was estimated for the three residential compo-
Ehts: urban, rural—nonfarm, and rural-farm. The multiple

1: . . . . .
eg’Ic‘eSSion equation may be written in the follOWing general

Y = a + blxil = b2X12 + ... + ng19 + ul

1 = l, 2, ...... N

j = l, 2, ...... 9

Yi is the ith observed value of the dependent variable;

Xij is the ith value of the jth independent variable;
\

 

10Donald J. Bogue and Dorothy Harris, Comparative
‘g‘9SEZBplation and Urban Research Via Multiple Regression and
‘g‘9~§:§3riance Analysis (Miami, Ohio: Scripps Foundation for

SSearch in Population Problems, 1954).

 

ZQLES

247

is the ith random disturbance term. It is assumed
that the ui are independent and come from a normal
distribution with zero mean and finite variance.

is the general constant term. This represents the
value of Y which may be expected when each of the
independent variables has a value of zero.

is the partial regression coefficient of the jpp
independent variable. These coefficients show the
average effect on Y of one unit change in the
accompanying independent variable when allowance
has been made for the other independent variables
of the equation.

:i_n any regression analysis of this sort, it is necessary

‘t1171;511t; the following conditions be assumed for the universe

1321:.<:>IT1 which the data are drawn: (1) the deviations from

ITeai‘Ea‘lrz‘ession are normally distributed about the regression;

(2)

the variance from regression is constant throughout the

ant ire range of the independent variables; and (3) the re—

g-ression is assumed to be of a linear form. The following

\7‘ . . . . .
EBLITTLables were inserted in each of the regreSSion equations:

The dependent variable:

Y'measures the number of children ever born per

1,000 ever-married white females age 15—44 in
the residence component of a county in 1960.

The independent variables:

X

l

was the value assigned the county by the size-
distance measure developed to approximate metro-
politan dominance for the residence component of
a county in 1960.

measures the percent of the white male employed
labor force who were farmers and farm managers in
the residence component of a county in 1960.

measures the percent of the white male employed
labor force who were farm laborers and farm foremen
in the residence component of a county in 1960.

248

4 measures the median years of school completed by
white males and females, age 25 and over, in the
residence component of a county in 1960.

X5 measures the median white family income in the
residence component of a county in 1959.

X6 measures the median white female personal income
for a county in 1959.

X7 measures the percent of white females, age 14 and
over, who were employed in the county in 1960.

X8 measures the percent of ever-married white females,
age 15-44, who were age 15-24 in the residence
component of a county in 1960.

X9 measures the percent of ever-married white females,

age 15-44, who were age 25-34 in the residence
component of a county in 1960.

In the estimation of each of the prescribed regres-
ES jL—<:>Jri equations the following statistics were computed:
Si jLJITII;>le correlation coefficients between all variables, the
e St imated partial regression coefficients, the estimated
Es«b-h-Eat‘ilridard error of the partial regression coefficients, the
eaisstlzimated standard partial regression coefficient (also
17% ferred to as beta coefficient or beta weight), the stane
(BJE‘JC’d error of estimate, the multiple correlation coefficient,
E37r1451 the coefficient of multiple determination. Let us give
SSIFJEicial consideration to some of these statistics in terms
(unis the nature of the statistic and how it will be employed

In the analysis.

:EB
“~S§Lfii§ Coefficients

 

In the testing of our hypothesis with respect to

‘A?
‘IRHEit.independent variables influence fertility behavior, it

249

_i;;5; inecessary that we be able to come to some conclusion as
-t:.<:> 110w important an independent variable is compared with
others in its relative ability to produce change in the
dependent variable. Partial regression coefficients cannot
1C><E= <ilirectly compared because they vary with the units in
‘A71i1:i_<::h.they are measured. Comparison can be made, however,
jLiff eeaach partial regression coefficient is expressed in terms
(3)355 :i.ts own standard deviation. Beta coefficients, then, are
I;>Eau1:*1:;ial regression coefficients expressed in standard mea-
53“;13ET<E:. They state the average change in the dependent vari-
ea:k:,:1_<s: that may be expected per standard unit change in one of
tzulfil‘si independent variables. Beta coefficients are obtained
by multiplying the partial regression coefficient by the

J:‘EEV'tleo of the standard deviation of the independent variable

‘:;<:> t:hat of the dependent variable. The formula used to

<:: . .
<:>Itlpute beta coeffiCients may be expressed as

SX.
B. = b. ——l
3 3 5y
1P‘alfiere

B. is the beta coefficient of the jth independent
3 variable;

b. is the estimated partial regression coefficient
3 of the jth independent variable;

8 is the standard deviation of the jth independent
j variable;

8 is the standard deviation of the dependent vari-
y able.

m.

1‘“:

250

Be ta coefficients are pure numbers which take into account

the variation in the independent variable relative to the

dependent variable. The beta coefficients are summapy mea-

s 1.11: es of the relative importance of an independent variable

in accounting for the variation in the dependent variable.

The sign of the beta coefficient indicates the direction of

the effect. Beta coefficients were estimated for all inde-
pendent variables in all equations and are used to compare
the independent variables of metropolitan dominance and

c OI‘Itur'nunity social structure in their relative importance in

de t ermining community fertility levels.

gmficient of Multiple Determination

In our analysis of differential fertility compari-
8’ one for urban and rural communities, we shall want to know
how successful the combined effect of the independent vari-
a‘b les is in determining fertility variation. The success of
the multiple-variable estimating equation in accounting for
the variation in the dependent variable may be summarized by
~Elle coefficient of determination (R2) which is simply the
11111 ltiple correlation coefficient (R) squared. In regression
tQIms the square of the correlation coefficient is an esti—
rt“ate of the proportion of the variance in the dependent
Vériable that is accounted for by the regression of the

erendent variable on the independent variables inserted in

1:Iie equat ion .

251

Ze r o—Order Correlation Coefficients

 

Zero—order correlation coefficients (Pearsonian r's)

“A7<E=:ET£3 computed in each of the regression equations to measure
1:;liiesa association between each of the independent variables
aaxeréi. the dependent variables as well as between each of the
:i_1:1<fi<32pendent variables themselves. We shall consider the
Zero—order correlation coefficients between the independent
‘J'EBJI?.jLables and the dependent variable as a first step in the
Ea-I'1Ei.3l_ysis to consider whether metropolitan dominance and
<:=<:>1111‘Inunity social structure variables do have a general
as S ociation pattern with fertility. However we should also
‘:=‘<:>1:1£3ider the possibility of intercorrelation of our indepen—
Gelit: variables to make some assessment as to whether high
it‘lrlitlearcorrelation among the independent variables might
<51Ii~11121nish the success in accounting for a large portion of
itljbl‘st variation in the dependent variable. As Blalock points
(:>‘;l1t in.his discussion of multiple correlation, "if we wish

it;<:> explain as much variation in the dependent variable as
stsible, we should look for independent variables which are
IDESVlatively unrelated to each other but which have at least
It‘Qderately high correlations with the dependent variable."11
rI“}1.is is an especially difficult task to perform in any socio—
ngical research. If high intercorrelations are discovered,
\

llHubert M. Blalock, Jr., Social Statistics (New
York :

McGraw-Hill, 1960), p. 348.

 

252

we must admit to this deficiency and make some suggestions
23. s to which variables would be more appropriate for explain-
ing a larger proportion of fertility behavior for future
study.12 Nevertheless, we shall consider zero—order corre—
lat; ion only as a preliminary analysis of what factors may
influence community fertility behavior and then move on to
more rigorous measures of association, measures which will
not only account for variation in the dependent variable by
a S ingle independent variable alone, but will also hold con—
Star-11; the influence that other independent variables may

eXert on the dependent variable.

_\

\nr 12Intercorrelations among the independent variables
care generally low. The highest intercorrelations, of
hourse, were found between the paired variables employed to
de 1E lect aspects of community social structure, such as,
negree of urbanization, socio-economic status, wife's alter-
iat ive opportunities, and demographic age structure. Lowest
intercorrelations are found for the paired variables reflect—
Q g socio—economic status (education and family income) and
wimographic age structure (ever-married females, age 15-44,
2L Q were 15—24 or 25—34) . Intercorrelations for family
mhqome and education for all divisions and the nation were
Qsmly below a value of i.20; intercorrelations for demo-
hl: Ephic age structure variables were on the average slightly
tatgher, although generally lower than —.30. Intercorrela-
tbns among variables reflecting employment in agricultural
lQQupations were high in urban and rural—nonfarm areas but
QQVV in rural-farm areas for the nine divisions. For urban
I: eas zero—order correlations ranged from .32 and .64;
gural-nonfarm .24 to .79; rural-farm coefficients were
therally below _-l_-_.l4. The most highly correlated paired
mfiriables occurred between female income and female employ-
WQ ht. The coefficients ranged from .49 to .83. No attempt
{\QS made to remove highly correlated variables from the
thression equations, though it would appear that overall
Ghere are low levels of intercorrelation among the indepen—
Qnt variables.

253

Partial Correlation Coefficients

Partial correlation coefficients are measures of the
degree of relationship between a dependent variable and a
S ingle independent variable, controlling for one or more
other independent variables. These should not be confused
With partial regression coefficients which are mentioned
above in our discussion of the multiple regression equation.
The formula for computing partial correlation coefficients
may be expressed generally as

r. . - (r. ) (r. )
rij.k 11 1k 4k
2

\/(l "' rik) (1 - rjk)

 

P a12“I:.ial correlation coefficients will be computed directly
fr Om zero—order correlation coefficients. It should be
r10ted that the square of the partial correlation coefficient
represents the proportion of variation in the virst variable

( Qependent variable) left unexplained by the third variable,

but which can be explained by the second variable. Partial
QQit‘relation coefficients will be used sparingly in the
é‘rlélysis. As an example, we shall want to use partial corre-
3‘ aHtion coefficients to consider the association between com-
hzllll-it-xity social structure variables and fertility in urban and
rural communities, holding constant the effects of metropol-
itan dominance. After considering such data, however, we
ghall want to move to more rigorous measures of association

Dr‘ovided in the multiple regression analysis.

254

Mu ltiple-Partial Correlation
C oe fficients

At certain points in the analysis it will be desir—
ab le to compute a multiple—partial correlation between a

dependent variable and more than one independent variable,

 

controlling for one or more other independent variables.
The multiple—partial correlation coefficient has not been
use 6 very frequently in sociological research, 13 but it can
]:><E= ‘lnsed at times to handle both multiple and partial corre—

lat ion problems simultaneously. The general formula for

determining the multiple-partial may be expressed as

 

 

2 2
2 _ R. . - r.
ri(jklm).n — 1.3klmn 2 in
l — r.
in
<2)::‘
2
r2 = Ri.jklmno - Ri.no
1(3klm).no l _ R2
i.no

r‘Ir‘lrlfie square of the multiple—partial correlation coefficient
Innisily be interpreted as that part of the variation in the
cEhsilpendent variable not explained by the control variable(s)
which is accounted for by the independent variables enclosed
lefil parentheses. This measure can be used in evaluating the
:IT‘EEJative importance in predicting the dependent variable of

6‘ .select number of independent variables from the original

\-

l3Hubert M. Blalock, Jr., op. cit., p. 351.

255

multiple regression equation, controlling for the effects of

the remaining independent variable(s) . For example, in com—

paring the importance of community social structure vari—

ables with metropolitan dominance, we may wish to determine

the amount of variation explained by the independent vari-

ables reflecting community social structure of the original

var iation explained by all independent variables (R2), con-

trolling for metropolitan dominance. Actually any combin-

at ion of independent variables from the original multiple

regression equation may be used. The second term of both

numerator and denominator includes only the dependent vari-
ab le and those independent variables being held constant.
The independent variables being correlated with the depen—

de r11: variable are left out of the second term of both

numerator and denominator. The first term of the numerator

8‘ lWays includes the entire battery of variables employed in

the original multiple regression equation. Multiple-

lballrtial correlation coefficients compared with the beta
QQefficients give us a more flexible approach to considering
the relative importance of independent variables in account—
ing for variation in fertility behavior because sets rather

than individual variables are considered simultaneously in

\

 

14For a comparison of the multiple-partial with the

Eartial correlation coefficient, see C. Horace Hamilton,
opulation Pressure and Other Factors Affecting Net Rural-
1951) ,

UP
2t‘loan Migration, " Social Forces, XXX, No. 2 (December,
13-14.

256

the statistic. Beta coefficients measure the relative impor-
tance always for a single independent variable in predicting
the dependent variable, controlling for all other indepen—
dent variables included in the regression equation. Multi-
gale—partial coefficients are especially useful in problems
<2<3ntaining variables which may be highly intercorrelated
k>eacause the combined effects of independent variables are
t:£aJ<en into account. Because beta coefficients are not inde-
I;><Ezridently additive, they are not the best indicators of the
JE‘tEBLLative importance contributed by two or more, perhaps

h ighly correlated, independent variables.

8“: a tistical Tests
Differences of opinion seem to exist among social
53 itliaitisticians as to whether tests of significance should be
Eitlnan310yed when all units in a limited universe have been
<:>:k:>sserved (a 100 percent sample) and what the interpretation
Es"Iii-(Duld be of these tests of significance when they are

35‘1E3191ied.15 Though we shall follow Bogue and Harris' advice

\

15This seems to be a very common problem confronted
UhDEZ’ almost any researcher using demographic data. Bogue and
1E1Eiacris suggest that no tests of significance are needed with
E3 ILOO percent sample if there is no attempt to generalize to
.larger universe in time or space. See Donald Bogue and
erothy Harris, op. cit., p. 12. Hagood and Price suggest
}1Ert such total populations be considered as "samples from‘
EstZilllarger hypothetical universes of possibilities." See
ailr'garet J. Hagood and Daniel 0. Price, Statistics for Soci-
JElebgists (Rev. ed.; New York: Holt, Rinehart and Winston,
J~S15Q), pp. 286-94. Selvin points out, however, that one must
63c:ognize that sampling error is only one type of random
€321i‘Ilz‘or. "Even where there is no sampling in the usual sense,

 

257

and not wish to generalize the findings of this study to a
larger universe than for the counties of the nation in 1960,
we shall proceed to apply tests of significance for the
statistics employed in this study due to possible random
error in response and processing as well as in sampling,
erince the census data employed in this study were collected
(>11 the basis of a 25 percent sample. However, it should be
eezrqohasized that the main concern of this thesis is locating
eaLrlci partitioning sources of variation in a dependent vari-
aafll:>.lxe, hence the use of correlation statistics. Tests of
S ignificance are of minor importance and will be used only
't:<::> Iirge caution when small statistical differences are dis-
c—‘—C>\7ered.
The "t" test was used to ascertain whether zero—
‘:’3=:‘<fieer correlation coefficients were significantly different
fibr:‘<:>rn zero. The chosen level of significance was .05. The

S t a tistic

 

has "Student's" distribution with N-2 degrees of freedom.

\

d‘:L'S<:repancies between the true situation and the observed

€25Sults may be produced by random errors of response or
- ITCDCessing. It might seem, therefore, that tests of signif-

tqance could be used to compare total populations, if the
62551:5 are interpreted as dealing with random errors of

QeSponse or processing." See Hanan C. Selvin, "A Critique

$3:ff Trests of Significance in Survey Research," American

'-JEEE§iological Review, XXII (October, 1957), 525.

258

The "F" test was used to determine the significance
of the partial correlation coefficients, multiple—partial
correlation coefficients, and multiple correlation coeffi—
cients. With respect to partial correlation coefficients we
txasted the hypothesis that the coefficient was significantly

CLifferent from zero at the .05 level. The general formula

for testing the partial rij.kl 18

2
r..
1].}(1 (N _ k _ l)

Fl,N—k—l = 1 _ r2
ij.kl

Where the total number of variables is k + l and N is the
t24C2>tzzal number of observations. This same formula may also be
1.18 e <1 to test whether the multiple-partial correlation coef-
ff-ji——<:=:ient is significantly different from zero. The .05 level
of significance is used in this test also. The general
:EE<:3fiI?Inula employed to test the null hypothesis that the multi-
I;>:1~€3: correlation coefficient (R) is not significantly differ—

eailfil‘: from zero may be expressed as

Fk,N—k—l = 1 R2 k
VVIIEire k represents the number of independent variables and
1S: tihe number of observations.
The "t" test was also used to ascertain whether the
eesstlimated regression coefficients were significantly differ—
earltl from zero. Again the level of significance chosen was

t:}1€3 .05 level. The form of the "t" test used was

CL

 

nos

 

§

»

259

b. - O
t = _J_____
Sb.
3
which has N — k — 1 degrees of freedom and where
b. is the estimated partial regression coefficient of
3 the jth independent variable;
5b is the estimated standard deviation of bj.
j

A final statistical test, the multiple comparison
tzeast, was employed to determine whether there were differ—
ences in the effects of the independent variables on the
dependent variable between residential parts of counties
(’;i4ritra-divisional) and between geographic divisions (inter—
dii:igxifiisional) for the same residential parts of counties.
iTIIfiL;i_ss involves testing the equality of partial regression
C! (De fficients between the various multiple regression equa—
t.jL.<::uras. The question which is asked in this test is whether
‘t171<EB :respective independent variables have the same impact on
f‘E=::7'tZ:Llity in all residential classifications: urban, rural—
n<:)3:13fféirm and rural farm. The question was also asked whether
the respective independent variables had the same effect on
fieancht:.leity in.all the geographic divisions. ‘Recall that our
:ijjilEB-Il. theoretical hypothesis predicts that differential
j'ITIZED‘EiCxt will be found largely between residential communities
wjL-tflj-Iicn the same division and that a comparison of the same
tbe of residential communities interdivisionally will

re .
.\7E3=Eil simularities rather than dissimilarities of impact.

260

The test used was the multiple comparison test of

Scheffe.l6

As an example of the test suppose we have k

equations and we wish to test the equality of the regression

coefficients
Let:
b. be
c.. be
of

be

T 0 test:

E?<:>:::1m;

k
a. nd
1221

SSUk

of variable Xj in the k equations.

the estimated partial regression coefficient of
in the equation k (k = l,....,u,....,v,.....k);

l

the jth diagonal element of the (X'X)— matrix

equation k;
the sum of squares of residuals from equation k;
the degrees of freedom from equation k (Qk =

- p , when N is the number of observations and
the number of parameters in equation k).

H: B‘?-BY=0
O J 3

1.1 V
Han—B- O,
l J 37,

= SSU.

the test statistic

 

 

 

 

SSDuV / (k — l)
k
SSU / Z Qk
k=l
\
C>:L. l6See K. A. Brownlee, Statistical Theory and Method-
TTE§SE§SEIy in Science and Engineering (New York: John Wiley,
0), pp. 3l6ff.

261

k

follows the 13‘ distribution with k - l and Z Qk degrees of
k=l

freedom. The chosen level of significance of the test was

.05. One assumption which must be met for the test to be

valid is that the residual variances in the k equations be

equal; that is,

This chapter has presented a discussion of the con—

<::eaptual and the statistical frameworks employed in this

1:.Ijfiesis., Concepts from the theoretical framework were trans-

;1f}:>n:med.into empirical measures, each based on an interval

S cale. Also discussed have been the multiple regression

nr1<:><iLel employed, the additional statistics considered neces—

Es £51:z:jy for the analysis, as well as the statistical tests

p><E:::::formed. In the following chapter the results of the

ane lysis will be presented, interpreted and discussed.

 

 

l7Bartlett's test of homogeneity of variance was
used to test the validity of this assumption. See G. W.
sn?decor, Statistical Methods (5th ed.; Ames: Iowa State
”n lVersity Press, 1956), p. 285.

CHAPTER V
ANALYSIS OF DATA: NATION AND DIVISION

In this chapter the results of our regression anal-
.yses will be employed to test the seven hypotheses which
vveme developed in the chapter dealing with the theoretical
:Ehramework of this study. The theoretical framework which
.i.s; used to explain rural and urban differential fertility
.1;>£at:terns for the nation and the geographic divisions of the
Jrlnait:ion is metropolitan dominance theory. The basic proposi—
‘tZL:i.c>n running through this study is that different patterns
<:>:1E’ urban and rural fertility variation are associated with
“t: ITL<E£ emergence of metropolitan centers as organizing agents
<:>:13: ihinterland communities. The test of differential urban
and rural fertility patterns, then, is also a test of metro-
I=:><:>-:‘Litandominance theory in general. The organization of
this chapter will be provided by the seven hypotheses. Each
I‘r“":.2":E.:>-::>thesis will be individually introduced followed by a
Dre Sentation and discussion of the types of data available
jET:?<:’ITI our regression analyses which are most relevant to the

t:
e S t ing of the particular hypothesis.

B1 -
W

Community social structure is a function of distance
and size of a dominating metropolitan center.

262

-¥

263

a. Distance and size of a dominating metropolitan

center will manifest a different impact on community

social structures in the urban hinterland than in

the rural hinterland.

To test this hypothesis we shall consider the zero—
order correlation coefficients which reflect the degree and
direction of association between metropolitan dominance and
'the various indices of community social structure. If metro—

Exolitan centers are organizing agents of their respective
lefban and rural hinterlands, we should expect to find signif—
.j_c:ant correlations between distance from metropolitan cen-
1::€31:s by size and community social structure characteristics.
51?.53131e 17 presents the zero-order correlation coefficients
:E?<:>J: metropolitan dominance and employment in agricultural
C:><::<::upations, measures which are assumed to reflect urbaniza-
‘t: :i.<:>n. Generally the coefficients are rather moderate in
53 :i_:2:e. Overall there appears to be a negative correlation of
‘fiujLuSS'tance from metropolitan centers by size and employment in
E‘-<EJ"31='—":i.<::ultural occupations, such as farmers and farm managers,
13523;::111 laborers and foremen. Employment in these occupations
Ciii-ITI-Iignishes as one approaches a metropolitan center. Metro-
E><:’:L—:Lctan centers seem to have more of an influence on employ—
Mel—1t of farmers and farm managers than on farm laborers and
fs<:xtr‘531nen.
In comparing urban and rural areas, there appears to
be 1. . . . . . .
ittle difference in direction of assoc1ation, however,
Vv.1“;1'1 respect to degree of association correlations of farm-

62
a: 53' and farm managers seem to be slightly higher in the

264

 

 

 

 

 

 

Table 17. Zero—order correlation of metropolitan dominance
and employment in agricultural occupations
(urbanization) for conterminous United States and
divisions, by residence: 1960
Correlation of Correlation of
Metropolitan Domi— Metropolitan
nance and Farmers Dominance and Farm
and Farm Managers Laborers and Foremen
(Zonterminous
IJnited States Urb RNF RF Urb RNF RF
51nd DiViSions r12 r12 r12 r13 r13 r13
(JIQITED STATES —.348* —.390* —.287* -.218* -.273* -.133*
.ISTeavvjEngland —.376* —.401* —.399* —.368* -.407* -.288*
lxflfljicidle Atlantic -.l76* —.372* —.428* —.378* —.392* -.287*
.12: .. N) Central -.ll6* —.ll9* -.207* —.151* -.122* —.096
1797 .. N. Central -.387* -.484* -.609* —.292* -.276* -.219*
:E; <:>NL1th Atlantic —.309* -.219* —.127* -.l9l* -.158* .045
E - s. Central -.095 -.014 .160* .112 .052 .174*
5‘7 .. £3. Central —.322* -.089 —.272* -.273* -.26l* -.283*
D<I<:>1;111tain —.249* -.246* -.205* -.189* -.100 .021
Pacific -.094 -.253* —.429* -.025 .087 .211*
\
*Significantly different from zero at the .05 level.

11153 X Metropolitan dominance (distance and size of
t :I: opolifan center) .

if
Eadtrltlers

X Percent of male
and farm managers.

labor force who are employed as

X Percent of male labor force who are employed as

if
arm lab 3

orers and foremen.

265

rural hinterland than the urban. In West North Central,
e.g., 37 percent of the variation in employment of farmers
and farm managers is explained by metropolitan dominance
among rural—farm communities, 23 percent among rural-nonfarm
communities, but only 15 percent among urban hinterland com—
munities. This pattern persists also for five other divi—
sions, especially the more metropolitan ones. For employ—
ment of farm laborers and foremen correlations are much
lower and there appears to be a reversal of pattern in terms
of the significance of metropolitan dominance in determining
the pattern of variation among urban and rural hinterland
communities, i.e., coefficients appear highest in urban
areas.

Table 18 provides the zero—order correlation coeffi-
cients between metropolitan dominance and socio-economic
status characteristics of urban and rural hinterland commu—
nities. The coefficients for education are surprisingly low,
suggesting that metropolitan centers have little influence on
education levels of hinterland communities, although coeffi-
cients tend to be higher among both rural-nonfarm and rural—
farm communities than among urban. The direction of associa—
tion is not consistent among the divisions, although for
rural-nonfarm areas the greater tendency is for educational
levels of communities to increase as one approaches metro-

politan centers.

266

 

 

 

 

 

 

Table 18. Zero-order correlation of metropolitan dominance
and socio—economic status for conterminous United
States and divisions, by residence: 1960
Correlation of Correlation of
Metropolitan Metropolitan
Dominance and Dominance and
Education Family Income
Conterminous
United States Urb RNF RF Urb RNF RF
and DiViSions 414 r14 r14 r15 r15 r15
UNITED STATES —.046* .075* —.046* .414* .472* .408*
New England -.127 .286* .283* .812* .772* .705*
Middle Atlantic —.035 .299* -.075 .297* .098 .364*
E. N. Central .105 .359* .194* .015 .043 -.026
W. N. Central -.296* -.258* -.306* -.109* .312* -.017
South Atlantic -.017 .095* .063 .703* .744* .706*
E. S. Central -.375* —.266* —.294* —.312* —.045 .362*
W. S. Central .071 -.181* -.329* .107 —.020 .001
Mountain .161* .200* —.039 .283* -.094 -.074
Pacific .053 .107 —.l60 .668* .636* .652*
*
Significantly different from zero at the .05 level.

X Metropolitan dominance (distance and size of met—
ropolitan center).

X Median years of school completed by males and

females,

X5 Median family income in 1959.

age 25 and over.

267

Of all the measures of community social structure
employed in this study, metropolitan centers seem to have
the greatest influence on family income levels in both urban
and rural hinterland communities, when comparing zero-order
correlation coefficients. The direction of association in
most cases is positive so that the greater the proximity of
a community to a metropolitan center, the higher will be its
composite level of family income. There is a high positive
correlation of metropolitan dominance and family income for
New England, South Atlantic, and Pacific divisions, but
among these there is no consistent pattern as to whether in
the urban or rural hinterland metropolitan dominance has
more influence in determining family income.

Table 19 considers the influence of metropolitan
dominance on wife's alternative opportunities in a community.
Female personal income is generally positively related to
distance from metropolitan centers by size. Female employ-
ment reflects the same association pattern although the
direction of association is more mixed. Correlation coeffi—
cients are relatively low, although metropolitan dominance
seems to influence female personal income levels more than
the female employment rate. In the comparison of urban and
rural hinterland communities, nationally metropolitan domi-
nance is relatively more important in accounting for female
personal income and female employment levels in urban areas

than rural, though among divisions this pattern is mixed.

Table 19 .

Conterxin
United 3:
and Divis

\

3:11:23 8':

1' .

268

Table 19. Zero-order correlation of metropolitan dominance
and wife's alternative opportunities for contermi-
nous United States and divisions, by residence:

 

 

 

 

 

1960
Correlation of Correlation of‘
Metropolitan Domi— Metropolitan
nance and Female Dominance and
Personal Income Female Employment
Conterminous
United States Urb RNF RF Urb RNF RF
and DiViSions r16 r16 r16 r17 r17 r17
UNITED STATES .232* .191* .195* .177* —.l3l* .139*
New England .352* .373* .457* .350* .379* .413*
Middle Atlantic -.094 —.218* -.227* .051 -.024 -.067
E. N. Central .155* .143* .105* .217* .241* .183*
W. N. Central .258* .266* .277* .048 .037 .069
South Atlantic .067 -.203* —.223* —.033 -.l77* -.182*
E. S. Central -.069 —.066 -.078 -.157* —.209* -.192*
W. S. Central .237* .204* .200* .061 —.017 -.004
Mountain .478* .382* .215* .418* .331* .209*
Pacific .301* .240* .277* .177 .135 .161

 

*Significantly different from zero at the .05 level.

X Metropolitan dominance (distance and size of met-
ropolitan center).

X Median female personal income in 1959.

2

X3 Percent females, age 14 and over, who are employed.

In New I
for 2]- E
of femal
cent f0I
jivisior
:iominanc

farm c3:

001212311:

r '3 "

269

In New England, for example, metropolitan dominance accounts
for 21 percent of variation in female income and 17 percent

of female employment in rural-farm communities, but 12 per-

cent for both cases in urban communities. For the Mountain

division percent of variation accounted for by metropolitan

dominance is greater for urban than rural—farm or rural non—
farm communities.

Table 20 portrays the relationship of metropolitan
dominance to the demographic age structure of women in the
reproductive age span of urban and rural hinterland commu-
nities. Distance from a metropolitan center by size is
generally negatively associated with ever-married females
ages 15—24 but the pattern is mixed for ever-married females
ages 25-34. Metropolitan dominance nationally is more influ—
ential in determining the distribution of females ages 15—24
in urban communities than rural, especially in the more
metropolitan divisions, but for females ages 25—34 the
influence of metropolitan dominance is rather insignificant.

In summarizing these tables we might conclude that
community social structure is only slightly a function of
metropolitan dominance. The amount of variation in commu—
nity social structure characteristics accounted for by
distance from metropolitan centers by size is highest for
family income, employment of farmers and farm managers, and
ever-married females ages 15-24. Family income is positively
associated with metropolitan dominance while employment of

farmers and farm managers and ever—married females ages

Table 20

H

Comte mi:
United S t
and Divis

\

DEED s:

 

270

 

 

 

 

 

 

Table 20. Zero—order correlation of metropolitan dominance
and demographic age structure for onterminous
United States and divisions, by residence: 1960
Correlation of Correlation of
Metropolitan Metropolitan
Dominance and Dominance and
Females, Ages 15—24 Females, Ages 25-34
Conterminous
United States Urb . RNF RF Urb RNF RF
and DiViSions r18 r18 r18 r19 r19 r19
UNITED STATES —.320 -.241 —.127 .004 .081 —.095
New England —.447 —.697 -.268 -.201 .306 .002
Middle Atlantic —.522 -.511 -.219 .356 .162 -.082
E. N. Central -.159 —.152 .025 .152 .127 -.015
W. N. Central —.076 .103 —.134 -.067 -.097 —.377
South Atlantic —.l65 —.297 —.005 .039 .094 .047
E. S. Central .011 —.001 .125 —.088 -.054 —.096
W. S. Central —.042 —.081 -.l73 -.065 .025 —.121
Mountain —.172 —.087 .139 .061 .020 -.139
Pacific -.333 -.144 -.008 .157 -.066 -.104
*Significantly different from zero at the .05 level.

X Metropolitan dominance (distance and size of met-
ropolitan center).
ages 15—44, who are

X8 Percent ever—married females,

age 15-24.
ages 15-44, who are

X9 Percent ever-married females,

age 25-34.

271

15—24 are negatively associated. We cannot conclude that
metropolitan centers have a more determinative influence on
any single broad dimension of community social structure,
such as employment in agricultural occupations, socio-
economic status, wife's alternative opportunities, or demo-
graphic age structure, since in each of these pairs of vari-
ables one of the variables stands above the other in terms
of its degree of association with metropolitan dominance.
The data included in these tables have suggested only slight
differences in the impact of metropolitan dominance on urban
vs. rural community social structures. For employment of
farmers and farm managers and education, metropolitan dom—
inance seems to be more influential in rural communities,
but for female personal income and ever—married females

ages 15-24, metropolitan dominance seems to be more impor-
tant in urban communities. However, the more obvious con-
clusion from the zero-order correlation coefficients seems
to be the similarity of impact that metropolitan dominance
exerts on urban and rural communities. In other words,
among the various divisions, where coefficients were rela—
tively large, they were large for both urban and rural areas,
and where low or nonsignificant, they were for both urban
and rural communities. This pattern also is true when con—
sidering the direction of association. Generally the data
suggest that the influence of metropolitan centers is not as

pervasive as one might expect from the theory. This suggests

politan
conditi

cially

Eveothe

272

further that additional factors to be considered are differ-
ences among the divisions due to differential rates of metro—
politanization as well as the influence of local community
conditions in determining urban and rural patterns, espe-

cially differential fertility patterns.

Hypothesis 2

 

Fertility behavior is a function of community social
structure and distance and size of a dominating metro-
politan center.

a. Community social structure and distance and size

of a dominating metropolitan center will manifest

a different impact on fertility behavior in the

urban hinterland than in the rural hinterland.

To test this hypothesis let us consider again zero—
order correlation coefficients, but we shift to a considera-
tion of fertility as the dependent variable. Is community
fertility behavior a function of both metropolitan dominance
and community social structure, regardless of whether commu-
nity social structure is a reflection of metropolitan domi—
nance or its own local environment? In considering this
hypothesis we are interested in the existence and direction
of association of fertility with metropolitan dominance and
community social structure indices as well as the impact or
importance each independent variable represents in account—
ing for intercommunity variation of fertility levels in
urban and rural hinterlands.

Table 21 offers the zero—order correlation coeffi-

cients of fertility and metropolitan dominance for each dhdsion

H

I
4 l

ropolit.

273

Table 21. Zero-order correlation of fertility and metropol—

itan dominance for conterminous United States and
divisions, by residence:

 

 

Correlation of Fertility
and Metropolitan Dominance

 

 

Conterminous United Urb RNF RF
States and Divisions rY1 rYl rYl
UNITED STATES -.211* -.268* -.l60*
New England —.490* —.676* —.451*
Middle Atlantic -.397* -.539* —.394*
East North Central -.285* —.319* —.247*
West North Central -.255* -.l65* —.l94*
South Atlantic —.264* -.213* -.182*
East South Central —.038 —.O3l -.047
West South Central —.294* -.122* -.014
Mountain -.309* -.372* -.159*
Pacific -.426* -.395* -.275*

 

*Significantly different from zero at the .05 level.
Y Cumulative fertility ratio.

X Metropolitan dominance (distance and size of met-
ropolitan center).

274

and the nation. All divisions with the exception of East
South Central indicate the existence of a significant asso-
ciation of fertility and metropolitan dominance. In all
cases, urban and rural fertility is negatively related to
distance and size of a dominating metropolitan center.
Especially in the more metropolitan divisions metropolitan
dominance is significant in accounting for fertility varia-
tion. In New England, for example, 24 percent of fertility
variation among urban communities is accounted for by metro—
politan dominance, 46 percent among rural—nonfarm, and 20
percent among rural-farm communities. Lowest coefficients
appear in the least metropolitan divisions such as East and
West South Central. Unexpectedly, while urban correlation
coefficients for the most part exceed rural-farm coeffi-
cients, the rural-nonfarm sector reflects a tendency to
exceed the other two sectors. There is enough evidence to
suggest that metropolitan dominance does have a differential
impact on fertility among the different types of residential
hinterland communities, though the direction of association
is the same throughout and the predominance of rural-nonfarm
coefficients is unexpected.

Table 22 contains the zero-order correlation coeffi-
cients for fertility and employment in agricultural occupa—
tixbns. Looking first at the association between fertility

arnfl employment of farmers and farm managers, slightly over
hallf the coefficients are significant in the residential

hirl'terlands. The direction of association among the

275

Table 22. Zero—order correlation of fertility and employment
in agricultural occupations (urbanization) for
conterminous United States and divisions, by
residence: 1960

 

 

 

 

 

Correlation of Correlation of
'Fertility' ' errtility
and Farmers _" '“2 and Farm

and Farm Managers Laborers and Foremen
Conterminous
United States Urb RNF RF Urb RNF RF
and DiViSions rY2 rY2 rY2 rY3 rY3 rY3
UNITED STATES .200* .172* .007 .358* .298* .129*
New England .149 .504* .463* .407* .674* .407*
Middle Atlantic .095 .236* .103 .143 .291* .255*
E. N. Central —.013 -.129* -.127* -.061 -.026 .185*
W. N. Central .139* .076 .083 .074 .045 .269*
South Atlantic .201* .166* —.139* .216* .229* .066
E. S. Central —.050 —.028 -.314* —.l76* .131* .081
W. S. Central .179* -.025 —.254* .654* .524* .126*
Mountain .340* .303* —.166* .216* .220* -.108
Pacific .347* .087 -.097 .307* .282* —.097

 

*Significantly different from zero at the .05 level.
Y Cumulative fertility ratio.

X Percent of male labor force who are employed as
.farners and farm managers.

X Percent of male labor force who are employed as
farm laborers and foremen.

276

significant coefficients of urban areas tends toward the
positive while in rural areas, especially the rural-farm,
the relationship tends toward the negative among the sig-
nificant coefficients. We might make reference here to
Table 15 (in Chapter II) which summarized the expected
relationships of fertility to the independent variables
established on the basis of previous research on differen-
tial fertility. This table proposed an inverse association
in rural areas and a nonsignificant association in urban
areas, but our findings suggest the greater contrast of
negative in the rural and positive in the urban. It does
hold, however, that correlation coefficients are larger in
the rural communities, suggesting that employment of farmers
and farm managers is more determinant of rural than urban
fertility levels.

With respect to employment of farm laborers and
foremen, on the basis of the empirical propositions estab-
lished in Table 15 of Chapter II, it was expected that it be
positively associated with fertility in rural areas but non-
significant in urban areas. Only three of the urban coeffi-
cients are nonsignificant, however, whereas four of the
rural-farm coefficients are nonsignificant. According to
the data of Table 22 the positive relationship holds gener—
Eilly in both urban and rural areas. However in the North-

euastern and North Central divisions the rural coefficients

53133 larger than urban, but for the divisions in the South

277

and West the urban coefficients appear to be more important
in accounting for fertility variation. For example, employ—
ment of farm laborers and foremen accounts for 17 percent of
rural-farm fertility, 45 percent of rural—nonfarm and 17 per-
cent of urban in New England, but in the West South Central
fertility variation is accounted for by this independent
variable 43 percent in urban communities, 27 percent rural—
nonfarm, and only 2 percent rural—farm. On the basis of
these findings we conclude that employment in agricultural
occupations as an index of community social structure does
have a different impact on fertility in urban than in rural
hinterlands, though the pattern is not consistent for farm
laborers and foremen.

In Table 23 are given the values for the zero-order
correlation coefficients for fertility and socio-economic
status. Education is for the most part negatively corre—
lated with fertility, although moreso in the rural residen-
tial areas than urban. Education is not a significant pre-
dictor of fertility levels in all residential sectors of the
Middle Atlantic division. For the urban hinterland of the
more metropolitan divisions, i.e., New England, Middle
Atlantic, and East North Central, the coefficients are
.positive and nonsignificant. Generally education is more
iJifluential in determining fertility levels in the rural

hinterland communities, except in the case of the West South

(Zenntral and Mountain divisions where education accounts for

 

U

Contez
United
and Di

.1 ‘ .

Pacific

 

 

278

 

 

 

 

 

 

Table 23. Zero-order correlation of fertility and socio—
economic status for conterminous United States
and divisions, by residence: 1960
Correlation of Correlation of
'Fertility >Fertility
‘ ' and ’ ' and: ‘
Education Family Income
Conterminous
United States Urb RNF RF Urb RNF RF
and DlVlSlonS rY4 rY4 rY4 rY5 rY5 rY5
UNITED STATES -.l77* -.236* —.159* —.032 —.052* .042*
New England .095 —.443* —.247 -.473* -.666* -.443*
Middle Atlantic .092 -.160 .014 —.272* -.037 -.l37
E. N. Central .027 —.272* —.426* .279* -.272* -.234*
W. N. Central -.086 -.222* -.318* —.283* -.216* —.091*
South Atlantic —.263* —.395* -.274* —.265* -.059 -.043
E. S. Central -.l62* -.374* —.222* -.024 -.169* -.027
W. S. Central —.621* -.528* —.464* .017 —.095* -.289*
Mountain -.507* -.455* .013 .047 -.230* .064
Pacific -.434* —.586* -.294* —.311* —.318* -.353*
*Significantly different from zero at the .05 level.

Y Cumulative fertility ratio.

X Median years of school completed by males and

females,

X Median family income in 1959.

5

age 25 and over.

279

39 percent and 26 percent of fertility variation in the
urban hinterland communities. These findings are in general
agreement with the empirical propositions established in
Chapter II in terms of direction of association and differen—
tial impact of the importance of education in accounting for
fertility level variation in rural and urban areas.

Family income, the second index of socio—economic
status of communities, is also negatively associated with
fertility, although moreso in rural than urban areas.

Family income is especially significantly related to fertil-
ity in the more metropolitan divisions. The empirical prop-
ositions from Chapter II relating to this variable suggested
the possibility of a positive association of family income
and fertility in the urban hinterland, but this is not sup—
ported by the zero—order correlation coefficient presented
here. Furthermore, on the basis of the empirical proposi-
tions established in Chapter II, it was expected that family
income coefficients would be higher in urban than rural sec-
tors, but this is supported only in the case of four divi—
sions: Middle Atlantic, East North Central, West North
Central and South Atlantic. Hence the differential impact
of family income on urban and rural fertility is not as
Clear as education. Family income appears to be a better
Eiredictor of fertility levels in the urban hinterlands of
tile more metropolitan divisions and education a better pre-
Ciixztor in the urban hinterlands of the less metropolitan

(iii\fisions, while both education and family income are

280

moderate predictors of fertility in the rural sectors of
most divisions, education being relatively more important
than family income.

Correlation coefficients of fertility and wife's
alternative opportunities are listed in Table 24. In the
case of both female personal income and female employment
there is a negative association with fertility. This is
again in agreement with the empirical propositions of
Chapter II. It was expected, however, since working women
is more an urban characteristic than rural, that the indices
of wife's alternative opportunities would be nonsignificant
in rural areas but significant in urban areas. With respect
to female personal income this does not appear to be the
case except for Middle Atlantic. Generally the coefficients
indicate that female personal income is significantly re-
lated to fertility in all residential sectors. In six
divisions, however, female personal income in the urban
hinterland does exceed either of the rural hinterlands in
accounting for variation in fertility and they are Middle
Atlantic, West North Central, South Atlantic, West South
Central, Mountain, and Pacific. Hence one might argue that
the impact of female personal income on fertility is slightly
different, i.e., greater, in the urban hinterland than the

Ifilral. In the case of female employment, on the other hand,
ififive of the rural-farm coefficients among the divisions are

nOnsignificant and, hence, concur with the empirical

Table

II

 

281

 

 

 

 

 

 

Table 24. Zero-order correlation of fertility and wife's
alternative opportunities for conterminous United
States and divisions, by residence: 1960
Correlation of Correlation of
-‘Fertility‘T -Fertility
t: and Female .7 and: -
Personal Income Female Employment
Conterminous
United States Urb RNF RF Urb RNF RF
and DiViSions rY6 rY6 rY6 rY7 rY7 rY7
UNITED STATES —.269* -.245* —.l72* -.210* —.301* -.l63*
New England -.248 —.490* —.4l7* -.231 —.388* -.252
Middle Atlantic -.286* —.097 -.011 —.239* -.235* -.008
E. N. Central —.l95* -.229* -.134* —.088 -.182* -.039
W. N. Central -.387* -.150* —.244* -.l72* -.044 -.036
South Atlantic -.301* —.233* -.131* —.227* -.355* —.183*
E. S. Central -.185* -.208* -.199* -.320* -.510* -.413*
W. S. Central —.288* -.l69* -.l96* -.157* -.l94* -.205*
Mountain —.458* —.343* —.l99* -.522* -.466* -.237*
Pacific —.454* —.201* —.066 -.240* -.133 —.063
*Significantly different from zero at the .05 level.

Y Cumulative fertility ratio.

X Median female personal income in 1959.

6

X Percent females,

7

age 14 and over,

who are employed.

.D

282

propositions of Chapter II. West North Central and Pacific
reveal both rural-farm and rural-nonfarm coefficients as
nonsignificant. However female employment does not appear
to be as important in predicting urban fertility as expected.
Furthermore female employment in rural-nonfarm hinterland
communities often reveals a greater amount of fertility
variation eXplained than in urban communities. While there
is indication of a differential impact of female personal
income on urban and rural fertility, female employment does
not suggest a clear pattern.

Table 25 contains the zero—order correlation coef—
ficients of fertility and demographic age distribution of
women in the child—bearing period. Generally the tendency
is toward a negative association of fertility and the prev-
alence of ever-married females ages 15-24 and a positive
association of fertility and ever—married females ages 25—34.
This concurs with the empirical propositions of Chapter II.
With respect to ever-married females ages 15-24 there seem
to be exceptions to the expectation that it would be a
better predictor of fertility levels in the urban than the
rural hinterland, and these exceptions appear in three
highly metropolitanized divisions: New England, Middle
.Atlantic, and Pacific. In these three divisions the urban
Cnaefficient is not significant. Generally, however, females

aEJes 15-24 is more frequently a significant variable in
(iertermining fertility levels for urban than rural areas.

TI'1O'Ugh most of the coefficients for ever—married females

283

Table 25. Zero-order correlation of fertility and demo—
graphic age structure for conterminous United

 

 

 

 

 

States and divisions, by residence: 1960
Correlation of Correlation of
’ Fertility . “Fertility‘
. .,.and " -* p and;.;-’
Females, Ages 15-24 Females, Ages 25—34
Conterminous
United States Urb RNF RF Urb RNF RF
and DiViSions rY8 rY8 rY8 rY9 rY9 rY9
UNITED STATES -.271* —.055* -.ll3* .203* —.004 .110*
New England .106 .619* -.073 .190 —.244* .065
Middle Atlantic .081 .308* .073 .032 —.l63 -.016
E. N. Central —.455* -.125* —.269* .129* .021 .028
W. N. Central -.478* —.269* —.l33* .238* .110* .155*
South Atlantic —.290* -.027 —.069 .169* —.ll6* .089
E. S. Central -.454* -.039 —.019 .174* —.101 -.009
W. S. Central -.282* —.042 ..051 .349* .133* .036
Mountain —.291* —.074 —.368* .047 -.009 .036
Pacific -.127 -.044 -.l99* .164 .070 .023

*Significantly different from zero at
Y Cumulative fertility ratio.

X8 Percent ever—married females, ages
age 15-24.

X9
age 25—34.

Percent ever—married females, ages

the .05 level.

15—44, who are

15-44, who are

284

ages 25—34 are relatively small and especially nonsignifi—
cant in the rural hinterlands, the pattern seems to be a
positive relationship to fertility in the urban hinterlands
and a positive but nonsignificant relationship in the rural
sectors. These coefficients seem to support the expecta-
tions based on the empirical propositions of Chapter II with
ever-married females ages 25—34 accounting for more of urban
fertility variation than rural. It is not as significant a
variable, however, as the prevalence of ever-married females
ages 15-24. The differential impact of demographic age
structure on urban and rural fertility is not clear. These
variables, however, are more significant in urban hinter-
lands than rural in determining fertility levels.

Before proceeding to a consideration of additional
data let us summarize what has been discovered to this point.
With respect to metropolitan dominance there is a consis—
tently significant negative association with fertility. It
is especially important in the more metropolitan divisions.
A differential impact on urban and rural fertility is sup—
ported, in that urban coefficients exceed rural-farm, but
rural—nonfarm coefficients were found to exceed both. With
respect to employment in agricultural occupations, farmers
and farm managers are negatively correlated with fertility
in rural areas but positively in urban areas. The differen—
tial impact hypothesis is supported in direction of associa-

tion and importance of the variable in that rural coefficients

excee
assoc
the s
secto:
the V5
status
ciated
nities
the fac
Family
the mgr
are lar
Opportu;
ment are
rural CC

on urban

to exCee

 

 

285

exceed urban. Farm laborers and foremen are positively
associated with fertility in urban and rural communities and
the size of coefficients vary randomly among the residential
sectors which suggests no consistent difference in impact of
the variable on fertility. Both measures of socio—economic
status, education and family income, are negatively asso-
ciated with fertility, moreso among rural than urban commu-
nities. A differential impact of education is supported by
the fact that coefficients are larger in rural areas.

Family income reflects a differential impact on fertility in
the more metropolitan divisions in that urban coefficients
are larger than rural. With respect to wife's alternative
opportunities both female personal income and female employ-
ment are negatively associated with fertility for urban and
rural communities. Female income has a differential impact
on urban and rural fertility in that urban coefficients tend
to exceed rural. For female employment, however, the differ-
ential impact is not clear. Urban coefficients exceed rural-
farm coefficients, which are mostly nonsignificant, but
highest coefficients appear for rural-nonfarm communities.
Finally considering the demographic age structure of commu-
nities, the prevalence of ever-married females ages 15-24 is
negatively related to fertility but for those ages 25-34
there is a positive association. The differential impact of
ever-married females ages 15—24 on urban and rural fertility
is partially supported by the occurrence of higher coeffi—

Cients in urban areas, but mostly for the less metropolitan

divisi<
divisi<

are qui

Variable

existenC

 

 

286

divisions. Urban coefficients of the more metropolitan
divisions are significant but coefficients in rural areas
are quite frequently nonsignificant. Ever-married females
ages 25-34 reflects a clearer pattern of differential impact
on fertility in that the urban coefficients exceed the rural.
To conclude this summary it is asserted that fertility is

a function of metropolitan dominance and community social
structure. For the most part many of the empirical proposi—
tions established in Chapter II were supported. Differen-
tial impact of the independent variables on fertility by
residential type of hinterland, as suggested already in
Chapter II, is to be discovered more in the importance of
variables in accounting for fertility variation than in the
existence or direction of association.

To further test the hypothesis of the existence of
an association and a differential impact of community sOcial
structure and metropolitan dominance on fertility, let us
consider coefficients of multiple determination (R2) which
will demonstrate the relative importance of the combined
effects of the independent variables on fertility. We shall
review only the total correlation of the combined effects of
all independent variables on fertility.

Table 26 diSplays the coefficients of multiple deter-
mination which indicate the amount of variation explained in
fertility by all independent variables, metropolitan dominance

and community social structure. In all cases the coefficient

Table 26.

287

Coefficients of multiple determination (R2) of

fertility and variables in multiple regression
equations for conterminous United States and

 

 

 

 

 

divisions, by residence: 1960
R2
Y.123456789
Conterminous United
States and Divisions Urb RNF RF
UNITED STATES .3376* .2715* .1657*
New England .4588* .7307* .4548*
Middle Atlantic .3751* .3832* .2297*
East North Central .4097* .2815* .3597*
West North Central .4212* .1912* .3325*
South Atlantic .3966* .3634* .2017*
East South Central .3409* .3955* .3389*
West South Central .6598* .4687* .3433*
Mountain .5738* .4237* .2653*
Pacific .6270* .5728* .3064*
Average Value .4736 .4234 .3136
*Significantly different from zero at the .05 level.

Y Cumulative fertility ratio.

288

is significantly different from zero. For urban hinterlands
among the divisions the amount of variation accounted for by
the independent variables ranges from a low of 34 percent
for East South Central to a high of 60 percent for West
South Central. For the rural-nonfarm sector the range is
from a low of 19 percent to a high of 73 percent for West
North Central and New England, respectively. The range in
the rural—farm sector is from a low of 20 percent in South
Atlantic to a high of 45 percent in New England. Nationally
the independent variables account for more variation in fer-
tility levels among urban communities than rural—farm, with
rural-nonfarm intermediate. If we consider the average
values of the coefficients among the divisions the pattern
remains the same, with urban coefficients explaining an
average of 47 percent of the variation in fertility, rural—
nonfarm 42 percent, and rural-farm only 32 percent. The
combined effects of the independent variables on fertility
is in all divisions greater in the urban sector than the
rural-farm. This suggests that fertility is a function of
community social structure and metropolitan dominance com-
bined and that these factors exert different effects on
urban fertility than rural—farm fertility. But the rural-
nonfarm coefficients are not consistently intermediate
between urban and rural-nonfarm. The coefficient of multi—
ple determination in the rural-nonfarm sector exceeds the

urban coefficient in three divisions: New England, Middle

an
N: u

L

v“

,,r~

289

Atlantic, and East South Central. More frequently, however,
the rural—nonfarm coefficient is less than the urban. Hence
with some reservation, it is concluded that the hypothesis
of a differential effect by the independent variables on

urban and rural fertility is supported.

Hypothesis 3

 

Community fertility behavior in both urban and rural
hinterlands is not only a function of the size and
distance of a dominating metropolitan center, but also
a function of conditions of its own immediate locality,
since all local communities in the metropolitan region
possess some degree of dominance over some portion of
the hinterland.

In order to test this hypothesis what is required is
to determine the direction and degree of association between
fertility and the various indices of community social struc-
ture, holding constant the influence of metrgpolitan domi—
nance. In other words, if fertility continues to be signif-
icantly related to community social structure variables
after the influence of a dominating metropolitan center has
been controlled, then we must conclude that this hypothesis
is supported. It is assumed, of course, that the associa-
tion of fertility and the indices of community social struc-
ture, after partialling out the effects that may be due to
metropolitan dominance, is a reflection of the influence of
local community conditions alone. The types of statistics
Which can be employed to test this hypothesis are (1) first-

order partial correlation coefficients for fertility and the

Various indices of community social structure controlling for

290

distance and size of a dominating metropolitan center and

(2) multiple—partial correlation coefficients for fertility
and combinations of the variables reflecting community social
structure again controlling for metropolitan dominance.

The partial correlation coefficients are contained
in the Appendix (Tables 49—52) and are only briefly dis-
cussed at this point. Generally the partials portray the
same patterns as the zero-order correlation coefficients.
The partials indicate that the statistical association
between community social structure variables and fertility
is not wholly due to the influence of metropolitan centers
on their hinterlands. Even after metropolitan dominance is
controlled, the first—order partials reflect for the most
part a significant relationship between fertility and com—
munity social structure. Some significant patterns among
the partials are worth mentioning. For employment of farm-
ers and farm managers only three urban coefficients among
the divisions reflect a significant association with fertil-
ity and these are positive. Highest coefficients are found
in the rural—farm hinterland and these are mostly signifi-
cant and negative in direction. The pattern for employment
of farm laborers and foremen remains inconsistent. The
direction of association and significance of the relation-
ships remain almost unchanged also for the indices of socio—
economic status as well as wife's alternative opportunities.
With respect to the distribution of ever—married females in

the child-bearing ages, by partialling out the effects of

it

291

metropolitan dominance, the tendency for urban coefficients
for both ever—married females ages 15—24 and 25-34 to exceed
the rural coefficients is increased.

Additional evidence of the existence of an indig-
enous relationship between fertility and local community
conditions, extraneous to metropolitan dominance, is ob-
tained from a review of multiple-partial correlation coef—
ficients controlling for distance and size of a dominating
metropolitan center. Table 27 arrays the multiple-partial
correlation coefficients (r2) which indicate the proportion
of variation in fertility explained by the combined effects
of all community social structure variables after metropol-
itan dominance has been partialled out. All coefficients
are significantly different from zero, hence, it is assumed
that fertility is a function of the local conditions of
rural and urban hinterland communities apart from the dom-
inance of metropolitan centers. Urban coefficients, how-
,ever, exceed both rural sectors in seven of the nine divi—
sions. Tables 28 through 31 provide the multiple-partial
correlation coefficients of fertility and the paired vari-
ables which represent the four broad dimensions of community
social structure, each coefficient controlled for the influ-
ence of the metropolitan dominance variable. These tables
indicate further that the combined effect of these paired
variables reflecting community social structure exert a sig—

nificant influence on fertility over and above metropolitan

I1-IIIE

Lf‘

t C C -.H Una. L L 000

L O

FRO

.1. Irr R. l ‘1‘ TL F¥DI.U
mu m... -H 3 02:3 .0

292

.Aumucmo smuflaomouuma mo mNHm tam mocmumfipv mosmcHEOU Cmuflaomouumz

ax

.OHumn muflafluumm m>aumassso.w

. #CQHUHWMQOU GOHHMHMHHOU

amauummlmamfluass tam coflumsflfiumump mHQHuHSE mo ucmflowmmmoo cmmzumn mocmummmwmaa

.Hm>ma mo. mzu um oumN Eoum ucmHmMMHU kHUGMUHMHQmHme

 

omo. vow. wam.
omo. tomm. *QOM.
mac. towm. *mwm.
ooo. amwm. amwm.
moo. ahmm. xmmm.
mmo. awha. *Nom.
owe. them. *mmm.
Nvo. *mam. «00m.
Nfia. ammo. tomm.
mma. aoam. *mmfl.
NNO. awwa. «mod.

mmo. mmm. mmd.
who. tflmv. *mnm.
moo. *Hmm. *fimw.
moo. *Hofi. *mmfi.
H00. Rmmm. swam.
omo. ammm. amom.
NNO. amoa. *HmH.
Nmo. *oom. *NmN.
NmN. *HMH. tmwm.
bmm. teem. *Hmn.
hmo. *mam. *th.

 

H.im..mva

4*m m m

 

Enablamudm

H. azswa

*«Q N N

 

 

Enmwsozlamusm

ll

moo. .moe. ewe. msam> mmmum><
mmo. amem. ammo. ogweomm
meo. ammm. *eem. cemucsoz
mmo. «map. , *ooo. Hmuucmo nusom ummz
aoo. *ovm. aawm. Hmuucmo nusom ummm
oeo. *Hmm. «mam. oflucmaua nusom
ovo. *Hmm. *Hmw. Hmuucmo :uuoz ummz
mmo. ammm. toae. Hmuucmo npuoz ummm
baa. «mmm. :mem. oeucmHum mavens
Neg. *smm. tome. pcmamcm 3mz
Hmo. anom.. «wmm. mmeasm omeHzp
Ram H Am NVNH m mcoflmfl>flm 0cm
m m mmumum capes:

chHD wDOCHEHmuaoo

 

coma "mocmpflmmu an .mcowmw>flp

tam mmumum pmuflcb msocflsumucoo How musmsHEOU cmuflaomouuma How wGHHHOHpGOU
musuosnum HMHUOm muHcsEEoo paw muflafluumm mo ANHV mucmaoflmmmoo coaumamnuoo
Hafiunmmlmamfluasa tsp Ammv coaumsflfiumumo mamfluHSE mo usmflowmmmou mo comanmmaoo .hm magma

.t -... r.. J a: c

.34- 1 .fi a s 3 p:

I A“
“H 3

I- \ 4r; ah .p 7.» m. Hr“ 0;..qu

a» C U -.H U ..H ...» pH .3 C U

MN. at

C «U ...H M .... ls
:nv T... H. (4

.I I‘ll urn

.- mm. :4. -40 U.

..w. .. :CP. Luv

INN

C. x «~53?

293

.Aumucmo smuflaomouums mo mNHm tcm mocmumfltv mocmsHEOG cmuflaomouumz

ax

.OHpmu muflaflpumm m>wuma5§do M

.pdmwowmmmou coaumamuuou

HMfluHmmlmHmfluada tam coflumsafiumump mHmHuHSE mo ucmHOmemoo cmmBqu mo:mnmwma0**%

 

 

 

 

 

MNH.M

Nmaa

.Hm>ma mo. opp um oumn Eoum quHmMMHp MaucmoHMHcmHme
000. who. mmH. mHH. HHH. mNN. boo. moo. Nma. msam> mmmnm>¢
ano. *Hoo. «Nma. Hma. «oma. *nmN. mma. «oma. *mOM. oawaomm
mNo. thoo. tome. mNH. ammo. emom. hmo. ammo. aowa. camucsoz
ooo. *Nmo. *Nmo. Hao. «owN. *th. Nmo. team. twee. Hmuucmo nudom pmmB
Noo. *QHH. *mHH. Hoo. ammo. *mmo. Hoo. *Omo. tame. Hmuucmu nusom ummm
mmo. ammo. ammo. mwo. awwo. ammo. boo. ammo. awoa. UHpsmau¢ nusom
omo. ammo. *Nmo. hmo. ooo. «nNo. moo. moo. ammo. Hmuucmo nuuoz ummz
emo. *OHH. xfloa. moo. admo. *NMH. omo. emao. *Nmo. Hmuusmo nuuoz ummm
mva. twee. *Nma. owm. Mao. Room. mma. Hoo. amma. oaucmau< matte:
mha. «Nwa. eham. 0mm. *Nmm. amvo. mom. aaea. «mem. tamamcm 3oz
mNo. aeao. ammo. moo. ammo. *hNH. oeo. anoa. ahea. mme¢em QMBHZD
team H.AMNVMH m seam H.Ava%H m Reta H.5mmvwn m chHmH>HQ tam

N **N N ¥¥N N %*N

mmumum UmuHCD
Eummlamusm Emmmcozlamusm chHD msosflsumcoo

 

 

owma

"mocmpflmmu >9 .mQOHmfl>fl© paw mmumum pmpflcb

msocflfinwucoo mom mocmsflEOG smuflaomouumfi How mGHHHOHuGOU AcoflumNHstHsv chfiumm
Isuuo Hmnsuasoflumm Ga ucmahonEm tam muflafluumm mo ANHV unmaoflmmmou coflumamunoo

Hmwpummlmamwuasfi 6cm Ammv coflpmcHEHmumU mamfluass mo usmflofiwmmoo mo cemwummEou

.mN mHQmB

r .. . .r _, witLCL LC AIL \ .:.~U.H.HVJ.HL..u CCU :Qfiu m..~.U.u-..~CU
1 w r. .r -.. FIFE CCU AJEV :QH J...:-.::<A.........7UTV 04.1% ufizzu LMV UCUflUflLL UCU LO ..FcIflhfiJfitOU -0 N mafiQnurH.
C

 

 

294

.Aumucoo smuflaooouuos mo mnem osm mummumflov oucmsflsoo cmuflaomonumz ax
.OHumn huflafluumm m>Humanfiso.w
.ucmfloflmwmoo sowpmamuuoo
Hmfluummloamfluasa ocm coflumcflsuouoo mHmHuHDE mo ucmfloflmmwoo cmmBqu oUGmHmMMHosea

mva.%
Nm
.Hm>ma mo. msu um OHoN Scum usmHmMMflo mauGMUHMHcoflma

¥%

 

 

Hoo. oHH. oha. NHH. oma. ooN. moo. Hoa. omN. m5am> momum><
Hoo. Roma. *mfiN. moo. «mom. *omo. ova. «mON. *Nmm. Damaomm
mNo. moo. *mNo. ooa. sva. towm. Noo. towN. *mam. chuQSOS
ooo. amoN. amoN. oao. RONm. tomm. mmo. toom. toow. Hmuusmo Susom ummz
Noo. *Noo. *«oo. Hoo. Roma. *ooa. Noo. *mmo. «ooo. Hmuucmv npsom ummm
Hmo. tamo. *NHH. mmo. ahma. *moa. moo. *moa. *moa. UHquHufi Susom
Nmo. *ooa. *HON. wNo. thoo. *HNH. mmo. «mNH. eawa. Hmuucmu zuHOZ umoB
Hmo. «moa. aoHN. Noo. tooo. *Nma. moo. «moo. *ooa. Hmuucmo Spuoz ummm
oma. ooo. *oma. ooN. Hoo. *HoN. NmH. towo. RNoH» Uflpcmaufi maooflz
boa. tamo. emoN. oom. *mmN. *dmm. mmN. NNo. *th. ocmaocm 3oz
vNo. ammo. *Noa. ooo. *mmo. *HmH. mvo. «ooo. *omo. mMBdBm QMBHZD
aeao H.Ameku m ««*o H.Amvku m *a«o H.Ameku m muoflmg>flo ppm
N atN N *tN N «%N

 

 

 

I. mmumum panama
EHmmIHMHSM EHMWQOZIHMHSM GMQHD WSOCHEHOHCOU

 

 

oooH "mocooflmou

an .chHmH>flo ocm moumum omuacb msocflaumucoo Mom mosmcflsoo cmpflaomouuma How

ocflaaonucoo mSumum UHEOGOUoIOHUOm ocm muflaflunom mo A no usmfioflmmmoo coaumamnuoo
ammuummlmamfluasa osm Ammo coflumcflfiumuoo mHmHuHDE mm acmAUHmmooo mo somflnmmaoo .oN manme

A

N

y:

231w U CZH EMU...“ O?

y P

... Turf» t5.

OHQHuHSE

’N(‘\ ..‘fhriurcor.l“f.\([

LO uCQfl UHLL ECU LO

alof~ F i$ «H h f\.-h lbf\p(\

C0 WHH CQECU

on.

NJHIQLMFNw

295

.Aumucoo smuflaomonuma mo mNHm ocm oosmumflov oucmcflfioo swpflaomouumz ax
.oHumH mpaaflunom m>fiumassso %
.ucmwoflmmmou coaumamunoo
HMfluummlmHmfluHSE ocm coflumsflsumuoo mamfluasa mo pcmfluammmoo GmmBumn oquummmaQast

boH.M
N

.Ho>ma mo. opp um oumn Eonm usmuommao mausmowmflcowma

mat

 

 

Noo. moo. mNH. mNH. ooa. omN. ooo. voa. NON. ®5Hm> momnm>¢
ooo. Hoo. ammo. me. wao. *moa. wma. *moa. «NOm. UHMflUmm
«No. «moo. *Nho. baa. *mma. tooN. moo. amHN. amoN. Gwmucfioz
Hoo. ammo. avmo. wao. *Nvo. aomo. Noo. aomo. amma. Hmnpcmu flusom ummB
Noo. awoa. *ooa. ooo. abom. abom. Hoo. *VHH. *mHH. Hmnucmo nusom ummm
Hmo. *Hmo. *Nmo. mmo. aboa. *mON. woo. ammo. *NmH. UHpcmau< nusom
mmo. tomo. *omo. mNo. «wao. saoo. hmo. tmNH. *mma. Hmuusmo QUHOZ ummB
omo. *hNo. aomo. moo. *omo. thmH. who. aawo. *oHH. Hmuucmo SuHOZ ummm
Nma. mNo. amha. moN. ammo. *mmm. hmH. «NmH. *ooN. UHquau< GHUUHE
owH. aoHH. *ooN. oom. amma. aon. ooN. boo. tth. pcmamcm 302
mNo. «mNo. *omo. boo. ammo. *mva. Noo. tomo. tooo. mMB<Bm QMBHZD
#tRQ H.AhoVWH m %«¥Q H.Ahov%H m *xe H.ABOVNH m mQOHmH>HQ Uflm
N **N N «*N N aaN

 

 

 

mmpmum pangs:
Eummlamusm EHMmCOZIHmmsm swamp mSOCHEHmuaOO

 

 

oooH "mosmoflmon an .mQOHmH>Ho

ocm mmumum omufisb wsosflauwusoo Mom mocmsflEoo cmuflaomouumfi Mow oceaaonusoo

mofluflasuuommo m>wumcumuam m.oma3 ocm wuflafluumm mo ANHV usmfloflmwmoo cowumawnuoo
Hafiuummlmamfluasfi ocm Ammo Goaumsflsnouoo mamauasfi mo ucmwoflmmmoo mo comHummEoo .om manme

... l ... ...» k (C re L. C CO H. A. .... ONZZUU . H. Mn 0 N. «N H. ..N

H

.Anmpcmo Cmuflaomouums mo mNHm ocm mosmumflov oUCMGHEoo cmuflaomouuoz X

.oflpmu suflaguumm m>npmassao »

296

.usmfloflmmmou COHumaoHHoo
Hafiuummlmamfluasfi ocm coaumcafiuouoo mamfluHSE mo ucmfloflmmooo cmmemQ mocmummmwmaa«

 

 

 

 

 

 

 

omH MM*¥
N
.Ho>oH mo. may no onmn Eoum uconoMMHo NHuCMUNMHcoflmt
moo. omo. moa. Hma. Hmo. Noa. ooo. hma. mmN. osHm> momum>¢
Nho. tooo. *oHH. wma. mHo. aboa. mma. *Hma. tomm. oamaomm
NNo. tome. *Nma. omH. omo. *oma. Noo. ahma. *oHN. Camussoz
ooo.. ooo. ooo. oao. aomo. evmo. mno. Roma. *NmN. Hmuucmo nudom umoB
Noo. Hoo. moo. Hoo. *hHo. «mac. Hoo. anom. emom. Hmuucmo Susom ummm
mmo. *Nao. amoo. woo. tomo. emoo. ooo. amva. *Nom. newsmaud susom
omo. ammo. *ooo. mNo. «boo. *Noo. moo. «ooN. *on. Hmuusoo Quuoz “mm:
omo. eono. *Hma. moo. *omo. avma. omo. *mwN. «mom. Hmuusmu sunoz ummm
mma. ooo. Roma. mom. omo. anom. ova. ammo. *oom. oausmaum oaooflz
moa. Hmo. teem. Hmw. eoho. «oom. mmN. oNo. *NoN. ocmaocm 3oz
mNo. ammo. amvo. omo. «0mo. aooa. mmo. amoa. tome. mmemam QMBHZD
*RtQ H.Aowv%H Nm ¥¥%Q H.Aomv%H m *ttm H.AomVNH Nm mGOHmfl>HQ GEM
m ...... l m 3m! m .3. mmumum ©3ch
Eummlamnsm EummCOZIHmHsm chHD msosHEHmucoo
oooa "oocmowmmu an

.mQOHmH>flo ocm mmumum omuwcb msosflfinmpcoo How musmcflfioo chHHomouumE How ocwa
IHouucoo musuosuum mom oanmmuooamo ocm muwafluumm mo ANHV ucmHUHmmmoo QOHuMHmHHoo

HmfiuummlmamfluHSE ocm Ammo coflumsflaumumo mHmHuHDE mo ucoflowmmmoo mo comflumgeou .Hm magma

(U

rh

the
(De:
C011

Rura
\

297

dominance. Employment in agricultural occupations is statis-
tically significant, especially in the rural-farm hinter-
lands. Higher coefficients are found in the rural sectors
than the urban except for the West South Central, Mountain,
and Pacific divisions. Socio-economic status is also sig—
nificant in most cases with rural coefficients exceeding the
urban with exception of the Middle Atlantic, West South
Central, and Mountain divisions. Wife's alternative oppor-
tunities and demographic age structure coefficients are
clearly higher in the urban hinterland than the rural.
Coefficients are significant in most cases except for the
demographic age structure variables in rural—farm areas which
tend toward being nonsignificant. These data suggest the
support of the hypothesis that fertility is also a function
of the conditions of the local community as well as metropol—
itan dominance. This finding raises some doubt, then, as to
the pervasive influence of metropolitan centers as organizing
agents of characteristics of hinterland communities. This
finding lends support to the ideas expressed in the articles
by Grigg and by Anderson and Collier.l In other words, if
metropolitan dominance was the primary determinant of inter-

community variation in urban and rural hinterlands, one would

 

1Charles M. Grigg, "A Proposed Model for Measuring
the Ecological Process of Dominance," Social Forces, XXXVI
(December, 1957), 128—31; and Theodore R. Anderson and Jane
Collier, "Metropolitan Dominance and the Rural Hinterland,"
Rural Sociology, XXI (June, 1956), 157.

 

298

expect that coefficients, controlling for metropolitan
dominance, would prove to be nonsignificant. This is not
the case with respect to the influence of internal community

characteristics on fertility behavior.

Hypothesis 4

 

Community fertility behavior is more a function of dis—
tance and size of a dominating metropolitan center in
the urban hinterland, but more a function of local com-
munity social structure in the rural hinterland, when
distance and size of the dominating metropolitan center
are controlled.

We have established that community social structure
is partially determined by metropolitan dominance and that
community fertility behavior is a function of both metropol-
itan dominance and the indigenous characteristics of local
hinterland communities. The question which is raised by
this fourth hypothesis is whether fertility levels of urban
communities is more a function of metropolitan dominance than
the fertility of rural communities? This hypothesis is based
on the fact that urban hinterland communities tend to concen-
trate in closer proximity to metropolitan centers than rural
hinterland communities. It is assumed that as distance in-
creases from metropolitan centers, the ordering influence of
metropolitan centers diminishes and the influence of local
community conditions increases.

To test this hypothesis two types of statistics will

be employed: (1) multiple-partial correlation coefficients

of fertility and community social structure controlling for

299

metropolitan dominance and (2) the standard partial regres-
sion coefficients (beta coefficients) of each of the inde-
pendent variables estimated from the multiple regression
equations. The multiple-partials will indicate the amount
of variation in fertility explained by community social
structure over and above metropolitan dominance. The beta
coefficients, on the other hand, will reflect the relative
importance of each of the independent variables in account-
ing for fertility variation, holding constant the effects of
all other independent variables. With respect to the multi—
ple—partials, the combined effects of more than one variable
on fertility is considered, holding constant only one vari—
able, viz., metropolitan dominance. On the other hand, with
respect to the beta coefficients, only the effect of one
independent variable is estimated, holding constant several
other variables, including metropolitan dominance as well as
other community social structure variables. By ranking the
beta coefficients in each regression equation in terms of
their relative size, we can come to some conclusion as to
which single variables are relatively more important in
determining fertility levels. In comparing these rankings
for urban and rural areas, we would expect to find metropol-
itan dominance as one of the most important determinants of
urban fertility, but for rural fertility we would expect that
some or all of the single variables reflecting community

social structure would be more important than metropolitan

300

dominance. With respect to the multiple-partial correlation
coefficients, our expectation is that controlling for metro—
politan dominance will reduce the correlation of fertility
and community structure to such an extent in urban areas
that coefficients in the rural areas would be considerably
larger. Obviously because of the fact that two different
types of statistics are employed to test the same hypothesis,
it might occur that the results of the two analyses will not
be the same. Although both types of statistics actually
test the hypothesis, it is assumed that the analyses of beta
coefficients will be the better test because of the fact
that more variables are being controlled. In the multiple-
partial analysis only the effect of metropolitan dominance
on fertility is controlled for each coefficient. Further-
more, we can already at this point expect some difficulty

in supporting this hypothesis because of the fact that on
the basis of the preceding analyses wife's alternative
opportunities and demographic age structure were found to
have coefficients larger in urban areas than rural, even
when metropolitan dominance was controlled. For this reason,
the beta coefficients are expected to be the better test of
the hypothesis.

First, then, let us consider the multiple—partial
correlation coefficients of community social structure and
fertility. Does the elimination of the effects of metropol-
itan dominance reduce urban coefficients to levels below

rural coefficients? Table 27 provides the multiple-partial

DOHfarm

partiall

VariableS

me“013011"

 

301

coefficients of fertility and community social structure
variables combined, controlling for metropolitan dominance.
In all divisions but two, New England and East South Central,
the urban coefficient exceeds the rural. The average value
of the multiple-partials for urban areas is .409, for rural—
nonfarm .335 and rural-farm .264. The hypothesis is not
supported by these data. This same table, however, provides
a comparison of the multiple-partial and the coefficient of
multiple determination for each division and residential
sector. The difference between these two coefficients indi—
cates the proportion of explained variation of fertility
which is lost when metropolitan dominance is partialled out.
For most of the divisions this value (the difference between
the coefficient of multiple determination and the multiple-
partial) is higher in the urban areas than the rural-farm,
but in the case of four divisions rural-nonfarm exceeds the
urban, and these are in three highly metropolitanized divi-
sions (New England, Middle Atlantic, and East North Central)
and the Mountain division. Actually, on the basis of aver-
age size of the difference between the coefficients of mul—
tiple determination and the multiple-partials, the rural-
nonfarm sector shows the greatest amount of loss doe to the
partialling of metropolitan dominance.

Tables 28 through 31 compare the multiple-partial
correlation coefficients involving fertility and paired
variables of community social structure, controlling for

metropolitan dominance, and the coefficients of multiple

(D

302

determination for the three variables on fertility. For the
variables measuring employment in agricultural occupations
(Table 28) the multiple-partial correlations tend to be
greater for rural areas, with the exception of West South
Central, Mountain, and Pacific divisions. The average value
of the multiple-partial coefficients is largest for rural-
nonfarm sectors, but lowest for rural-farm. Although the
urban difference between the multiple—partial and the coef—
ficient of multiple determination is almost always greater
than the rural-farm difference, and sometimes greater than
the rural-nonfarm, the proportion lost in urban areas due to
the partialling of metropolitan dominance is not enough to
reduce the multiple-partial coefficients clearly below the
level of rural-farm and rural—nonfarm coefficients. It
should be noted, however, that whereas for the coefficients
of multiple determination four of the divisions revealed
urban coefficients higher than either rural coefficients,
for the multiple-partial only two urban coefficients exceed
either rural sectors. Hence, for employment in agricultural
occupations the hypothesis is partially supported.
Socio-economic status (Table 29) similarly indicates
more frequently higher multiple-partial coefficients in
either rural-nonfarm or rural-farm areas. But again, the
highest average value is found in the rural—nonfarm areas
and the lowest the rural-farm rather than urban. But this
pattern is also true for the coefficients of multiple deter—

mination where metropolitan dominance is allowed to influence

(‘1'

CO

wi

Cie
Pol.

Cie:

303

fertility. In fact, for the total correlation the urban
coefficient is larger than either rural coefficient only for
West South Central. For the multiple-partials three divi-
sions are found to have urban coefficients exceeding either
rural coefficient (Middle Atlantic, West South Central and
Mountain). In the case of the urban sector of four divi—
sions, in comparing the multiple—partial and coefficient of
multiple determination, a greater proportion of fertility
variation explained is lost by partialling out metropolitan
dominance, but it is not enough to reduce the urban multiple—
partials below the rural. Hence the hypothesis is only par-
tially supported by the socio—economic status dimension of
community social structure.

Table 30 compares the multiple-partial correlation
coefficients and coefficients of multiple determination for
wife's alternative opportunities. With respect to the
latter coefficient four of the divisions have urban coeffi-
cients larger than the rural sectors, but partialling metro—
politan dominance seems to affect the rural—nonfarm coeffi-
cients more than the urban. With respect to the multiple—
partials six of the divisions have urban coefficients larger
than the rural. The average value of the multiple—partial
is still highest for rural-nonfarm and lowest for rural—farm
rather than urban. While rural—farm coefficients lose rela-
tively little (6 percent) of the variation in fertility
explained by controlling metropolitan dominance, the urban

coefficients are not significantly reduced from the level of

304

coefficients of multiple determination to bring them below
rural-farm levels. The hypothesis is not supported in the
case of wife's alternative opportunities.

With respect to demographic age structure of women
of childbearing age (Table 31), the coefficients of multiple
determination are much too high and the effects of metropol-
itan dominance too little to alter the basic pattern of
highest coefficients in the urban hinterlands of the divi—
sions. Eight divisions indicate urban coefficients of
multiple determination exceeding either rural coefficients
and, when metropolitan dominance is partialled, nine divi-
sions have urban coefficients higher than the rural. The
hypothesis is not supported by demographic age structure.

Of the four broad categories of community social structure,
only employment in agricultural occupations and socio—
economic status tend to indicate some support of the hypoth—
esis. In fact the evidence seems to suggest that the rural-
nonfarm hinterland is more affected by the partialling of
metropolitan dominance than the urban since for all four
categories of community social structure the greatest pro—
portions of explained fertility variation are lost in this
sector.

A more sensitive test of the relative importance of
metropolitan dominance and community social structure in
each of the residential types of hinterlands, however, is
provided by the beta coefficients. The relative size of the

beta coefficients compared with the size of the beta

(I)

83C

0111'

305

coefficients for all other independent variables employed in
the same multiple regression equation is indicative of the
relative importance of that particular independent variable
in accounting for variation in the dependent variable,
fertility. According to the hypothesis, then, we should
expect to find a tendency for metropolitan dominance to be
ranked first in determining fertility levels for urban
hinterland communities, whereas the community social struc-
ture variables should tend to be ranked above metropolitan
dominance for rural hinterland communities.

Let us consider the rankings of beta coefficients
for each division, comparing the urban and rural patterns.
In this manner we use the results for each division as a
test case of the hypothesis. After considering each divi-
sion separately we shall attempt to summarize the patterns
among the divisions with the use of average rank values for
each independent variable. Tables 32 through 40 contain the
beta coefficients and rankings by residential hinterland for
each division.

According to Table 32 New England is a poor fit of
our expectations. Though metropolitan dominance is rela-
tively more important in urban than rural-farm areas, none
of the coefficients are statistically significant for the
three residential sectors. For urban and rural-nonfarm com-
munities in New England four social structure variables are
more important than metropolitan dominance and in rural-farm

areas seven social structure variables exceed metropolitan

LITvMHUHHLHML

r

fhh.

NV MW

306

.GOHumsom coflmmouoou

mHmHuHsE “masofluumm CH ooosaosfl mmHQmHHm> ucmocmmmoCH Hmnuo Ham nufl3 ooummeou Muflafluumm ca
ocflussooom CM mocmuHomEH m>flumamu >9 manmwnm> unmocmmmosfl mo xcmmat

COHuMHHm> Mom

.Hm>ma mo. ozu um OHmN Eonm usmuommao mausmoflwflcoflma

 

 

 

 

 

 

 

o nma.: m omo.: o ooo.: Amxo emumm mom mmamsmm

m mmH.n k eeo.u a *oom.u imxv wanna mom mmHmEmm
musuosuum 0mm oanmmuoosmn

m mom. a ems. m aoo.u Akxv pcmENOHQEm mamsmm

H *koa.u m ammm.u n meo.u onv msoocH panama
mofluficsuuommo m>flumsuouam m.mMH3

N *Nmm.n H ammv.u m amev.u Amxv msooaH seesaw

a mac. o mmo.u m Hmo. Aexv coflumospm
msumum UHEOQOUMIONUOm

v omm. m Room. a *onm. Amxv cmsmuom cam mumuonmq Enmm

h nNH. o omo. m omo.: ANXV mumomcmz Eumm ocm mumaumm
macaummdooo amusuasofluod QH usofimmmem

o Nmo.l m oma.l m wNN.| Aaxv mocmcflsoo cmuflaomonuoz
tam hm aam mm **m mm manmflum> pcmocommocH

EHMhIHMHSm EHMHQOZIHMHUZ SMQHD

 

ocmaocm 3oz

 

 

oooH

"mosmoamou an .GONmfl>Ho ocmaocm 3oz Mom muflafluumm Ga sofluMHHm> How ocflucsooom SH
mommuuomsfl m>aumamu >9 oHQmHHm> unmocmmooaw mo xcmu osm Ammo ucmfloflmmmoo mumm .Nm magma

 

307

dominance in accounting for fertility variation. Family
income and farm laborers and foremen are consistently more
important than metropolitan dominance in all three residen-
tial hinterland communities. Metropolitan dominance in
New England appears to have relatively little influence in
determining fertility levels. New England, then, does not
support the hypothesis.

The beta coefficients and rankings for Middle
Atlantic division are given in Table 33. In contrast with
New England, for Middle Atlantic metropolitan dominance is
extremely important in determining fertility levels. In all
residential sectors metropolitan dominance ranks first. How—
ever, this pattern does not support our hypothesis either,
since there are no community social structure variables in
the rural areas which are more important than metropolitan
dominance.

Table 34 presents the results for East North Central
division. Metropolitan dominance is again significant in
accounting for fertility variation and in all three residen—
tial sectors, although it is ranked first in rural-nonfarm,
second urban, and third rural—farm. The hypothesis, then,
is partially supported in the contrast of urban and rural—
farm areas, but not in the comparison of urban and rural-
nonfarm. In rural-farm areas two social structure variables
are more important than metropolitan dominance: education

and employment of farmers and farm managers. For urban

OCOH «OUCUUHWQH
>3 .COHNH>NU UHUCGHJ< UACUNZ MOM XQHNNUHUH :fi CONUUHHJ> MOM WZflOCZOUUQ :4
TUCUHHCQEH U>HU5HUL k2 UHQCfiH3> UCEUCUQUTCfl LO MCCM 3:5 «.my JCUflUflLLUCU SOUS -Mfi QNQZB

308

.GOHumsoo conmmHomH
oHQHuHDE HMHdoHuumm CH omosHUGH mmHQMHHm> ucmoaommocH Hmnuo HHm nuHB oonmmeoo huHHHuHmm cH
soHumHHm> How ocHucsooom CH mocmunomEH o>HHMHmH >9 oHQMHHm> usmonmmwocH mo mem*«

.Hm>oH mo. on“ um oumN Eoum ucmHmMMHo hHuchHMHGonR

 

 

 

 

 

 

 

o oMH.I m MNo.I m #50. onv fimlmN mom mOHmem

s ¢MH.I n mNo.: a ammH.u lmxo emumH mom mmHmemm
mucuosnpm mom UHnmmuooEmQ

m moH. N «moN.| o ooo. Ahxv usmENOHQEm meEom

m mvN.I m mNo.: m *ooq.u ioxv msoocH mHmsmm
mmHuHcsunommo o>HumchuH< m_mMH3

m Noo.- m mao.u m ammm.: Amxo msoocH HHHamm

m who.: a omo. m oHo. Hexv coHumospm
msumum UHEOQOUMIOHUOm

m *Hmm. m aon. H mNo.- meo cafimuom ppm mumuoan sump

v th.I o m¢o.1 o mHo. ANxv muoomsmz Eumm ocm mHmEHmm
mcoHpmmduoO HMHSHHSUHHo¢ CH ucmENONmEm

H *Hee.- H «aHm.n H *mka.u HHxV mocmcHsop ampHHoQOHums

n d m
**m .m aem .m **m .m oHQMHum> ucoocommocH
Enmmleusm EummsozIHmusm swamp

 

UHuCMHu¢ mHUUHS

 

 

oooH uousmonmu
an .GOHmH>Ho UHucmem mHoon Mom NHHHHHHQM :H coHumHum> How osHussooom :H
mocmuuomEH o>HHMHmH >9 mHQMHHm> ucoocomoosH mo xcmu ocm A.mv usmHonmmoo muom .mm mHQmB

l..\ .\ r\ F ‘ r\ r\ '1‘ l u! F.l\ fi "U1 kwtv'H
- . . - . r H 1 l l i C C i . LEE . U ...E MC H. > a H H H HAUL. C H 20H. U OH HZ> (HO 3H .
. - . . - . .. -. : .HH.EZOU u
OUCGHHOQEH C>H UGHOH >3 m. L . ... Uh CH

UHQSHHwa ..HCOUCUQOTCH LO VEHHQH 3C3 «.m: HZOHUHLHOOU H.033

.vm aNQmS

 

 

309

.COHumCom COHmmmComC

mHmHuHCE CMHCUHuHmm CH omoCHoCH moHQMHHm> quoCmmooCH Hono HHm CuHB omummfioo NHHHHpHmm CH
COHHMHCm> Com oCHuCCOUUm CH wUCMHHomEH m>HHMHmC >9 oHQMHCm> quoCmmwoCH mo MCmmtt

.Hm>mH mo. mCu um 0Com Eoum HConMMHo mHquUHMHCone

 

onv omlmN mom mmHmEmm

 

 

 

 

 

 

m emo. m moo. H mac.
a *mmm.n o HemH.u H HOHe.u meo emumH mom mmHmEmm
musuosuum mom UHCMMHmOEoQ
m emoN. h who. m omo. Anxv quENOHmEm mHmEom
H .moH.u a .HmH.u a .mmH.u ioxv meoocH mHmEmm
moHuHCCuHQmQO o>HCMCHmuH< m.omH3
m omo.: m aomm.n m *oom. lmxv psoocH sHHsmm
H *on.u m *mHH.u m *NoH. Hexv commmuspm
msumum UHEOCOUmIOHUom
o HemH. m moo. o mmo.u mev swamuom ppm mumuopmH sump
N «omN.I N *me.I o omo.! ANXV mumoMsz 8Com oCm mumfiumm
mCOHummCooO HmCCuHCUHCo¢ CH quENMHmEm
m emmm.n H ammm.u m amom.u HHxv mocmcHsop cmpHHooouumz
m n H
*«m .m tam .m tam .m mHQmHCm> quoCmmmoCH
EHMWIHMHSNL EHMHCOZIHMHSM CGQHD.

 

HmuuCoU Cuuoz ummm

 

 

oooH "moCmonmC

Np .ConH>Ho HmuuCoU Cuuoz ummm Com NCHHHuHom CH COHCMHCm> Com oCHuCCOUUm CH
GUCmuHomEH m>HumHmH >9 oHQMHCm> HCooConoCH mo meu oCm A.mv quHonmmoo mumm .wm oHQMB

cie:
tho:
rura
inar
exce
marr
sect
more
poli

OHly

‘

310

areas the percent of ever-married females in the ages 15-24,
as a reflection of local community conditions, is more
important than metropolitan dominance in accounting for
fertility variation. Hence the case of East North Central
division only hints at the support of the hypothesis.

Table 35 compares the urban and rural beta coeffi-
cients for West North Central division. In this case,
though there are community social structure variables in the
rural sectors which are more important than metropolitan dom-
inance, metropolitan dominance in urban areas as well is
exceeded by the relative importance of two variables: ever—
married females ages 15-24 and female income. In both rural
sectors education and ever-married females ages 15-24 are
more influential in determining fertility levels than metro-
politan dominance. The case of West North Central suggests
only a tendency to support the hypothesis.

The relative importance of the independent variables
for South Atlantic are indicated in Table 36. In this case
the observed pattern for the relative importance of metropol-
itan dominance in affecting fertility among the residential
sectors is opposite the expected. Metropolitan dominance is
very important in rural-farm and rural—nonfarm areas, but
relatively unimportant for urban. For urban communities the
social structure variables which seem to influence fertility
more than metropolitan dominance are females ages 15-24,

socio—economic status, and wife's alternative opportunities.

SOON "SUCOUHWQH
>3 .EOHNH>HU HGHJCOU CUHOZ UHOS MOM NUHNHUHOL CH COHUGHMW> HOL UCHUCSOUUE 3H
@UCWUHOAEH ®>HUQHOH AD OHQEflHfl> UCOUCUQOWCH LO XCQH DEW «.mv HCWHUHWNUOU Gumm .mm UNQQE

 

311

.COHumCom COHmmmComC

mHmHuHCE HMHCUHuHmm CH omoCHUCH mmHQmHCm> quoCmmoUCH HoCuo HHm CHHB omummfioo NCHHHqum CH
COHCMHHm> Com oCHuCCooom CH ooCmuComEH m>HumHoH >9 oHQmHHm> quoCmmooCH mo memaa

.Hm>mH mo. oCu pm OHmN EOHM quCwmeo hHquUHMHCona

 

 

 

 

 

 

 

o ooo. m ooo. H Hmo. onv omImN mow monEom
m Hemm.u m aomm.n H amoe.u Amxv emumH mom mmHmEmpll
mCCuoCCum mod UHCmmummEmQ

o moo. H NHo. m HHMH. ino mamsHOHosm mHmEmm

m *omH.| o oHo.I N eHmN.I A Xv mEOUCH mHmEmm
mmHHHCCuCommo m>HHMCCouH¢ m.mMH3

H meo.u e amvH.u a aoom.u Amxv msoucH HHHEMC

H *Hmm.u H amom.u o «omo.: Hexv C0pr056m
msumum UHEOCOUMIOHoom

N aeom. m omo. m mNo. Amxo cmsmuom ppm mCmHOCMH spam

m mNo.I m Homo.l o oHo.I ANxv mumoMsz Eumm oCm mCmEHmm
. mCOHummmUOO HCHCHHCUHCod CH quENOHmEm

e *Hom.u m aeHH.: m amom.n inv mommaHCOp amuHHomouumz

m .|d m
**m .m *«m .m **m .m oHQMHCm> uCooCmmmoCH
summlensm EHMMCOZIHmHCm CmQCD

 

HmuquO CHCOZ ummz

 

 

oooH "mUCoonoC

an .ConH>Ho HmuquO Cuuoz ummz Cow NCHHHpCom CH COHCMHHm> How oCHuCCOUUm CH
ooCmuuomEH m>HumHoC NC oHQMHHm> uCoUCmmmoCH mo meH oCm A.mv quHUHmmmou mpom .mm mHQmB

. u L. tllluruirirrIKFKKfi PAW
flirt; r.. +ircrL»: m. n....\/ —. 1v .... _. .r..~.H >A~ UH 3* ”...—m .H ”UNV “FL mUPoHL mUflwflLLF~ H. w G VHFL 5th.. MVNMWJ «hm» ULLMIUHUHHkW “WOO Wavy muMK 0 0M: UNQNVE

 

312

.COHumCom ConmoComH

mHmHuHCE CmHCoHuumm CH omoCHoCH mmHQMHHm> “CooCommoCH HmCuo HHm CCH3 omnmmeoo muHHHuHmm CH
COHCMHCm> Com oCHuCCooum CH moCmuHomEH o>HumeC an oHQmHHm> uCooCmmmoCH mo mem**

 

 

 

 

 

 

 

.Hm>mH mo. oCu um ouon Eouw quCoHMHo NHquUHMHCona
w NHo. m ooo.: H *NOH. onv omImN mom mmHmEmm
H *mOH.u a .Hem.n H emmm.u Amxv «mumH mom mmHmamm
mCCHUCHum mom UHflmemmEmQ
o 88H.. m .mHm.- m HHOH.I AHxv wcmsHoHCEH mHmsmm
o omo.: o omo.: w *me.I A xv msoomm.mHmEmm
moHuHCCuCommo m>HumCCouH< m.mMH3
m .NHH. H oHo. m «ooN.: Amxv wsoocH HHHamm
N amoN.I H «hmm.l N *HHm.I w.xv COHCMODUM
mCumpm UHEOCoomIOHoom
v .NNH. m *mHH. o oqo. Amxv :msmuom ppm mumuoan Eumm
m mHN.I o moo.l m ooo. ANxv mumomsz 8Com oCm mHmEme
mCOHummCuoo HmCCuHCUHCo< CH pCmsNonEm
H HOHm.| m aHHm.u o «NQH.I Aon mocmcHsop cmpHHOCOCHmz
n n .dl
*«m .m ¥«m .m tam .m oHQmHHm> quoCommoCH
Eummleusm EumeOZIHmCCm CmQCD

 

oHucmHua nusom

 

 

oooH "moCoonmH

mp .COHmH>Ho OHHCMHCC Cuzco Com NuHHHuHom CH COHHMHCm> Mom oCHuCCooom CH
moCmuComEH m>HHMHmC an oHQmHCm> quoCmmooCH mo meu oCm A.mv quHOHmmoOU mumm .om oHQme

Educat:
Howeve'

data i

also
types
in ur

polit

rank
Table
tant ;
Variak
foreme
metro;
fertn;
ables E
emPlOym
importa,
fails tc

metropol

313

Education is consistently important in both rural sectors.
However, again the hypothesis is not substantiated by the
data for the division.

Table 37 relates to the relative importance of the
independent variables in accounting for fertility levels in
the East South Central division. Metropolitan dominance
does not rank first for rural areas, but this is the case
also for urban areas. Education is in both rural hinterland
types more important than metropolitan dominance. However
in urban areas four social structure variables exceed metro-
politan dominance in influencing fertility levels. The
hypothesis is again not supported.

The beta coefficients and relative importance by
rank are provided for the West South Central division in
Table 38. In this case metropolitan dominance is unimpor—
tant in the rural sectors and superceded by social structure
variables, specifically employment of farm laborers and
foremen and the socio—economic status variables. However
metropolitan dominance is not important in determining urban
fertility levels either, since four social structure vari-
ables exceed metropolitan dominance. Both education and
employment of farm laborers and foremen are consistently
important in all residential sectors. The hypothesis again
fails to be verified.

With respect to the Mountain division (Table 39)

metropolitan dominance again fails to be ranked as the most

\A u: ..Hmu >|H U MMMH. FHOU n. U~#AUMUA U.Ww~..nm~ “And.“ >..u.n.H..N my...*.\-.nHH..L (a? lull .l . . P 1 x .
0H o. a. .H H a H HoHoCH MO icon Coco Om: HCQHUHCMQOU ouom -.Hm.

314

.COHumsom COHmmmComH
onHuHCE CMHCoHqum CH omoCHoCH moHQMHCm> quoCmmmoCH HmCuo HHm CuH3 ooummfioo NHHHHuHmm CH
CoHumHCm> Com oCHuCCooUm CH moCmuComEH m>HHMHoH an oHanCm> uCooCmmooCH mo mem«*

.Ho>oH mo. mCu um OHmN Eoum uCoHoMMHo mHquoHMHCone

 

 

 

 

 

 

 

m HHo.- m emo.u H amo. Amxo amumm mom mmHmEmm
a .omH.- m .mom.u H .Noa.n Amxv amumH mom mmHmsmm
mHCuUCHpm 0mm UHCmmmmOEmo
H *oom.l H «ooo.: N aoom.l AHXV uCoENOHmEm mHmEmm
H Hmo. a ammH. o HHH. onv mEoocH mHmsmm
mmHuHCCuuommO m>HHMCHmuH¢ m.mMH3
m mmo.u m .oHo.n m ave. Amxv meoocH HHHemm
o *eHH.I m .mHH.I m «amH.: 1:. Aqxv cOHumospm
msumum UHEOCOUMIOHoom
m .mom. o .OHH. a .HmH.n Amxo cmsmnom 6cm mumuoan EHmm
N *mNm.I H HomH.I o mHo.I ANXV mumoMsz Eumm oCm mHoECmm
mCoHumoCuoo HMCCHHCUHHoC CH uCoENmflmEm
m .oHH.u m .OHN.I m .mHH.u AHxv mommcHEOp cmpHHoooupmz
.N n 1H,
*«m .m eam .m Ham .m oHQmHCm> uCooComooCH
Enmmleusm ECMMCOZIHMCCm CmCHD

 

HmuuCoO Cusom ummm

 

 

oooH "moCmonoC
an .COHmH>Ho HmuuCoO Cusom ummm Com NCHHHuCom CH COHumHCm> How oCHuCCouum CH
ooCmuHOQEH o>HCMHmH an oHQMHCm> quoCommoCH mo meC oCm A.mv uCoHUHmwooo mumm .mm oHQmB

V —q fin“ J1“! HUI". J AN) v \J .I v
.. nwlwflfl.V’.V ml w.vM~ .QMII
.LF LFL ..
5.- .Plh . o n
.r I.‘
'1‘ o . . V
t .. u IVNH A MN ~
I . "IL: M
I . a - MO
,,
I OH
. ... m“ a
. a! w ~nV~ v— uv .v I
.II ..UlH
« fit
‘ H An.
CHhE-
o: ...H .

GHQQS

315

.GOHumsvm scammmnmmu
mHQHuHSE Haasoflpumm CH wmwsaocfl mmaflmauw> pcmwcmmmwcfl HmSpo Ham QMAB Umummaou muHHHunm CH
QOHUMHHm> How mcfluasooom SH mucmuuogefl m>aumamu an magmaum> ucmwcmmmocw mo xamm«*

.Hm>ma mo. mnu um OHwN Eoum ucmHmMMHU mauchHmwnmflm¥

 

 

 

 

 

 

 

m mmo.u w *oaa. m *qwm. Amxv «mumm mmm mmamamm
n omo.: m Imoa.u w IB¢H.I Amxv wmnma mmm mmamamm
mnsuusnum mmm Uflcmmmmoamo
m mmo.u m Immo.u m omo.: Anxv wcmemoamam mHmEmm
o m¢o.: m mmo.n o Imoa.u A xv mEoocH mHmEmm
mmfluflcsuuommo m>HpmcwmuH¢ m.mmwz
N *Hmm.n m Imma.l m Imoo.u Amxv waoocH maflamm
H Imo¢.: m Imam.n m *Hom.n hixv coflumosom
mSuMpm UHEOQOUMIOHUom
m Imam. H *wmv. H *omm. Amxv swamgom mam mumuonmq summ
fl *oma.l o «moa.l n *ooo.l Amxv mummmamz Enmm cam mHmEHmm
mQOHummdooO HmHSuHSUHHmm Cw ucmENOfimam
m *vma.u n Inmo.u m ¥¢¢H.I Aaxv mocmcHEOG cmuflaomouumz
n HI. ml
¥*m .m **m .m **m .m manmflum> ucmwcmmmmcH
Eummlamnsm Eummcozlamusm GmQHD

 

Hmuucmo nusom ummz

 

 

coma "mocmwflmmu
an .qoamfl>H© Hmuucmo susom ummz How wuflafluumm CH COHuMHHm> How mcflucsouom GM
mocmuuomafl m>wumamu an mHQMflum> ucmwcmmmwcfl mo xcmu 6am A.mv ucmfloflmmmoo muwm .mm manme

5 I . I.I.‘I.IIIIIIVI.\III ril-
L ‘

OUCGQHUQEH O>HJSHGH \AQ @HCG..~H$> UCOUCOLUCGH Mmu MESH C:m..~nm3 UCBHUWMQQOU SHOE .mm, mwfinxi~

 

 

316

.COHumsvm COHmmmummH
mHmHuHCE HMHCoHunmm CH UmUCHoCH mmHQMHHm> quowmmUCH HmCuo Ham CuHB wmummfiou MuHHHuHmm CH
CoHumHnm> How mCHuCCooum CH mUCmuHOQEH m>Hanmu an mHQMHHm> quflCmmmUCH mo mem**

.Hm>mH mo. mCu um OHmN EOHM quHmMMHw mHquoHMHCmHm*

 

 

 

 

 

 

 

 

m moo.u m Noo.: m HHo. Amxv «mumm mom mmamamm

H *va.I o *oma.u m *mom.u A xv wmnma mmm mmamamm
musuosnum 0mm UHCmmummEmm

m «mH.u m «OHN.I ¢ H¢H.I AHxv wamaaonam mamamm

m moa.| m moo. o mmH.I A xv maooCH mHmEmm
mmHuHCCpHommo m>HumCHmpH< m.mMH3

w *Nma. v *Hma.u H *Hoa. Amxv wfioocH HHHEmm

H mHo. a *mom.u H *mmm.u A xv coHumosom
mSHMu. m UHEOCOUMIOH 00m

m mmo.n H moo.u m oao. Amxv swamuom cam mumuonmq Eumm

N *mHN.I m *mmH. m *Hmm. Amxv mummMsz Eumm UCm mumaumm
mCOHummCUUO HmnsuHCOHHm¢ CH quENOdem

o ¢OH.I N *mmm.l m HomH.I Aaxv mocmcHaoc cmgHHomouumz

WI m
**m .m **m .m **m hm mHQMHum> quUCmmmUCH
EHmhIHMHSm EHMWGOZIHMHSM CMQHD
CHmuCCOZ

oomH
umUCmUHmmH >9 ~COHmH>H© CHmuCCOE How muHHHuumm CH COHHMHHm> How mCHpCCOUUm CH
moCMCHOQEH m>HumHmH an mHQMHHm> quowmmUCH mo meu UCm A.mv quHUHmmmoo mumm .mm magma

impor
It is
socia
relat

DORLE

317

important variable in determining urban fertility levels.

It is surpassed by four other variables reflecting community
social structure. While metropolitan dominance is also
relatively unimportant in rural-farm areas, for the rural—
nonfarm communities it ranks second.

The Pacific division (Table 40) is the only case
among the divisions which indicates that metropolitan dom-
inance is relatively more important in determining fertility
levels in urban areas than rural—farm or rural-nonfarm. But
even in this case there are community social structure vari-
ables which surpass metropolitan dominance in relative impor-
tance. In urban areas of the Pacific division metropolitan
dominance is overshadowed by education and female income and
in both rural sectors, education, family income and ever-
married females ages 15—24 are more important predictors of
fertility than metropolitan dominance. Though this suggests
the support of the hypothesis, again it does not follow
exactly the expected pattern.

Overall it appears that where the hypothesis fails
is too great an expectation for metropolitan dominance as a
relatively important variable in determining urban fertility
levels compared with community social structure variables.
In several divisions metropolitan dominance ranks below
several community social structure variables for both rural
sectors, but generally in these same cases metropolitan dom—

inance did not prove to be the most important variable in

318

.COHumsvm ConmmHmmH
mHmHuHCE CmHCUHuHmm CH UmUCHUCH mmHQMHHm> qupCmmmUCH Hmnuo HHm CuHB cmummEoo NCHHHuHmm CH
CoHumHHm> How mCHuCCooom CH moCmuuomEH m>HHMHmH an mHQmHum> “CmoCmmmoCH mo mem««

.Hm>mH mo. msu um OHmN Eoum quHmMMHU mHquoHMHCmHm*

 

 

 

 

 

 

 

m mwo.l m mmo. H moo. Amxv wmlmm mmm mmHmEmm
m *HmN.n N .mmN.- q *HmN.u Amxv «NImH mom mmHmamm
mnsuusuum 0mm UHCmmumOEmQ
m NQH. o .mmH. o *meN. AHxV mamaHoHCEm mHmaom
o mmo.l v *mom.l H *Hmw.l A xv mEOUCH meEmm
mmHuHCCuHommO m>HHMCHmuH< m.mMH3
N .omN.I m *mvN.n m Noo.: Amxv maoocH HHHamm
H NOHm.I H Hemm.n N .mmm.u Aaxv coHumoseMIl
mspmum UHEOCOUMIOHoom
H omo. H mmH. m Noo.: Amxv swamuom mam mumuonmg sums
m H¢Q.I m mNo.I m *Hmm. Amxv mummMszlEumm UCm mHmEHmm
mCoHummdooO HmHCuHSUHHm< CH pCQflNOHmEm
d mmH.I m H¢ON.I m *me.l AHXV mUCMCHEO© CmuHHomonumz
n n .d
**m .m **m .m xxm .m mHQMHHm> qupCmmmpCH
Enmmleusm EHMMCOZIHmHCm CmQCD

 

oHHHomm

 

ommH
"mUCmUHmmH m9 .COHmH>H© UHwHomm How auHHHuumm CH COHHMHHm> How mCHuCCOUUm CH
moCmuHomEH m>HHMHmH an mHQMHHm> qupCmmmpCH mo meu ow A.mv quHonmmoo mumm .om mHQma

319

accounting for urban fertility. Where metropolitan domi-
nance proved to be an important variable in urban areas, it
was also important in the rural sectors. Although there
were tendencies among the divisions to support the hypoth—
esis, generally the hypothesis failed to be substantiated.
Another way of looking at the pattern of rank order
of the independent variables in determining fertility levels
within urban, rural-nonfarm, and rural-farm areas, however,
is to consider the average rank received by each independent
variable among all the divisions. It is difficult to sum-
marize the patterns which are found among the divisions when
considered individually, and the use of average rank value
provides at least a summary measure by which we may consider
the comparison of relative importance of metropolitan domi—
nance and community social structure together. Table 41 pro—
vides the average rank values of each independent variable
among the divisions. While these data indicate that on the
average metropolitan dominance is a relatively important
variable in determining fertility levels, this is true for
all three residential hinterland types. On the average met-
ropolitan dominance is more important in rural-nonfarm areas
than urban and more important in urban areas than rural-farm.
However, as the table indicates, in no residential sector
is metropolitan dominance on the average ranked first among
the independent variables. In urban areas ever-married
females ages 15-24 exceeds metropolitan dominance in average

rank value and in both rural sectors education exceeds it.

Table 41

320

Table 41. Average rank among divisions of relative impor—
tance of independent variables in accounting for
variation in fertility measured by beta coeffi-
cients, by residence: 1960

 

 

Average Rank of
Beta Coefficients

 

 

 

 

 

 

Rural- Rural-

Independent Variable Urban Nonfarm Farm

Metropolitan dominance (X1) 3.9 3.2 4.1
Employment in Agricultural Occup.

Farmers and Farm Managers (X2) 6.9 6.3 4.4

Farm Laborers and Foremen (X3) 5.9 5.3 4.3
Socio-Economic Status

Education (X4) 4.1 2.9 4.0

Family Income (X5) 5.2 4.2 5.2
Wife's Alternative Qpportunities

Female Income (X ) 4.2 6.0 5.4

Female Employmené (X7) 6.0 4.4 4.8
Demographic Age Structure

Females age 15-24 (X8) 2.4 4.7 4.6

,Females age 25-34 (X9) 6.3 7.9 8.1

 

partial

we must

tiated.

rural a:
ables aI
determir
Suggest:
in rura:
These 6;
as a pe:
all him
hinterl
Communi
hinterl
each t3
ior, tk
StrUCtt

321

On the basis of the data provided by the multiple—
partial correlation coefficients and the beta coefficients
we must conclude that the fourth hypothesis is not substan-
tiated. with both types of statistics it was shown that in
rural 22$ urban hinterlands community social structure vari-
ables are more important than metropolitan dominance in
determining fertility levels. These data also concur in
suggesting that metropolitan dominance is most influential
in rural—nonfarm hinterland communities and not urban.

These data further cast some doubt on metropolitan dominance
as a pervasive influence and organizing process throughout
all hinterland areas vis—a-vis the influence that local
hinterland communities themselves exert in determining inter—
community spatial variation. In contrasting urban and rural
hinterland communities, however, in terms of the impact that
each type of hinterland community exerts on fertility behav-
ior, these data seem to suggest that of the community social
structure variables which are found to influence community
fertility behavior, employment in agricultural occupations
and socio—economic status are more important in rural areas
whereas wife's alternative opportunities and demographic age

structure are more influential in urban hinterland areas.

exi

not

t1

5

.1‘

1"

322

fiypothesis 5
In the more metropolitan geographic divisions compared
with the less metropolitan geographic divisions, the
size and distance of a dominating metropolitan center
is more important in accounting for variation in commu—
nity social structure and fertility behavior in both
urban and rural hinterland communities.

This hypothesis, of course, stems from a considera-
tion of the different rates of metropolitanization which
exist among the various divisions. To this point we have
not considered the effect that differential rates of metro-
politanization may have upon the pattern of hinterland com—
munity variation as well as community fertility variation.
It is assumed that the nation is moving toward a situation
in which metropolitan centers become the chief "organizing"
influence on intercommunity variation. Those divisions,
therefore, which indicate higher levels of metropolitaniza—
tion may be considered prototypes of intercommunity varia—
tion patterns which will eventually emerge as the dominant
pattern for all geographic divisions. Upon subdividing the
nine geographic divisions into two categories, i.e., those
divisions which indicate metropolitanization levels above
the national average and those divisions below, we should
expect to find intercommunity variation patterns more
clearly the result of the "organizing" effect of metropol—
itan centers in the more metropolitan divisions. In the
less metropolitan divisions, the effect of metropolitan

centers should not be as clear. The expected pattern in the

more metropolitan divisions is spelled out in the hypothesis

presex
fiivis
exert
well

varia
of CC

hints

323

presented above. In other words, for the more metropolitan
divisions size and distance of metropolitan centers should
exert greater influence on community social structure as
well as community fertility behavior compared with the same
variable in the less metropolitan divisions. This pattern,
of course, is anticipated similarly for all residential
hinterland types.

Tables 42 through 44 provide a test of the impor-
tance of metropolitan dominance in determining community
social structure variation in the more and less metropolitan
divisions for urban, rural-nonfarm, and rural-farm hinter-
land areas respectively. Table 42 provides the zero-order
correlation squared (an estimate of the proportion of vari-
ation in the dependent variable explained by the independent
variable, which in this case is metropolitan dominance)
between metropolitan dominance and each of the eight indi-
vidual indicators of community social structure for the
urban communities in each of the nine geographic divisions.
The divisions are ordered according to percent of their
population residing in metropolitan areas. To facilitate
the difficult task of making comparisons for such data, an
average value of the squared correlation coefficient for
each community social structure variable is also provided
for the two categories of divisions. For urban coefficients
‘we see that for five of the average values, the more metro—

politan divisions exceed the less metropolitan divisions,

.COHHMNHCOQHHOQOHUOE MO HO>®A %Q WQOHWH>HU MOM OMSUUSHUW HMHUOW XOHCZEEOU

UGO OUCGCHEOU COHHHCQOHUUE HO A My mUHHOflUflnHMOOU COH.UGH®IH«HOU .HmumOIHCIOHON

N .Nv OHQGE

 

324

.Hm>mH mo. mCu um oumN Scum HCmeMMHU mHquoHMHCmme

 

woo. MHo. Hfio. Nmo. mMH. Nmo. moo. vmo. ¢.ov mSHm> mmmum>¢

moo. ooo. ammo. moo. smmo. *HVH. MHo. moo. 0.0m HmanmU fiudom ummm
ooo. ooo. Noo. sumo. *NHo. *mmo. *mmo. *omH. m.m¢ HMHUCmo SCHOZ ummB

 

.woo. «0mo. *mHH. *mNN. *omo. *oNo. «omo. *Noo. m.mv CHmpCCOE
Noo. *HNo. Hoo. goo. some. ooo. eomo. *mmo. N.om UHHCmHuC Cusom
woo. Noo. woo. *omo. HHo. moo. ewHo. eon. m.mm HmupCmU Cusom ummz l
CmuHHommuumz mmmH
emo. NmH. Hmo. Noo. mmN. moo. oHo. mvo. o.¢H mCHm> mmmnm>¢
*mNo. *mNo. vao. egNo. ooo. HHo. *mNo. *mHo. H.Ho HmHquU Cunoz ummm
ovo. NOON. *mNH. ewNH. emmo. oHo. emmH. *HvH. m.0H oCMHmCm 3oz
mNo. *HHH. Hmo. «Hmo. some. moo. Hoo. moo. N.mH UHMHomm
*HNH. HNHN. moo. moo. *mmo. Hoo. eme. eHmo. m.Hm oHuCMHuC mHoon

 

CmuHHomoCuwz muoz

NHC .ouumz CoHummHCmuHHomouumz
N N N N N N N N X mo Hm>mq an mConH>HQ

 

 

 

oomH "CmQHC
.CoHummHCmuHHomoupmfi mo Hm>mH an mConH>Ho Mom mHCuUCHpm HmHUOm huHCCEEoo
oCm mUCmCHEoo CmuHHomonumfi mo ANCV mquHUHmwmoo COHHMHmHHoo CmoHOIOHmN .Nv mHQmB

325

but for three average values the pattern is reversed. On
the average, then, metropolitan dominance is more important
in accounting for variation in employment of farm laborers
and foremen, family income, female employment and ever-
married females ages 15-24 and 25-34, in the more metropol—
itan divisions than in the less metropolitan divisions. For
employment of farmers and farm managers, education, and
female income the pattern is reversed. Average differences
in coefficient size are especially large for family income
and ever—married females ages 15-24. In the more metropol-
itan divisions, metropolitan dominance explains on the aver-
age 30 percent of family income variation and 15 percent of
variation for ever-married females ages 15-24, whereas for
the less metropolitan divisions the average percentages are
14 and 1 respectively. Of course, using an average value
for comparison does gross over several inconsistencies in
the data. In several cases the coefficients of the more
metropolitan divisions fall below the highest coefficients
of the less metropolitan divisions, and this especially
occurs for East North Central and Pacific divisions. Fur—
thermore, it should be pointed out that the difference
between the average values for the two categories of divi—
sions is often very slight. Nevertheless on the average
metropolitan dominance appears more frequently to be more
important in determining community social structure varia-

tion in the more metropolitan divisions than the less

netropc

esis is

rural-1
cated

farm c.
sisten'

Wherea.

326

metropolitan divisions. Hence we conclude that the hypoth—
esis is supported to some extent by the urban sector.

The hypothesis is more consistently supported by the
rural-nonfarm and rural-farm average coefficients as indi—
cated in Tables 43 and 44. The average value of rural—non-
farm coefficients for the more metropolitan divisions con-
sistently exceeds those for the less metropolitan divisions,
whereas in the case of rural-farm coefficient averages six
of the eight comparisons are consistent with the hypothesis.
In the case of rural—nonfarm average coefficient values,
again the largest difference appears for family income and
ever-married females ages 15-24. On the average metropol-
itan dominance in the more metropolitan divisions explains
25 percent of the variation in family income and 20 percent
of the variation in ever—married females ages 15-24. In the
less metropolitan divisions the percentages are 13 and 2
respectively. In the case of rural-farm average coeffi—
cients, only family income seems to manifest a large differ-
ence. As in urban areas, so in the rural-nonfarm and rural-
farm, the two divisions which seem to push the average value
for the more metropolitan divisions above the less metropol—
itan divisions are Middle Atlantic and New England. Though
again.one must point out that differences between the two
levels of average values are slight and that several incon—
sistencies occur among indivisual divisions, it must be

«granted that the data seem to support the hypothesis that

327

.Hm>mH mo. mCu um oumN Eouw quHmMMHo >HpCmUHMHCmHme

 

woo. mmo. hmo. Hoo. NMH. vwo. omo. omo. ¢.o¢ mDHm> mmmum>¢

moo. ooo. ammo. #00. N00. *Hmo. moo. ooo. o.om HmuquU nusom ummm
*moo. *HHo. H00. *H50. sumo. *moo. *omo. *fimm. m.m¢ HMHUCmU SHHOZ ummz

 

ooo. moo. NOHH. eowH. moo. *omo. 0H0. eHoo. m.mv CHmuCCoz
«moo. *moo. Hmo. *Hvo. *vmm. emoo. *mNo. emwo. N.om UHuCMHpfi Cadom
Hoo. boo. ooo. *Nwo. ooo. *mmo. ammo. moo. m.mm HmuquU Cusom pmmz
CmuHHomouumz mmmq
mmo. mmH. mmo. ooo. mmN. mHo. omo. vmo. o.vh mCHm> mmmum><
*oHo. ammo. ammo. somo. Noo. *mNH. *mHo. *flHo. H.ho HmuquU xuuoz ummm
sfimo. somv. «va. *mMH. somm. *Nwo. «00H. «HOH. m.0H oCmHmCm sz
woo. HNo. mHo. «omo. «wow. HHo. moo. ammo. N.mm UHMHomm
oNo. *HoN. Hoo. *mvo. OHo. emmo. eme. «mMH. o.Hm UHuCMHu¢ mHooHE

 

CmuHHomouumz who:

 

H H .ouumz CoHumeCmuHHomouumS
N N N N N N N N R MD Hm>mH >9 mCOHmH>HQ

 

EHMMCOZIHMHCm

 

 

oomH "EHMMCOCIHMHCH
.CoHumuHCmuHHomoume mo Hm>mH an mCOHmH>Ho now musuusuum HmHUOm muHCCEEOU
UCm mUCMCHEoo CmuHHomouumE mo ANHV mquHonmmoo COHHMHmHHoo HmoHOIOHmN .mm mHQmB

It‘l.l*alsF(.(ll1\1\ll*(F‘

u(

INI5\

I’- l‘ ’1. f!‘ or E 0' u. u {\ ( PI\

1‘ \7 LEE
rh nlul+rF r F

It n.~h\y~f\lICL fox

.va.

0 HQ arm.

328

 

 

 

 

 

 

 

.Hm>mH mo. mCu um oumN Eoum quCmMMHo mHuCMUHMHCmHme
mmo. hHo. mNo. vvo. hNH. mmo. Nmo. 00H. ¢.ov mCHm> mmmnm>¢
moo. eoHo. sumo. moo. aHmH. somo. somo. eoNo. o.mm HmnquU Cusom ummm
*NwH. emHo. moo. ammo. ooo. ammo. *mwo. *Hhm. m.m¢ Hmuquo CUHOZ ummz
mHo. mHo. *wwo. *mvo. moo. Noo. ooo. *Nvo. m.m¢ CHmuCCOS
Noo. ooo. *mmo. «omo. *mmw. woo. Noo. HoHo. N.om UHHCmHuC Cusom
*mHo. somo. ooo. *ovo. ooo. «mOH. Homo. ewho. m.mm Hmuquo Cusom ummz
CmuHHommwumE mmmq
moo. omo. mmo. Hmo. ooN. mmo. mmo. NvH. o.¢n mCHm> mmmum>¢
ooo. Hoo. *mmo. *HHo. Hoo. *mmo. emoo. *mwo. H.ho HmuquU Cuuoz ummm
ooo. *Nmo. *HBH. *mom. *hmv. «omo. *mmo. *mmH. m.om UCMHmCm 3oz
HHo. ooo. mNo. ammo. *mNo. mNo. xmvo. xde. N.mH UHMHumm
Hoo. ammo. woo. *Nmo. «NNH. moo. *Nmo. emmH. w.Hm UHquHu< mHoon
CmuHHomouumz who:
mHH mHu hHH mHH mHH ¢HH MHH NHH .ouumz COHumeCmuHHomouumz
N IN N N N N N N R mo Hm>mH an mCOHmH>HQ
EummnHmusm
oomH "ECMMIHMHCH

.CoHumuHCMUHHomonumE mo Hm>mH an mConH>Ho How mHCuoCuum HMHUom muHCCEEOU

oCm mUCMCHEOU CMHHHomouumE mo ANCV mquHUHmwmoo COHHMHmHHOU HmoHOIOHmN

.aa mHnme

TRUE

don

beh-

est

Cie
nan.
See}
FIEt;

in:

329

metropolitan dominance is a more important determinant of
community social structure variation in the more metropol-
itan divisions than the less metropolitan divisions.

But to conclude the testing of this hypothesis, we
must also consider the relative importance of metropolitan
dominance in determining fertility behavior by the extent of
metropolitan development among the divisions. Table 45 con-
tains two sets of data to test the relative importance of
metropolitan dominance in determining community fertility
behavior: rankings of the beta coefficients from the
estimated multiple regression equations and the squared
zero-order correlation coefficients of fertility and metro—
politan dominance. For both sets of statistics the hypoth-
esized pattern is supported for urban, rural-nonfarm and
rural-farm areas, although the pattern is clearer using the
zero—order correlation coefficients than the beta coeffi-
cients. Compared with the influence of metropolitan domi—
nance on community social structure, metropolitan dominance
seems to have a greater effect on fertility. On the average
metropolitan dominance explains 17 percent of the variation
in fertility in urban areas of the more metropolitan divi-
sions, 25 percent in the rural-nonfarm areas, and 12 percent
in the rural—farm areas. In the less metropolitan divisions
the average coefficient values are 6, 5, and 2 percent re-
spectively. For the comparison of the beta coefficients,

again the average values of ranks for metropolitan dominance

Table 45.

330

Rank of metropolitan dominance by relative impor-

tance in accounting for variation in fertility

based on beta coefficients and zero-order corre-
lation coefficient of metropolitan dominance and
fertility for divisions by level of metropolitan—

 

 

 

 

 

 

 

 

ization, by residence: 1960
Rank of
Metropolitan 2
Dominance by rlY
Divisions by Beta Coef.
Level of Metro- %
politanization Metro. Urb RNF RF Urb RNF RF
More Metropolitan
Middle Atlantic 81.8 1 l l .158* .291* .155*
Pacific 79.2 3 5 4 .181* .156* .076*
New England 70.3 5 5 8 .240* .457* .203*
E. N. Central 67.1 2 l 3 .081* .102* .061*
Average Value 74.6 2.8 3.0 4.0 .165 .252 .124
Less Metropolitan
W. S. Central 53.5 5 7 5 .086* .015* .000
South Atlantic 50.2 6 3 l .070* .045* .033*
Mountain 48.8 5 2 6 .095* .138* .025*
W. N. Central 43.3 3 3 4 .065* .027* .038*
E. S. Central 36.0 5 2 5 .001 .001 .002
Average Value 46.4 4.8 3.4 4.2 .063 .045 .020
*Significantly different from zero at the .05 level.

it

le

CC

331

indicate that it is on the average a more important variable
in determining fertility variation among the more metropol-
itan divisions than the less metropolitan divisions. The
average rank of metropolitan dominance in urban areas is 2.8,
rural—nonfarm 3.0, and rural-farm 4.0 for the more metropol-
itan divisions, and 4.8, 3.4, and 4.2 respectively for the
less metropolitan divisions. There are inconsistencies, of
course, when individual divisions are compared rather than
average values of the two categories of divisions. Further-
more, the two sets of data, beta and zero-order correlation
coefficients, do not portray the same patterns. For example,
the zero-order correlation coefficients for New England are
highest among the divisions, but in terms of the rankings by
beta coefficients metropolitan dominance is relatively less
important for New England compared with the other more metro-
politan divisions. Beta coefficients reflect the effect of
metropolitan dominance on fertility with the influence of the
other independent variables partialled. The zero-order corre—
lation coefficients, of course, do not control for other
variables. Hence this difference perhaps explains the incon-
sistency for the two types of data. Nevertheless on the
average the hypothesis is supported so that we may conclude
that metropolitan dominance is relatively more important in
the more metropolitan divisions in accounting for fertility
variation, as well as community social structure, in both
rural and urban areas, than in the less metropolitan divi-

sions.

en

ti

111
i1

16

332

Hypothesis 6

 

In the more metropolitan geographic divisions, size and
distance of a dominating metropolitan center will be
more important in accounting for variation in community
fertility behavior, in both urban and rural hinterlands,
than local community social structure, when controlling
for the influence of metropolitan centers; in less met-
ropolitan geographic divisions, local community social
structure will be more important in accounting for vari-
ation in community fertility behavior than size and
distance of a dominating metropolitan center.

We have considered previously the competing influ-
ence of metropolitan dominance and local community condi-
tions in determining intercommunity variation in fertility
behavior, but this consideration did not include the effects
of differential levels of metropolitanization among the
divisions. The comparison in a previous hypothesis was
primarily among the residential hinterland types, i.e.,
urban vs. rural. It was anticipated that metropolitan dom—
inance would be more important in determining fertility
levels in urban hinterland communities than local community
conditions, but less important in rural communities. The
hypothesis was not supported. The comparison being made in
this section, however, is inter-divisional rather than inter-
residential. In the previous hypothesis it was found that
metropolitan dominance is more important in determining
urban and rural fertility behavior in the more metropolitan
divisions. But is metropolitan dominance so important in
these divisions that it overshadows the effects of local

community social structure? Furthermore in the less metro—

politan divisions is metropolitan dominance so low in

333

importance that local community social structure exceeds it
in determining fertility variation in urban and rural hinter—
land communities?

According to the average rank values of the beta
coefficients for each of the independent variables presented
in Table 46, the hypothesis seems to be true. Comparing the
urban hinterlands of the two categories of divisions, metro-
politan dominance on the average is more important in
accounting for fertility variation than community social
structure in the more metropolitan divisions, but exceeded
in the less metropolitan divisions by ever—married females
ages 15-24 and education. The average rank of metropolitan
dominance in urban areas for the more metropolitan divisions
is 2.8, but only 4.8 for the less metropolitan divisions.
Comparing rural—nonfarm areas, metropolitan dominance shares
first place in relative importance with family income for
the more metropolitan divisions but is exceeded by education
in the less metropolitan divisions. For rural—farm areas,
though the average rank values of metropolitan dominance are
very similar for the two categories of divisions, in the more
metropolitan divisions on the average metropolitan dominance
is more important in determining fertility variation but in
the less metropolitan divisions it is exceeded by education,
employment of farmers and farm managers, and employment of
farm laborers and foremen. The consistency of education as

a more important determinant of intercommunity fertility

334

.CHmuCCOZ oCm .HmHquU Cusom ummz .HmuquU Cusom
ummm .UHuCMHud Cpsom .HmnquU Cuuoz ummz mUCHUCH mCOHmH>HU CmuHHomouumE mmmqee

.UHMHUmm oCm .Hmuquo
Cuuoz ummm .UHuCMHu4 mHooHE .UCmHmCm BmZ moCHUCH mConH>Ho CmuHHomouumE muoze

 

 

 

 

 

 

 

 

o.m ¢.H «.6 m.H m.m m.m Amxo emImN mmm mmHmEmm
e.¢ o.e m.N m.¢ m.m m.m A xv «NumH mmm mmHmsmm
musuosuum mmm UHCmMHmmEmQ
o.m N.e o.m m.¢ m.¢ m.H AHxv ucmsHoHCEm mHmsmm
m.m m.m m.m m.m o.m m.m onv mEOUCH mHmEmm
mmHHHCCuHQmmo m>HHMCHmuH¢ m.meB
N.m N.m o.m m.m o.m m.¢ Amxv msoocH HHHEmm
a.m o.N m.N m.¢ o.a m.m Aexo aoHumoswm
msumpm UHEOCoomIOHUom
o.m m.m N.m m.m o.m m.m Amxv CmEmuom oCm mHmCOQMH Eumm
m.m m.m m.m m.m m.m m.m ANXV mummmsz Eumm UCm mHmEHmm
mCoHummono HmusuHsoHum< CH quENmHmam
N.a v.m m.e o.¢ o.m m.N Aon mommaHsoo cmpHHomouumz
mm mzm QHD mm mzm nub mmHQMHHm> quoCmmmoCH
quHonmmoo mumm mo quHonmmoo mumm mo
mem mmmum>< MCmm mmmum>¢
mConH>HQ mCOHmH>Ho
eeCmuHHomouuwz mmmH eCmuHHomouumz who:

 

 

oomH "mUCmonmu ma oCm mCOHmH>Ho mo CoHuMNHCmuHHom
louumE mo Hm>mH >9 mquHonwmoo mumm >3 omusmmmfi muHHHqum CH COHuwHHm> How
mCHuCCooom CH moCmuHomEH m>HHMHmH an meQMHHm> quoCmmmoCH mo meH mmwnm>< .ov mHQmB

335

variation in all three residential sectors of the less metro-
politan divisions is to be especially noted, whereas for the
more metropolitan divisions the most important determinant

is distance and size of a dominating metropolitan center.

Hypothesis 7

 

The impact of community social structure and metropol-
itan dominance on fertility behavior will manifest fewer
differences when comparing the same type of hinterland
communities (urban or rural) on an inter-divisional
basis than when comparing different types of hinterland
communities (urban vs. rural) on an intra—divisional
basis.

The basic theoretical assumption underlying this

hypothesis is that metropolitan centers exert a differential

 

impact on the different residential hinterland communities
within their metropolitan regions. As metropolitanization
continues to emerge as a key process in the determination of
intercommunity variation patterns, we should expect to find

that differences will increase among residential hinterland

 

communities. Though our analysis focuses primarily on fac—
tors which influence fertility behavior, it is possible to
test this differential impact hypothesis by statistical
testing. We have estimated multiple regression equations
for each of the residential sectors of all geographic divi—
8ions of the nation. In each equation fertility is the
dependent variable and indices of metropolitan dominance and
Community social structure are the independent varibles. By
means of the multiple comparison test it is possible to

determine whether each of the independent variables has the

336

same effect on fertility in the different resiential sectors
of each of the divisions. An example of the kind of ques-
tion posed in the application of the multiple comparison
test is whether the effect of metropolitan dominance upon
fertility is the same for the rural-farm communities as for
the urban communities of a given division. The multiple
comparison test involves the comparison of the effects of
the partial regression coefficients on fertility in two
different multiple regression equations. Hence the test can
be employed to determine whether there is a statistically
significant difference between only two partial regression
coefficients at a time. For each independent variable rep—
resented in the multiple regression equation we may for any
division make comparisons of rural-farm vs. rural—nonfarm,
rural-farm vs. urban, and rural—nonfarm vs. urban. If the
hypothesis is at all true, i.e., if metropolitan centers do
exert a differential impact on the residential hinterland
communities, we should eXpect to find significant differences
between residential sectors for most of the partial regres-
sion coefficients reflecting the effects of the independent
variables on fertility behavior. Furthermore, given differ-
ent rates of metropolitanization among the divisions, a sig-
nificant difference should be found between partial regres-

sion coefficients more frequently for the more metropolitan

divisions.

337

In addition, however, the multiple comparison test
can be employed to determine whether the independent vari-
ables exert a differential impact on fertility among the
geographic divisions of the nation. An example of the kind
of question posed by this test would be whether the effect
of metropolitan dominance on fertility is the same for the
urban communities of one division as for the urban commu—
nities of another division. In contrast to the first com—
parison described above, i.e., an inter-residential compar-
ison for a given division, the comparison called for here is
an inter-divisional one within the same residential sector.
The theoretical framework of this study suggests that as
metropolitanization emerges as a dominant process within the
nation, that divisional differences will disappear while
residential differences will be enhanced. If this observa-
tion is valid, after applying the multiple comparison test
at the two levels suggested above, i.e., inter—divisional
and inter-residential, we should expect to find a pattern of
significant differences for the inter-residential comparisons
but homogeneity for the inter-divisional comparisons. Hence
this test is a test of the differentiating effects of metro-
politan centers on residential hinterland communities.

First, let us consider the results of the multiple
comparison tests between residential sectors. The results
of these statistical tests are given in Table 47. Overall

it appears that there is enough evidence to assert that there

3EH3

Table 47. Summary of results for multiple comparison tests among residential
sectors of conterminous United States and divisions

 

 

 

 

 

 

 

f ‘——‘

 

 

Independent Variables

 

Residential Sectors Compared
for Nation and Divisions X1 X2 X3 X4 X5 X6 x7 x8 x9

 

UNITED STATES

 

 

 

 

 

Rura -Parm vs. Rural-Nonfarm 0 0 1 1 1 l l 1 0

Rural-Farm vs. Urban 1 0 1 1 l 0 l 1 1

Rural-Nonfarm vs. Urban 1 0 1 1 1 1 1 1 1
New En land

RuraE-Farm vs. Rural-Nonfarm O 0 l l 0 1 l 1 0

Rural-Farm vs. Urban 0 1 l o 0 1 1 O 0

Rural-Nonfarm vs. Urban 0 1 1 1 0 1 1 l 0
Middle Atlantic

RuraIeFarm vs. Rural-Nonfarm l 1 1 1 0 1 1 O l

Rural-Farm vs. Urban 1 0 l 1 0 0 l 0 l

Rural-Nonfarm vs. Urban 1 l 1 l 1 1 l 1 1
East North Central

Rura -Parm vs. Rural-Nonfarm 1 1 0 1 0 0 1 l O

Rural-Farm vs. Urban 0 0 1 1 1 0 l O 0

Rural-Nonfarm vs. Urban 1 1 1 1 l 0 0 1 0
West North Central

Rural-Farm vs. Rural-Nonfarm l 1 1 1 0 0 O 1 l

Rural-Farm vs. Urban 0 0 O 1 O 0 O O 0

Rural-Nonfarm vs. Urban 0 0 O 1 0 1 O 1 1
South Atlantic

Rural-Farm vs. Rural-Nonfarm 0 l 0 0 0 0 1 l 1

Rural-Farm vs. Urban 1 l O 1 1 0 0 1 0

Rural-Nonfarm vs. Urban 0 l O l l 0 l 0 1
East South Central

Rural-Farm vs. Rural-Nonfarm O 0 0 l 0 0 0 1 l

Rural-Farm vs. Urban 0 O 1 O o 0 l O O

Rural-Nonfarm vs. Urban 1 O 1 0 0 1 l 0 0
West South Central

Rural-Farm vs. Rural-Nonfarm 0 0 l l l O 0 0 1

Rural-Farm vs. Urban 0 0 l l 1 O O l 1

Rural-Nonfarm vs. Urban 0 0 1 O 0 O l 1 1
Mountain

Rural-Farm vs. Rural-Nonfarm O l 0 1 l 0 0 1 O

Rural-Farm vs. urban 0 l O 1 0 O 0 l 0

Rural-Nonfarm vs. Urban 1 l 0 0 0 0 l l 0
Pacific

Rura -Farm vs. Rural-Nonfarm O O O 0 0 0 O O O

Rural-Farm vs. Urban 0 1 O 0 1 1 0 0 0

Rural-Nonfarm vs. Urban 1 1 0 0 1 1 0 0 O

 

"1" denotes that there is a significant difference between the regres-
sion coefficients of the independent variable for the two sectors compared.

"0" denotes that there is 29 significant difference between the regres-
sion coefficients.

X1 Metropolitan dominance. x6 Median female personal income
X2 Percent male labor force employed in 1959‘
as farmers and farm managers. x7 Percent females, age 14 and over,
x3 Percent male labor force employed employed.
as farm laborers and foremen. x8 Percent ever-married females,
x4 Median years school completed by ages 15'44' who are age 15'24°
males and females, age 25 and X9 Percent ever-married females,
over. ages 15-44, who are age 25-34.

x5 Median family income in 1959.

339

are consistently significant differences among the residen-
tial sectors to support the hypothesis. Differential
effects for all_three comparisons occur frequently among the
various divisions: farm laborers and foremen, female income,
and female employment in New England; metropolitan dominance,
farm laborers and foremen, education, female employment and
females ages 25-34 in Middle Atlantic; education in East
North Central and West North Central; farmers and farm man-
agers in South Atlantic and Mountain: farm laborers and
foremen in West South Central. Few differences occur in
East South Central, the least metropolitan division, and
unexpectantly in Pacific, a highly metropolitan division.
Education shows the most frequent occurrence of a signif-
icant differential impact in comparing the residential sec-
tors, followed by females ages 15-24, farmers and farm man-
agers, farm laborers and foremen, and female employment.

It is difficult to conclude which residential hinter—
land contrast indicates the more frequent significant compar—
isons. It appears to be the comparison of the rural-nonfarm
and urban sectors. Interestingly for this comparison, the
differences seem to concentrate more consistently in the
more metropolitan divisions, which more accurately is an
urban-suburban comparison rather than a rural—urban compar-
ison. For the rural-farm vs. urban comparison education
indicates most frequently a substantiation of the differen-

tial effects hypothesis while metropolitan dominance, female

s ig
the
var

feI

the

340

employment and females ages 25-34 least frequently reveal
significant differences of their effects on fertility. For
the rural—nonfarm vs. urban comparison the independent
variables which most often indicate differential effects on
fertility are farmers and farm managers, female employment,
and females ages 15—24. In the case of comparing the rural
sectors the variables most frequently revealing a signif—
icant difference in their impact on fertility are education
and females ages 15-24. Family income, female income, and
metropolitan dominance indicate few significant differences
for this residential comparison.

To summarize, the differential impact of the inde-
pendent variables on fertility between residential hinter-
land areas is substantial. These differences are especially
pronounoaiamong the more metropolitan divisions, although of
these divisions one exception seems to be the Pacific. For
this division the rural sectors appear to be quite homoge-
neous with respect to the impact of the independent vari-
ables on fertility. Greater contrasts for this division are
primarily rural vs. urban. It could be in the case of the
Pacific division that the percentage of population residing
in metropolitan areas is deceptive of its actual level of
metropolitanization.

Let us at this point turn to the inter—divisional
comparisons within the residential sectors. The results of

the multiple comparison tests between divisions are provided

in

ta

si

a:

CE

OC

fc

fa

e1

()4

C l

f1

341

in Table 48. It is obvious from a quick perusal of this
table that the independent variables reveal relatively
similar effects on fertility among the divisions. In com-
paring these results with the inter-residential comparisons
we conclude immediately that the seventh hypothesis of this
chapter is substantiated. The more significant evidence of
differentiation, then, is among residential hinterlands, not
among divisions. Interestingly most of the significant
cases of differential impact of the independent variables
occur within the urban areas. The differences exist mostly
for the socio—economic status variables of education and
family income. The few cases of significant differential
effects in the rural sectors appear to be female employment
and metropolitan dominance in rural—nonfarm areas and female
employment and females ages 15-24 in rural-farm areas.
Interestingly relatively few differences between independent
variables occur in the comparisons of the more metropolitan
divisions. More frequently the significant contrasts occur
when comparing the impact of the independent variables on
fertility among the less metropolitan divisions and as well
between the less and more metropolitan divisions.

Our conclusion for this hypothesis, then, is one of
confirmation. In accord with metropolitan dominance theory,
divisional differences are disappearing while residential
hinterland differences are becoming prominent. This is

further supported when considering the results of the multiple

o i
w
[T

“C .’

.
9

Li
:.

Ea:
Ea:
Ea!
Ea:
Ea:
Eel

3:233:
«warns!»

,v

(DUDCDU)
0000

('1 P11 [II
0! w

(11'

P411

Table 48.

342

Summary of results for multiple comparison tests between divisions of contermi-
nous United States by residential sector

 

 

Divisions Compared for Residential Sectors

Independent Variables

 

X
...

x3

X

X

X

X

 

URBAN

 

England
England
England
England
England
England
England
England

V3.
V8.
V8.
V8.
V8.
V8.
VB.
vs.

Middle Atlantic
East North Central
West North Central
South Atlantic
East South Central
West South Central
Mountain

Pacific

Middle
Middle
Middle
Middle
Middle
Middle
Middle

Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic

V8.
V8.
V8.
V8.
V3.
V3.
V8.

East North Central
West North Central
South Atlantic
East South Central
West South Central
Mountain

Pacific

East
East
East
East
East
East

West
West
West
West
West

North
North
North
North
North
North

North
North
North
North
North

Central
Central
Central
Central
Central
Central

Central
Central
Central
Central
Central

V8.
V3.
V8.
V8.
V8.
V8.

V8.
V3.
V8.
V8.
V8.

West North Central
South Atlantic
East South Central
West South Central
Mountain

Pacific

South Atlantic
East South Central
West South Central
Mountain

Pacific

South Atlantic

VS.

East South Central

South Atlantic
South Atlantic
South Atlantic

East
East
East

West
West

South Central vs. West South Central

vs. West South Central

vs. Mountain
vs. Pacific

South Central vs. Mountain
South Central vs. Pacific

South Central vs. Mountain
South Central vs. Pacific

Mountain vs. Pacific

RURAL-NONFARM

New
New
New
New
New
New
New
New

England
England
England
England
England
England
England
England

VB.
VB.
VI.
V8.
V3.
V3.
V8.
V3.

Middle Atlantic
East North Central
west North Central
South Atlantic
East South Central
West South Central
Mountain

Pacific

Middle
Middle
Middle
Middle
Middle
Middle
Middle

Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic

V8.
VI.
V8.
V8.
V8.
V8.
V8.

East North Central
West North Central
South Atlantic
East South Central
West South Central
Mountain

Pacific

0 00 000 0000 OOOOH 0000H0 0000000 00000000

0000000 0H000000

0 00 000 0000 00000 000000 0000000 HHO0P000

0000000 00000000

0 00 000 0000 00000 000000 0000000 l-‘l-‘P-‘l-‘OHOH

0000000 00000000

0H00000 00000000

0 00 OO-‘H OHOO CPD-'00 HHPHO—‘H 0HH0000 0H000000

0 00 000 0H00 CHOP-‘0 0000I-‘H 0H0000H 00000000

0000000 00000000

0 00 H00 0000 000PH 000000 0000000 00000000

0000000 00000000

0 00 HOP 0000 0000-‘0 000000 0000000 00000000

0000000 00000000

0 00 000 0000 00HOH 00POH0 000000.‘ 00000000

0000000 00000000

0 00 GOP oor-oo COD-‘00 00P000 0000000 00000000

0000000 00000000

 

m.r ~.w «so Alw ALL _,_

HKV

Asu a\;

.x.

To

A.. TL

 
 
 
 

.5; r u

... H.L up;

A.

H
u

343

Table 48—-Continued

 

 

Independent Variables

 

Divisions Compared for Residential Sectors X X X X X

X
...

X
N

3 4 5 6 7

X
(D

 

East North Central vs. West North Central
East North Central vs. South Atlantic
East North Central vs. East South Central
East North Central vs. West South Central
East North Central vs. Mountain

East North Central vs. Pacific

West North Central vs. South Atlantic
West North Central vs. East South Central
West North Central vs. West South Central
West North Central vs. Mountain

West North Central vs. Pacific

South Atlantic vs. East South Central
South Atlantic vs. West South Central
South Atlantic vs. Mountain

South Atlantic vs. Pacific

East South Central vs. West South Central
East South Central vs. Mountain
East South Central vs. Pacific

West South Central vs. Mountain
West South Central vs. Pacific

H OH 000 0000 0P000 000000
0 00 000 0000 00000 000000
0 00 000 0000 00000 000000
0 00 000 0000 00000 000000
0 00 000 OHOO 00000 0000D-‘0
0 00 000 0000 000t-‘0 000H00
0 00 HOD-J 000R OPOHP COCO-'00

Mountain vs. Pacific

RURAL-FARM

New England vs. Middle Atlantic
New England vs. East North Central
New England vs. West North Central
New England vs. South Atlantic

New England vs. East South Central
New England vs. West South Central
New England vs. Mountain

New England vs. Pacific

Middle Atlantic vs. East North Central
Middle Atlantic vs. West North Central
Middle Atlantic vs. South Atlantic
Middle Atlantic vs. East South Central
Middle Atlantic vs. West South Central
Middle Atlantic vs. Mountain

Middle Atlantic vs. Pacific

East North Central vs. West North Central
East North Central vs. South Atlantic
East North Central vs. East South Central
East North Central vs. West South Central
East North Central vs. Mountain

East North Central vs. Pacific

West North Central vs. South Atlantic
West North Central vs. East South Central
West North Central vs. West South Central
West North Central vs. Mountain

West North Central vs. Pacific

00000 000000 0000000 00000000
00000 000000 0000000 00000000
00000 000000 0000000 00000000
0H000 OHOOOO 0000000 00000000
COD-‘00 000000 0000000 00000000
00000 000000 0000000 00000000
000l-‘0 OHOHHO 0000000 00000000

0 00 000 0000 00000 000000
0 00 000 0000 00000 000000

OOHOH 00H000 0000000 00000000
00000 000000 0000000 00000000

344

Table 48--Continued

 

l

Independent Variables

 

X X X

X

 

Divisions Compared for Residential Sectors X1 X2 X3 X4 5 6 7 8 9
South Atlantic vs. East South Central 0 O 0 O 0 O l 0 0
South Atlantic vs. West South Central 0 O 0 0 1 0 0 O 0
South Atlantic vs. Mountain 0 0 0 l 0 O O 1 0
South Atlantic vs. Pacific 0 0 0 O 0 0 0 0 0
East South Central vs. West South Central 0 0 0 0 l O l J 0
East South Central vs. Mountain 0 0 0 U 0 0 a I 0
East South Central vs. Pacific 0 0 0 O 0 0 0 O 0
West South Central vs. Mountain 0 0 0 l l 0 0 l 0
West South Central vs. Pacific 0 0 0 0 0 0 0 0 0
Mountain vs. Pacific 0 O 0 0 0 O O O O

 

"1" denotes that there is a significant difference between the regression coef‘

cients of the independent variable for the two divisions compared.

”0“ denotes that there is 22 significant difference between the rtJressi»n
coefficients.

Metropolitan dominance.

Percent male labor force employed as farmers and farm managers.
Percent male labor force employed as farm laborers and foremen.
Median years school completed by males and females, age 25 and over.
Median family income in 1959.

Median female personal income in 1959.

Percent females, age 14 and over, employed.

Percent ever-married females, ages 15-44, who are age 15-24.

Percent ever-married females, ages 15-44, who are age 25-34.

~o

lrn

FH

D—d

If)

345

comparison tests for only the more metropolitan divisions,
the divisions which may be considered the models for all
divisions of the nation as metropolitanization becomes a
prevailing mode'of organization. Divisional differences
will be superceded by differences among the residential
hinterland sectors effected primarily by the differentiating

influences of metropolitan centers.

Summaryyof Findings

 

Seven basic hypotheses, each emanating from a metro-
politan dominance theoretical framework, have been tested
against various types of data. Of these seven hypotheses,
six have been substantiated and one rejected. Let us review
the results of the testing of these seven hypotheses very
briefly.

It has been determined that community social struc-
ture is a function of distance and size of a dominating met-
ropolitan center. Community social structure variables
highly influenced by metropolitan dominance are family in—
come, employment of farmers and farm managers, and ever-
married females ages 15—24, although significant correla—
tions were also found for employment of farm laborers and
foremen, education, and female employment. Differential
impact of metropolitan dominance is supported more by the
existence and degree of association with the community social
structure variables than by direction of association. For

education and employment of farmers and farm managers,

346

metropolitan dominance is more influential in rural hinter-
land communities, but for female personal income and ever—
married females ages 15-24, metropolitan dominance is more
influential in urban hinterland communities. The size of
the correlation coefficients suggests, however, that metro—
politan dominance is not as pervasive in determining inter-
community variation of social structure as expected.

It has also been shown that fertility behavior is a
function of both metropolitan dominance and community social
structure. Differential impact of metropolitan dominance is
supported only by the degree of association within the resi-
dential sectors. In all sectors it is negatively associated
with fertility. The only community social structure variable
which supports a differential impact with respect to direc-
tion of association is employment of farmers and farm man—
agers. In terms of degree of association, community social
structure variables which are more determinate of urban fer-
tility are family income, female income, ever-married females
ages 15—24 and 25-34. Community social structure variables
highly influential in rural areas are employment of farmers
and farm managers and education. Combined, however, the
community social structure variables and metropolitan dom-
inance reflect greatest influence in determining fertility
variation in urban areas, lowest in rural-farm areas, and

intermediate in rural-nonfarm.

347

The third hypothesis, which was also supported, was
tested by means of partial and multiple-partial correlation
coefficients. It was found that fertility behavior con-
tinues to be significantly related to community social
structure after the influence of metropolitan centers has
been controlled. Employment in agricultural occupations
and socio—economic status were especially important in deter-
mining fertility levels in rural hinterland communities
while wife's alternative opportunities and demographic age
structure tend to be more influential in urban hinterland
communities. The fact that the correlation coefficients of
fertility and community social structure variables remained
statistically significant for the most part, after metropol—
itan dominance had been partialled, provides additional
evidence by which to doubt the pervasiveness of metropolitan
dominance in determining inter-community fertility variation.
Fertility levels in urban and rural areas are the product of
both metropolitan dominance and local community conditions.

The fourth hypothesis of this study was not supported
by the data employed. On the basis of the development of a
metropolitan dominance theoretical framework it was proposed
that community fertility behavior would be more a function
of metropolitan dominance in urban hinterland communities,
but more a function of local community social structure in
rural hinterland communities. Both multiple-partial correla—

tion coefficients and beta coefficients were employed to

348

determine the relative importance of metropolitan dominance
and community social structure in determining urban and
rural fertility variation. For urban areas it was expected
that metropolitan dominance would reveal coefficients exceed-
ing those for community social structure variables in deter—
mining fertility variation, but that community social struc—
ture coefficients would be more prominent in rural areas.
The data indicated no consistent pattern in this respect.
Even when considering the average rank values of the inde-
pendent variables among the divisions in determining urban
and rural fertility levels, it was discovered that in all
residential sectors a community social structure variable
exceeded metropolitan dominance for relative importance in
determining fertility variation. The most important commu-
nity social structure variable in urban areas was ever-
married females ages 15—24 and in both rural sectors, educa—
tion.

The fifth and sixth hypotheses are similar to the
previous hypotheses but different in that they considered
the possibility that patterns of association among the
dependent and independent variables were blurred by the
existence of different levels of metropolitanization among
the divisions of the nation. In the testing of these two
hypotheses some control over the differential metropolitan—
ization of the divisions was attempted by subdividing the

divisions into two categories: (1) those indicating a

349

percentage of population residing in metropolitan areas

above the level of the nation and (2) those with percentage
metropolitan below the level of the nation. It was expected,
then, that the patterns predicted on the basis of metropol-
itan dominance theory would hold true more for the more
metropolitan divisions than the less metropolitan divisions.
It was determined that metropolitan dominance is a more
important determinant of community social structure in the
more metropolitan divisions than the less metropolitan divi—
sions in all three residential sectors. On the average cor—
relation coefficients between metropolitan dominance and
community social structure were higher for the more metropol—
itan divisions. This pattern was true for all the average
coefficients in the rural-nonfarm sector, six of the eight
coefficients in the rural-farm, and five of the eight in the
urban sector. The pattern of more relative importance of
metropolitan dominance in the more metropolitan divisions

was much stronger when considering the influence of metropol-
itan dominance on urban, rural—nonfarm, and rural—farm fer-
tility levels. Though inconsistencies in the predicted
pattern do occur in the comparison of individual divisions

by level of metropolitanization, it appears that a universal
metropolitan dominance pattern has not as yet emerged among
all divisions of the nation and, therefore, level of metro-
politanization is a significant factor to consider in explor-
ing further hypotheses derived from metropolitan dominance

theory.

350

The sixth hypothesis attempted to test a more
specific pattern in the dominating influence of metropolitan
centers on their hinterlands. This hypothesis is a repeat
of the fourth hypothesis, which was rejected, but with the
added factor of level of metropolitanization among the divi-
sions somewhat controlled. The finding of this test indi—
cates that in all residential sectors metropolitan dominance
is on the average more important in determining fertility
levels than community social structure in the more metropol—
itan divisions, but superceded in the less metropolitan
divisions by various community social structure variables.
Though previous results suggested the questioning of the
pervasiveness of metropolitan centers as organizing agents
of hinterland inter-community variation for the nation, the
finding here suggests the reinstatement of this pattern par—
ticularly for the divisions of the nation which are more
advanced in the development of metropolitanization. The
data indicated that hinterlands of all three residential
types do come more under the dominance of metropolitan cen-
ters than the influence of local conditions in the more
metropolitan divisions. Apparently in the less metropolitan
divisions local community conditions still exert a consider-
able impact in determining inter—community variation in
urban, rural—nonfarm, and rural—farm hinterlands compared

with metropolitan dominance.

351

Finally the verification of the seventh and last
hypothesis of this study also adds considerably to the sup-
port of metropolitan dominance theory. In the chapter deal-
ing with the theoretical framework of this study, it was
mentioned that one of the primary points which distinguishes
urban dominance and metropolitan dominance theory has to do
with the pattern of effects exerted by cities upon their
hinterland region. Urban dominance theory predicts the
disappearance of urban and rural residential differences on
the basis of the assumption that cities exert a homogeneous
impact on hinterland communities. Metropolitan dominance
theory, on the other hand, leads to the conclusion that
urban-rural differences will persist primarily due to the
differential impact of metropolitan centers upon hinterland
communities. By means of the multiple comparison test it
was possible to test whether the differential effects pat—
tern among residential sectors of metropolitan hinterland
regions was true with respect to the comparison of the
effects of the independent variables employed in the multi-
ple regression equation on fertility variation. It was
found that among the divisions of the nation, the indepen—
dent variables in many cases do exert differential effects
on fertility in comparing the residential sectors. Hence we
conclude that fertility differences do exist between the
residential areas of the divisions, though these differences

were shown to be in terms of the pattern of association of

352

the independent variables in determining fertility levels
themselves. In other words, this means that a factor, such
as education, has a significantly different effect on fertil—
ity in urban areas than rural areas. It is suggested, then,
that the chief cause of this pattern of differential effects
is the influence of metropolitan centers on the different
types of residential hinterlands. The fact that the pattern
of differential effects between residential sectors was
found to be more pronounced in the more metropolitan divi-
sions provides added support to this particular hypothesis
derived from metropolitan dominance theory. Furthermore,
the finding that few differences exist for the effect of the
independent variables on fertility in the comparison of the
divisions also adds support to the metropolitan dominance
theory. What this implies, of course, is that divisional
and regional differences, which were formerly significant
sources of areal differences within the nation are giving
way in the face of the increasingly greater influence of
metropolitan centers in determining inter-community and
interareal variation in the nation. This is supported by
the finding that similarity of effects on fertility by the
independent variables is to be found more clearly among the
more metropolitan divisions, the divisions in which the most
support for the differential impact among residential sec-
tors was expected to be found.

In conclusion, we have found that fertility behavior

is influenced both by distance and size of a dominating

353

metropolitan center and community social structure. In
speaking of community social structure, we have in mind the
influence of local conditions indigenous to the respective
hinterland communities, since inter-community social struc-
ture variation produced by the influence of metropolitan
centers had been partialled. Among all the divisions it
appeared that local community conditions exert a consider-
able influence on fertility behavior in all residential sec-
tors, more than the metropolitan dominance variable. However,
in considering the more metropolitan divisions distinct from
the less metropolitan divisions, it was found that metropol—
itan dominance is a very important determinant of community
social structure variation and fertility behavior. Further-
more for these same more metropolitan divisions, it was found
that in all residential sectors metropolitan dominance ex-
ceeded community social structure in determining fertility
levels, whereas in the less metropolitan divisions education
and females age 15-24 exceeded metropolitan dominance in
rural and urban hinterland communities respectively. A dif-
ferentiating process is operative among the residential sec-
tors, especially among the more metropolitan divisions, where—
as divisional differences are diminishing. It is concluded
that the discovery of such a pattern is proof of a gradual
emergence of metropolitan centers as the primary organizing
agent of inter-community hinterland variation. As all divi-
sions continue to become more metropolitan, the influence of

metropolitan centers will likewise increase.

CHAPTER VI

REFLECTIONS AND IMPLICATIONS

FOR FURTHER RESEARCH

New approaches to the study of differential fertil—
ity are definitely needed in demographic research today. It
is my hope that if any contribution has been made by this
thesis, it has been in the direction of demonstrating a new
approach to the study of the urban-rural fertility differen-
tial, or for that matter, any of the fertility differentials.
Indeed the prediction by many demographers that the gap
between urban and rural fertility levels in the United
States is disappearing is not sufficient justification to
end our investigations in this area. Rather new approaches,
methodological and theoretical, must be sought in order to
increase our understanding of the true and complex dynamics
of fertility.

Methodologically the present study has attempted to
point out the inadequacies of such conventional approaches
to differential fertility as trend analysis and the "aggre-
gate" approach. Longitudinal and cross-sectional analyses
which attempt merely to document whether there exists a sig-
nificant association between various socio-economic and

demographic variables and fertility are not sensitive enough

354

355

to represent accurately and comprehensively current differen—
tial fertility patterns. The degree of association as well
as the relative importance of such variables in determining
fertility patterns should be investigated also. Although
there are other methodological approaches available which
could solve such problems, the present study found a dis—
tributive approach operationalized in the form of multiple
regression analysis extremely valuable for such purposes.
Employing new methodological techniques, however, does not
necessitate a search for new sources of data. Though a
major source of information for studies employing trend
analysis and the aggregate approach has been the census,
and though the vogue in fertility analysis has been large
sample surveys, the present study has attempted to demon-
strate that census data, likewise, are very amenable to
entirely different methodological approaches such as
multiple regression analysis. Methodological ingenuity and
experimentation are urgently needed in differential fertil—
ity analysis at a time when conventional methods seem to be
providing only fruitless repetition and little new knowledge.
Theoretically we must also turn in new directions.
Differential fertility theory has been as deficient as its
methodology. Indeed there is much evidence of emerging new
middle range theories of differential fertility and these
are contributing much to a bridging of the gap which formerly

existed between "theory" and "empirical data." Demographic

356

transition theory, which for many years served as puny proof
that demographers were interested in theory, has undergone
repeated attack. Many of its generalizing principles have
been shown to be inadequate to explain current as well as
past fertility patterns. It has failed to generate new
hypotheses. In addition this thesis has attempted to demon—
strate that the theory itself, when applied to the current
situation in American society, leads to inaccurate and ques—
tionable conclusions. Implicit in the demographic transi-
tion theory is the concept of urban dominance. It is this
concept which has led many researchers to the conclusion
that the urban—rural fertility differential is doomed to
dissolution. The present study has argued that urban domi—
nance is not an adequate description of the dynamics of
American society, but metropolitan dominance is. It is
suggested that metropolitan dominance theory must somehow

be incorporated into the framework of demographic transition
theory if the latter is to continue as a useful theoretical
framework by which to understand currently modern complex
societies.

By substituting metropolitan dominance theory for
urban dominance theory in differential fertility analysis,
new questions and problems are produced and new dimensions
for further research are uncovered. But, of course, there
have been deficiencies in the manner by which metropolitan
dominance theory has been researched as well. Many re—

searchers apparently have failed to perceive that these two

357

theoretical positions, urban and metropolitan dominance,
lead to conflicting conclusions. This oversight has been
due primarily to an overemphasis on the gradient principle
and an equally underemphasis of the differentiating prin—
ciple. The gradient principle, it has been noted previously,
is an element common to both urban dominance and metropol—
itan dominance theory. Hence the successful testing of this
principle has usually found the researcher confused as to
whether his findings actually supported one theory or the
other. All too frequently the researcher would choose to
avoid the issue altogether rather than deal with it directly.
It is the differentiating principle, however, which is the
distinguishing characteristic of the two theories, but
little research has been performed directed by this prin-
ciple. In the application of metropolitan dominance theory
to differential fertility analysis, it is the differentiating
principle which has lead to the conclusion that urban and
rural fertility levels may not eventually converge. Even if
the levels themselves do converge, on the basis of the ideas
presented on differential fertility in this thesis, it is
possible that the patterns of differential fertility found
in urban and rural areas may not necessarily resemble each
other.

Much more research, then, is needed to determine the
differentiating effects of metropolitan centers on the hinter—
land communities comprising the metropolitan region. The

present study attempted to do just that and the findings

358

seem to indicate that urban and rural differential fertility
patterns are distinct, especially in the more metropolitan
geographic divisions of the nation. In a sense, then, the
findings of this study have lent some support to the "pro"
side of the current argument as to whether the "urban" and
"rural" categories provide a meaningful distinction in the
analysis of American society. The present study has assumed,
of course, that one of the main dimensions of the differen-
tiating effect of metropolitan dominance on hinterland com—
munities is along urban-rural lines. The findings of this
study seem to bear this out, although, obviously, more
research on this problem is needed. I am not wholly satis-
fied with the design of this thesis in testing this question,
however. A weakness of the present study, it seems to me,
is the failure to consider the differentiating effects of
metropolitan dominance for areas smaller than a geographic
division. It seems to me that the next logical step is to
research the hypotheses of this study for individual metro-
politan regions. At this level it would be possible to
contrast the differentiating effect of metropolitan domi-
nance on urban and rural hinterland communities even within
the same distance zone from a given metropolitan center. In
such a study, of course, differential fertility patterns
would provide only one of numerous dimensions which should
be investigated to determine the differentiating effect of
metropolitan dominance on urban and rural hinterland commu-

nities. In addition, other classification schemes for

359

hinterland communities must also be devised other than the
traditional urban—rural categories.

In conclusion, I wish to point out that these are
only a few of the problems which could profitably be re—
searched in the future. As the pattern is in any intensive
struggle with a given problem, I find this study gives rise
to many more questions than it answers. In fact I am not
sure myself that the answers which have been presented in
this thesis are valid. I do not intend, however, to follow
the example of Malthus and spend the rest of my life produc—
ing improved editions of the same old, overworked ideas. If
the ideas presented in this thesis are of any value, they

should lead to the production of even fresher ideas.

BIBLIOGRAPHY

Abu-Lughod, Janet. "Urban—Rural Differences as a Function
of the Demographic Transition: Egyptian Data and an
Analytical Model," American Journal of Sociologyi LXIX
(March, 1964), 476-90.

 

Anderson, Theodore R., and Jane Collier. "Metropolitan
Dominance and the Rural Hinterland," Rural Sociology,
XXI (June, 1956), 152-57.

 

Barclay, George W. Techniques of Population Analysis.
New York: John Wiley, 1958.

Bash, Wendell. "Differential Fertility in Madison County,
New York, 1865," Milbank Memorial Fund Quarterly, XXXIII
(April, 1955), 161-86.

 

. "Changing Birth Rates in Developing America: New
York State, 1840—1875," Milbank Memorial Fund Quarterly,
XLI (April, 1963), 161-82.

 

Beegle, J. Allan. "Social Structure and Changing Fertility
of the Farm Population," Rural SociologX. XXXI, No. 4
(December, 1966), 415-27.

 

Blake, Judith. "Ideal Family Size among White Americans:
A Quarter Century of Evidence," Demography, III, No. l
(1966), 154-73.

 

Blalock, Hubert M., Jr. Social Statistics. New York:
McGraw-Hill, 1960.

 

Bogue, Donald J. The Structure of the Metropolitan Commu—
nity. Ann Arbor: Horace H. Rackham School of Graduate
Studies, University of Michigan, 1950.

 

_*_______ “Urbanism in the United States, 1950," American
Journal of SociologX; LX (March, 1955), 471—86.

 

 

. "Population Distribution," in P. Hauser and
O. D. Duncan, The Study of Population. Chicago:
University of Chicago Press, 1959, pp. 383-99.

 

, and Dorothy L. Harris. Comparative Population and
Urban Research Via Multiple Regression and Covariance
Analysis. Oxford, Ohio: Scripps Foundation, Miami
University, 1954.

 

 

360

361

Bogue, Donald J., and James A. Palmore. "Some Empirical and
Analytic Relations among Demographic Fertility Measures,
with Regression Models for Fertility Estimation,"
Demography, I, No. l (1964), 316-38.

Brownlee, K. A. Statistical Theory and Methodology in
Science and Engineerihg. New York: John Wiley, 1960.

Brunner, Edmund deS., and J. H. Kolb. Rural Social Trends.
New York: McGraw-Hill, 1933, Ch. 5.

Campbell, Arthur A. "Design and Scope of the 1960 Study of
Growth of American Families," in C. V. Kiser, Research

in Family Plannihg. Princeton: Princeton University
Press, 1960, pp. 167-83.

 

Comhaire, Jean, and Werner J. Cahnman. How Cities Grow.
Madison, New Jersey: The Floram Park Press, 1959.

 

Cowgill, Donald 0. "Transition Theory as General Population
Theory," Social Forces, XLI (March, 1963), 270-74.

Davis, Kingsley. "The World Demographic Transition," The
Annals, CCXXXVII (January, 1945), 1—11.

. The Human Society. New York: Macmillan, 1949.

 

 

. "The Sociology of Demographic Behavior," in
Robert K. Merton, Leonard Broom, and Leonard S. Cottrell,

Jr., Sociolognyodgy; Problems and Prospects. New YOrk:
Basic Books, 1959, pp. 309-33.

, and Judith Blake. "Social Structure and Fertility:
An Analytical Framework," Economic Development and
Cultural Change, IV (1956), 211-35.

 

Dinkel, Robert M. "Occupation and Fertility in the United

States," American Sociological Review, XVII (April,
1952), 178-83.

 

"Fertility in Mid-Century America," Eugenics
Quarterly, III (March, 1956), 22-31.

 

Duncan, Otis Dudley. "Fertility of the Village Population

in Pennsylvania, 1940," Social Forces, XXVIII (March,
1950), 304-9.

. "Note on Farm Tenancy and Urbanization," Journal

of Farm Economics, XXXVIII, No. 4 (November, 1956),
1043-47.

 

D1

Fe

F]

P:

362

Duncan, Otis Dudley. "Gradients of Urban Influence on the
Rural Population," The Midwest Sociologist, XVIII
(Winter, 1956), 27-30.

 

. "Human Ecology and Population Studies," in P.
Hauser and O. D. Duncan, The Stugy of Population.
Chicago: University of Chicago Press, 1959, pp. 678-716.

 

 

. "From Social System to Ecosystem," Sociological
Inguiry, XXXI (Spring, 1961), 140-49.

 

 

. "Residential Areas and Differential Fertility,"
Eugenics Quarterly, XI (June, 1964), 82—89.

 

 

. "Farm Background and Differential Fertility,"
Demography, II (1965), 240-49.

 

 

, Ray P. Cuzzort, and Beverly Duncan. Statistical
Geography. Glencoe: The Free Press, 1961.

 

 

 

 

 

, and Beverly Davis. "An Alternative to Ecological
Correlation," American Sociological Review, XVIII (1953),
665-66.

, and A. J. Reiss, Jr. Social Characteristics of
Urban and Rural Communities, 1950. New York: John
Wiley, 1956.

 

 

 

, and Leo Schnore. "Cultural, Behavioral, and
Ecological Perspectives in the Study of Social Organiza-
tion," American Journal of Sociology, XLV (September,
1959), 132-46.

 

Feldman, Arnold S., and Charles Tilly. "The Interaction of
Social and Physical Space," American Sociological Review,
XXV, No. 6 (December, 1960), 877—84.

 

Firey, Walter, Charles P. Loomis, and J. Allan Beegle. "The
Fusion of Urban and Rural," in Jean Labatat and Wheaton
J. Lane, HighWays in Our National Life: A Symposium.
Princeton: Princeton University Press, 1950, pp. 154-63.
Reprinted in P. K. Hatt and A. J. Reiss, Jr., Cities and
Society. 2nd ed. Glencoe: The Free Press, 1957,

 

 

 

pp. 214—22.

Freedman, Deborah. "The Relation of Economic Status to
Fertility," American Economic Review, LIII (June, 1963),
414-26.

Freedman, Ronald. "The Sociology of Human Fertility,"
Current Sociology, X—XI, No. 2 (1961-62), 35—119.

 

363

Freedman, Ronald. "American Studies of Family Planning and
Fertility: A Review of Major Trends and Issues," in
C. V. Kiser, Research in Family Planning. Princeton:
Princeton University Press, 1962, pp. 211-27.

, and Deborah Freedman. "Farm-Related Elements in
the Nonfarm Population," Rural Sociology, XXI (March,
1956), 50-61.

 

 

, David Goldberg, and Doris Slesinger. "Current
Fertility Expectations of Married Couples in the United
States," Population Index, XXIX (October, 1963), 366-91.

 

 

, David Goldberg, and Doris Slesinger. "Fertility
Expectations in the United States: 1963," Population
Index, XXX (April, 1964), 171-75.

 

 

, and Doris P. Slesinger. "Fertility Differentials
for the Indigenous Non—Farm Population of the United
States," Population Studies, XV (November, 1961), 161-73.

 

 

, P. K. Whelpton, and A. A. Campbell. Family Plan—
ning, Sterilipy and Population Growth. New York:
McGraw-Hill, 1959.

 

 

 

Gini, C. "The Cyclical Rise and Fall of Population," in
Population, Harris Foundation Lectures. Chicago:
University of Chicago Press, 1929.

 

Goldberg, David. "The Fertility of Two Generation Urbanites,"
Population Studies, XII (March, 1959), 214-22.

 

. "Another Look at the Indianapolis Fertility Data,"
Milbank Memorial Fund Quarterly, XXXVIII (January, 1960),
23-36. '

 

 

. "Some Recent Developments in American Fertility
Research," in National Bureau of Economic Research,
oomographic and Economic Change in Developed Countries.
Princeton: University of Princeton Press, 1960, pp. 137-
51.

 

 

. "Fertility and Fertility Differentials: Some
Observations on Recent Changes in the United States,"

in M. C. Sheps and J. C. Ridley, Public Health and
Population Change. Pittsburgh: University of Pittsburgh
Press, 1965.

 

 

 

Goldsmith, Harold F., and James H. Copp. "Metropolitan
Dominance and Agriculture," Rural SociologY. XXIX, No. 4
(December, 1964), 385-95.

 

364

Goldstein, Sidney. "Changing Patterns of Fertility in
Norristown, Pennsylvania, 1920-1950," Social Forces,
XXXIV (October, 1955), 72-76.

 

, and Kurt B. Mayer. "Residence and Status Differ-
ences in Fertility," Milbank Memorial Fund Quarterly,
LXIII, No. 3 (July, 1965), 291-310.

 

 

 

Goodman, Leo A. "Ecological Regression and Behavior of
Individuals," American Sociological Review, XVIII (1953),
663-64.

. "Some Alternatives to Ecological Correlation,"

 

American Sociologlcal Review, XXIV (1959), 610-25.

, and William H. Kruskal. "Measures of Association
for Cross-Classification," Journal of the American
Statistical Association, XLIX (1953), 732-64.

 

 

Grabill, W. H. "The Fertility of the United States Popula-
tion," in Donald J. Bogue, The Population of the Unitoo
States. Glencoe: The Free Press, 1959, pp. 288—324.

 

, C. V. Kiser, and P. K. Whelpton. The Fertility of
American Women. New YOrk: John Wiley, 1958.

 

 

Gras, N. S. B. An Introduction to Economic Histopy, New
York: Harper, 1922.

Grigg, Charles M. "A Proposed Model for Measuring the
Ecological Process of Dominance," Social Forces, XXXVI,
No. 1 (December, 1957), 128—31.

 

, and Rupert Vance. "Regionalism and Ecology: A
Synthesis?" Florida State University Research Roporpg
in Social Science, III, No. 2 (August, 1960), l—ll.

 

 

 

Gutman, Robert E. "Comment on Kiser's Paper," in National
Bureau of Economic Research, Demographic and Economic
Change in Developed Countries. Princeton: Princeton
University Press, 1960, pp. 113-116.

 

 

"In Defense of Population Theory," American
Sociological Review, XXV (June, 1960), 325-33.

 

 

, and Irving Bender. "Some Sources of Variation in
Family Size of College Graduates," Milbank Memorial Fund
gparterlyj XXXV (July, 1957), 287-301.

 

 

Hagood, Margaret J., and Daniel 0. Price. Statistics for
Sociologists. Rev. ed. New York: Holt, Rinehart and
Winston, 1952.

 

365

Hajnal, John. "The Study of Fertility and Reproduction:
A Survey of Thirty Years," in Thirty Years of Research
in Human Fertility: Retrospect and Prospect. New York:
Milbank Memorial Fund, 1959, pp. 11-37.

Hamilton, C. Horace. "Population Pressure and Other Factors
Affecting Net Rural-Urban Migration," Social Forces,
XXX, No. 2 (December, 1951), 209—15.

Hashmi, Sultan S. "Trends and Factors in Urban Fertility
Differences in the United States.“ Unpublished Ph.D.
dissertation, Department of Sociology, University of
Chicago, 1962.

"Factors in Urban Fertility Differences in the
United States," in Ernest W. Burgess and Donald J. Bogue,
Contributions to Urban Sociology. Chicago: University
of Chicago Press, 1964, pp. 42—58.

 

Hathaway, Dale, J. Allan Beegle, and Keith Bryant, Rural
America, Census Monograph in press.

Hatt, P. K., N. L. Farr, and E. Weinstein. "Types of
Population Balance," American Sociological Review, XX
(February, 1950), 14-21.

Hauser, Philip M. "Present Status and Prospects of Research
in Population," American Sociological Review, XIII
(August, 1948), 371-82.

"On Design for Experiment and Research in

 

 

 

Fertility Control,‘I in C. V. Kiser, Research in Family_
Plannipg. Princeton: Princeton University Press, 1960,
pp. 463-74.

, and Otis Dudley Duncan. The Study_of Population.
Chicago: University of Chicago Press, 1959.

 

Hawley, Amos. Human Ecology. New YOrk: Ronald Press, 1950.

 

Hiller, E. T. "Extension' of Urban Characteristics into
Rural Areas," Rural Sociology;§QL(September¢{1941),;
3242—57 0'. I V :5, .

Hillery, George, Jr. "Toward a Conceptualization of
Demography," Social Forces, XXXVII (October, 1958),

45-51.

Hughes, Rufus B., Jr. "Human Fertility Differentials: The
Influence of Industrial-Urban Development on Birth Rates,"
Population Review, III (July, 1959), 58—69.

366

Hutchhmmm” Edward P. "Comment on Ryder's Paper," in
National Bureau of Economic Research, Demographic and
Economic Change in Developed Countries. Princeton:
Princeton University Press, 1960, pp. 131-33.

 

 

Jaffe, A. J. "Urbanization and Fertility," American Journal
of Sociology, XLVIII (July, 1942), 48—60.

 

 

 

Jonassen, Christian. "Functional Unities in 88 Community
Systems," American Sociological Review, XXVI (June, 1961),
399-407.

Jones, Lewis. "The Hinterland Reconsidered," American

Sociological Review, XX (February, 1955), 388-98.

 

Kiser, Clyde V. Group Differences in Urban Fertility.
Baltimore: Williams and Wilkins, 1942.

 

. "Changes in Fertility by Socio—Economic Status
during 1940-50," Milbank Memorial Fund Quarterly, XXXIII
(October, 1955), 393-429.

 

"Fertility Trends and Differentials in the United
States," Journal of the American Statistical Association,
XLVII (March, 1956), 162-83.

 

"Methodological Lessons of the Indianapolis
Fertility Study," Eugenics Quarterly, III (September,
1956), 152-56.

 

 

"Differential Fertility in the United States," in
National Bureau of Economic Research, Demogpaphic and
Economic Chapge in Developod Countries. Princeton:
Princeton University Press, 1960, pp. 77-113.

 

, and P. K. Whelpton. "Fertility Planning and
Fertility Rates by Socio—Economic Status," Milbank
Memorial Fund Quarterly, XXVII, No. 2 (April, 1949),
188-244.

 

, and P. K. Whelpton. "Resume of the Indianapolis
Study of Social and Psychological Factors Affecting
Fertility," Population Studies, VII (1953), 95—110.

Kish, Leslie. "Differentiation in Metropolitan Areas,"
American Sociological Review, XIX (August, 1954), 388-98.

 

Kitagawa, Evelyn M., and P. M. Hauser. "Trends in Differen—
tial Fertility and Mortality in a Metropolis--Chicago,"
in E. Burgess and D. Bogue, Contributions to Urban
Sociology. Chicago: University of Chicago Press, 1964,
pp. 59-85.

 

 

367

Kunz, Phillip R. "The Relation of Income and Fertility,"
Journal of Marriage and the Family, XXVII (November,
1965), 509-13.

Lorimer, Frank. Culture and Human Fertility. New YOrk:
International Documents Service, Columbia University
Press, 1955.

McKenzie, R. D. The Metropolitan Community. New York:
McGraw-Hill, 1933.

Martin, Walter T. "Ecological Change in Satellite Rural
Areas," American Sociological Review, XXII (April, 1957),
178-83.

Mayer, Kurt. "Fertility Changes and Population Forecasts
in the United States," Social Research, XXVI, No. 3
(Autumn, 1959), 347—66.

 

Merton, Robert K. Social Theory and Social Structure.
Glencoe, Illinois: The Free Press, 1957.

Mishler, Elliot, and Charles Westoff. "A Proposal for
Research on Social Psychological Factors Affecting
Fertility: Concepts and Hypotheses," in Current Research
in Human Fertilipy, New YOrk: Milbank Memorial Fund,
1955.

 

 

Moore, Wilbert. "Sociology and Demography," in P. Hauser
and O. D. Duncan, The Study of Popplation. Chicago:
University of Chicago Press, 1959, pp. 832-51.

 

Notestein, Frank W. "Class Differences in Fertility," The
Annals, CLXXXVIII (November, 1936), 26-36.

. "Population-~The Long View," in T. W. Schultz,
Food for the World. Chicago: University of Chicago
Press, 1945.

 

, and X. Sallume. "The Fertility of Specific
Occupation Groups in an Urban Population," Milbank
Memorial Fund Quarterly, X (April, 1932), 120-30.

 

Okun, Bernard. Trends in Birth Rates in the United States
since 1870. Baltimore: The Johns Hopkins Press, 1958.

 

Pearl, Raymond. The Biology of Population Growth. New York:
Knopf, 1925.

Quinn, James A. Human Ecology, New York: Prentice-Hall,
1950.

368

 

Robinson, W. S. "Ecological Correlations and the Behavior
of Individuals," American Sociological Review, XV (1950),
351-57.

Robinson, Warren C. "Urbanization and Fertility: The Non—

Western Experience," Milbank Memorial Fund Quarterly,
XLI, No. 3 (July, 1963), 291-308.

 

Rosenquist, Carl M., and Alvin H. Schafft. "Differential
Fertility in Rural Texas," Rural Sociology, XII (March,
1947), 21—26.

 

Ruggles, Richard, and Nancy Ruggles. ”Differential Fertility
in the United States Census Data," in National Bureau of
Economic Research, Demogpaphic and Economic Change in
Developed Countries. wPrinceton: Princeton University
Press, 1960, pp. 155-206.

 

 

Ryder, N. B. "Fertility," in P. Hauser and O. D. Duncan,
The Study of Population. Chicago: University of
Chicago Press, 1959, pp. 400—36.

 

Ryder, Norman. "Variability and Convergence in the American
Population," Phi Delta Kappgp, XLI (June, 1960), 379-383.

 

Schnore, Leo. ”Social Morphology and Human Ecology,"
American Journal of Sociology! LXIII (May, 1958), 620-34.

 

 

 

. "Social Mobility in Demographic Perspective,"
American Sociological Review, XXVI, No. 3 (June, 1961),
407-23.

. "The Rural-Urban Variable: An Urbanite's
Perspective," Rural SociologY. XXXI, No. 2 (June, 1966),
131—43.

 

 

Selvin, Hanan C. "A Critique of Tests of Significant in
Survey Research," American Sociological Review, XXII
(October, 1957), 519—27.

 

Snedecor, G. W. Statistical Methods. 5th ed. Ames: Iowa
State University Press, 1956.

 

Sorokin, Pitrim A. Social Mobilipy. New York: Harper, 1927.

 

, and C. C. Zimmerman. Principles of Rural-Urban
Sociology. New YOrk: Henry Holt, 1929.

 

 

 

Stoeckel, John, and J. Allan Beegle. "The Relationship
between the Rural-Farm Age Structure and Distance from
a Metropolitan Area," Rural Sociology, XXXI, No. 3.
(September, 1966), 346-54.

 

369

Sydenstricker, Edgar S., and Frank W. Notestein. "Differen—
tial Fertility According to Social Class," Journal of
the American Statistical Association, XXV (March, 1930),
9—32 0

 

 

Thompson, Warren S. "Population," American Journal of
Sociology; XXXI (May, 1929), 959-75.

 

 

, and Nelle E. Jackson. "Fertility in Rural Areas
in Relation to Their Distance from Cities, 1930," Rural
Sociology, V (June, 1940), 143-62.

 

Tisdale, Hope. "The Process of Urbanization," Social Forces,
XX (1942), 311—16.

 

United Nations. The Determinants and Consequences of
Population Trends. New YOrk: United Nations, Depart-
ment of Social Affairs, Population Division, 1953.

 

 

U. S. Bureau of the Census. U.S. Census of Population:
1960. Number of Inhabitants. United States Summary,
Final Report PC (1)-1A. Washington, D.C.: Government
Printing Office, 1961.

 

 

 

U. S. Bureau of the Census. U.S. Census of Population:
1960. General Social and Economic Characteristics.
United States Summary, Final Report PC(1)-lC. Washing-
ton, D.C.: Government Printing Office, 1962.

 

 

 

U. S. Bureau of the Census. U.S. Census of Population:
1960. Subject Reports. Women by Number of Children
Ever Born. Final Report PC(2)-3A. Washington, D.C.:
Government Printing Office, 1964.

 

 

 

 

Vance, Rupert. "Is Theory for Demographers?" Social Forces,
XXXI (1952), 9—13.

 

, and Sara Smith. l'Metropolitan Dominance and
Integration," in R. B. Vance and N. J. Demerath (eds.),
The Urban South. Chapel Hill: University of North
Carolina Press, 1954, pp. 114-34.

 

Westoff, C. F. "The Changing Focus of Differential Fertility
Research: The Social Mobility Hypothesis," Milbank
Memorial Fund Quarterly, XXXI, No. 1 (January, 1953),
24—38 0

 

. "Differential Fertility in the United States:
1900 to 1952," American Sociological Review, XIX, No. 5
(October, 1954), 549-61.

 

 

 

 

370

Westoff, C. F., E. G. Mishler, and E. L. Kelly. "Preferences
in Size of Family and Eventual Fertility Twenty Years
After," American Journal of Sociology, LXII (March,

1957), 491-97.

, R. G. Potter, Jr., P. C. Sagi, and E. G. Mishler.
Family Growth in Metropolitan America. Princeton:
Princeton University Press, 1961.

 

, Robert G. Potter, Jr., and Philip C. Sagi. The
Third Child: A Stugy in the Prediction of Fertilipy.
Princeton: Princeton University Press, 1963.

 

, P. C. Sagi, and L. E. Kelly. "Fertility Through
Twenty Years of Marriage," American Sociological Review,
XXIII, No. 5 (October, 1958), 549-56.

 

Whelpton, P. K. "Geographic and Economic Differentials in
Fertility," The Annals, CLXXXVIII (November, 1936),

37—55.

, and C. V. Kiser. "Social and Psychological
Factors Affecting Fertility: I. Differential Fertility
among 41,498 Native—White Couples in Indianapolis,"
Milbank Memorial Fund Quarterly, XXI (July, 1943),
221-80.

 

Williams, Robin M., Jr. "American Society in Transition:
Trends and Emerging Development in Social and Cultural
Systems," in James H. Copp, Our Changing Rural Society:
Perspectives and Trends. Ames: Iowa State University

Press, 1964.

Woofter, T. J. "Trends in Rural and Urban Fertility Rates,"
Rural Sociology, XIII (March, 1948), 3-9.

Wrong, Dennis. Population. New York: Random House, 1956.

 

APPENDIX A

DETAILED RESULTS OF MULTIPLE REGRESSION EQUATIONS

372

Table 49. First—order partial correlation of fertility and
employment in agricultural occupations (urbaniza-
tion) controlling for metropolitan dominance for
conterminous United States and division, by
residence: 1960

 

 

 

 

 

 

 

Correlation of Correlation of

Fertility and Farm- Fertility and Farm

ers & Farm Managers Laborers & Foremen
Conterminous
United States Urb RNF RF Urb RNF RF
and DIVISlonS rY2.1 rY2.1 rY2.1 rY3.1 rY3.1 rY3.1
UNITED STATES .138* .076* -.041* .327* .243* .110*
New England -.043 .345* .346* .279* .593* .324
Middle Atlantic .028 .046 -.080 -.008 .104 .161
E. N. Central —.049 -.l78* -.188* -.110* -.069 .167*
W. N. Central .045 -.005 -.045 .000 -.001 .237*
South Atlantic .130* .125* -.166* .174* .202* .075
E. S. Central —.054 —.029 -.311* —.l73* .133* .091
W. S. Central .092 —.036 -.268* .624* .513* .128*
Mountain .286* .235* -.205* .168* .198* -.107
Pacific .340* -.014 —.247* .328* .345* -.041

 

*Significantly different from zero at the .05 level.
Y Cumulative fertility ratio.

X Percent of male labor force who are employed as
farmers and farm managers.

X Percent of male labor force who are employed as
farm laborers and foremen.

11"

373

 

 

 

 

 

 

Table 50. First-order partial correlation of fertility and

socio—economic status controlling for metropolitan

dominance for conterminous United States and divi—

sions, by residence: 1960

Correlation of Correlation of
Fertility and Fertility and
Education Family Income
Conterminous
United States Urb RNF RF Urb RNF RF
and D1V151°n5 rY4.l rY4.1 rY4.1 rY5.1 rY5.1 rY5.1
UNITED STATES -.l91* —.225* -.l69* .063* .088* .119*
New England .038 —.353* -.l40 —.l47 —.308* -.198
Middle Atlantic .085 .001 -.017 -.l76* .019 .007
E. N. Central .060 -.178* -.398* .296* -.273* -.248*
W. N. Central -.175* -.278* -.404* -.323* -.l76* —.096*
South Atlantic —.278* -.385* -.268* -.116* .152* .123*
E. S. Central -.l90* -.397* -.247* -.037 -.l70* -.010
W. S. Central -.629* -.564* -.496* .051 -.098* -.289*
Mountain -.487* -.4l9* .007 .147 —.286* .053
Pacific -.456* -.595* —.356* -.040 -.095 -.238*
*Significantly different from zero at the .05 level.

Y Cumulative fertility ratio.

X Median years of school completed by males and
females, age 25 and over.

X Median family income in 1959.

5

l

374

 

 

 

 

 

 

Table 51. First-order partial correlation of fertility and
wife's alternative opportunities controlling for
metropolitan dominance for conterminous United
States and divisions, by residence: 1960

Correlation of Correlation of
Fertility and Female Fertility and
Personal Income Female Employment

Conterminous __

United States Urb RNF RF Urb RNF RF

and D1V1Sl°ns rY6.1 rY6.1 rY6.1 rY7.l rY7.1 rY7.l

UNITED STATES —.231* -.205* -.l45* -.179* —.279* —.l44*

New England —.092 —.348* -.265* —.073 —.l93 —.081

Middle Atlantic —.354* -.260* -.112 —.239* —.294* -.037

E. N. Central —.l59* —.195* —.1l3* -.028 -.ll4* .007

W. N. Central —.343* -.112* -.202* -.l65* -.039 -.023

South Atlantic -.295* -.289* -.l79* -.245* -.408* -.224*

E. S. Central -.l89* -.211* -.203* -.330* -.528* -.430*

W. S. Central -.235* -.148* —.l97* -.l46* -.l97* -.205*

Mountain -.372* -.234* -.l7l* —.455* -.391* -.211*

Pacific -.378* -.119 .011 -.185 -.087 -.020

*Significantly different from zero at the .05 level.

Y Cumulative fertility ratio.

x6

X Percent females,

7

Median female personal income in 1959.

age 14 and over, who are employed.

375

 

 

 

 

 

 

Table 52. First-order partial correlation of fertility and
demographic age structure controlling for metro-
politan dominance for conterminous United States
and divisions, by residence: 1960

Correlation of Correlation of
Fertility and Fertility and
Females, Ages 15—24 Females, Ages 25-34

Conterminous

United States Urb RNF RF Urb RNF RF

and D1V1Sl°n5 rY8.1 rY8.1 rY8.1 rY9.1 rY9.1 rY9.1

UNITED STATES -.366* -.l28* -.l37* .208* .018* .097*

New England -.l45 .280* -.226 .106 -.054 .074

Middle Atlantic -.l61 .045 —.015 .202* —.091 -.052

E. N. Central —.529* -.185* —.272* .189* .066 .025

W. N. Central —.516* —.257* -.l64* .229* .096* .090*

South Atlantic -.351* —.097* -.071 .186* -.099* .099*

E. S. Central -.454* —.039 -.014 .171* -.103 —.013

W. S. Central -.308* —.052 .049 .346* .137* .034

Mountain —.367* -.115 -.354* .069 -.002 .014

Pacific -.316* -.111 -.210* .259* .048 -.005

*Significantly different from.zero at the .05 level.
Y Cumulative fertility ratio.
X8 Percent ever-married females, ages 15-44, who are

age 15-24.

X9 Percent ever—married females, ages 15-44, who are

age 25-34.

376

 

 

 

.Hm>mH mo. mCu um omen EOHM quHmMMHo mHquoHMHCmHma

ammm.m HoH. HoH. HmH. . . . . .Amxv «mImN mmm mam 0:3 .oolmH mmm
.memEmm muH£3 omHHHmEIHm>m mo quoumm

«mmo.HHI Nom.l oNH. NHH.NI . . . . .Amxv oNImH mmm mam 0:3 .oolmH mmm
.mmHmEmm muHCB omHHHmEIHm>m mo UCmUHmm

Nmm.HI mao.u eaH. mNN.I . . . . . .AHxv emHoHasm mum 0:3 Hanson 2H
. .um>o oCm oH mmm .mmHmEmm muHCS mo pCmoumm

*mmmemlu ooufiel moo. OMOOI e s e e e e e e e e e e e e e e onv hpcgou
How mEooCH HmComumm mHmEmw muHCB CmHomz
some.a mmo. Hoo. moo. . . w . . . .Amxv msooaH HHHsmm mquz amHamz

emon.ol Hmo.I Hog. OON.NI . .A Xv Hm>o oCm mN mmm .mmHmEmm oCm mmHmE
muHCB ha omumHmEoo HOOCUm mo mums» CMHomE

*mHm.aH mmN. maN. eNm.m . . . . . . Amxv amsmnoa cam mnmuoan sums
was 0:3 muuom HOQmH mHmE mpHCB mo “Cmoumm

mmo.: OHo.I NHm. moH.I . . . . . . . Nx mummMCmE Eumm oCm mnmfiumm
new 033 mUCOM HOQMH mHmE mpHCB mo quUHmm
*MNN.HHI HmN.I mmm. Hoo.0HI . . . . . . . . .AHxV mommaHsoe amuHHoaoupmz
amoH.mN .... Hmo.mOH Noo.oooN . . . . . . . . . . . . . . . Sump quumCOU
mmsHm> u mquHo COHHMH>mQ mquHonmmoo mmHQMHum> quoCmmmoCH

ompCmEoo IHmmeOU osmoCmum COHmmmHmmm
mpmm HmHuHmm

ammm. . . . . . . . . . . . . . . . . . COHHMCHEHmqu mHmHnHCS mo quHonmmoo
vNOOMMN s s e s e e e s s e s e e e s e s e e 0 e e s mumgflpmm MO .HOHHm Unggmpm
*Hmme e e e e e e e e e e e e s e s e e e s e pamHUHmmmOU COHuMHmHHOU mam-MUHSE

 

 

IHEHmpCoollhuCCOU mo whom Cones How
Hem CCOQ um>m CmHoHHCU mo HmQECC mCHUCmCHmCH muouomm mo mHmmHmCm esp mo muHCmmm

.aaumH mam

oomH “mmumum ompHCD mCOC
.mmHmEmm muHCB omHHHmEIHm>m ooo.H
.mm THQMB

377

.Hm>mH mo. may we OHoN Eoum “CmumwwHo mHquUHMHCmHms

 

 

 

«omm.N mao. NNH. Nom. . . . .ono emImN mom mum 0:3 .aeumH mom
.mmHmEmm wuHC3 omHuHmEIHm>m mo HCmoumm

*NmH.oHI NHH.u HNH. NNN.H- . . . .Amxo vNumH mom mum 0:3 .eaumH mom
.mmHmEmm muHCB omHunmEIHm>m mo quoemm

eoNo.oHu oNN.: HNH. mmN.Hu . . . . .AHxv omHoHosm mum 0:3 Hanson aH
.Hm>o oCm oH 0mm .memEmm muHCB mo quonm

mONeH' @NOe' moo. 0000' e e e s e e s s s e e e e e AQXV hpcsoo
Com mEOOCH HMCOmumm mHmEmm muHCS CMHomz
emoa.HH eNN. Hoo. moo. . . . . . Amxo msouaH HHHsmH maHaz cmHomz

sSmN.HHI oON.I mmo. ooo.¢l .onv Hm>o oCm mN wmm .mmHmEmm oCm mmHmE
muHCB >9 omumHmEOU Hoonom mo meme» CMHomZ

*HHH.NH oNN. NoH. oom.H . . . . . Amxo aoamHOH one mumCOQmH same
mum 0:3 mouom Hoan mHmE muHCB mo quonm

mHm.I mHo.I mHH. Hmo.l . . . . . ANxv mummmCmE Eumm oCm mHmEHmm
mum 0:3 monom HOQmH mHmE muHCB mo quoumm
imam.oH- «Hm.u moo. mHo.aHu . . . . . . . AHxv mocmcHsoo cmpHHooonumz
aHOH.H¢ .... ¢MH.oS mmH.oMHm . . . . . . . . . . . . . . Eumu quumCOU
mmCHm> u mquHo COHHMH>mo mquHonmmou mmHQMHam> quoCmmmoCH

omusmaoo IHmmmoo osmoCmum COHmmmHmmm
seem HMHuumm

sNHN. . . . . . . . . . . . . . COHHMCHEHmqu mHmHuHCE mo quHUHmmeU
mom.moN . . . . . . . . . . . . . . . . . . . msmEHumm mo nouam osmoCmum
*HNm. . . . . . . . . . . . . . . . . “CmHUHmmmoo COHHMHmHHOU mHmHuHCE

 

 

IlmuCCOU mo sumo EnmmCOCIHmCCC Com
Hem CHOQ Hm>m CmHoHHCo mo anECC mCHUCmCHmCH muouomm mo memHMCm may mo muHCmmm

.aaumH mom

ommH "mmumum omuHCD mCOCHEHmuCoo
.mmHmEmw muHCB omHHHmEICm>m ooo.H

.am mHnme

378

 

 

 

.Hm>mH mo. one on ones Scum pCmHmHMHo mHquonHCmHma

eNoH.m ooo. eNH. amm. . . . .ono amumN 0am 0am 0:3 .eean 0mm
.mmHmEmm muH£3 omHHHmEIHm>m mo quonm

emom.on ooH.I HoH. oom.Hu . . . .Amxo eNumH mom mam 0:3 .aean 00:
.mmHmEmm muHCB omHHHmEIHm>m m0 quoumm

*HmH.mI Noo.: moH. HHo.u . . . . AHxV 00H0H020 0:: 0:3 Hacsoo :H
.Hm>o oCm oH mmm .mmHmEmm muHCB mo quonmm

*mmmem| HfiOe| moo. NNOOI s e e s e e s e e e s e e e wav “fiasco
now mEOUCH HMComHmm mHmEmw muHCB CMHomz
*mmH.OH mNN. Hoo. mHo. . . . . . Amxv mEOUCH SHHEmH muHC3 CMHomz

ammH.NHI mmN.I NHm. mHm.HI .onv Cm>o oCm mN mmm .mmHmEmm oCm mmHmE
muHCB an omumHmEoo Hoonom m0 mammh CMHomZ

emoo.H omH. ooH. meH. . . . . . Amxo amsmnom ocm mumuoan Esme
was 0:3 monom COQmH mHmE muHCB mo quuumm

aoom.N| Hmo.l omo. HmH.I . . . . . ANXV mummmCmE Snow oCm mumfiumm
was 0:3 monom HOQMH mHmE mpHCB Ho quonm
*me.HHI omN.I H¢N.H oom.oHI . . . . . . . AHxV mUCMCHEoo CmuHHomouumz
aNmo.mo .... mmo.mH Hmo.mHmm . . . . . . . . . . . . . . Sump quumCoo
mmsHm> u mquHo CoHumH>mQ mquHonmmoo mmHQmHum> quoCmmmoCH

omquEOU IHmmmoo osmoCmem Conmmummm
mumm HmHuHmm

somH. . . . . . . . . . . . . . . . COHHMCHEamqu mHmHuHCE mo quHUHmmeU
omm.mmm . . . . . . . . . . . . . . . . . . . . . mumEHumm mo nonum osmoCmum
*50fie e e e e e e s e e e e e e e e e s e pamfloflwmmou COHmefimHHOU mHmnflgflsz

 

 

IlhuCCoo mo puma EHmHIHmuCH Cow .ooImH mmm
Com CHOQ Hm>m CmHoHHCU mo HmQECC mCHUCmCHmCH weepomm mo mHmaHMCm one mo mpHCmmm .mm mHQmB

oomH "mmumum omuHCD mCOCHEHmuCOU
.mmHmEmm muHCB omHHHmEIHm>m ooo.H

379

 

.Hm>mH

mo. 0:: 0m 000a gone 0:mu0mmH0 HHuamoHHHcon*

 

 

 

OHm.I mmo.: mmm. mNo.I . . . Amxv omImN mmm one 0:3 .oolmH mmm
.mmHmEmm muHCB omHHHmEIHm>m mo quUHmm

*mHo.NI ooN.: oao. mmH.Hu . . . onv aNImH 00: 0am 0:3 .aaumH 00:
.mmHmEMm muHCS omHHHmEIHm>m mo pCmonm

mHo.u ooo.: mNH. NHo.n . . . . A xv o0H0H0a0 0:: 0:3 Nucsoo 0H
.Hm>o oCm VH mmm .mmHmEmm muHCB mo quUHmm

Nmmol mfiool HNO. m00.l . o 0 0 o . o o o o o o o . AONV hfiCDOU
How mEOUCm HmCOmHmm mHmEmm muHCB CMHomE
*eHN.NI mea.u moo. oHo.u . . . . . A xv msouaH HHHsma 00H:3 :mHomz

NHN. Hmo. mem.H «om. .Aexo 0030 cam mN 00m .mmHmsme ocm mmHms
muHC3 >9 omumHmEoo HOOCUm mo meme» CMHomS

eaoo.m on. mNN.m emo.HH . . . . . Amxo :msmnoa oam mnmnoan sumo
mum 0:3 wUHOm HOQMH mHmE muHCB mo pCmUHmm

:Heo.Nu oNa.: Noo.N Hem.HI . . . . . ANxo mnmomama same on: mumsumm
was 0:3 monom HOQmH mHmE muHCB mo HCmUHmm
moH.HI mNN.: HoH.m ome.en . . . . . . . AHxv mocmaHsoo :muHHooouumz
emHN.N .... mom.mooH mma.mmHm . . . . . . . . . . . . . . sum» panamaoo
mmDHm> u mquHo COHHMH>mQ mquHUmemOU mmHQMHHm> quoCmmmoCH

omusmfioo IHmmmOU osmoCmum COHmmmummm
mumm HMHuumm

*mmve e e e s e e e e s e e e e e e s e e e COHumCHEHmHmQ mnfimflplfigz HO chmfloflwmmoo
owmshNH e e s e s e e e s e e e e e e e e e e s e s e s e wpmgflgmm MO “OH-HM GHMUfiHmum
*5500 s e s e e e s s e s e e s s e e e s e e s s “cmfloflmmmou GOHHMHthOU mflmflulfisz

 

 

BmZIIXHCsoo mo puma Conn: now .oolmH mmm
Hem CCOQ Hm>m CmHoHHCU mo HmQECC mCHUCmCHmCH muonomm mo mHmmHMCm mCu mo muHCmmm .om mHQmB

ommH uConH>Ho oConCm

.mmHmEmm mHHCB UmHHHmEIHm>m ooo.H

380

 

 

 

.Hm>mH mo. esp um oumu Eouw quumHMHo mHquonHCmHma

mmo.l mmo.: Hmm. NmN.I . . . Amxv omImN mmm mum 0:3 .oolmH mmm
.mmHmEmm muHCB omHHHmEICm>m Ho quonm

mom.u eao.u oem. omH.u . . . Amxv aNImH 0a: 00m 0:3 .aeamH 00m
.mmHmEmm muHCS omHuumEIum>m mo quonmm

aao.H emH. mos. mHa. . . . . .Aon o0H0Hosm mam 0:3 Haasoo :H
.Hm>o oCm vH mmm .mmHmEmm muHCB mo UCmUHmm

*ovhenfi' Nmmel mHOO mNo.-I e s e e s e e e e e e e s e onv %“GSOU
How mEooCH HmCOmumm mHmEmm muHCB CMHomz
*Hom.mu Nma.u woo. mHo.u . . . . . Amxo msoocH HHHsma 00H:3 :mHomz

Hoo.HI mmo.l mmm.H Hmm.HI .onv nm>o oCm mN mmm .memem oCm mmHmE
muHCB an omumHmEoo HOOCUm mo mums» CMHowS

«mHm.m mom. Nmm. mNm.N . . . . . Amxv CmEmHom oCm mumHOQwH Snow
mum 0C3 moaom Hoan mHmE muHC3 mo quonm

Hmm. mmo. Nom.H ONm. . . . . . ANXV mummmCmE Snow oCm mumfiamm
mum 0:3 mUCOM COQMH mHmE muHCB mo quuumm
omo.: oNH.I oeH.N moH.NI . . . . . . . .Aon 00cmaHsoo :muHHoaouumz
amom.N .... oNH.mmmH Nmo.ooom . . . . . . . . . . . . . . Sump qunmCOU
mmCHm> u mquHu COHHMH>mo mquHonmmoo mmHQmHum> quoCmmmoCH

omusmsoo IHmmmou osmoCmum ConmmHmmm
mpmm HmHuumm

*HMH. . . . . . . . . . . . . . COHHMCHECmqu mHmHuHCz mo quHonmmoo
mMOCGm O O O 0 O O O O O O O O O O O O O O I mpmglcflpmm MO HOHHm UHmvgmpm
*mmms e s s e s e e e e s e e e e e “cm-WUHHmmOU COflpMI—UmHHOU mHmfluHsz

 

IlmuCCoo mo puma EHmmCOCIHmHCH now
use CCOQ nm>m CmHoHHno mo HmQECC mCHUCmCHwCH muowomm mo mHmeMCm one mo muHCmmm

.aaan 0a:

ommH uCOHmH>Ho CCMHmCm 3oz

.mmHmEmm muHCB omHHHmEIHm>w ooo.H

.mm mHQmB

381

 

 

 

 

 

.Hm>mH mo. are no omen Scum quummmHo mHquoHMHCmHms
mNH.HI HMH.I mom. omo.: . . . .Amxv «MImN mmo mum 0:3 .oolmH mmm
.mmHmEmm wuHsz owHHHmEIHm>m mo quonm
mma.Hu ooH.I Noo. oNe.Hu . . . .Amxo eNan mom 00: 0:3 .aeumH 0mm
.mmHmEmm muHCB omHHHmEIHm>m mo quonmm
eeN.H mom. NHH.H HNH.N . . . . .Aon omsoHosm 0:0 0:3 Huasoo :H
.Hm>o oCm oH mm .mmHmEmm muHCB mo quoumm
*NmmoHl hwvol @mOo mOHoI o o o o o o o o o o o o o o Amxv WUCSOU
Mom mEooCH HMCOmHmm meEmm muHCB CMHomz
somH.Nn Nmm.u mHo. oao.u . w . . . Amxo msooaH HHHsmH 00H:3 :mHooz
NOH. mHo. Hmm.m Hmm. .A xv Hm>o oCm mN mmm .mmHmEmm oCm mmHmE
muHCB an omumeEoo Hoosom mo mums» CmHomz
omm.H mNN. mom. momN.H Amxv CmEmHoH oCm mHmHOQMH Enmm
mum 0:3 mouom HOQMH mHmE wuHCB mo quUHmm
ooh. HNH. NHm. NNo. . . . . . ANXV mummmCmE Eumm oCm mHmEHmm
mum 0:3 monom HOQMH mHmE muHCB mo quoumm
NHH.: Nmo.- Noe.m ooa.Hu . . . . . . . Aon 00cquaoo :muHHooonumz
*Nmm.m .... mmm.0HHH oom.NNmm . . . . . . . . . . . . . . Esme pCmmeOU
mmCHm> u mquHo COHHMH>mo mHCmHUHmmmou mmHQMHHm> quoCmmmoCH
omnnmfioo IHmmmoo osmoCmum ConmmHmmm
mumm HmHuumm
«mmo. . . . . . . . . . . . . . . . COHumCHEHmCmo mHmHuHCE mo quHUHmmmoo
hmOemmN e s e e o e e e o e e s s e e e e e e s e mumgflumm mo HOHHM UHqumpm
*wN-Oe e e e e e e s e e e e e e e e e e e “GmHUHmmmOU COHNMHmHHOO mI—HQHUHSE
ommH ”ConH>Ho oCmHmCm 3oz
IlmuCCoo mo puma EummleHCH How .ooImH mmm .mmHmEmm mpHCB omHHHmalum>m ooo.H

Hem Cuon Hm>m CmHoHHCo mo HmQECC mCHUCmCHmCH mnouomm mo mHthow mnu mo muHCmmm

.mm mHQmB

 

382

 

 

 

 

.Hm>mH mo. may no oumn Eoum quHCMMHo NHHCMUHMHCmHms

mNm. omo. moo. omo. . . . .Amxv vamN mmm mum 0:3 .ooImH mmm
.mmHmEmm muHCB omHHHmEIum>m mo quoumm

*Hoo.HI omH.u mos. ooH.u . . . .Amxo eNImH mom mum 0:3 .aaImH 0am
.mmHmEmm muHCB omHuumEIHm>m mo quUumm

Nom. ooo. Hoe. oNN. . . . . .Aon omsoHosm mum 0:3 Huasoo :H
.nm>o oCm oH mmm .mmHmEmm muHCB mo quonm

*amH.mu ooa.u HHo. aeo.u . . . . . . . . . . . . . . .ono N::000
How mEOUCH HMCOmHmQ meEmm muHCB CMHomZ
ammo.MI mmN.: moo. Hmo.l . . . . . .Amxo mEOUCH XHHEmm muHCB CmHomz

mHN. oHo. mmm.H mmN. .onv Hm>o oCm mN mmm .mmHmEmm oCm mmHmE
mpHCS >9 omumHmEoo Hoonom mo meme» CMHomz

mom.- HNo.- NHH.N eao.n . . . . . Amxo casmnoa ocm mnmuoamH 50mm
was 0:3 mUHOH COQMH mHmE muHCB mo DCmoumm

omH. NHo. mmH.N ooo. . . . . . ANXV mummMCmE Eumm oCm mHmEHmm
mum 0C3 mUHOM HOQmH mHmE muHCB mo quUHmm
*HmN.mu mHe.u oNo.H HNH.oHI . . . . . . . .Aon 00cm:Hsoo :00HH000u002
aHmo.o .... oNo.mmm NmN.oHNo . . . . . . . . . . . . . . Esme quumCOU
mmCHm> u mpCmHo CoHumH>mo mquHUHmmmoo mmHQmHum> quoCmmmoCH

ompsmaou Imemou UnmoCmpm ConmmHmmm
mumm HmHuumm

*mHm. . . . . . . . . . . . . . . . CoHumCHEHmqu mHQHuHCZ mo UCmHUHmmmoo
NmfieOMH e e e e e s e e o e s e e s e e s e e e e mUMEHumm mo HOHHm ©HMUGM“W
*NHO. . 0 0 o . . . o o o o o . o o o o . UCGflUHWMwOU COH#MHOHHOU OHQHHHSZ

 

 

mHoonIImuCCoo mo puma Cones Cow .ooImH mmm
Hem CHOQ Cm>m CmHoHHCo mo HmQECC mCHUCmCHwCH muouowm mo mHmemCm was mo muHCmmm .mm mHQmB

ooNH HSEC/Ho 0H0:mH0:

.mmHmEmw muHCB omHHHmEIHm>m ooo.H

383

 

 

 

.Hm>mH mo. mCu um omen Scum quummmHo mHquoHMHCmHma

eom.u mNo.: omo. HNH.: . . . .ono amImN mew mew 0:3 .aeumH 0mm
.mmHmEmw muHCS omHHHmEIHm>m mo quonm

oHN.- mNo.: oom. oaH.I . . . .ono «NImH mom on: 0:3 .aaumH 0o:
.mmHmEmm muHCB omHHHmEIHm>m mo quonm

wooN.NI mmN.- Ham. mmN.HI . . . . .AHxv omsoHasm 00: 0:3 Huasoo :H
.Hm>0 oCm «H mmm .mmHmEmm muHCB mo quUHmm

HmHe' NNOOI Mun—”Os MOOOII e s e e e e e o s e e e e s e Amxv ”HQSOU
now mEooCH HMComeQ meEmm wuHCB CMHomE
Hmo.u oao.u oHo. NHo.I . . . . . .Amxo msooaH HHHsma m0H:3 :mHoms

Hmo.H omo. OHm.H moH.H .onv Hm>o oCm mN mmm .mmHmEmm oCm mmHmE
muHCB an omumHmEoo HOOCUm mo memo» CMHomE

amHm.H mHN. mmm.H oom.N . . . . . Amxv CmEmHom oCm mHmHOQMH Eumm
was 0:3 muuom HOQMH mHmE muHC3 mo pCmonm

Nmm.l moo.l mom.N omo.: . . . . . ANxV mummMCmE Snow oCm mumEHmm
. was 053 moaom HOQMH mHmE muHCB mo quuamm
«mHo.mI on.I ONo.N moo.oHI . . . . . . . .AHXV moCmCHEoo CmpHHomouumz
amom.m .... Hmm.NHm omh.mmmm . . . . . . . . . . . . . . Esme quumCOU
mmsHm> u mquHo CoHDMH>mo mquHonmmoo mmHQMHHm> quoCmmmoCH

omquEOU IHmmmou osmoCmum Conmmammm
mumm HMHpHmm

*mmm. s e e e s e e o o e s s s s e e GOHHMH‘HHEHmflmg mflmncfl “H52 “0 “cmlfiolfimmmoo
mmm. flmlfi O O O 0 O O O O O O O O 0 O O O O O O O O 0 mflmElofi“mm mo HOHHW Unmwgmum
swam. . . . . . . . . . . . . . . . o . . . UCOHUHNMOOU H.HO._H...Hm._HOHHOU OHQHUHSE

 

 

IlmuCCOU mo puma EummCOCICmHCH How .ooImH mmm
Hem Cuon um>m CmuoHHCo mo CmQECC mCHUCmCHHCH mnouomm mo mHmmHMCm mCu mo muHCmmm .om mHQme

oooH ":0HmH3Ho 0HucmH0< 0Hoon

.mmHmEmm GHHCB omHHHmEIHm>m ooo.H

 

—_—=—.———..H F

 

 

 

 

 

 

 

 

.Hm>mH mo. one on ones Eouw quHmmmHo mHquoHMHCmHma
Hmo.HI mmH.I moo. oom.l . . . .Amxv omImN mmm was 0:3 .oolmH mmm
.mmHmEmm muHCB omHHHmEIHm>m mo quUHmm
HNm.Hu emH.I Nmo. moN.Hu . . . .Amxv eNImH 0mm 000 0:3 .eean 0mm
.mmHmEmm muH£3 omHHHmEIHm>m mo quoumm
ooo. meH. ooN.H HaH.H . . . . .AHxv omsoHoam 00: 0:3 N:::00 :H
.Hm>o oCm oH mmm .mmHmEmm muHCB mo quUHmm
Ammo-HI- mvNe' omo. vael 0 e e s e e e e e e 0 e e e s AOXV “pcsoo
Com mEooCH HmComHmm mHmEmm ouHCB CmHomz
mHo.I Noo.- mmo. Hoo.u . w . . . .Amxo msoocH HHHsma 00H:3 :mHomz
oem.u NHo.I omN.m omH.NI .A xv 0030 ac: mN 0mm .mmHmsmm ocm mad...a
mpHCB an omumHmEoo Hoonom mo mums» CMHomS
NQNo.N HmN. ooo. oom.H . . . . . Amxv CmEmuom oCm mumHOQMH Show
4 was 0:3 mouom COQmH mHmE muHCB mo quoamm
8 mNm.HI HoH.I mmo. mmo.: . . . . . ANxv mammMCmE Eamm oCm mumfiumm
3 mam 0:3 monom HOQMH mHmE muHCB Ho quunmm
sooh.ol HH¢.I mNo.m mmm.mNI . . . . . . . .AHXV mUCmCHEoo CmuHHomouumz
ammH.m .... omH.omo mmo.mmHm . . . . . . . . . . . . . . Esme quumCoo
mmCHm> u mquHo mCOHumH>mo mquHonmmoo mmHQMHHm> quoCmmmoCH
omusosoo IHmmmoo osmoCmum COHmmmHmmm
mumm HmHuHmm
*0mm. 0 e e o e s e o e e s e e s e e e COHflvMCHEHmme mHQHquz MO “cmfloflmwmoo
ommOmom O O 0 O O O O O O O O O O O 0 0 O O D O O O l mflmaflpmm MO .HOHHm UHmwfiHmpm
*mhve e e e e e s e e e e e e s e s e e e e e “GmHUHmmmOU COHpmHmHHOU mHmeHsz
ommH uCOHmH>Ho UHquHu< mHoon
Ilmquoo No news EnmmleuCH now .ooImH mmm .mmHmEmm mnHCB omHuumEInm>m ooo.H

Hem CHOQ Hm>m CmHoHHCU mo HmQECC mCHUCmCHwCH muouumw mo mHmmHMCm ens mo muHCmmm .Ho mHQMB

385

 

.Hm>wH

mo. one no omen EOHH quHmMMHo XHHCMUHMHCmHms

 

 

 

«om. Mao. mNm. HHm. . . . . .ono emImN mew mew 0:3 .aean 00m
.mmHmEmm muHCB omHuHmEIHm>m mo quUHmm

aHoH.mI OHo.I HmN. Nmm.NI . . . . .Amxv oNImH mmm one 0:3 .oolmH mmm
.mmHmEmm muHCB omHuHmEIHm>m mo quonmm

woe. oNo. mom. oeH. . . . . . .Aon omsoHosm 0:: 0:3 Haasoo :H
.um>o oCm vH mm .mmHmew wuHCB mo quUHmm

*mmOeMI mmIfie| qHOe NfiOel e s s e e e e o e e s s e e e e Amxv >HQDOU
Com mEooCH HMComem mHmEmm muHCB CMHomE
aomm.o OON. moo. mHo. . . . . . . .Amxv mEooCH XHHEmm muHCS CMHomz

amom.m NoH. mNH.H mNo.m . .onv um>o oCm mN mmm .mmHmEmm oCm mmHmE
mHHCS an omumHmEoo Hoocom mo mums» CmHomE

NHH.H: mmo.- omm.H mNH.H- . . . . . . Amxo :mamuom oam m0000:MH same
was 0:3 mouom COQMH mHmE muHCB mo quoumm

mom.l oNo.I mHm.H mom.l . . . . . . ANxv mammMCmE Snow oCm mumaumm
mum 0C3 muaom HOQmH mHmE wuHCB Ho quUumm
*emo.o- ooN.- HNo.H HNo.oH- . . . . . . . . .Aon 00:m:Hsoo :muHHooosumz
*NmN.N .... mNm.HOH mmo.ommH . . . . . . . . . . . . . . . Sump quumCOU
mmCHm> u mquHo COHumH>mQ mquHonmmoo mmHQmHnm> quoCmmmoCH

omquEoo IHmmmou osmoCmum Conmmummm
mumm HMHuumm

*OHo. . . . . . . . . . . . . . . . . . . COHHMCHEumqu mHmHuHCE mo quHUmemOU
mmOOth 0 0 O O O O O O O O O O O I O O O O O I O O O O mumgflumm MO .HOHHm GHMGGmum
*Oflme s e e e e e e e e e e s e e s e s e s s e “cmfloflmmmoo COHumHmHHOU mHQHpHsz

 

 

oomH ":0HmH3Ho Hmnuamu :0002

ummmllmpCCOU mo puma Cones How .oolmH mmm .mmHmEmm muHCB omHuHmEIHm>m ooo.H
Com CCOQ um>m CmHoHHCU mo HmQECC mCHUCmCHwCH mnouumw m0 mHmmHMCm mCu Ho muHCmmm .Nm mHQmB

386

 

 

 

.Hm>wH mo. msu um omen Scum quHmmeo .mHquoHMHCmHmin

ova.H moo. oHN. Noe. . . . . .Amxo amImN mom mum 0:3 .aeImH 00m
.mmHmEmm muHCB omHHumfilum>m mo quUHmm

imam.mu emH.I oHN. Hmm.u . . . . .ono «NImH 00m 00m 0:3 .eaumH mom
.mmHmEmm muHCB omHHHmEIHm>m mo quUHmm

HoH.H oHo. mHm. mom. . . . . . .AHxV omHOHmEm was 0:3 muCCoo CH
.Hm>o oCm oH mm .mmHmEmm wHHCB mo quUHmm

*PNHOMI hmI—Hel MHOe HflOs' e e e e e e e e e e C e s e e s wav mpgsoo
mom wEooCH HMComumm mHmEmm muHCB CMHomz
*mNm.mI oNN.: eHo. oHo.I . . w . . . .Amxo msooaH HHHsmH 00H:3 :mHomz

emoH.mI oHH.I HoN.H oam.mu . .A xv 0030 ea: mN 00: .mme50H 00m mmHma
muHCB an omumHmEoo Hoosom mo meme» CMHomE

mmN.H mmo. mHm. OHH. Amxv CmEmHOM oCm mumHOQmH Emmm
mum 0C3 monom HOQmH mHmE mpHCS mo quonmm

amoH.oI mMN.I oom. omN.NI . . . . . . ANxV mummmCmE Eumm on mumfiumm
mum 0C3 moHOM HOQmH mHmE muHCB mo quonmm
eoHe.oI mmN.: omo.H meH.NHI . . . . . . . . .AHxv 00::cHaoo :muHHoaouumz
amHm.NH .... mNm.oom mom.HooH . . . . . . . . . . . . . . . and» pCmumCoo
mmCHm> u mquHo COHumH>mQ qumHonmmoo mmHQMHHm> quoCmmmoCH

omusmaoo IHmmmou UnmoCmum COHmmmammm
mumm HMHuHmm

stN. . . . . . . . . . . . . . . . . COHmeHEnmumo mHmHuHCE mo quHonmmoo
“NIH O mmlfi O O O O O O O 0 0 O O O O O O O . O O O O O O muggioflpmm mo MOM-Hm UHMUQmpm
aHmm. . . . . . . . . . . . . . . . . . . . pCmHUHmmmoo COHumHmHHOU mHmHuHCE

 

 

IlmuCCoo Mo puma EHMHCOCIHMHCH How .ooImH wmm

Hem CHOQ Hm>m CmHoHHCU mo HmQECC mCHUCmCHmCH muouomm mo mHmmHmCm was mo muHCmmm

oomH “COHmH>Ho HmnquU Cuuoz ummm
.mmHmEmm manz omHHHmE Hm>m ooo.H

.mo UHQMB

387

.Hm>mH mo. mCu um ones 800w quHmHMHo mHquonHCmHma

 

 

 

Hom. 0N0. Hmm. moN. . . . .ono emumN 000 000 0:3 .aeumH 000
.mmHmEmm mUHCB omHHHmEIHm>m mo quUHmm

0000.0- mNN.: 000. mom.Nu . . . .ono «NImH 000 000 0:3 .eaan 000
.mmHmEmm muHCB omHuumfilum>m mo pCmoumm

0oma.m moN. Ham. N00.H . . . . .Aon 00H0Hos0 000 0:3 H::300 :H
. .Hm>0 oCm oH mm .mmHmEmm 00HC® mo quUHmm

3000.NI 00H.I NNo. ooo.: . . . . . . . . . . . . . . .A xv H0::00
00m mEOUCH HMCOmHmm mHmEmm muHCB CMHomz
00NH.NI ooo.: «No. Noo.- . w . . . .Amxo 0:00:H HHHs00 000:3 :00002

0Nom.ou 0Hm.u 0H0.H Nmo.oHn .A xv 0030 0:0 0N 000 .00H0s00 0:0 00H0:
muHflz >9 ompmHmfioo HOOCUm m0 mums» CmHomE

000H.¢ 00H. 000. oem.H . . . . . Amxo :0s0000 0:0 00000:0H :00:
mum 033 monow HOQMH mHmE muHCB mo quUHmm

ammm.mu omN.u 00H. HHo.: . . . . . ANxv 00000:0s 5000 0:0 0005000
000 0:3 mouom HOQMH mHmE muHCB mo quonm
eaNm.mu mmN.: 000.m 00H.0HI . . . . . . . .Aon 00:0:0500 :000H000000s
00H0.NH .... mmH.0oa HH0.mHmo . . . . . . . . . . . . . . 500: 0:000:00
mmCHm> u mquHU COHHMH>mQ mquHUHmmmoo mmHQMHHm> quoCmmmoCH

omesmfiou IHmmmoo oumoCmum COHmmmHmmm
mumm HMHuumm

soom. . . . . . . . . . . . . . . . . . COHDMCHEHmme.mHmHuHCZ mo quHonmmoo
mnfimemom e e e e e e e e e e e 0 e e 0 e e 0 e e e e e mymgflumm mo .HOHaHm UHmcflHmpm
*0000 e e s e e e e e s e e e e e e e s e e e “cmnfloflmmmoo COHUMHmHHOU mHmeHsz

 

 

IlhuCCOU mo 000m EHmMIHmCCH 00m .ooImH 0mm
00m C009 0m>m CwaoHHCo mo HmQECC mCHoCmCHmCH meouomm mo mHmmHMCm 0C0 mo muHCmmm .om mHQme

ommH "ConH>Ho HmnquU Cuuoz ummm
.mmHmEmm mans omHHHmE um>m ooo.H

388

 

 

.H0>0H

mo. 03“ #6 OHMN EOHM HQTHQMHHU \wHUGMUHNHCmHmum

 

 

 

ooo. Hmo. Hmm. mmN. . . . . .Amxv 0mImN 0mm 000 0:3 .oolmH 0mm
.mmHmamm 00HC3 00H00mfil00>m 00 0C000mm

0NN0.0- mo0.- 00N. omm.N- . . . . .ono 0NImH 000 000 0:3 .00ImH 000
.nonamm 00HCB ©0H0005I00>0 00 0C000mm

000H.N omH. 00m. mom. . . . . . .Aon 0000Has0 000 0:3 00:00: :0
.00>0 0C0 0H 0mm .mmHmamm muHCB mo 0C0000m

0Hmm.0u HmN.u mHo. 000.: . . . . . . . . . . . . . . . .ono 00:000
000 0800CH H0C000mm 0Hmamm 00HC3 CmHomz
0mmm.0n ooN.: moo. NNo.- . . . . . . Anxo 0:00:H HHHs00 000:3 :00002

:Nmm.HI mmo.I HmH.H mNm.NI . .A0xv 0030 0C0 mN 0mm .mmHmamw 0C0 mmHma
00HCB >9 omumHQEOU Hoosom Mo 0000% CmHomz

H00. NNo. o0o.H H00. . . . . . . Amxo :050000 0:0 00000:0H E00
000 053 00000 0030H 0Hma 00HC3 00 0C000mm

NHm.I mHo.I mNH. HHN.I . . . . . . ANxv m00m0CmE E000 oCm 000E000
000 0:3 00000 0090H 0H0E 00HCB 00 0C000mm
000m.mn ooN.- on.N NN0.NH- . . . . . . . . .Aon 00:0:0500 :000H0000002
amoN.o .... Nmm.mmo mom.NmN0 . . . . . . . . . . . . . . . E000 0C000C00
mmCH0> 0 mquHo C0H00H>0o mnCmHonmeou 00HQ0H00> 0C00C0000CH

ompsmaoo IHmmmoo o0moCoum COHmmm0m0m
000m H0H0000

*HNo. . . . . . . . . . . . . . . . . . . . COHumCHE0wumo 0HQH0H52 00 quHonmeoo
.NNmOQON O O O O O O O O O O O O O C O O O O O O O O I C O mpMEIoflpmm MO HOHHM “Haggmum
*mfloe e e e e e e e e e e o e e e s e e s e e s e “cmfloflmmmou COHpmufimHHOU mflmfluHsz

 

 

ummzllmquoo 00 0000 C090: 000 .oolmH 0mm
000 C009 00>0 C00©HHC0 mo 00085C mCHUCmsHmCH m0ouomm mo mHmmHMCm 0:0 00 muHCmmm .mm 0HQMB

oomH "COHmH>Ho H000C00 £0002
.mmHmemm 00HC3 o0H00mEI00>0 ooo.H

 

389

 

 

 

.H0>0H mo. 090 00 000m 800m 0C00000Ho hHquuHMHCmHms

HNN.H 00o. mmN. mmN. . . . . .Amxo 0mumN 000 000 0:3 .00ImH 000
.00H080m 00H93 o0H0008I00>0 mo 0C0000m

0mHH.mI omN.I NoN. mHm.H- . . . . .wav 0Nan 000 000 0:3 .00umH 000
.00H080m 00H93 o0H0008I00>0 mo 0C0000m

0m0.H Noo. ooN. moo. . . . . . . . ASXV 00%0HQE0 000 093 muCCOU
CH .00>0 0C0 0H 0m .0090600 00H93 mo 0C0000m

mmoefil thel wnfiOe MNOOI e e s e e e e e s e e e e e e s onv hugsoo
000 0EOUCH H0C0000m 0H080m 00H93 C0Ho0z
amNo.mu 00H.: HHo. omo.: . . . . . . .Amxo 0:00:H HHHs00 000:3 :00002

:HN0.oI mmN.: mmo.H omm.oI . .onv 00>0 oCm mN 0mm .00H080H oCm 00H08
00H93 >9 U000HQEOU Hoonom 00 0000» C0H©0z

oom. omo. 00m. 0Hm. . . . . . . Amxo :0s0000 0:0 00000:0H E00.0
000 093 00000 0090H 0HmE 00H93 00 0C0000m

ammm.HI Hmo.l mmN. mm0.I . . . . . . ANxv 000m0Cmfi E000 0C0 0008000
000 093 00000 0090H 0H08 00H93 00 uC0000m
0Hmm.mu 0HH.I H00.N 0NH.mu . . . . . . . . .Aon 00:0:0000 :000H000000z
*mHm.HH .... NNH.om0 mmN.omHm . . . . . . . . . . . . . . . E000 quumCoo
m0CHm> u 00C0H0 C0H00H>0o 00C0H0me000 00H90H00> 0C0oC0m0oCH

00050800 IHmm0OU o0moC00m C0H0000m0m
000m H0H00mm

*HOHO e e e s s s e e e s e s e e e O s s GOHuMfiHHEHmUmQ mHQHHng MO “GmHUIoflmmmOU
q0m0m¢N O O O O O O O O O O O O O O O O O O O C I O O O mpmpHHuflmm MO HOHHm GHmwcmpm
*hm¢e e e e e e e e e e e e e e e e e e e s e e “cmnflU-flwmmou HMO-WNWMHmHHOU GHQHpHSE

 

 

oomH "C0H0H>Ho H000C0U 99002 0003

IlhuCCOO 00 00mm E000C0CIH00C0 000 .oolmH 0mm
00m C009 00>0 C00oHHCU 00 009ECC mCH0C0CH0CH 0000000 00 mHmmHmCm 090 00 00HCm0m .mm 0H9ma

.00H080w 00H93 o0H00mEI00>0 ooo.H

390

 

 

.00>00 mo. 000 00 0000 E000 000000000 00000000000000n

 

 

 

0mm. moo. mom. mmo. . . . . .0000 00:00 000 000 003 .00100 000
.0000E00 00003 000000E|00>0 00 0000000

0000.0- 000.: 0mm. 000.0: . . . . .0000 0Nnm0 000 000 on; .00um0 000
‘0000000 00003 000000EI00>0 00 0000000

000.0 moo. 00m. N00. . . . . . .0000 00000000 000 on; 000000 :0
.00>o 000 00 00 .0000E00 00003 00 0000000

*HOMOM' omHol mHOo mm00| o o o o o o o o o o o o o o o o onv ”UGSOU
000 0E000A 00000000 000E00 00003 000002
000.0- 000.- 000. omo.: . . w . . . .0000 000000 000000 00003 000000

0000.001 00m.l 000.0 000.00: . .A xv 00>0 000 mm 000 .0000E00 000 0000E
00003 00 000000E00 000000 00 00000 000002

0m00.0 00m. 000. 000.0 . . . . . . Amxv 0000000 000 00000000 0000
000 003 00000 00000 000E 00003 00 0000000

00m.| mmo.l 000. 000.: . . . . . . Amxv 0000000E E000 000 000E000
000 003 00000 00000 000E 00003 00 0000000
0m0m.mu 000.: mm0.m 000.00: . . . . . . . . .0000 0000:0000 000000000002
00m0.00 .... 00m.mmm m00.m000 . . . . . . . . . . . . . . . E000 00000000
00000> 0 000000 000000>00 000000000000 00000000> 00000000000

00000E00 I000000 00000000 0000000000
0000 0000000

*mmm. . o o o . . o . o o . . . o o o . o o . COHHMCHEHQfiGQ mHQH#H52 MO UCGHUHMMGOU
“wooflom o o o o o o o o o 0 o o o o o o I o o o o o O o o o mumgflumm MO HOHHm UHmwcmum
0‘5me 0 o o 0 o o 0 o o o o o o o o O o I o I o O O UHH®HHUUHMH$OU QOHUMHmHHOU ml—Hmflpnfigz

 

 

0000 "00000>00 0000000 00002 0003
1:000:00 00 0000 E000I00000 000 .001m0 000 .0000E00 00003 000000EI00>0 000.0
000 0000 00>0 00000000 00 000E00 00000000000 0000000 00 00000000 000 00 0000000 .00 00008

 

391

 

.Hm>ma mo. map

um OHmN Eonm ucmanMflU haucmoflwflcmflm«

 

 

 

 

 

*Hmm.m moa. Hmm. 0mm. . . . .Amxv «mumm mom mHm 0:3 ~¢¢nma mom
.mmHmEmm muߣ3 UmHuHmEIHm>m m0 Humoumm

*mmo.hu mNm.- mma. mom.au . . . .Amxv «Numa mam mum 0:3 .wvuma mmm
.mmHmEmm muHQB cwHHHmEIHm>m mo ucmuumm

*mo>.al moa.l mma. mam.| . . . . .Anxv wmonmEm mum 0:3 mucsoo CH
~uw>o 6cm @H mm .mmamamm mufln3 mo unmoumm

*hamom' mmnfio' ooo. PHOOI o o O o o o o o o o o o o O o wav ”HGSOU
How mEooca HMCOmumm mHmEmm muHSB amflvmz
*mm¢.¢u ooN.: woo. mao.u . . . . . .Amxv maoocfl mafiamm muflnz cmflcmz

«mo¢.hl Ham.l con. mma.ml .Afixv um>0 cam mm mom ~mmHmEmm cam memE
muHSB an Umumamaoo Hoonum mo mummh mafiwmz

mmo.a mwo. an». mmm. . . . . . Amxv cwEmnom mum mumuonma Eumm
mum 0:3 mouom Henna mama wuflnz mo unmoumm

oom.a «we. 0mm. ¢MH.H . . . . . Amxv mummmnma Eumm cam mumsumm
mum 0:3 muuom Honma mHmE muHLB mo unmoumm
*omn.au Noa.u noo.a www.mu . . . . . . . .Aaxv mocmcHEOw cmuflaomouumz
*hv©.m .... mah.mon nmm.hoov . . . . . . . . . . . . . . Sump ucmumcou
mmsam> u mucmflo COHuMH>mQ mucmHUHmmmoo mmanmflnm> ucmwcmmmccH

Umusmfiou Iflmmmou Unmwcmum coflmmmumwm
mumm Hmflunmm
*hmm. . . . . . . . . . . . . . . coflumcflaumumo mamfluasz mo ucmHUHmmmou
“whommnfi o o o o o O o o o o o 0 o o o o O o o 0 mumEflumm mo .HOHHm @Hmmvcmum
*omoo o o o o o o o o o o o o o o o o o UCQHUH—uwmmoo COHHMHmHHOU mHm-flpnfigz
coma “conH>HU oaucmaufi
Susomllmucsoo mo puma cmnus How .fiwlma mom ~mmam8mw mpflnz Umfluumfilum>m ooo.H

Hmm GHOQ Hm>m cmuwaflno mo Hmnfisc mcflocmdamcfl muouoww mo mflmhamcm mnu mo muasmmm

.mm magma

392

11354.
. .

.Hm>ma mo. ms“ um oumN Scum ucmeMMHG hauchflMHcmHm¥

 

 

woa.u woo.u mom. mmo.- . . . . .Amxv «mumm mom mum 0:3 .kuma mom
.memEmw mufl£3 Umfluumfilum>m mo ucmuuwm

¥mmm.ou H¢N.u Hum. mo>.Hu . . . . .Amxv wmumfl mom mum 0:3 ~¢¢nmfl mom
~mmamamm muHLB Umfluumfilum>m m0 unmoumm

¥mm>.ml mam.| mNN. Hmm.HI . . . . . .Anxv UmmoHQEm mum 0:3 mundoo CH
.um>o cam va mm .mmHmEmm muflLB mo unmoumm

mNo.-HI omOol 0H0. OHOo-I o o o o o o o o o o o o o o o o onv huasou
How mEouca anaemmmm wamfimm muflnz cmflwmz
mmN. mac. ooo. moo. . . w . . . .Amxv macoafl maflamm ouflzz amflwmz

*oom.ml nmm.l com. www.ml . .A xv Hm>0 Gum mm mmm ~mmHmEmm Ucm mmHmE
muH£3 an Umumamfiou Hoonom mo mummm amHUmz

*mqm.~ MHH. ohm. mam. . . . . . . Amxv swamu0m cam mumuonma gum“
mum 0:3 mouom Honma mama muߣ3 mo unmoumm

mmo.: moo.l mmN. ©HO.I . . . . . . Amxv mummwzmfi Eumm cam mHmEHmw
mum 0:3 wouom Honma mama mufln3 mo pcmoumm
*Hmh.ml Ham.- mom.m Hmm.¢au . . . . . . . . .Aaxv mocmcHEoc cmuflaomouumz
¥mmm.h .... www.mam moo.mmmm . . . . . . . . . . . . . . . Emmy ucmumcou
mmsHm> u mucmwo COHMMH>mQ mucmaoflmmmou mmHQMHHm> pcmwcmmmvcH

CmudeOU lawmmoo Cumccmum COHmmmummm
mumm Hmfluumm

*mom. 0 o I o o o o o o o o o o o o o COHgmGHEHGme mufimflpnfisz MO “Gmnfioflmmmou
,mPMO HMN o o o o o o o o o o o o o o o o o o I o o o mgmgflumm MO .HOHHm Cngamum
*moo. O O O I O O O O O O O I O O O C O O O “cmlofioloflmwwou QOflpmHmHHOU mHQIoHUHSS

 

 

luquDOU mo puma Enmmcoclamusu MOM ~v¢ImH mmm
Hmm CHOQ Hm>m cmHUHHSU mo umnfidc mafiuzmsamcfl muouomm mo mflmwamcm may mo muHSmmm .mm magma

coma "coflmfl>fln oflucmapm nusom
.mmHMEmM O‘WHSZ» UmflHHMEImewm 0007—”

HUS CHOQ H®>W COHUHHSU MO HwflESC OCHUCQSfiMCH mHOUUQW MO QHWN&GCQ @flu MO MUHDMOE .Qh GHQQE

 

 

.Hm>ma mo. mnu um oumu Scum ucmeMMflw waucmoHMHcmHm*

 

 

 

mNm.H who. 5mm. haw. . . . . .Amxv wmumm mom mum 0:3 .¢¢uma mom
.mmHmEmm mnHSB CmHHHmEIHm>m mo ucmonmm
«mmm.mn moa.u aqm. Non.| . . . . .Amxv wmuma mmm mum 0:3 ~¢¢nma mom
‘mmamamm mufl£3 Umfluumfilum>m mo unmoumm
mao.al ooa.u mwm. Nmm.- . . . . . .Anxv ammoamsm mgm 0:3 mucsoo afl
~Hm>o mam wa mom ~mmHmEmm muHSB mo pcmoumm
mN-mol mMOol mHOo 0000' o o o o o 0 0 o o o O o O o o o Amxv mflcsou
How mEoocA HMQOmHmm mHmEmm muH£3 amawmz
*wmm.a NHH. moo. mac. . . . . . . .Amxv mEooafl maflamm mpflnz madam:
*mnn.m| mmN.: mmo.a mom.ml . .A¢XV Hm>o cam mm mmm ~memEmm 0cm mmHmE
mgߣ3 >9 wmumamaoo Hoonum mo mummm Gwavmz
*mnn.m NNH. mmN. mmn. . . . . . . Amxv swamu0m can mumuonma gumm
3 mum 0:3 mouom Momma mama muwfiB mo ucmuumm
9 «mom.¢l mam.l mva. mon.| . . . . . . Amxv mummmame Eumm Usm mHmEumw
3 mum 033 mouom Honma mama muH£3 mo unmoumm
*mmo.ml cam.u Hoh.m www.mau . . . . . . . . .Aaxv moamcHEow amuflaomouumz
*Hwh.m .... vmv.anm No¢.mmmm . . . . . . . . . . . . . . . Sump ucmumcoo
mmsam> u mucmflo coaumfl>mm mucmfloflmwmoo mwanmflum> ucmwcmmmwcH
mmusmaoo Iflmmmou Unmwcmum scammmummm
mumm HMflpumm
*NONI 0 o 0 o o o o o o o o o o o o o o o o o COflgmgHEHmqu mnfimflunflsz MO pamncfinvnofimmmoo
HN®OHmm O O O O O O I O O O O O O O .0 O O O O O O O O O O O mumEI-Humm mo “OH-Hm Ugmwgmpm
*m.v¢0 o o o o o 0 o 0 o I o o o o o o o o o O o o o “cmHUHmmwou COHUMHmHHOU mHmflgflsz

 

 

coma "cowmfl>flw Uflucmaum nusom
Ilmgcsoo mo #Hmm EHmMIHMHSH Mom ~vVImH mmm ~mmHmEmm muH£3 Umfluumalum>m ooo.H
umm QHOQ Hm>m cmuwaflno mo Hmnfisc mGHUGmSHmQfl muouomw Mo mamhamcm mflu m0 muasmmm .om GHQMB

 

394

 

.am>ma mc. mnu um OHmN Eoum pcmummmao maucmoamacmam*

 

 

 

 

moo. omo. mmm. oom. . . . . .onv omumm mom mum 0:3 .oouma mom
.mmamamm mua£3 UmaHHmEIHm>m wo ucmuumm

*ooo.ou Noo.- mom. ooo.au . . . . .onv omuma mom mum 0:3 .vouma mom
.mmamEmm muanz Umauumfilum>m mo uamuumm

*ovm.mu oom.u 5mm. wom.au . . . . . .onv om>0aoam mum 0:3 mucsou ca
.Hm>o cam wa mmm .mmamamm muasz mo ucmonmm

mQMOn—H HHHO 0H0. NNo. o o o O O I o o o I o O o O O o onv mpgsoo
How mEooca amQOmumm mamEmm muanz cmaomz
oao. woo. moo. ooo. . . w . . . .Amxv maooca maaamm moanz cmaomz

*noo.mu ooa.u oma.a ooo.m- . .A xv um>o cam mm mom .mmamEmm ocm mmama
mua£3 >9 Umumamaou aoosom mo mummm amaomz

oama.mn nma.u oom.a moo.mu . . . . . . Amxo swamuom oam mumuonma ammo
mum 0:3 mouom Honma mama mua£3 mo unmommm

amm.l mac.l wmm. cam.l . . . . . . Amxv mummmcmfi Eumm ocm mumfinmm
mum 0:3 mUHOM Honma mama man3 m0 udmoumm
*Noo.au maa.u vmm.m oom.ou . . . . . . . . .Aaxo muomaasoo cmuaaooouumz
*mmmom 0000 mmooomm NOhomomN o 0 o o o o 0 o o I o O o o o EHmp “cmpmcoo
mmSam> u mucmao coauma>mm mpcmaoammmou mmanaHm> ucmocmmmoca

omusmfioo Iammmou oumocmum Goammmummm
mumm amapumm
*avm. . . . . . . . . . . . . . . coaumcaEHmqu mamauasz mo usmauawmmou
Nmmoomafi o o o o O o o o o o o o o o o o o a 0 o mumgflpmm mo .HOHHM Gumccmpw
*wme o o o o o o o o o o o o o o o o mpCQHUHmmwOU COH“MH$HHOU m-HAWHUHHSZ
coma "coama>ao amuucmo zpsom
umthlmuasoo mo puma Conn: now .vwlma mmm .mmamEmm muaSB omauumfilum>m ccc.a

umm anon Hm>m cmuoaano mo Hmnasc maaUGmSamca muouomm mo mammamcm mnu mo mpasmmm .ah manB

 

 

 

 

 

.am>ma mc. may pm oumm Eoum ucmumwwao Manamoamaamam*

mmo.: omo.: ooo. omo.: . . . . . ono omumm mom mum 0:3 .oouma
mmm .mmamEmm mua£3 omauumfilum>m mo unmoumm

*m¢¢.¢l mcm.l cmd. mom.al . . . .Amxv ¢m|ma wmm mum 0:3 .fifilma mmm
.mmamfimw muanz omauumEInm>m mo ucmonmm

*moo.on ooo.: mam. ooa.mu . . . . .ono ommoaoam mum on: mpcsou ca
.Hm>o cam va mmm .mmamEmm mua£3 mo unmoumm

*Nmmom NmHo HNO. @000 o o o o 0 o o o o o o o o o AQXV >¥CSOU
Mom mEOUGA amemem mamamm muanz cmaomz
*ooo.an ooo.] ooo. oao.u . w . . . .Amxo maouca maflEmm moans cmaomz

*flmm.ml moa.l omo.a mmv.ml .A NV Hw>0 Ucm mm mmm .mmamEmm Una mmamE
muaSB >9 omumamaoo aoonom mo mummm :maomz

*mho.m ooa. mom. moo.a . . . . . Amxo swamuom ocm mumuonma ammo
% mum 0:3 monom Honma mama muanz mo ucmoumm

3 *aoo.mu oma.u moo. oom.an . . . . . Amxv mumomcme Eumo mom mumeumm
mum 0:3 monom uonma mama muanz mo ucmuuwm
*cam.¢| cam.l onm.¢ mmo.cml . . . . . . . .Aaxv mocmcaaoo cmuaaomouumz
«mom.m .... cmv.¢mm m¢c.mmm¢ . . . . . . . . . . . . . . Enmu ucmumcoo
mmSam> u mucmao coauma>mm mucmaoammmoo mmanwanm> pcmocmmmoca

omusmfioo Iammmoo oumocmum ceammmnmmm
mumm amauumm

*00m. 0 o o o o o 0 o o o o o o o o o o COHHMQHEHmme mufimflpnfififlz mo “gm-WO-Wmmmou
HovI th I I I I I I I I I I I I I I I I I I I I I I I mpmgflpmm MO .HOHnHm wumwcmpm
*mmoo o o o a 0 o o o o O o I o o o 0 o o o o “amfloflmmmoo QOHng—HQHHOU mflmflgnfisz

 

 

coma "coama>ao amuucmu nusom ummm
Ilmucsou mo puma Eummcozlamnsu Mom .fifilma mmm ~mmamEmm mpasz omaHHmEIum>m ccc.a

Hmm anon Hm>m Cmuoaano mo Hmnfisc mcHUCmsawca muouomm mo mammamcm wan mo muasmmm .mm manme

396

 

 

Illm. aria”,

.am>ma

mo. mop pm oumu scum ucmummmao waocmoamaamamo

 

 

 

 

 

omo.: aao.u mom. ooo.: . . . .ono omumm mom mum 0:3 .ovuma mom
.mmamfimm muaLB omauumfilum>m mo ucmonmm
*hmo.ml oma.| mmo. mao.al . . . .A xv fimlma mmm mum 0:3 ~¢¢Ima mmm
.mmamEmm mua£3 omaHHmEIHm>m mo unmonmm
*mam.ns ooo.: mum. oao.m- . . . . .ono ommoaoEm mom 0:3 apnoea ca
.Hw>o ocm wa mmm .mmamfimm mua£3 mo unmoumm
womanfi H000 dNOo mmo. o o O o o o o o o o o o o o o onv ”UHHSOU
Mom mEooca amcomumm mamfimm muasz cmaomz
mmo.l mmc.l cmc. mac.l . . . . . .Amxv mfiooza maaEmm mua£3 Gmaomz
*mac.ml vaa.l ¢mv.m amm.¢l .avxv Hm>o cam mm mmm .mmamEmm cam mmamE
mpanz >9 omumamaou aoonom mo mumm» cmaomz
*ofla.¢ mcm. mmm. mc¢.a . . . . . Amxv cwEmHOM oam mumuonma Eumm
mum 0:3 mouom Honma mama muaSB mo unmonmm
*omo.on mNm.: ova. Noo.a- . . . . . Amxo mumomama ammo osm mumsumm
mum 053 monom Honma mama muazz mo pcmoumm
*oo~.~- oaa.n mo¢.m oom.mau . . . . . . . .Aaxo mocmaaaoo cmpaaooouumz
*mmo.aa .... mmo.omv ¢¢m.mmmv . . . . . . . . . . . . . . Bump ucmumcoo
mmsam> u mucmao coauma>ma wunmaoammmoo mmanmaum> ucmocmmmocH
Umusmaou lawnmoo Unmozmum coammmnmmm
mumm amaunmm
*mmm. o o O o o o o o o o o 0 o o o o o COHHQCHEHmme mHQHUHDz mo pgmflUHMMOOU
QOOOHNM O O o o o o o o o o o o o o o o O o o O o o o mpmgflpmm mo .HOaHaHm GHmUfiHmpm
*Nmmo o O o o O o o o o o o o o o O o o o o o “Gmflouflmmmou GOHU.MHm«H.HOU mn—Hmufiu-HHHE
coma "coama>ao amuucmo nusom
pmmmllaucsoo mo puma EHMMIamusu How .¢¢Ima mmw .mmamamm mpanz omauume Hm>m ccc.a

umm anon Hm>m cmuoaano mo HmQESG mcaocmSamCa muouumw mo mammamcm may mo muasmmm

.mm magma

 

 

 

.am>ma mc. may um oumN EOHM ucmummmao maucmoamacmam*

 

 

 

*ooo.o ooo. ooo. oqo.a . . . . .ono omumm mom mum 0:3 .oonma mom
~mmamEmm wuana omauumelum>m mo pamonmm

*moa.qn noa.n mmN. omo.au . . . . .Amxo omuma mom mum 0:3 .oouma mom
.mmamfimm muan3 UmaHHmEIHm>m mo unwonwm

mmo. omo. oom. ooN. . . . . . .A xv ommoaoam mum 0:3 oucsoo ca
.Hm>o Cam fia mmm .mmamfimm muaLB m0 ucmoumm

*HOOINI mOHII mHOI NHVOII I I I I I I I I I I I I I I I I wav \Agcsoo
How mEooca amcomumm mamEmm mua£3 cmaomz
*mmm.al moc.l cmc. mmc.l . . . . . . .Amxv mEooca maaEmm muaLB cmaomz

*mva.ml aom.l cmm. chc.ml . .avxv Hm>o mam mm mmm .mmamEmm ocm mmamE
muaSB >9 omamamfioo aoonom mo mummm cmaomz

*moo.o oom. mom. omm.m . . . . . . Amxv swamuom cam mnmuonma sumo
mum 0:3 mouom Honma mama mua£3 mo ucmuumm

*amm.al ooc.I cmm. mwh.l . . . . . . amxv mummmcma Eumm cam mumEumm
7 mum 0:3 mouom Honma mama mua£3 mo ucmoumm
w *ooo.m- ooa.u omo.m owm.man . . . . . . . . .Aaxv mocmcaaoo cmpaaooonumz
*mhm.¢ .... omm.maaa mco.omnv . . . . . . . . . . . . . . . Enmu unmumcou
mmsam> u mucmao ceauma>mm mucmaoaomwoo mmanmaum> pcmocmmmoca

omudmfiou Iammmou oumocmum coammmummm
amauumm

*OOQI I I I I I I I I I I I I I I I I I I I I COHHMCHEHmU‘mQ mHQHpHSE MO “cmHUHmmmoo
mam I NmH I I I I I I I I I I I I I I I I I I I I I I I I I I mpmglofipmm mo “OR-Hm wumvcmpm
*NHmI I I I I I I I I I I I I I I I I I I I I I I I “Gmfloflwmmou GOflpMHmHHOU mflmflpufisz

 

 

coma "coama>ao amuucmo nusom
ummZIImuCSOU mo puma Conn: How .wvlma mmm .mmamEmm muanz omaHHmEIHm>m ccc.a
umm anon Hm>m cmuoaano o0 umnadc mcaucmSamca muouomm mo mammamam mnu mo muasmmm .fin magma

 

398

 

.am>ma mc. msu pm oumN EOHM ucmummmao xaucmuamacmamo

 

. . . Amxv fimlmm wmm mum 0:3 .vwlma mmm

 

 

*wmc.m caa. mom. mob. .
.mmamfimm muaSB omanumalum>w mo usmoumm
«cmc.m| mca.l mom. mon.l . . . . .Amxv vmlma mmm mum 0:3 .fiwlma mmm
.mmamawm mua:3 UmaHHmEIHm>m mo unmoumm
oama.mu ooo.: mom. omo.: . . . . . .ono om»0aoam mum 0:3 mucsoo ca
.Hm>o mam wa mm .mmamsmm muaSB mo ucmonmm
mth' NMOI' ONOI OHOII I I I I I I I I I I I I I I I I onv wpasoo
How mEOUCH amGOmnmm oamEmm mua£3 cmaomz
*aoo.mu oma.u oao. omo.: . . . . . . .Amxv macuaa maaamm muanz cmaomz
*omm.ol cmm.: mma.a 5cm.m| . .avxv um>o cam mm mmm .mmamamm ocm mmama
muanz >9 omumamaoo aoonom mo mummm cmaomz
¥om¢.m wmv. cma. mmo.a . . . . . . A xv :mEmHOM Una mumuonma Sumo
mum 0:3 mUHOm Honma mama muass Mo unwonmm
*omo.mu ooa.: mom. omo.: . . . . . . Amxv mumomame ammo ocm mumaumm
mum 0:3 muuow Honma mama mua£3 mo unmoumm
*oom.mu ooo.: mma.o moo.ou . . . . . . . . .Aaxv mommaaeoo cmuaaooouumz
omwc.m .... acm.cmo amc.wmom . . . . . . . . . . . . . . . Sump ucmumcoo
mmsam> u mucmao coauma>mo mucmaoammmoo mmanmaum> unmocmmmoca
UmusmEOU Iammmou oumocmum coammmummm
mumm amauumm
*mofiI I I I I I I I I I I I I I I I I I I I I COHHMGHEHmqu mHQHUHHflz mo “cmHU-flmmmov
MHQI “MN I I I I I I I I I I I I I I I I I I I I I I I I I I mpmgflumm MO IHOHHm GHMUGmum
*mmOI I I I I I I I I I I I I I I I I I I I I I I I “cmfloflmwmou conflum-fimHHOU GHQHHHSE

 

 

coma "Geama>ao amuucwo sysom ummz

Ilmucsoo mo puma EummGOCIamusu How .wvlma mmm
umm smog um>m cmuoaano mo umnfisc mcaocmSawca muouomm mo mammamcm mnp mo muaSmmm .mm manB

.mmamamm muaSB omannmelum>w ccc.a

 

 

 

.am>ma mc. may um oumm Eouw ucmnmmmao maucmoamacmam«

 

 

 

mmo.: mmo.- «mm. ooa.n . . . . .ono omnmm mom mum 0:3 .ovuma mom
.mmamem muanz omauumfilum>m mo ucmuumm

omo.- omo.- oom. ooo.- . . . . .onv qmuma mom mum 0:3 .owuma mom
.mwamamw mua£3 omaunmfilnm>m mo unmoumm

omo.: mmo.: mmo. ooN.: . . . . . .Anxv ommoaQEm mum 0:3 mucsoo ca
.um>o cam wa mm .mmamfimm mua£3 mo unmoumm

¢¢mI| NflOII mNOI mNOI' I I I I I I I I I I I I I I I I Amxv qusoo
How wEOUCH amflOmHmm mamEmm mua£3 Cmaomz
*moo.mu aom.n omo. oaa.n . . w . . . .Amxo maooca >aa5mm muaoz :maomz

¥m©¢.wl mo¢.I mmm.H omm.NHI . .A Xv Hm>0 Uflm mN mom .mmHmEmm wflm mmHmE
muaSB >9 omumameou aoonom mo mumm» :maomz

*mnh.m mam. ooN. ooo. . . . . . . Amxv cmsmuom ocm mumnonma sumo
mum 0:3 muuom Honma mama mua£3 mo ucmuumm

*cmm.ml oma.l wma. mom.l . . . . . . amxv mummmcme Sumo cum mumaumm
% mum 0:3 muuom Honma mama muanz mo unmoumm
3 *oam.mn oma.u ooq.o ooa.oa- . . . . . . . . .Aaxo mucmaaaoo amuaaooonpmz
*maa.¢a .... ooo.oam omo.ommn . . . . . . . . . . . . . . . aump ucmumcoo
mmSam> u mpcmao coauma>mm mucmaoammmoo mmanmaum> unmocmmmoca

omusmfioo lawnmou oumocmum coammmnmmm
mumm amauumm

*m¢MI I I I I I I I I I I I I I I I I I I I I COflHMCHEHmumD mufimnflunfigz mo ucmflunfiwmmou
womIHNm I I I I I I I I I I I I I I I I I I I I I I I I I I mumgflgmm MO HOHHm UHmwcmpm
*ommI I I I I I I I I I I I I I I I I I I I I I I I “cmHHV-ofimmmov COHpmHmHHOU mHfiHfluHsz

 

 

coma "coama>ao amnucmu nusom
ummzllaucsoo mo puma EHmMIamHSH How .fivlma mmm .mmamEmm mua£3 omauumfilum>m ccc.a
Hmm GHOQ Hm>m cmuoaano mo Hmflfidc mcaocmsamca mnouumm mo mammamcm m3» m0 muaSmmm .oo magma

 

400

 

.am>ma

 

mo. mop um oumN Eoum ucmummmao aaucmoamacmamo

 

 

 

nmm.a who. «on. oao. . . . .onv «mumm mom mum 0:3 .oouma mom
.mmamEmm muaSB Umauumalum>m mo ucmoumm

*hno.m- oom.s mom. moo.au . . . .onv «muma mom mum 0:3 .owuma mom
~mmamfimm muanz omauumfilum>m mo unmoumm

mom.al awa.| mom. mmo.: . . . . .onv ommOaQEm mum 0:3 mucdoo Ca
.Hm>o 6cm wa mmm .mmamEmm manB m0 ucmuumm

oom.au mma.u oao. omo.: . . . . . . . . . . . . . . .onv apnoea
Mom mEooGa amQOmHmm mamfimm mua£3 Gmaomz
*omn.a boa. mao. mmo. . . . . . .Amxo maooca maHEmm muaog cmaomz

*moo.mu omo.: mNo.m omo.mau .ono um>o ocm mm mom .mmamemw ocm mmama
muazz an omumamEou aoosom mo mummm Cmaomz

hma. oao. moo. oma. . . . . . Amxv swamH0m ocm mnmgonma ago“
mum 0:3 mUHOm Honma mamE muHSB m0 unmoumm

*oha.o mom. omo. moo.o . . . . . Amxo mumomcma sumo oom mumaumm
mum 0:3 mouow Honma mamE mwa£3 mo unmoumm
*moo.mu oma.u ooo.o oom.nau . . . . . . . .Aaxo mucmaaaoo zmuaaooouumz
mom.a .... www.mmom ooo.ooom . . . . . . . . . . . . . . ammo ucmumcoo
mmsam> u mucmao coauma>mo mucmaoammmoo mmaflmaum> ucmocmmmocH

owusmEOU Iammmoo Cumocmum coammmummm
MHmm HMHUHMQ

*whmI I I I I I I I I I I I I I I I I I COHumA‘HHEHmw-wma mHQHUHDS MO “cmfloflmwmov
HHOIQON I I I I I I I I I I I I I I I I I I I I I I I mgmgflumm mo HOHHm UHmwcmum
*NuleI I I I I I I I I I I I I I I I I I I I I “amufioflmmmoo COHUMHmHHOU mn—HQHUHSE

 

 

Camucsozllmucsoo mo pumm cmnus Mom ~Eulma mmm
Hmm GHOQ Hm>m cmuoaazu mo umflEdc mzaocmsamca muouomm mo mamwamcm man we muasmmm .om magma

oooa ”scama>ao

.mmamEmm mua£3 omaunmfilum>m ccc.a

401

 

.am>ma mc. msu um OHwN Eoum ucmummmao maucmoamacmamI

 

 

 

ooo.: Noo.: ooo. mao.u . . . . . . .ono omnmm mom mum 0:3 .oouma
mmm .mmamfimm muaSB omaHHmEIum>w mo ucmuumm

*noo.mu oma.u mom. ooo.au . . . . . . .onv «muma mom mum 0:3 .oouma
0mm .mmamEmm mua£3 omaHHmEIHm>m mo unmoumm

*ooo.mu oam.n mom. mom.au . . . . . .onv omwoaQEm mum 0:3 apnoea ca
.Hm>o ocm va mmm ~mmamEmm muanz mo ucmouwm

NMOI NOOI MNOI HOOI I I I I I I I I I I I I I I I I onv %ҤOU
How mEooca aMCOmHmm mamEmm manB cmaomz
*omo.mu aoa.n amo. omo.- . . . . . . .Amxv meooca maaamm moanz amaomz

*aom.mu mom.u ooo.a oom.oau . . . . um>o ocm mm mom .mmamamm oqm mmama
mua£3 >9 omumamEoo aoonom mo mummm amaomz

Noa.n ooo.: amm. ooo.- . . . . . . Amxo swamuom oqm mumuonma Eumm
mum 0:3 mouom nonma mamE muaSB mo unmonmm

*moo.m aoa. omm. ooa.a . . . . . . Amxo mumomcme aumm oam mumaumm
mum osB mouom Honma mama muank mo unmoumm
*noa.ou omo.: omo.oa Na~.~¢: . . . . . . . . .Aaxo mucmcaaoo amuaaoooupmz
*mom.m .... oao.ooma mom.mmmo . . . . . . . . . . . . . . . sump ucmumcoo
mmSam> u mucmao aoauma>mo mucmaoawmmoo mmaflmaum> ucmocmmmoca

omudmaoo Iammmoo Unmocmum coammmummm
mpmm amauumm

*flNfiI I I I I I I I I I I I I I I I I I I I I COHU.MCI.HEHmme mHQHuHSS mo “cmflUflmmmOU
flmm I mmm I I I I I I I I I I I I I I I I I I I I I I I I I I mumafloumm m0 HOHHm Uhmncmum
*n—HmoI I I I I I I I I I I I I I I I I I I I I I I I “gmHUHmmeU QOHum-fimHHOU mn—HfimfluHsz

 

coma "coama>ao Camucsoz
Ilwuasoo mo puma EumchCIamnsn How .vvlma mmm .mmamamm mua£3 omauumfilum>o cccoa
“mm anon Hm>m cmuoaano m0 Hmnfidc mQaUCmSamca muouomm mo mamwamcm mflu mo muaSmmm .mm magma

402

 

 

 

 

 

 

.am>ma mc. mzu um OHmN EOHM ucmumwmao MaucmoamacmamI

ooo.: ooo.: moo. aoa.u . . . . . .Amxo omumm mom mum 0:3 .oouma
mmm .mmamEmm mua£3 omaHHmEIHm>m mo ucmuumm

*nam.mu ooo.- ohm. Noo.m- . . . . . .onv omuma mom mum 0:3 .oouma
mmm .mmamEmw muanz omauumfilnm>m mo udmoumm

oom.au oma.u mmo. ooa.au . . . . .Anxo ommoaoEm mum 0:3 mucsoo ca
.um>o ocm «a mmm .mmamamm muanz mo ucmunmm

mQNIHI mOHII mMOI w¢OIII I I I I I I I I I I I I I I I Amxv hugoo
How wEooca amCOmumm mamem muaSB Gmaomz
«ooo.m mma. moo. ooo. . . . . . .Amxv maouca maHEmm moans qmaomz

oma.a ooo. omo.m vam.m onv um>o ocm mm mom .mmamamm ocm mmame
mua£3 an omumamaoo aoonom mo mummm :maomz

hoo.u mmo.: mom. ooa.u . . . . . Amxo swamuom ocm mumuonma Sumo
mum 0£3 muuom Honma mama mua£3 mo unmoumm

*omo.¢n ooo.- 5mm. mmo.: . . . . . Amxv mumomsma.aumw ocm mumsumm
mum 0:3 muuom Honma mama muanz m0 unwoumm
oom.an ooa.u mmo.ma oom.¢mn . . . . . . . .Aaxv mocmcaaoo cmuaaooouumz
OMMI IIII OhmImbmfl HNOINmm I I I I I I I I I I I I I I EHmp ucmumgoo
mmsam> u mucmao GOauma>mQ mucmauawmmou mmaflmaum> unmocmmmoca

omusmsoo Iammmou Unmocmum coammmummm
mumm amaunmm

Imom. . . . . . . . . . . . . . . . . . COHHmGaEHmuma mamauasz mo unmaoammmoo
omNINOq I I I I I I I I I I I I I I I I I I I I I I I mpmgflumm mo .HOHRm wumccmym
*mHmI I I I I I I I I I I I I I I I I I I I I pamfloflmmmoo cOHuwHGHHOU mn—HQHUHSZ

 

 

Ilmuqsoo mo puma EamonamHSH “Ow .wvlma mmm

oooa "conH>ao composes

.mmamamm mua£3 omauumalum>m ccc.a
Hmm cuon Hm>m cmuoaano mo HmQESQ mcaUGmSaMCa muouomm mo mammamcm ms» mo muasmmm

.om magma

403

 

 

 

 

 

 

;Ja4,
.am>ma mc. m99 um OHmN Eonm ucwummmao >aucmoamacmam¥

woo. woo. woo. omo. . . . .ono omumm mom mum 0:3 .oonma mom
.mwamEmw mpanz omaHHmEIum>m mo udmonmm

*Nao.mu ooo.- oom. oom.au . . . .ono omuma mom mum on: .oquma mom
.mmamfimw mua93 omaHHmEIHm>m mo unmonmm

*ohh.m mom. ooo. oom.a . . . . .Anxv ommoaoam mum 0:3 mpasoo ca
.Hm>o ocm wa mmm .mmamfimm m9a93 mo ucmuumm

*fiHNImI- Hmel GOOI omOIl I I I I I I I I I I I I I I I onv hygou
How QEOUGH amQOmumm mawEmm mpasz cmaomz
mmo.- ooo.- oao. aoo.- . . . . . .Anxv mEoooa maaamm moanz cmaomz

*omo.¢l mom.u ohm.m oom.aau .onv um>o ocm mm mom .mmamamm ocm mmame
muanz >9 Umumamfioo aoo9om m0 mumm> cmaomz

omo.- ooo.: ooo. omo.- . . . . . Amxo cmamg0m ocm mumuonma sumo
mum 093 mouom Ho9ma mama wua93 mo ucmuumm

*aoo.¢ aom. omo. oom.m . . . . . Amxo mumomcme aumm ocm mumsumw
mum 093 mouom 909ma mama mua93 mo ucwoumm
*oom.¢n ooo.: omo.a oom.ou . . . . . . . .Aaxv mocmcaaoo cmuaaooouumz
*oom.m .... mmm.mooa mmo.oomm . . . . . . . . . . . . . . ammo ucmumcoo
modam> u mucmao ceauma>mo mucmaoammmou mma9maum> ucmocmmmoca

omusmEoo Iammmoo Unmocmum coammmummm
mumm amauumm

*FNQI I I I I I I I I I I I I I I I I I gOHpmsflaumme mHm-Wplfisz m0 “cmfloflmmmoo
mquflflH I I I I I I I I I I I I I I I I I I I I I I I mumgflpmm MO HOHHM chmmgmpm
*thI I I I I I I I I I I I I I I I I I I I I pgmfloflmmmoo GOHU‘MHmHHOU mnfimlofl“nfisz

 

 

oamaommll>uzdoo mo 99mm cm9us How .¢¢Ima mmm
umm cuo9 Hm>m cmuoaano mo Hm9as: mCaocmSamca muouumm mo mam>amcm mnu mo muaSmwm .cm ma9mB

coma "coama>aw

.mmamEmm mua93 omaunmfilum>m ccc.a

 

 

.am>ma mc. may 9m OHmN Eoum ucmummwao >apcmoamacmamI

 

 

 

ooo. mmo. pom. oam. . . . . .ono omumm mom mum 0:3 .oouma mom
.mmamEmm muanz omauumelum>m mo usmoumm

*ooo.mu omo.: mom. oom.an . . . . .A xv omuma mom mum 0:3 .oouma mom
.mmamfimm mua93 omanumfilum>m mo ucmonmm

*Noa.m moa. ooo. ooo. . . . . . .Anxo ommoaoam mum 0:3 >ucooo ca
.um>o ocm «a mmm .mmamEmm wuacz mo unmoumm

*mmfiINI mONI' HHOI ONOIII I I I I I I I I I I I I I I I I AQXV ”“GSOU
How mEouca aMGOmHmQ wamEmm m9a93 Gmaomz
«omm.mu ooN.: mac. mvc.| . . . . . . .Amxv mEooca >aaEmm mua93 cmaomz

*cmm.o| vmm.l omm.a mom.ma| . .Avxv um>o cam mm mmm .mmamamo ocm mmamE
mua93 >9 omumamEoo aoonom mo mumm> cmaomz

oom.a mma. mom. oom. . . . . . . Amxo swamu0m ocm mumuonma aumm
mum 093 monom no9ma mama mua93 mo ucmoumm

4 mom.l mmo.: am¢. oma.l . . . . . . Amxv mummmcma Eumm can mHmEHmm
O mum 093 mouow 909ma mama mua93 mo unmonmm
4 *aom.mn ooo.- omo.m mmo.ou . . . . . . . . .Aaxc mocmcaEOo cmoaaoooupmz
*cmm.n .... moa.a¢o mNm.¢mmo . . . . . . . . . . . . . . . Emma ucmumcoo
mmsam> u mucmao coauma>mc mucmaoammmoo mma9maum> unmocmmwoca

owusmfioo Iammmov Unmocmum coammmummm
mumm amauumm

*mnm I I I I I I I I I I I I I I I I I I I I I golfiymcnflgumumg mHQHpA—Hsz MO “Gmfl UHMMGOU
mao.cma . . . . . . . . . . . . . . . . . . . . . . . . . . mumEaumm o0 uouum oumocmum
*thII I I I I I I I I I I I I I I I I I I I I I I I “cmfloflmmmoo COflgmHmHuoo mHQHUHSS

 

 

coma "Geama>ao oamaumm
Il>ucsoo mo 99mm EHMMQOCIamusu How ovvlma mmm .mmamEmm m9a93 omauumEIum>m ccc.a
umm cuo9 Hm>m cmnoaa9u mo Hm965: maaocmSamca muouumm mo mam>amcm map mo muaSmmm .am ma9me

 

 

.am>ma no. 099 um oumN Souo ucmumooao >aucmoaoacmam*

 

 

 

>m¢.I mfic.l aaw. wcm.l . . . . .Amxv mmlmm 0mm mum 093 .wfilma 0mm
~mmamSmo mua93 omauumSIum>m o0 unmoumm

«ovm.m| mmN.: mmo. mmo.a| . . . . .Amxv wmlma 0mm mum 093 ~Sulma 0mm
.mmamSmo moa93 omouumSIum>m o0 ucmoumm

uoo.u oou. moo. ooo. . . . . . .Ahxo omuoaoam mum 0:3 upcsoo so
.um>o ocm va mm .mmamSmo mua93 o0 ucmuumm

QBBII mmOIl HNOI GHOII I I I I I I I I I I I I I I I I onv hpgou
uoo mSOUQu amCOmumm mamSmo mua93 cmaomz
oooo.mu ooo.: omo. ooo.: . . . . . . .Amxo maoocfl >auamo mounz amuomz

*mmc.¢l cmm.: aoo.m mmn.ca| . .Avxv um>0 ocm mm 0mm .mmamSmo ocm mmamS
mua93 >9 omumamSoo a009om o0 mumm> cmaomz

aoo.. cmc. ccm. amm. . . . . . . Amxv cmSmuoo mam mumu09ma Sumo
mum 093 mouoo u09ma mamS mua93 o0 unmoumm

5 omv.l omo.! hmm. mca.l . . . . . . Amxv mummmcmS Sumo Ucm mumSumo
% mum 093 mouoo u09ma mamS mua93 o0 unmoumm
ooo.uu mou.s moo.m moo.mu . . . . . . . . .Aaxo muomcuaoo cmuoaooouumz
*©¢0Im I II I mmHIfimh MBMImmONn I I I I I I I I I I I I I I I gum“ pgmumcoo
mmsam> u mucmao GOHuma>mQ mucmaoaoomou mma9maum> uzmocmmmoca

omuSQSoo Iaoomoo Uumocmum COHmmmummm
mumm amauumm

*OOMI I I I I I I I I I I I I I I I I I I I I nonfipmgflaumumm mlflmnflgnfigz MO “QmHUHmmmOU
MFHINNN I I I I I I I I I I I I I I I I I I I I I I I I I I mHMEHgmm mo HOHHm UHMUgmum
*flmmI I I I I I I I I I I I I I I I I I I I I I I I pamnofionoflmmmoo COflpmHmHHO-flv QHQHHHHJZ

 

 

coma ":0ama>ao oaoaomm
I|>ucsoo o0 uumm Sumolamusu uoo .fidlma 0mm ~mmamSmo muu93 omuuumSIum>m ccc.a
umm cu09 um>m cmuoaa90 o0 u09S5: mcuocmSaoCa muouumo o0 mam>amcm 099 o0 muaSmmm .mm ma9mB

APPENDIX B

CHECKLIST FOR SELECTING STUDIES FOR INTENSIVE REVIEW

ll.

407

CHECKLIST FOR SELECTING STUDIES FOR INTENSIVE REVIEW

Reference:

Type of study: Causal (attempts to determine factors
affecting fertility)
Descriptive (attempts to establish
relationships with fertility)
Other:

Source of data: Census:
Survey sample (size and nature of
S:

 

 

Other:
Unit of analysis:

What country (U.S.?) or parts of country included in
study:

Period of time or date of study (when data originally
collected):

The basic problem:

General hypothesis(es):

Measure(s) of fertility:

Independent variables (used in present thesis):

Relationship
Measure: with Fertility

 

(a) Occupation:

(b) Female Employment:

(c) Education:

(d) Income:

(e) Age:

(f) Distance:

Independent variables (extraneous to present thesis):

. Relationship
Variable Measure: with Fertility:

 

 

 

 

12.

l3.

14.

15.

l6.

17.

408

Are residence categories used?
What are the categories?

Findings:

Contributions to rural or urban social structure:
Are color categories used?

What are the categories?

Findings:

Control variables:

How used and findings:

Statistical techniques:

Selected bibliographic references (empirical studies
only):

Remarks, summary and/or evaluation of study, quotable
quotes (give page no.):