THE ADOPTION or NEW AGRICULTURAL vmcnczs I

 

IN NORTHEAST BRAZIL: NN EXAMINATION or-FI‘IIIIIIEI‘I _' f ‘ ‘ ‘t

DECISION - MAKING

Thesis for the Degree of PII. D.
MICHIGAN STATE UNIVERSITY
DAVID LEWIS PEACOCK
1972

 

 

III "III III I I

III/III III/III III I I III/IIIIII

I III ~
3 1293 00852 6836 ‘ L

L I}? ‘5? I RY

BAIx-‘f‘n-C' -

   

This is to certify that the
thesis entitled
The Adoption of New Agricultural Practices in Northeast

Brazil: An examination of Farmer Decision-Making

presented by

David Lewis Peacock

has been accepted towards fulfillment
of the requirements for

 

 

Ph.D. degree in Agricultural Economics

 

MAI/£23;

Date Zlaw’ J’: /97‘;

 

0-7639

 

 

 

 

 

This is to certify that the
thesis entitled
The Adoption of New Agricultural Practices in Northeast

Brazil: An examination of Farmer Decision-Making

presented by

David Lewis Peacock

has been accepted towards fulfillment
of the requirements for

Ph.D. Jegreein Agricultural Economics

 

 

 

MAI/m}

Date 2260’ J); /97;

 

0-7639

 

 

r‘

..uov§“ \”
r'l . ‘ --
0 .OA‘I‘

a-
..-u-I

v-\ v .\
\.I..I...-..-.

.I
u.>”
- .
A.‘.-y,..‘
9
F.
I
.J‘
‘ ‘V
' ""vu- .
' F
a.

‘ ~u-“_

‘
’1‘
.0 ..-..‘~. .
.
I.
‘u. .‘.. r
I." ’a.

'7‘. .'-. ..\,
' m
..-..,_ s..-‘
... .. "“‘-.~

"H” . \

.., -.&.‘~‘
A
I ' ‘
'-.. . ..
.,r_
‘5.

., _. ‘ ‘

"Ul4‘ '
.

4 :‘A"».. ‘

.H.

1-.‘... ‘~

., -, ~

, . .

.I""\ ‘ :. ~~

.\ “

I.‘

A "b ‘:‘

'IM‘“ ' ".v.

. A‘.~“

I‘ h F
I s "~..
." ‘

A.‘A‘\

't
'0. ‘

-.‘N~ -'-’
~.. -

~1‘;

 

‘ .
.‘~
. .~ ‘
.. '— ..
hl“: c ’-
C‘v‘
.
.
a.‘ a...
n .A‘
-"‘C n
‘5.
I ‘.
I... _..
' A
1‘: \I‘V.
~ A
2-.»
.
H. ."A‘ .
‘-;. \-‘
". f\
“.
a
v
u
‘ t u
.l.‘ . .‘
._ .
~A1 I:
.I
I §“
"' 1g“
5“ “v
5‘ ‘1
'.

 

I
.’ A.
:_p"
..“.(~"‘
\- l
J r
\‘-4
.0
' LA
'~:‘ ‘1
‘ L ..
..¥

 

ABSTRACT
THE ADOPTION OF NEW AGRICULTURAL PRACTICES
IN NORTHEAST BRAZIL: AN EXAMINATION OF
FARMER DECISION~MAKING

BY

David Lewis Peacock

It is generally agreed that farmers' adoption of
improved practices is essential for agricultural development,
but there are differing views of the process by which
farmers adopt these new practices. This study examines
the adoption process in terms of individual decision-making
under conditions of uncertainty and involving the interaction

' 1

or social, economic, ane personality variables.

Various theoretical concepts, obtained from a review of
the relevant literature, Were incorporated into a conceptual
framework of farmer decision-making with respect to
innovations. Based upon this conceptual framework,
relationships were hypothesized between the farmer's use of
new practices and selected variables amenable to analysis
using data collected in a survey of farmers in Northeast
Brazil. This survey was one part of a cooperative research
project involving Michigan State University and SUDENE (the
Northeast Brazil Economic Development Agency) which examined
the production and marketing of agricultural commodities in
the Recife area. .

The Chi-square and analysis of variance techniques were

Imed to test the hypothesized relationships and to examine

Wnsible inter-relationships between the independent

.l‘_

‘ .gnp own

0 ,‘ 4
‘5), Ant-h anti:

- D
. a-a-O-C O o-

': \.\ " .
n! - .e... A

I

C!
.4. pun-.nnun 9;.
’
..... .vK-aben-L .o»
- .

‘ a c
n:-..-. 30;... Av'
..... .»-....

#—

""’ ‘.....hrn

' 4 :3

"‘I‘ 0'--vlb—..5\
a

n ‘ I I

U." n.A.-. .- H.

u ouv...~.. ~'
.

"I '1'“, 0‘-..
~._ ,
—." .Oouv

.
V‘A‘
A._‘

r-.,,

"h
as

(/1
n a

variables.

David Lewis Peacock

These analyses provided a statistically based

Idew of a system of variables influencing the farmers'

decision regarding new agricultural practices. ‘AutOmatic

 

Interaction Detection was then utilized to determine the

 

relative importance of the variables which appeared to

umtribute to an explanation of the variation among farmers

hitheir adoption of innovations.

The major findings of this study are that:

1.

Although certain attitudes seemed to be
related to a greater knowledge of new
practices, the effect of attitudes (to the
extent they were measured by the variables
included in this study) was very small as
compared to other variables.

Better educated farmers, as expected, were
more knowledgeable about yield improving
methods and were more likely to use the
agronomist or mass media as a source of
information about new practices.
Surprisingly, the achievement of a very few
years of formal education (1 to 3 years) was
found to have a substantial impact upon the
farmer's level of innovativeness.

Farmers who used the agronomist and/or the mass
media were significantly more innovative than

their counterparts who depended upon neighbors

David Lewis Peacock
and relatives as a source of information about
farm practices.

Lower income farmers were found to be
considerably less likely to use new practices
than their wealthier counterparts and there was
sufficient reason to believe that their
reluctance to accept "risks” was importantly
involved in this relationship.

The local availability of modern inputs was
found to be an important predictor of
innovativeness, especially among the smaller
and more poorly educated farmers.

Smaller farmers were more likely than larger
farmers to have either not borrowed money or
to have borrowed from non-commercial sources
(small amounts from relatives, neighbors,
landowners, etc.). They very frequently
indicated that their unwillingness to use
credit was due to the "risk" involved in
borrowing. Farmers who had obtained
sufficient "credit" (by their own assessment)
or had secured loans from banks and credit
cooperatives were more innovative than those
who either didn't borrow or had obtained loans
from non~commercial sources.

In general, landlords and tenants seemed to be

more innovative than owner~0perators. The

Tpol‘nv
A 6..\l s u a u

' 9
up 1 u T A
h» DIV-U

flv‘i‘A’._A‘
b'..y. .:

". ,.,,‘.< I“. .
.r .. k

I
u...“.“,:‘.
' ‘ '- ...|.

 

" u
"I..“

a - c
.........‘_..'
.

. . ..— *_ ' - . ' ’
, _ ‘ ‘

...‘u‘ -..‘~‘ u.
y

.. . _ _‘ . K

' ‘ ...v H. .

l-...... Us.-.
0

."s-. -n-‘.‘_..

:"‘ "\u.‘..
s

..T‘ .‘,.“ -:r

. . y. n . . - ‘ ..

~.

.._ D

. ..AV1,".

u A..'..~‘:‘-.‘.

.

v
' “‘1 :.-.-.‘,
..‘. ' v t \ — L
. I
‘5 -il.‘ ‘ n ‘
‘h.‘¥l.‘v “
3

I I
l q '

' ‘A'.’u~y-
U¢~m1\..“l
5, ::

h D

'A
L‘s I‘ _\ ‘
e n “ " ‘

a .
u *y.. ‘ .‘
8“ "OI

u e ‘ d L &‘

O
o
‘h

I ’l. '
‘A.'\_'C;:"

.~\.‘ ‘
V-..
. -
3],,“ ‘
i' I . ‘ ‘
a. . »‘
s v ‘ ‘ J
c
I, g ‘. '
cl,“ . n F‘. A
A
\_\.-
‘s
\
3‘.“
o
‘ file? ”h
. ‘ . ‘
h_‘ E

David Lewis Peacock
most reasonable explanation was that the
landlords in the sample were better educated
and had greater financial resources than
owner-operators, while tenants utilized new
practices either because landlords furnished

the requisite inputs or required their use.

The results of this study suggest that viewing the

ahmtion of new practices as individual decision-making under

mmertain conditions and involving the interaction of social,

eamomic and personality variables is a useful approach to

gaming a better understanding of technological change. A

stamegy for securing more rapid adoption of new practices in

Nonmeast Brazil might also be develOped on the basis of this

research.

(1)

(2)

(3)

(4)

 

Important elements of such a strategy would be:
the develOpment of institutional methods for
reducing the ”risk” involved in trying new methods;
broadening the use of the agronomist and the mass
media as sources of information about new
agricultural methods;
increasing the availability of commercial
agricultural credit (especially to smaller farmers)
and reducing the "risk" involved in borrowing, and
over the long run, providing at least a few years
of formal education to a greater share of the rural

pOpulation.

O"!- .usr.
-. ’

n-u. .\.- . ‘
\b.n—oq.

.

ts_
|tu Nn‘5.

 

THE ADOPTION OF NEW AGRICULTURAL
PRACTICES IN NORTHEAST BRAZIL:
AN EXAMINATION OF FARMER
DECISION-MAKING

BY

David Lewis Peacock

A THESIS

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

DOCTOR OF PHILOSOPHY
Department of Agricultural Economics

1972

 

 

 

I

 

 

.
k.
.
‘~.
‘ -\I
'.05
“. ~-'
“‘ ‘k\ .
s
v . ‘
‘ 'R 1...,
. ‘ \ ’ ‘
~I 0“.“ .‘>
...‘ .
‘ P
. '.'I‘
; .. ,.
.. bt~..-., .‘
'~«_.‘ V
' - " e.. _
.I I“ H .
"‘ vu‘ ~_
‘ 9 -
o
.P
v V.
I “\¢,_
0
.__‘_ .
' uh...”- -
‘~.' \ .. "Z
I ‘r-y“.
\._ -
-’ . .
'-..‘ ”"f‘~‘
- ‘H\’
"o1.'.
' Q
a .' v 5,“,-
.‘~‘.‘

67 ACKNOWLEDGMENTS

I would like to express my particular appreciation to
Du Harold M. Riley for his professional guidance and
pmsonal encouragement throughout the preparation of this
flmsis. I also wish to acknowledge the contribution of
Dn Carl Eicher, Dr. Donald Henley, and Dr. James Shaffer
mm reviewed the earlier drafts of this manuscript and
pnwided many helpful suggestions.

A very special thanks to my wife Carolyn for her
pafience and encouragement during the course of my graduate

sawies at Michigan State University.

ii

/)

 

o-

-4

““-Ao.o

’ .a

“'K" on.-
~

‘.

. I;

I,"

b.

TABLE OF CONTENTS

Page
Acknowledgements ii
Table of Contents iii-vii
List of Tables viii-xi
List of Figures xii-xiii
CHAPTER I. INTRODUCTION 1
The Re search Problem 1
The Approach: Adoption of Innovations As
Decision-Making Under Uncerta in Conditions 4
Plan of the Thesis 6
CHAPTER.II. REVIIRV OF TIRE LITTHP‘T'RE 10
Overview 10
The Decision-linking Behavior of Farmers in Less
Developed Countries 10

The Peasant Farmer

Non- Economic \Ian (lZ)--

as
The Peasant FarIer as Economic l'an (20)--The

Behavior of

A Summary

Farime s in Less Deve10ped Countries:

(27)--Diifc rent Approaches to

Securing Adoption of New Agricultural Practices

(33)

A Brief Review of "Diffusion" 33

New Idea (
System (37

'34)--Time (36)--Members of a Social
).

-Cemmunication (39)

Ibcision-Making with Incomplete Knowledge 43

Risk, Uncertainty, and Profit (43)--The Image
(49)--Aspirations and Satisficing (SS)--

Knowledge

Situa .ions (S7)--(‘ardinal Utility

(59)--Friedman- Savage Utility lunction {63)--
Focus- Les ss (67)--Schema tic Model of the Farm
”Iandger (72)

iii

.

wok.”
, , .
,-1.¢I

A'- ll. 1-!an

' u
y-
rut b.‘\u.

 

n‘iaub-

I
n . ‘-
.o-o
..-.'I.'. a

III“-nb ..‘_.H

" s . .

.‘vou
.

 

II .
-.
I -o,‘ .‘\
' .. y..
- D
.v......“-_
i ‘ ‘
......‘._”-
‘ C
n1... “1» '--
,
.n-Aot ... -:..
‘
l-fi.‘.~’.. 'r’
‘ a
.."'"o.:... ...
‘ I
.h“ .. .fl‘ . ‘
, . ‘
U."¢..o.u.

 

 

‘ n
-o. .,
" A' ‘t..
.‘-l
I‘I.U.A.
. I
.A.‘-““.’
H..' h u
..._.‘
»
-.H .
,
. 1-. ‘f‘
. I I N
‘ -4 ‘
"’I r.
l O
1
-. \ _‘
~.' .
‘b—
1.
~
.
. .
‘v. .
g.“ ’1'.
. “~4 .
_..

 

II._
. I
o ‘. ,
. .u, \
‘u‘C
‘ ~—
ha
p
\- ‘ ~‘ '
~ A
v‘.' '
.‘~ A
‘
'. 9
: ‘2‘. r
‘ n
‘~‘ Tl
‘A
.
.
..
- . ‘
H." ‘ 4 ~-
.. r
u s‘.
.- ~‘.
r‘s.
._ Hp.
~.:
.
‘F-D \o.
H. v._,
..
‘ a o
.. “
‘~ .‘I‘
. 4
, “\
I‘I‘
...(‘.:~.
‘I
l' S 5
V, ‘
.
‘ U
‘1 ~. '-
N‘, ‘
.' \
I\
\
”v

iv

Page
Summary 4
GMPTER III. A CONCEPTUALIZATION OF FARMER ADOPTION
OF NEW TECHNOLOGIES UNDER CONDITIONS
OF UNCERTAINTY 76
Sources of Uncertainty 76
Current Technology (77)-—New Technologies (78)—-
Human Relations (78)--Institutional Arrangements
(79)--Prices (8C)
Expectations 80

Utility for Income 89

Constraints Upon the Implementation of New
Agricultural Practices 93

The Adoption Process: Decision-Making Involving
the Interaction of Many Variables 94

Rationality (94)--Decision-Making: An
Interactive System (96)

CHAPTER IV. AGRICULTURE IN THE RECIPE ARE-1. OI? .‘xTOR'I‘IIEAST
BRAZIL: THE SOURCE AND Cil:—\R.-"\CTER1 STICS OF
THE DATA 103
Introduction 103

An Overview of the Agricultural Economy in
Northeast Brazil 103

Climatic Sub—Regions of the Northeast (103)--
Economic History of Brazil's Northeast (105)

The Data from Northeast Brazil 114

Selected Characteristics of the Sample Farms and
Farmers 117

Crop Yields (119)--Size of Farm and Farm Sales
(lZO)--Attitudes and Beliefs (lZZ)--Educational
Achievement (lSO)--Communication Behavior

E131)--Use of Selected Agricultural Practices
133)

 

 

O ' '
..'Al‘~fl.' Fr

.--.o'oiil A! u

‘Afl’.;.l' ‘00-
. .

.u—‘a... flux.
:

.. .
u-w‘u- w- .06

u...rouldo .‘.

 

9 . ,Q
.001- ... l
r A \ ‘
.u....- .
o i Q .
.":",~.‘ v‘
->-.o.~u . _\
‘ a
:n-u- ,. .
I-........; '
I ' -
00...... ,‘
j >4
o u... ,
\‘~.‘., V.
'0‘ _.I ‘
n .
.i
.- -..._ 9 .
u.
“"-.v..... ‘
.
...v ‘ .
."‘I‘n .

 

. '9“ ’
‘- "v-4
. ‘.
. "- \.
.
,‘.
. "o
" O‘-..
M l‘l ‘ .
.' h.
\uv... _
Y‘ \ -
I
A.\_
v.;
A so
H,“
.
I
'-
' It a
_\ a.
r H. V.
,‘r.
‘ r
‘- 't\r‘.
\. ’ u
n'. ‘
.\ P
«g ‘1.
I ‘.C-
in
,
~' -.\-
, C. .5-
P ' .
\-. .
“'
-‘_\.'F~-
.
.
v
.
'M. 3
I."
I
.
M ‘ . .
.‘~ ."o I
wit.‘
‘A
l
'xl..
'K 1“-
. ‘.
l
.-
I v
I “
\
l . ‘.
a
A. If
Q
1\
.
'I
"L
y
.-
VP v‘
“ ‘c .
"H‘ v."
Vi.‘
-

 

V

CHAPTER V. ANALYSIS OF SELECTED VARIABLES AND
THEIR RELATIONSHIP TO THE ADOPTION
OF AGRICULTURAL INNOVATIONS BY FARMERS
IN NORTHEAST BRAZIL
Introduction
Community Attitudes
Individual Attitudes

Fatalism (143)--Luck versus Knowledge (143)--
Deferred Income (154)--Attitude Toward New
Technology (156)--Interpersonal Trust (l59)~-
Attitude Toward Trying New Methods (160)--
Summary (164)

Educational Achievement
Communicated Information
Utility Function
Summary
UMPTER VI. ANALYSIS OF SELECTED VARIABLES WHICH
MIGHT SI’R‘UI.‘ AS C ,NSTRAINTS UPON
INNOVATIVE BEHAVIOR
Credit
Amount of Credit Used By Recife Area Farmers
(187)--Sources of Credit (192)~-Rcasons for
Not Borrowing (197)--Amonnts of Credit Used
by Farmers-~Regression Analysis (204)--
Summary (209)
Tenure
Input Availability
Summary
“REUNIVII. THE INTERACTION OF SELECTED VARIABLES
IN THE ADOPTION PROCESS: AN AUTOMATIC
INTERACTION DETECTION APPROACH TO THE
ANALYSIS OF INNOVATIVENESS

The Automatic Interaction Detection Program

Advantages and Disadvantages of A.I.D. (225)

Page

136
136
136
143

165

175
181

186

186

211
216
220

I . "
‘sr-r‘fi' ‘
V ‘ A'hln s

 

i
_ . .I"
n

 

 

I ‘
...5.. ...
\v- u .-
«wofibri' one. ,
I
v I‘
' v-n
'1'- ‘ u
I... a... .
I o r...
...
... _
.I‘.. ...
. u
c.
a
.. .. .
n. .
9v ~
. o..
' In.»
I ' ’-
...‘ _.‘ ‘
n . ‘
‘0‘...

 

 

 

 

 

p
n...,_ ‘.. .
.

o.

\‘ .-._

M‘-

.-..‘.‘ ‘l‘

. II

u.-.

.' ’0 ,

,._‘-

i. _‘

_'.._.-

..~

.. ‘

U
..

. . .

‘.. "I
.‘ .. .I

"; u’»—

.. . I‘v-
‘ -. -.\
o

u '1

1‘ . .
.‘§I .‘

o

.o-_A ‘ .
v -..A ‘
..‘_. \

4-‘. . .
. V

. 0-.

n . ..
-| .

.. .“ ‘

,‘P.

."" ‘-
‘D
‘ t.‘ \ _ .
.
‘a
4
r. .
. v.,
‘ I
-u
-0 Q.
_ O
.‘
-. ".
u.‘ _
u. .
I.
s
' ‘ '.v
s Q
-.~. .
‘ 0
4°.
.I ‘
'-a '_v..
‘- Cw-
. ~.
\~.
'u u“
‘..
.
'9 "o.
I
.w‘ iC‘ .
, 0.
K.‘
A.-
'v -1
s _.
~ ‘ k
‘
.._ .
. ..
k'. ~ ‘u
4“
'u
u.
A. ‘v-
V 'd
.I
(.‘l -
. ,
. «
its‘ .
_
“ ’
v“ »
. «,
l_ '1
o
k
s
.V
x
v
‘
~
~ \
I
‘\
‘0

 

Q
a
i.
'Q
I
3.1
'o“ .‘a
«" \.‘
--._~_

 

' I
e .V
.
\. _
. n" A
‘ u i
t
. '4 g
“s

vi
The Dependent Variables

General Approaches to Indices of Innovation
(227)--An Improved Innovation Index (229)--
The Index of Innovativeness Used in This
Study (Z32)--The Dichotomous Dependent
Variables Used in This Study (234)

Selected Independent Variables

Farm Size (235)--Tenure (236)--Income

Level (236)--Credit (237)--Attitude Toward
Trying New Practices (240)--Risk Preference
(240)-‘Price Responsiveness (240)--lnput
Availability (241)--Educationa1 Level (242)--
Information Sources-I (242)--Information
Sources-II (243)-—Attitude Variables (244)

Summary of Variables Included in the A.1.D.
Program

A.1.D.-I: Fertilizer Use by $50 Francisco
Rice Farmers

A.1.D.-II: Insecticide Use Among'SEo
Francisco Rice Farmers

A.I.D.-III: The Use of Tractors by $20
Francisco Rice Farmers

A.1.D.-IV: Innovativeness Among 550
Francisco Rice Farmers

A.1.D.-V: Innovativeness Among Recife
Bean Farmers

Summary

The Effect of Attitudes Upon Innovativeness
(288)~-Education, Information Sources, and
Innovativeness (289)--Innovativeness and the

Willingness to Accent "Risks" (291)--
Innovativeness and the Ability to Implement
Decisions (291)--Tenure and Innovativeness
296)

CHAPTER VIII. SUMMARY AND CONCLUSIONS

Review of Research Project
Analysis of Decision-Making Under Uncertainty

as an Approach to Studying the Adeption of
Innovations (300)

Page

227

244

247

252

{\J
O\
O

269

278
287

297
297

 

- \ ,
.yuboO-‘V“)

- '
\\..:¢:'~«\o .AF

.‘I u-IU"D‘

- a ‘
. t
I”... we»:

3......sv. at
a

' I O
roavda.~ ,9
.

Anti: ‘5.-

. ,‘ .
.o q. ‘
‘1
n4“ v. '
.
‘
.,... ‘uvp‘
-. A
s..iu .rl J.‘
- Q
.-
...
\ :T.'An .-
-¢~.~. 5“..- a
n»

v
1"...”

“’kbU-r.
v . .

v,
‘anu-

4....

vii

Page
Conclusions 302
A Strategy For Increasing the Adoption of New
Agricultural Practices 314
Informational Input (315)--Reducing the
"Risk" of New Methods (319)--Credit (321)--
Input Availability (322)--Education (323)
Suggestions For Further Research 324
Bibliography 327

Appendices 334

 

I
1
u

"“>n.,
\
~..“.I

‘ l
.. ‘.A‘ .
n. h“

 

"v'

‘5--

 

 

 

.. \‘V .,‘
h -_ .I‘
"w “ y .
I . .
\- .-.
y ' I.
D.“\ .
“
‘~‘
. .
5". ’3'“..-
., H
. .‘n\“
‘ V
L ln‘.
c “ |.
h.‘ ~‘LN
I
w.h“,~r
. _ _. _
bty“. ‘
\
7 ‘IV
I, . .
A. I ‘
‘ I» ~
0" O
‘. _‘
Iv.‘ .."
\
A.h\§““_‘-
' ~
._""‘
UH."§A_
‘ I
\ ‘gfl_‘\.‘
‘Iv‘..'-“
7‘ ~ .I
h- .
.U"'¢n
‘ u
- “‘
o!\“L‘
‘ ‘0
.I.‘ ..g'-
»_‘\. _
‘
'7‘..~.
“u .
. ‘.
I
I
. “
.2 -"~.
.
»~”l‘
\
.“;‘
4
.._~1

 

 

Table

10.

ll.

14.

LIST OF TABLES

Page
MSU/SUDENE Survey of Farmers in Northeast
Brazil ..... ............ ............ ......... 117

Average Yields, By Commodities, for the
MSU/SUDENF Survey Farms, and Other Areas .... 120

Distribution of Farm Size Over the Sample
Farms ............. .................. ........ 121

Distribution of Gross Farm Sales for Farmers
in the Recife Area ... ........... ..... ....... 122

The Percentage of Farmers Who Believed That
Their Neighbors Would Look Unfavorably Upon
Someone Who Tried Something New or Who
Progressed More Rapidly Than Themselves ..... 123

The Percentage of Recife Farmers Agreeing With
Selected Attitudinal Statements .......... ... 125

The Educational Achievement of Recife Farmers . 130

From Where or Whom Recife Farmers Had Learned
the Latest Farming Method They Were Using ... 131

Percentage of Recife Farmers Who Have Read a
Farm Magazine, Attended an Extension Meeting,
and Have a Functioning Radio in Their Home... 132

Percentage of Farmers Using Selected
Innovations By Commodity, 1967 ....... . ..... . 134

Information Sources Used by Recife Farmers
Holding Opposing Expectations of Community
Attitucjles 00.00.00.000.........OIOOOCOOOOIOOO 141

Percentage of Recife Farmers Who Expressed A
Fatalistic View By Income Levels ............ 145

Percentage of Recife Farmers Who Expressed A
Fatalistic View By Farm Size .............,.. 14s

The Percentage of Fatalistic and Non-Fatalistic

Recife Farmers Who Knew About Fertilizer and
Knew How to Improve Yields ..... . ..... ....... 147

viii

' F
...-"’0'5!‘ \

.
......nctbbol -
'-
. O
O. ‘fln. ”A.
Ia‘bn IU-
. .a ‘
IDQDD out.—
” 5d
ubboob an.

 

n' v , ‘-
-‘05;, '
“Os-‘0 Au».

.-
‘ \
‘ 'Uaobdy
9‘ O
- A .
'.'. - v ‘v-
"‘ ‘fl-unnn..'
.
v
‘."5' av-
"‘-u-u ¢ ..

 

O
.
““—... I
-
n. M N
...- A.‘
’ -
v
\: p-.-
’Ii. "‘ r
“5.,‘ .
.r
I
._ A
\y F. -‘
"- _. \H .
95“.
I ..
O‘-
M’.“
....~

. q .
\I'I ,v-F“
.' "

.‘-V“

-1
I. I '

o

c.,'_‘ ‘;-~._ .

"w”: t
' l

1..
3:“ ~‘0"
n k-

h

'-

‘-.: :iA.:

.‘b.‘ .
‘l .

 

Table
15.

16.

17.

18.

19.

20.

N
pd

22.

25.

26.

27.

ix

Page
Information Sources Used By Fatalistic
and Non-Fatalistic Farmers in the
ReCife Area0.0.0....OOOOOOOOOOOOOOOOOOOOOOOO 147

The Importance of Business Knowledge in Earning
Higher Incomes as Judged by Recife Farmers
at Various Income Levels .................... 149

The Importance of Business Knowledge in Earning
Higher Incomes as Judged By Recife Farmers
on Various Sized Farms ...... ................ 149

The Importance of Business Knowledge in Earning
Higher Incomes As Judged by Recife Farmers
With Various Levels of Education .......... .. 150

Information Sources Used by Recife Farmers
Differing In Their View of the Source of
Higher Earnings ............... .............. 152

Information Sources Used By Recife Farmers
Divided According to Their Attitude Toward
New Technology .............................. 159

Use of Selected Innovations By Recife Farmers
Differing In Their Attitudes Toward Trying
New Practices 0... ....... O ......... ......COOO 161

The Recife Farmer's Attitude Toward Trying
New Practices by Size of Farm ..... . ........ . 162

The Recife Farmer's Attitude Toward Trying
New Practices By Level of Income ............ 163

The Percentage of Rccife Farmers Knowing About
Fertilizer, Believing That It Was Difficult
to Use, and Knowing How to Improve Yields --
By Education Level .......................... 166

Percentage of Recife Farmers Who Had Read a
Farm Magazine or Attended a Meeting Where an
Agronomist Talked About New Production Methods
~- By Educational Levels .................... 167

The Source of Information Used by Recife Farmers
to Learn About Their Latest Practice According
to Educational Levels .................u..... 168

The Relationship Between Frequency of Listening
to Radio and the Educational Level of Recife
Farmers 0..........OCOOOOOOOOOOOOIOO00.00.... 169

v
'0'

‘: ...-nit“
I a I:
..‘.o.o-U ‘

.. 1
\..9 P: I

..i 'u’o‘

. .
">-O.‘F“
I A

.
....nsgun.‘

' '
“ .np—
U. guuv

o P
....y. A.
up .
o~b§‘I0. ~-
. .
cunt-II.-

o
I...- 1’5.

...-.0 ..
.. ..
I'lfii-65 V.

......o.
.
.‘oh.

 

 

' I
t:"-- _-v
A. 0.. v-
.4 D
I. .H. -
Op - ’.V
.. .
by _‘,L.
v.5“;u. F.‘
... ‘.~ ‘
.
A
‘-4.‘ '
--.‘-‘ ~
—
.1
' I." '-
l. ."S\.
..
. ‘ '0..
~-. \
‘H...
.0
A
‘ “gi.
U. V
"
-.
I. I
.. v.
. ._ . -..,
. -..
..
-.._— f.
'U
1‘.
‘1
‘h‘ u. .‘
’V ' 4-
_‘¢.. ‘5
.
._ I“?
I
§ ‘
‘ I~-.
".
~:v.,_‘
'
~
~ .
eu“{
‘ .
'J“
u "—
'A l.
‘vfu
1‘.
'u.“
‘n ’u",
‘ ~
<
.‘ ,
h ‘5.
, _
\ ~A ‘
“a f.
u. L
o
‘.
‘H‘A-
“ .
.,~. a
V” c
- "‘_
\ .‘
Vs. “f.
Q~ .‘
§
‘.
n
:h‘.’
V.‘\.“
..

Table
28.

29.

30.

31.

32.

33.

34.

35.

37.

38.

39.

40.

41.

42.

43.

Page
Educational Level of Recife Farmers By
Size of Farm ................................ 170

Educational Level of Recife Farmers By Level
of Income .......... .................. . ...... 171

Percent of Recife Farmers Using Selected
Innovations By Source of Information ........ 172

Percent of Recife Farmers Using Selected
Innovations by Use/Non-Use of Certain
Information Sources ................... . ..... 174

The Percentage of Recife Farmers Using Selected
Innovations By Income Level ........ ......... 177

Recife Farmers' Indication of the Consequences
of the Failure of an Innovation, According
to Levels of Income .... ...... . ..... .... ..... 178

Percent of Farmers Using Credit and Amount of
Credit Used ................................. 187

The Percentage of Recife Farmers Who Borrowed
Amounts Less than 3 Given Amount and the Range
of Amounts Borrowed by Commodity Area .. ..... 188

The Amount of Credit Used Per Acre Planted by
Commodity Area .............................. 189

Median Farm Size and Median Amounts Borrowed
Per Acre by Commodity Areas ................. 190

Percentage of Loans Received and Mean Amount
Borrowed Per Acre by Size of Farm ...... ..... 191

Percentage of Recife Farmers Using Credit by
Size of Farm .......... . ......... ....... ..... 192

Percentage of Farmers Borrowing From Given
Sources By Size of Farm . ........ ... ......... 193

Percentage of Farmers Borrowing From Given
sources By Tenure 0.0.0.0000.........OCOOCOOO 194

The Percentage of Total Loans Accounted for
by Various Sources of Credit and Their Mean
and Median Size of Loans .................... 196

Reasons Given by Non-Borrowing Recife Farmers
for Not Using Credit by Levels of Family
Income ........... ........................... 198

kHz":

~-.-u. \'o ‘ ¥

1":" five»:
U:$ut \rtt\~a
.

‘ . "I. -..
.n- g .
'b.\ . N
nun-cu Vbl\o¢

VI. ’ P

'1
I‘I" HALCA

-‘ q .
a . ‘5‘ .-
" '5' H .
l' Aiahv..‘lc
.
far. ...-..-
- nub .

 

', F’QA’.P
c. -,, , ‘
.... ....5...y
' v ..
"‘ .t-vA

..
“‘5 buoth

I

Her-...; " v
V p.

.»1.‘._. v-
Q

9‘“

.

(I.
’v

on.““

P o
o _ 9
t V‘_~.

VII

n.‘_‘.

n
1
0

A.
‘ 'h "
'D

‘b&.‘

M .
in.,’ 1

l
b
u».,““
.
....
. ,
..‘ CV

’1!

.‘ss‘

(11
-:

A-— A

Uw..

V-

Table

44.

45.

46.

47.

48.

49.

50.

xi

Reasons Given by Recife Farmers For Not

Using Credit By Size of Farm ..... ...........

Reasons Given by Recife Farmers For Not

Using Credit by Income Levels ..... . ...... ...

Reason Given By Recife Farmers For Not Using

More Credit by Size of Farm .................

The Percentage of Tenants, Landlords and
Owner—Operators Who Achieved Given Levels

of Formal Education ...... ...... ....... ......

The Percentage of Tenants, Owner-Operators,
and Landlords Using Selected Innovations

Percent of Rural Distributors Supplying
Selected Inputs and 3 Percentage of Recife
Farmers Using These Inputs .................

The Percentage of Recife Farmers Who Believed
Chemical Fertilizer, Insecticides and Rental
Tractors Were Available in Their Municipio,

By Commodity Area ...........................

Page
200

201

202

213

214

217

218

u
\a.
‘.."‘ '41 '
l I
...
......_ - -
\
......" g ‘4
_.._‘~‘-
- -
h.'.-..‘
- ‘ .
-~.__ _\
--'»... .“
..
‘ t
0‘ ‘ "
. v _
."‘-o. \
v» .
. _ .
.
a “
'."’-.._
.
.
.. .'
u- ., .
u
‘ .....“
.. ...
‘. -
‘v .f
....
'n u
< ”9.
"~.
-__
t.
.
.V._ ...
‘-.‘.¥l
.
K
-~-_‘
.‘ V
4 ~‘_~
I
D--‘ O
...- J
'-
.‘. ‘
'4. \ '.
v:‘e ..
....-

—‘I"
5“. - .A
‘

\ ~‘.

..
l~>-.‘.
~D
'
A ‘g
.‘ .
. , .
‘ \
a. ’
~ U‘ .|
I _
“e ."
q
‘5. ‘
Vi.
.
In I. v
« ' s Q,
.I ‘ ‘
u
C
- ‘
r...
.. .
I.
C
.
1
. ‘_-._ '
1 n‘.
'A
h \‘
.o
-

n‘ ‘,
'. -VI-
.‘~

.
‘ I
.> ;,
‘_" —
‘ o .
.o
4",
. ...
h‘ .‘
“

 

 

 

.
.
.
‘ I
0
‘
‘9
.
C.
.d“
.1
“~~
.
.
‘
I‘. ‘
v
. I
..
. K

LIST OF FIGURES

Figure
1. Illustration of Utility Analysis in Choices
Involving Risks ..... ....................... ...
2. Friedman-Savage Utility Function . ....... ........
3. Wharton's Diagram Representing the Differences in
"Expected” Outcomes from Traditional and New
TCChn010gies ......OOOOOOOO ...... 0.00.00.00.03.
4. Nielson's Model of the Farm Manager .............
5. Frequency Distributions Representing the "Degree
of Belief” (Probability) a Farmer Might Assign
to Various Crop Yields Under Present and New
Technologies ........ . .................... .....
6. Frequency Distribution Representing a Farmer's
Expectations of Various Yields with Increased
Accumulations of Information ............... ...
7. Suggested Income-Utility Function ......... ......
8. Schematic View of Farmer Decision-Making With
Respect to New Agricultural Practices .........
9. Map of Northeast Brazil Indicating Sample Areas
for congOdities StL’d)' 0.00.00.00.00...0.0.0....
10. Iso-Utility Curves for Choices in the Risk
Game 0.0.0.0.........OOIOOOCOOO0.0.0.0.........
11. Schematic View of the Decision-Making System of
Farmers in the Recife Area ....................
12. Supply and Demand for Credit by Large Recife
Farlncrs O......0000.........OOOOOOOOOOOO......O
13. Skhematic View of the Decision-Making System of
Farmers in the Recife Area-II ..........J......
14. A.1.D.-I: Fertilizer Use By 550 Francisco Rice
Farmers OO.......QOOOOOOIOOOOOOOOOO....00......
15.

AnI.D.-II: Insecticide Use by $50 Francisco
Rice Farmers ........................;. ....... .

xii

70
73

87

88

97

118

180

182

203

221

248

254

xiii

Figure Page
16. A.I.D.-III: The Use of Tractors by 550
Francisco Rice Farmers ........................ 261

17. A.1.D.~IV: Innovativeness Among 550 Francisco
Rice Farmers .......... ..... . ..... . ..... . ...... 269

18. A.1.D.-V: Innovativeness Among Recife Bean
Farmers ....................................... 279

 

.
"" Iva-Q‘ ..‘~
.
“- ‘Ai5sous.
.

o

'I... ."
.... .. ‘
mue..u.¢. :

 

 

 

I
.
"”‘l 0! 6...
..-.. g‘ .u‘
.
I Q
.., ”... .-. .
L~ ‘
c....-. e... ‘
.
,.v--._ ..-‘.
. . ..
4.....- ,g» ..
'. "f. "r-IA,,
- V“ “. ‘.I
. u
I..~, " . V ‘
. “.
..u‘; "f.
a
‘. .‘
.fiv.. 6.._'1 .
.\«
Ihdggd‘ “
.
"\
.. ‘n..‘
.... ...”. v
._ . .
V .I “A“~
...... “L.‘¥
s.
‘.
.'\~ ‘
._ . 0-,._
"'h‘ .-
c u"
c
I
...
"u r“.‘_‘
.. “ ‘
".__ .
‘v .
\ ‘s..
H ‘ " V-
13.0 \
‘U
O“ -
' I
3.. A.-
I
u.“
If - .
.
{M ‘ ‘ AI...
-‘ It,
-. 54 V
‘5‘
§"
.,‘ .
N'I‘ ‘rh‘
I“ .r‘ 0’
.‘ ‘
.
.J‘. ‘
.‘ h ‘ .
3’. ‘
v.“

 

CHAPTER I

INTRODUCTION

The Research Problem

 

It is generally agreed that farmers‘ use of new and
improved practices is critically important to the advancement
of agricultural productivity among the less developed
nations.1 There is considerably less agreement, however,
with regard to the process by which farmers adopt new
technologies and the strategies that should be employed to
secure rapid technological change in the rural sector.

At the moment there seems to exist at least two
conceptual viewpoints concerning the process of adeption of
new agricultural practices, each having evolved from a
different point of view about the behavior of farmers in
traditional agriculture. One viewpoint is that farmers in
a traditional agriculture behave quite differently from
their counterparts in modern agriculture. The reason given
fin this view is that the traditional farmer is primarily a
'vocial being" and that he lives by attitudes, beliefs,
values, and cultural norms which widely diverge from those of
thenmdern farmer. An alternative viewpoint conceives of the
fanum in traditional agriculture as no less an ”economic
man"than farmers in developed countries, being not a
prisoner of socio-cultural constraints but the captive of low

Productivity resulting from a lack of appropriate and

b“

}A Report of the President's Science Advisory Committee,
Thg;m§ga Food Problem (Report of the Panel on the World Food
upplr, Vol. 1, The White House, May 1967), pp. 1-6.

1

 

 

.
‘ ..‘u ‘I Q
O. "1‘ ' a“
-.. Losu..‘

d

l ‘
..a. N"
...,_ 4 I
U\ .‘
-::¢‘..L..yu .
'\

I .-
Inna...‘ ‘f."
' \

 

L" ‘ ---—....
’ I
""' " \.-..._ ,

u, ”a a“.

"‘~I‘ -3

 

VIA‘
' 1

.,a‘_ - ’ ‘.

...... ...u...
. ‘ ‘
.... “ ’ -

’ _ ' 7-
‘Q ‘04 ...: ‘~.
.
.. .., ‘f .‘
.....u. .' “
.
V.
3‘ ¢. V .».
.‘¥‘l "
'w._ r ‘ ‘

' ‘ ~ .4
|‘». . .“

. ‘
‘u.

‘, \ ‘y’_.‘
“T U‘ s.,;~
.

t, ‘

‘ul
‘ P'-
...-._;‘ rH“
s_‘

.... .

.;.- t,_ “V,

“ ...T u \I
. 4
‘
>I ~
'\
.. F
....1 :J3 A‘
.g .
‘1 J
‘
.r -. . .
k: \ .-1,4
l ._~‘l
_

.

I
n.“

.~ -

~ v

NWT 7:" a
.‘ ‘

._.._I

.K “ '
. 6: P‘

 

,4.

i
A
‘I

".v,
xi, ...-r‘ \(
a '-L .‘

Q.
‘E ".\

2

profitable alternatives to his present practices.

The strategies for increasing the use of improved
practices which emerge from these divergent points of view
differ greatly in emphasis. Concentration upon the social
behavior of traditional farmers leads to strategies, such as
those suggested by Niehoff2 and Rogers,3 which emphasize the
interactions between change agents and members of the
traditional community. Such strategies suggest that
programs of change should be adapted to fit the
socio-cultural environment found in the local communities.
This often involves designing change programs to serve what
local farmers believe to be their needs and adjusting
innovations to fit local cultural patterns. The alternative
strategy is to provide the farmer with an increased supply
of profitable technologies through investment in the
suppliers of these factors of production (such as public
agricultural research and extension services), or to
increase the profitability of modern inputs through
mmfipulation of the economic environment (through price
Ixflicies, subsidized inputs, improvements in land tenure,
eteJ.

Either of these strategies, taken alone, suffers from
someinadequacies. While there is no doubt of the need to

provhk:farmers in less developed nations with techiically

__

. 2Arthur H. Niehoff, A Casebook of Social Change, (Chicago:
aldine Publishing Company,1966), pp. ll~filt

 

3Everett M. Rogers, Diffusion of Innovations, (New York:
The anePress of Glencoe, 1962), pp. 278-282.

 

3
well-adapted and clearly profitable alternatives to their
present practices, it is quite certain that all farmers will
not adopt even the most profitable new inputs at exactly the
same time. Therefore, it would seem useful to try to
determine why rates of adoption vary and utilize the
resulting knowledge to deveIOp strategies for increasing the
overall pace of adoption.

On the other hand, strategies which emphasize the social
interactions involved in the adoption process tend to
overlook the relevant economic considerations involved in
technological change. Not only are micro-economic phenomena
often ignored, but these strategies which insist that
innovations must fit the socio-cultural milieu of traditional
communities seem to ignore the critical importance of
agricultural development. The urgency associated with
increasing agricultural productivity in many countries may
not permit the modernization of farming communities through
a gradual accommodation of the existing culture to modern
practices.

Given the shortcomings of the present strategies, it
would seem useful to re—examine the adoption process in an
attempt to move toward a somewhat more comprehensive
understanding of technological change. The frame of
reference chosen for this study is to regard the adeption of
new agricultural methods as an individual decision-making
process conducted under conditions of uncertainty and

involving the interaction of social, economic and personality

variables.

 

_ .
..,..nv-

.. .'...
2.3“...“
__-_-————

'
.-
{...-9;:ggv- H:
’ ' ...»u.buv

a
.. ¢~un. ouoqcvn‘
1“. ‘0‘... I a:

 

 

 

‘I‘uvina-..
'\\
0~v»s..u
no .
IA 'flA'.
- A
‘Mh “to" g
.
-0
. ’_ fl
. .. ‘
fi' 0... u
.
. _ _

.;...,.. '.. "
\ \ A
..v....., . “h
- I
... ._.- .. .

I t-
' "“q Cut.
.‘ .
': ‘0... -.‘_-
.o
.l. ......t_‘. .
Y.
..“- fl
. ‘ D-V .
"‘ \s.._. .
v
.
... .
‘ . ‘ ‘Fr‘ .
."~u. n
I .‘.““ .
...
. bp.‘-‘.
.‘I- I ‘
. “‘5~._
II, .. . ‘
a
: _“..-‘ r-
...n.“l‘\ .
'- 5

 

.0
'1: -..;
‘.._ \ .
”I. ‘.
. . . ‘
. ...
p
., a... 1
"Ma. A,
.__.‘
. . i‘
--'a ‘c.
'.‘ ‘ .‘_'
.. '9
..“\‘_ \
»
‘n 5

 

a '-
.. .~ ‘
~' ... .
e, .

.h \r‘
A ,0
u

, 4.
i

 

4

The Approach:' Adoption of Innovations
As Decision-Making Under Uncertain Conditions

 

 

Many researchers have recognized that adoption of
agricultural innovations is a matter of farmer decision-
making and have observed that uncertainty is clearly involved
in the process:

The adoption process is one type of decision-
making. The adoption of an innovation requires a
decision by an individual. He must begin using a
new idea, and in most cases cease using an idea that
the innovation replaces. (Rogers: 1962)4

In general, farmers who are limited to traditional
agricultural factors are more secure in what they know
about factors they use than farmers who are adopting
and learning how to use new factors of production.

The new types of risk and uncertainty about the yield
inherent in factors embodying an advance in knowledge
are of real concern to farmers. They could be of
critical importance to farmers who are producing so
little that there is barely,enoagh production for
survival. (Schultz: 1964)D

The subsistence farmers who must execute plans for
technological change must weigh the risks and potential
probabilities of gains and losses in any Change.

Thus in deciding upon change, we have not only the
economics of probabilities of gains and losses, but
also of the values of farmers regarding risks and risk-
taking. (Mellor: 1969)6

The innovation carries a subjective risk to the
individual. He is unsure of its results, . . . .
(Rogers: 1962)7

 

41bid., pp. 77-78.

5Theodore W. Schultz, Transforming Traditional Agriculture,
(New Haven: Yale University Press, 19647.

6John W. Mellor, "The Subsistence Farmer in Traditional
Economies," Clifton R. Wharton, Jr., (Ed.), Subsistence
Agriculture and Economic DeveIOpment, (Chicago: Aldine
Publishing Company, 1969), p. 214.

 

 

 

 

7Rogers, on. cit., p. 84.

L

 

 

 

A

nuke. cv.

p-. .
' .- ,
‘0‘ .~
tin-...»
.. .
'~‘-b -
lo...” _‘
~

-~.,‘ v
v

“T'J‘ 01

T'hoi

- . ‘
\ ”CIA;
" ‘.t.‘
...‘_

.V‘ '2.
b“"-. “
“

... “%r
s.'
I I -
o
.
“uph‘.~

b 5“
‘ A
“C.“
p
“5
a
‘3 iv.
‘.. \
V U
‘
.C :5.»
L“;

~
lip-..-
~ '
G.
q
aleh‘:
._ A
A ‘av
..
\'~.
‘u :‘u.
3:”3‘
Y
.
Frh‘t
. w,
:‘¢
\

 

5

Thus, one approach to reexamining the adoption of new
agricultural practices would be to view the process in terms
of farmer decision-making under conditions of uncertainty.
This study will be carried out within this general context.
An important advantage of such an approach is that it
provides sufficient flexibility to consider a wide range of
variables which might be expected to influence the farmer's
decisions about new technologies.

The objectives of this study were as follows:

1) To develop a conceptual framework of decision-making
under uncertain conditions which can be applied to
the process of adoption of new agricultural
practices in less developed countries.

2) Within the context of this conceptual framework, to
postulate relationships between certain variables
and the adoption of innovations, and test these
relationships using data provided by a survey of

farmers in the Recife area of Northeast Brazil.8
3) To examine the relative importance of the postulated
relationships as explanations of the variance in the
adoption of innovations among the Recife area farmers.
4) Suggest some elements of a strategy to hasten the

adoption of more productive agricultural practices

 

8This data is the result of a farmer survey'conducted

in a cooperative research project involving researchers from
Michigan State University and SUDEHE (the Northeast Brazil
Economic Development Agency). The research project examined
the production and marketing of food and certain other
tnoducts in the Recife area of Northeast Brazil. See Charles
‘SlaUNg Harold Riley, et. 31., Market PreceSses in the .eCife
Eleaffi3Northeast Br(:i_, (Michigan State University, Latin
1idmarman Studies Center, Research Report No. 2).

 

 

iv ""0 f

 

 

p
'
q
.
a
1 b- .-
vn-cv “
Hub». so..-
I
0 ‘o'F-.g
A .....y I .
' V
. o.: ...? -‘
a.

I 1
‘ .
‘loo.
k . "‘ on-
-"AI.‘ ..v
.
‘V. ‘- ," I
b ' ‘l‘

.... .
g '3 Fe .1
.._
. .. .
..
. .
1“.
.'-M.
.
‘t.._l
I
l
b-.}
'I

 

'
<

 

 

I
.
.
.“
‘\‘ I
‘ In.» \._
c ‘5‘
a
I
H .
v
" . u g L
c
‘ . . n _
u~ A \-
‘
~.
1‘ .
e ‘_
.
A.‘ ,
.
‘L 5“
d \
~
.
‘-.
n‘ l.“
., ‘ "r
‘n‘;‘~‘ ‘ ‘
. .j
l-
«5'
I
' it.
'I
, e
.
\
\.

 

 

6
by the farmers in Northeast Brazil.
5) Provide an asseSsment of decisionsmaking under
uncertainty as a framework for studying adoption
of innovations, and suggest further research which

might be conducted utilizing this approach.

Elan of_the Thesis

 

In the pursuit of these objectives, Chapter II provides
a foundation for the ensuing research with a review of the
literature which focuses upon the behavior of farmers in
less developed nations, the diffusion of innovations, and
the nature of the decision-making process under conditions
of uncertainty. It was believed that from the various ideas
presented, certain concepts would emerge that could provide
the basis for a theoretical framework of the adoption process
as it applies to production-increasing technologies in rural
areas still in the early stages of agricultural development.

Chapter III involves the development of a conceptual
framework based upon some of the concepts suggested by the
literature review. This framework, in turn, provides both
testable hypotheses and a theoretical counterpart to later
statistically-based models of the adoption process.

Chapter IV briefly describes the agricultural setting
in the Northeast of Brazil and examines the characteristics
of the farmers included in the MSU/SUDENE survey. This
survey, undertaken as a part of a marketing study conducted
in 1966s67, questioned the Recife farmers about their personal

duuecteristics (age, education, size of family, etc.), farm

.' \
,onvd'

.

. a a
.. 1. ‘l
(.... O“
... nu

..
- -\
o .. ,

 

9‘s.

““cvu

 

 

 

u u
a ‘31
o 5&4
.."
...
I....
n...
.4
u b-
--\v v
\
..‘t'

 

7

operations (size of farm, tenure, use of modern inputs),
attitudes and beliefs (fatalism, attitude toward new
technology, willingness to defer income), their responsive~
ness to price changes, use of credit, use of selected inputs,
information sources used, income~risk preferences, market
behavior, and their perceived availability of certain inputs.
Although these data did have certain limitations for the
present study, they provided a considerable amount of
information which seemed applicable to the problem at hand.
Undoubtedly, the greatest limitation was a lack of data
about the profitability of new methods.

Chapters V and VI involve an examination of relationships
postulated between innovativeness and certain variables
which were amenable to analysis using data from the MSU/
SUDENE survey of farmers in the Northeast of Brazil.
Innovativeness, in this case, can be thought of as the
relative rate at which individual farmers adopt new
technologies; those who readily accept new practices being
regarded as more innovative than those who are more
reluctant to change. The specific relationships examined may
be generalized into the seven hypotheses which follow:

1. Farmers' attitudes have an influence on their
innovativeness.

2. Farmer innovativeness is positively related to
educational achievement.

3. Innovativeness is related to the sources used by
farmers to learn about new agricultural practices.

4. Farmer innovativeness is positively related to the
financial capacity to accept "risks." (Poorer
farmers are less innovative because they are

, .
P"".‘
.00"‘

 

...;IL-

1
...-.54.
l

 

 

“

 

 

 

.5

'.‘: -._
w‘ .

\

'

. s

.\"_ .
ue<
§
._ ‘

8
reluctant to subject their families' wellmbeing to
a crisis involving the failure of an innovation.
Stated differently, they are unable to accept the
"risk" perceived in new methods.)

5. Farmer innovativeness is positively related to his
anticipated utility from increased income.

6. Farmer innovativeness is positively related to the
availability of inputs and credit.

7. Farmer innovativeness is related to land tenure
arrangements.

A more general hypothesis -- to which these seven are

subservient -- is that: an interactive system of social,

 

economic and personality variables such as proposed by the
conceptual framework developed in Chapter 111 can be identified
empirically, and that the relative importance of these
'ariables to the adoption of an innovation can be established.
Identification of such a gys££m_would aid in directing
attention toward the develOpment of more comprehensive
strategies for increasing the pace of technological change
within the agricultural sector of developing nations.
Chi-square and analysis of variance techniques were used
in these chapters to examine not only the relationship
between the variables and the adoption of innovations but the
relationships between the independent variables themselves.
It was believed that a clearer understanding of the adoption
process would be achieved by developing a knowledge of the
inter-relationships between the independent variables in
addition to discovering their relationship to innovativeness.
Chapter VII examines the relative impact and interaction
of variables which, on the basis of relationships discovered

in the preceding chapters, appeared to have an effect on farmer

 

.»I"‘
1"..:"“
nui"'

L‘-

y
_ nu

 

..
A‘VQ"-

_.’ ..\

. ...I-.bo
.

 

 

. .
.-.-o vac. "-
.-

t I M
‘0‘4' lhugu. afi

 

"‘F'Ol.\~AA

‘ ,.n

s.— .....A-u-’
.

 

‘Q
»
...
‘ I
‘1 .... _ '.
'_ O
a. , ‘
-¢-.._ '_-
.,.‘_ .
. ll...
., . ....
a...

"he; .‘:
. ‘v-

 

-
u..'.....

....
‘ .4

 

.__‘ ‘
t ...
‘ I
x ..
‘_..
7- b
s E‘ ‘
_\. ,___
‘ J-l
a
....
~l\\¥-‘~
n, \\ e-..
.‘y ‘
's.,
o
;- fl ‘ I
~,~ ,. ' ‘
It ">
. . ,
“s..
. ..' u
t. id .
..r‘ '- .,‘.‘,
..I I
I ~~
C I
u.
t. v
‘u_ AA o
u‘ Q
~. 1?
o-
“ .
.
\‘~ ‘ -
‘« {Va
.v‘ .- I
I
I
.
‘-
1~ "
.'- '1‘ ‘.
g! A
a
3‘
,
b .“
l'\t.’t— ‘
‘-C‘ .Q‘ .
.
y .

9

innovativeness. A statistical technique termed AutOmatic‘
Interaction Detection was used in the analysis. This
technique provides an explanation of the variance in the
dependent variable in terms of the interaction of the
independent variables. Using either the use/non-use of
certain practices or a weighted index of innovations as the
dependent variable, the computer was employed to generate
several models of the interaction between the independent
variables and farmer innovativeness. An examination of
these interaction models becomes the focal point of Chapter
VII.

Chapter VIII summarizes the research by providing:
(1) an evaluation of decision-making under uncertainty as an
approach to studying adoption of agricultural innovations;
(2) an assessment of the relationship between specific
variables and innovativeness; (3) conclusions regarding the
generalized hypotheses proposed; (4) a proposal of certain
elements to be included in a strategy for hastening the
adeption of more productive agricultural practices in the
Recife area of N rtheast Brazil; and (5) suggestions for

further research which might follow this approach.

 

 

 

.
I.“
.
k ‘1...
, .
-_‘
._ .
...
i
‘
‘e
.V ‘
. u
.3 3
‘-.
.
~7 ‘~
-._ \ Q
.'~ C
.‘V
._§
'.“\ Q
~ :1

‘1
".J
‘ an
I ‘ h n
.u ,
'e

 

 

CHAPTER II
REVIEW OF THE LITERATURE

Overview

 

Since this study seeks to examine the adoption of new
technologies in the context of an individual decision-making
process, a review of the literature should draw upon those
concepts which might contribute to an understanding of the
behavior of the "decision-maker” and the nature of the
decision process employed by the individuals involved.
Taking the farmer in less developed countries as the
"decision-maker" and the adoption of new agricultural
technologies as the decision process, there would seem to be
three general areas of the literature which bear investiga-
tion: (1) concepts about the behavior of farmers in less
developed nations (often under the heading of "subculture of
peasantry," "subsistence" or "traditional agriculture");

(2) concepts about the diffusion of innovations; and

(3) theories and empirical studies of individual decision-
making under uncertainty. It was believed that from the
various ideas presented, certain concepts would emerge that
can provide the basis for a conceptual framework of the
adoption process as decision-making under uncertain
conditions.

The Decisi0n~Making Behavior of
Farmers in Less Developed Countries

 

As with so many other important issues in agricultural

-developnmnt, there is less than complete agreement about the

10

-
.A.V"f

U
"" '.‘ II.

‘ -..-
..d
.
9D"

.- ‘

’ . on.
D... .a h.'00
..0"...

c
"no nod'a-
.

.a._.. §\~.. -

 

--o..-- ..

. -

."""Ib§ .e
.

""' .F'O 6.“

u" shoe .,

I. , ...
vu-..o,-‘ ...‘,

-....2-.-. ,.

 

C

"." «5o-
..
...

 

n
‘\
. ’ ‘ ‘
I;>..‘-- ‘: O
.....l._‘ .~ .
L"""- "'
......_‘~ .'h.
"‘u . _
‘ 2: ‘IA . .‘
..I-
. ' s. ,H‘

I
. t . .
J: A. --‘_
. D
i '4 I ."
.._ _ ,
‘.‘Q. .n,.
.t-t ‘-'.n_ :
"Q.. '
no
.. ." :_- 9r
a ‘_'_
.-,_
.
.Z I._ .
' h " on.
w.
""...‘._
1
._‘
u‘ . .‘
suryl‘ "k.
.
..-
. {val
» "~.¢" ~
“a. A
., ‘ 9‘. ‘0"
'VA A
. L.,

 

‘v, .
‘ .' +F. It-
,A_. 3-5‘
C
Q
‘A
3 L' “'4"
I
.‘ ‘: ‘3
. u .
. d‘ C- I“
t ‘y
H
§
"I V
. I
-. ..fl .'r’ _
s I
I
Q
'1
‘a‘

 

 

ll
behavior and motivations of the farmers in traditional
agriculture. With respect to the adoption of new agricultural
practices, one side of the present controversy argues that:

Despite all that has been written to show that
farmers in poor countries are subject to all manner of
cultural restraints that make them unresponsive to
normal economic incentives in accepting a new
agricultural factor, studies of observed lags in
acceptance of particular new agricultural factors
show that these lags are explained satisfactorily by
profitability. (Schultz: l964)9

While the other side contends that:

Available evidence seems to indicate that peasant
behavior is far from fully oriented toward rational and
economic considerations. Undoubtedly, however, the
degree to which peasants are efficiency-minded and
economically rational depends in a large part on their
level of modernization. It does not seem justified to
assume that subsistence farmers will be promptly
motivated to adopt agricultural innovations merely if
the pecuniary advantages of such acceptance are
pointed out. (Rogers: 1969)10

In general, this controversy has come to be focused upon
the behavior of the small farmer within the less developed
country -- the "peasant" or "subsistence farmer." There is,
of course, a compelling reason to concern ourselves with the
behavior of the "peasant." One estimate of their number
suggests that they may represent more than 50% of the world's

. ll .
population. If indeed "peasant farmers" are so numerous,

 

gSchultz, op. cit., p. 164.

10Everett M. Rogers, Modernization Among Peasants, (Holt,
Rinehart and Winston, Inc., 1909), pp. 31-32.

11John D. Rockefeller, 3rd, "The Challenge of Population
and Food," Clifton R. Wharton, Jr-, (Ed.)," Subsiscence
‘A ricultnue and Lconomic Development, (Chicago: Aldihe
Publishing Company, 1969), p. 3.

 

 

- \
. . ‘|,~
...... --.
, a ‘ fl.
.- n
u. u . ‘
out... U

.... p
.. \
.3 n..'. e
'
':

 

 

'- A-I
U! s-.
n- u. ‘
“" ..-
ue...‘_ .‘
\ \
0"-n.

.
.‘I ...-n'
a
N.

 

 

 

. ' _.
I .V".v-\
“ v....."‘

on. ..., “.‘

7‘ ., g
.. '.“\b—.

"., I .

' 2: o:
.. .. "‘
n"-_‘ ‘
..."~ ...
~- . .
\ e .. f-
- I... A“
._v
' P.‘
1‘ v1_“\
‘ n

u. ".
‘4; ~\‘
U ‘i ‘
". I
.-QA I
"a.“

-_“ -

— 1“ .
‘5
s 4. L
'a.‘
"b‘C~A‘_
I‘
.Q
a.‘\ ‘I. ‘
u 7-.
v_‘ '5:
‘
‘ .'
\“ .‘
' 4
.fiL
‘l s
\‘ .

-_ -

.-‘ ‘ V
“‘ it
.\

‘t

‘2

“ O V‘ ‘

Q-)

|

-

\

‘1 ‘r

q .
b"

 

12

they may be expected to represent the major proportion of
decision«makers in a traditional agricdlture. Consequently,
considerable attention will be given in this study to the
various views concerning the behavior of the "peasant."

The major question to be examined has been posed by
Wharton in the following manner:

Are subsistence farmers basically different than
nonsubsistence farmers? If so, in what way? Or are

the differences merely a matter of degree?12

The Peasant Farmer as Non-Economic Man

 

An expansion of Wharton's question, ”are subsistence
farmers basically different from nonsubsistence farmers?",
might be as follows: do peasant farmers behave as if they
were "economic men” (at least to the extent that commercial
farmers are assumed to do so) by attempting to equate
marginal costs and returns, responding to economic incentives
and being desirous of obtaining higher incomes; or are they
non-economic beings who are constrained in their decisions by
the force of the values, attitudes, traditions, and mores of
their socio-cultural environment?

Kusum Nair13 perceives the farmer in less developed
countries (at least in India) as not necessarily respondent
to economic incentive. Her view can best be illustrated in
her own words as she summarizes her study of Indian peasants.

Speaking in general terms about the problems of agricultural

 

12Clifton R. Wharton, Jr., "The Issues and Research ~
Agenda," Clifton R. Wharton, Jr., (Ed.), Subsistence Agriculture
andEcenomic DeveIOpment, (Chicago: Aldine PEblishihg Company,
1969), p. 458. ‘

13Kusum Nair, Blossoms in the Dust, (New York: Frederick
A. Praeger, 1961).

 

 

Ata-

 

 

.
...... - .
'1. 5‘

......ID. A

"I”. data .-

a
.-vu out. '

Q
'--.---1 0‘ .
o. .

40.. .. ..
,
"-..au. ..-

“co-..-.

 

"'---.-.v I.

' Q
"1-- ..

I.-.uu.. ., .,

- uvyn..~ ‘

' u
"D 15..-“.
- .

 

 

..
"‘- 6....
. .

‘ “’--.. .

' \

.... ..,_... ‘

o
' ‘n-
.4 2
~... ‘
., ‘-
~-
.. iv
;‘~-e._‘ .
..._‘..l

 

 

.~. _
.__ ‘5
v ‘ A.
"-... s
..
u ‘I
., ‘
.
.
‘I ‘ a
.._.. ~
-
‘.
..K’-
..\ h \.\:.
. .‘_~~
V.
.
I
a
..
.5 1‘
n ‘~

 

 

'\
‘. 9
, n
".
u.’ D
I I
c Q
....
b n
u
n V
.
n

13
development in Indian communities, Nair states that:

After talking to peasants all over the country ...
I came to the conclusion that the problem of material
resources is only one of several factors that must be
taken into consideration and to which any programme
designed to raise farm yields must be adjusted. And
it is not always the most important. A community's
attitude to (sic) work can be a more decisive
determinant for raising productivity in Indian
agriculture than material resources, or for that matter
even technology.14

After observing wide variations in the motivations and
accomplishments of Indian farmers, Mrs. Nair concludes that:

In the absence of common valuations, a uniform
response to common incentives and stimuli cannot be
expected. On the other hand, variations in the value
system can make all the difference to the extent of
success or failure of'a deveIOmeht scheme independently
of the material and natural resources.

 

 

Focusing her attention on the peasant's attitude toward
his standard of living and its improvement, Nair gives the
following assessment.

The upper level they [peasants] are prepared to
strive for is limited and it is the floor that is
bottomless. This does not mean that the desired
standard is always fixed at the subsistence level.

It varies with different communities. In some groups
it is very much higher than others, and it may be
considerably more than the minimum necessary to breed
and survive. But whatever the level, it tends to be
static, with a ceiling rather than a floor, and it is
socially determined. Generally, the lower the level,
the more static the aspirations tend to be.

If my observation is correct, it largely
invalidates one of the principle assumptions on which
present planning for economic development of the rural
sector is based.

14Ibid., p. 190.

151bid., p. 191.

16Ibid., pp. 192«193.

v
A
!" .:' H
t.“ ”“
Q
.
,0. 0-; “
'\ ...
~-- ...»
I ' '
... we. ‘L
5

.
nae-‘0" ‘

. .
... pb-qv-
\ .

havgflbebo~

 

'.-1" u. '1.

.
In.¢.~- ..
-

Ivy-.... .V

I
I Q A. »
IK ,1 ’*\
- I ...I
““.~'. . _~
"' '--; ‘..
u
‘ I‘ .
u
‘ 9-».
Ho‘,‘ ...;
-. l
sun... A.
.-
~-v._
. U
, c
.
- ‘ ..v
“' ‘ ~¢J
a. .
\ §““..
-.
"éei.
."u
“ ‘F..-.
.-. 'vkt;
.
.. .
‘I
-' ” ‘-
‘o “‘ .e
5
‘. x
j. . .
_ o
a... .
. ..,
I
I. \ P
. .
... v.
v
“'Q' l
“1,.
1! ,
.
.,‘ .
. «‘1
.".‘\...
t.
F
l .,
“A.‘.
‘..». ....
“
...
‘I-
N‘. i
\. ~«u
U
a
1
1‘1- V.
"b - '4
-...“ y.
'e
.\I
..
'1‘ s” .-
‘9' Pk
' x
"
~“v-.
i. ~V‘.
.
‘s‘s
‘.
l‘ M,
_. :t‘
l ‘» r
\-
..'u
' s
.‘ 2‘:
h‘.‘
~
~
‘1
7'.
\‘ ‘
‘v fl
..
h
n
p
I
‘I
J!
P‘-
. -
c t

14
From her observations of peasant farmers in India, Nair
reaches the following conclusions about the economics of
agricultural development:

... It is apparent that there is no economics in
isolation from sociology and social psychology. There
are many causal relationships and connections between
purely economic factors and social and cultural
conditions which cannot be ignored or excluded from
economic analysis .... 7
As a result of his basic research in Latin America and

supporting work in Asia and Africa, Rogers has also

concluded that peasant farmers are basically different from
our concept of commercial farmers. He perceives certain
behavioral traits as typical of peasants wherever they are
found. Further, Rogers argues that these common patterns of
behavior constitute a "Subculture of Peasantry" which can be
defined in terms of the following characteristics: (1) mutual
dISLTUSt in interpersonal relations; (2) perceived limited
good; (3) dependence and hostility toward government
authority; (4) familism; (5) lack of innovativeness;

(6) fatalism; (7) limited aspirations; (8) lack of deferred
gratification; (9) limited view of the world; and (10) low

empathy.18

According to Rogers, the "peasant" is the captive of a
mentality of mutual distrust, suspiciousness, and evasiveness
in interpersonal relations. He appears to be playing life's
game by attempting to maximize the shortrun advantage to

himself and assuming that others are doing likewise. A

 

l71bid., p. 194.

18Rogers,'Modernization Amon_ Peasants, pp. Z4~38.

 

 

..
no:

.9
b-

 

t
.-

...-.55‘
A

 

‘
,.. .,
.—
"r...
v .._ .
“'Ql .- .‘
‘ .
‘D‘
n. ,
‘1»
-.
.
-..
P-
.
u.
-.
'fi)“‘ I
.-
-. .
’-
I .‘
Ire. '
’-
. o
‘ h
0..
I 9-
..g -
‘ U
.

 

n

 

 

15
distrust syndrome of this type seems to preclude cooperative
action, casts suspicion upon anyone who advances more
rapidly than his neighbor (considering him to have cheated
or exploited his fellows), and claims as fraudulent any
individual who professes to have altruistic motives.

Closely related to the mutual distrust among peasants
is their concept of the world as "having only an absolute
quantity of that which is gggd," The perception of a finite
quantity of that which is desirable in life (wealth, health,
security, love), these things being always in short supply
and beyond the power of the peasant to eXpand, logically
leads the peasant to the conclusion that he must struggle
to maintain his share and that anything gained by others is
at his expense.

Rogers asserts that the peasant's view toward government
is one of ambivalence. He may exhibit suspiciousness,
evasiveness, and hostility toward government, yet look to
government for the solutions to his problems. While govern-
ment may seem distant, hypocritical, and exploitative to the

pea ant, Rogers finds this same peasant to be more strongly

'J)

imbued with the “help-me" philosophy than with the self-help
approach.

Familism is another trait which is assumed to be
characteristic of peasants. The individual's goals are
subordinated to those of the family. One's own family, or
perhaps extended family, may provide the security needed by

individuals to cepe with these elements of society which he

 

o
.—
..pofl

.oo-‘V' "

I-.. a
u . .
“' -I¢... u
e
‘ .
‘ ~—-.

.,
"-v..,,

u... -.. .

...e...‘ ‘.

.¢n

 

 

- ~-
., .5 -,
.. .' “
uh.
u'-.
s
a‘. ‘
‘ n
a ._-.
-.
....-
._‘
._ v
t .‘s..
._ ‘ '
. ~._
u.
- _‘~
17-.
, ‘d-_
-..V._.‘
’ .
,. ‘
l
‘8
-_ .
.,‘
’~. ~ .
.

 

 

 

n!
L—A

 

16

has learned to distrust. Rogers also notes that familism
may be an important contributor to the peasants' lack of
innovativeness. Under such a system the decision to accept
or reject new ideas must be made jointly by the entire
family, often with older members playing dominant roles, the
result being a lack of flexibility in decision—making and
the likelihood of little innovative behavior.

According to Rogers, peasants are generally not
responsive to new ideas. He suggests a few reasons, in
addition to familism, why this is the case.

First of all, his life pattern inclines the peasant
to follow these ways he knows will produce positive, even
though small scale, results rather than try a new idea
that might end in failure and thereby endanger his
existence.

It is often said that the lack of peasant
innovativeness is a function of scarce economic
resource and technology inapprOpriate for the village
setting. Peasants are poor, and a lack of ready
capital undoubtedly serves to discourage the adoption
of those 19' ideas which require cash outlay.

Peasants are also poor in technological resources and
know-how. ... Many agricultural innovations from
temperate climates have been introduced in tropical
settings without adequate adaptation to new conditions.
The result has been failure, and further negative
conditioning of peasants' attitudes toward innovation.

Available evidence seems to indicate that peasant
behavior is far from fully oriented toward rational and
economic considerations. J

Rogers finds fatalism widely reported as a characteristic
of peasants. Defined as "the degree to which an individual
recognizes a lack of ability to control his future," fatalism
(often intertwined with his religious beliefs) may serve both

as an explanation of the peasants' misery and a barrier to its

 

19Rogers, Modernization Among Peasants, p. 31.

 

 

I
no .
u. .

 

 

 

 

 

 

(IT

17
alleviation.
If one sees himself in control of his life
situation, he can be motivated to improve his

existence, but if he resigns himself to the hands

of Fate, he cannot be induced to seek a higher

standard of living.

Peasants are also thought to possess very low levels of
aspiration. Rogers believes that a sense of fatalism, the
image of the limited good, authoritarian child-rearing and
lack of significant opportunities have conditioned peasants
to expect little from life. In addition, the pervasiveness
of the view of a limited good is said to result in
inconspicuous consumption by those who do manage to improve
their level of living. The resultant lack of visible signs
of accomplishment would logically have a stagnation effect
on the level of aspirations in the community as a whole.

Lack of deferred gratification is the unwillingness to
postpone "immediate satisfaction in anticipation of future
rewards". Rogers' statement extends somewhat beyond the
belief that peasants are reluctant to finance higher earnings
in the future by reducing present consumption because of the
immediacy of needs at their present low level of living. He
suggests that the peasant's lack of deferred gratification
can reach a somewhat psychopathic state of preferring the
here-andsnew to even the immediate future.

Peasants, Rogers continues, view their world in terms of
imprecise time periods and a limited geographical perspective.

The minutes and hours of a day have little meaning to the

 

fif—f a V

2°1b1d., p. 33.

l
..A .5

 

 

 

1

iv!-

,_ 0-
‘fi' 4
. I
no .I
yo 5.

y I

«b. ca ‘0.

..
to .5 '1
;" " o ..
. ._ . .
0‘ an. .o—II
.. ..p,‘

n ‘
, "; ‘0
“'"V .:

 

,... .
'\
K. '-
.
.V .
I ...4- ‘
'-..u.x \
( 4
. .
. . _
A . ~"‘-u
I. ~‘l>‘\
| ' "
0.,. n
t.
".. 4
“hu.

 

 

   

 

.
a. ‘ F
.... i”
~
' tn - '
. 4
‘~‘~‘
4
.
.
0‘ ..-
"c.
.A
-i "
.t
U‘
I. ‘
r ‘s
g A“
t "
‘I...
u
.
A. ‘
“‘ 'A~.“~
1“
v n
n
:-

   

 

 

\
~. “.,
.
‘l 1 ‘
u r
. 'd
\_.
.‘ \‘- .
.. ‘5.“
v-‘t‘
..
v v V
.
.l
-. u
n‘ p“
‘l
v
y
.
.
.
q
s.
5
N
\
e
‘.
.
u A .‘
“~

 

18
peasant, who measures time instead by the sun and the moon
and the natural seasons of the year. Punctuality and
precision are largely foreign to village life. The villager's
knowledge of the world outside his own community is also
likely to be vague, and the compensating effect of the mass
media may remain unrealized because of his limited exposure.

The final trait of the peasant, in his conceptualization
by Rogers, is a low degree of empathy. The ability of the
peasant to mentally project himself into the role of others
(as empathy is defined) is severely circumscribed by his
limited exposure to other ways of life and the socio-
psychological distance between himself and those individuals
in unfamiliar roles -- the urban elite, the extension worker,
the governor or the president.

Banfield's21 examination of a rural community in Southern
Italy led him to conclude that cultural constraints were at
the center of its continued poverty and lack of social develop-
ment, although in a somewhat different manner than proposed by
Nair and Rogers. Since Banfield begins with the premise that
a modern economy and democratic political order depend upon
an ability to formulate and maintain a "high degree of
formal organization," he attempts to discover why organiza-
tions had not developed in this community as they had in
rural areas elsewhere. Why, for example, had this particular

community not evolved such organizations as a voluntary

 

21Edward C. Banfield, The Moral hasis_9f a Backward
Society, (The Free Press, Glencoe, lliinois, 1958).

 

 

 

..-A-:".y 6.

- .
lvl\~‘iib s

I
"-... .

H..-.v.

,
-.._‘

i" o
‘t.

“Ua.‘.-

 

 

 

 

 

.1
*3
. .
.I’.-. .
. ...‘
‘.“
.
v‘;
.
.
'\
I
.‘ .__
w.
.
.
I
. ‘ C
‘r
.
‘ 4“
4.
\
.
.
.‘.
K‘.
.
.
.V
n A.

V
:“a
fly

‘H‘
l
.V
.
.
.

 

19
ambulance service, a local newspaper, a parentmteacher
association, or a farmer c00perative which would seem to
the outsider to have obvious advantages for everyone
concerned.

After rejecting such commonly mentioned factors as
ignorance, poverty, class antagonisms, land tenure arrange~
ments, oppression of the State, and despairing fatalism as
insufficient to explain the behavior of the peasants and
the townspeople; Banfield submits what he believes to be a
useable predictive hypothesis. According to this hypothesis,
the decisions made by individuals in this particular
community were guided by the following principle:

Maximize the material, short-run advantage of

the nuclear family and assume that all others will
do likewise.32

Naming this principle "amoral iamilialism,“ Banfield makes
no claim for its universality among backward societies, but
does assert that it was useful in explaining the behavior
of both peasants and villagers in the community he studied.
There is, of course, no reason to assume that such a
decision—making principle would prohibit the adoption of new
practices on an individual basis. In fact, Banfield cites the
case of peasant farmers continually applying fertilizer in
hope of a larger yield. What it does prescribe, which was
of greater interest to the author, is organization for the
greater good of the community. In contrast to "enlightened

selfninterest," the practitioner of "amoral familialism"

Ibid., p. 85.

‘v

22

 

D... I. O

r..~..

 

 

 

 

 

20
will be expected to contribute nothing to further the
interest of the group or community, except as it is to his
immediate private advantage. Organizations will be difficult
to form and maintain because inducements are often to an
important degree unselfish and gains often non‘material.
Moreover, members of successful organizations must trust
each other and have a loyalty to the organization. This is
complicated when anyone who professes to be public spirited
will be suspected of being a fraud. Thus, Banfield concludes
that the community he studied was a prisoner of its family-
centered ethos and that their lack of ability to act
concertedly for the common good was the fundamental impediment
to their economic and social progress. Without a change in
the ethos of the people, which according to the author may
only be accomplished by outsiders, development will not
occur.23

The Peasant Farmer as Economic Man

 

Professor T. W. Schultz, among others, has argued that
too much has been made of the social and cultural constraints
limiting rural deve10pment. On the contrary, he believes that
the backwardness of agriculture in less developed nations can
best be explained in economic terms.

The laggardliness of agriculture in poor
communities is frequently attributed to particular
cultural values. These values relate to work, thrift,
industriousness, and aspirations for a higher standard
of living. They are then used to explain why there
is so little economic progress and why particular
economic deve10pment programs are unsuccessful in

.....

 

v‘

zslbid., pp. 163pl64.

 

.- .
.”"‘~0p
,

“cup-..“

1'.

 

 

 

21
practice. As a rule, however, it is not necessary
to appeal to differences in such cultural values, 24
because a Simple economic explanation W111 suffice.
Or, in addressing Wharton's question, Schultz states
that:
Thus, however relevant the cultural attributes
are in explaining some important classes of problems,
they do not provide a basis for distinguishing between
traditional and other types of agriculture.25
The crux of Schultz's argument lies with the "state of
the arts." The problem is not that farmers in traditional
agriculture allocate their efforts and resources badly, in
fact he argues that stagnant agricultures are likely to be
in better economic adjustment than those exhibiting a
pattern of dynamic growth; but that they lack knowledge of
new productive inputs or, even worse, that improved locally-
adopted technologies are not available to them. A "traditional
agriculture", as viewed by Schultz, is one in which any past
changes in relative prices and production technologies have
been completely adjusted to by the community such that it
lms long since reached an economic equilibrium -- marginal
cosusbeing equal to marginal returns even though incomes
arelow ("efficient but poor"). What is known about the
trmfitional factors of production —- their costs and their
lemmas -- has been known for some time, and there is little
OPDortunity for showing the farmer how to farm better unless

wefiTSt introduce technologies with which he has had no

\\

4
Schultz, op. cit., p. 26.

 

25; _
Tbid;, p. 29.

 

 

22
experience.26
Within this conceptual framework, of an economic
equilibrium uninterrupted by the introduction of improved
production factors, Schultz finds the explanation for
peasants' attitudes toward work, thrift, and investment:
Incentives to work more than these people do
are weak because the marginal productivity of labor
is very low; and the incentives to save more than
they do are weak because the marginal productivity
of capital is very lew.27
He emphasizes, particularly, the high costs of gaining
additional productivity through the employment of traditional
inputs. According to Schultz, the growth of an economy is
dependent upon growth in the number of income streams. But
new income streams must be purchased through savings and
investment, and when costs are high relative to the size of
the income stream (low returns to investment in traditional
inputs) it is hardly reasonable to eXpect much in the way
of economic growth. To stimulate growth in the agricultural
sector, then, it must be provided with lower cesr income
Mneams -- more productive inputs.28
Economic growth from the agricultural sector of
a poor country depends predominantly upon the
availability and price of modern (nontraditional)
agricultural factors. The suppliers of these factors

in a very real sense held the key to such growth.
When they succeed in producing and distributing

‘.

26Ibid., pp. 24-52.

 

2".
’gbid., p. 28.
2% .

8Ibi'd., pp. 71-102.

23
these factors cheaply, investment in agriculture
becomes profitable, and this then sets the stage
for farmers to accept modern factors and learn how
to use them. It is also an inducement to increase
savings and to develop institutions to provide
credit for investment in such factors.2
But once new productive factors are available will
farmers accept them? If these factors are indeed profitable,
Schultz believes they will. Adequate attention must be
given, however, to the conditions of profitability. To be
profitable the new inputs must increase yields sufficiently
to more than offset the (often high) cost of acquiring them.
It is absolute increases in yields that pay bills and provide
profits, Schultz warns, not relative increases. Further, the
matter of risk and uncertainty must be taken into
consideration. Variations in yields from traditional inputs
are well known from decades of experience. This, of course,
is not the case with new factors of production. Uncertainty
about possible variations in yield are particularly critical
in poor countries where low initial incomes allow very little
latitude for experimentation.30
The rate of adeptien of new inputs in poor countries,
Suvmldue allowances for risk and uncertainty, concludes
&munm, is dependent upon its profit; and in this respect
Uwirrespense is similar to that observed among the farmers
1°Hmodern agriculture. Thus, according to Schultz:
Once there are investment opportunities and -
., foicient incentives, farmers will turn sand into gold."1
“*__“__ .

29 ‘fi ‘ " ' -
Ibid., p. 145.
30

Ibid. , pp. 162-168.
31 «

‘Ibid. , p. s.

24
There is considerable empirical support for characterizing
the farmers in less deve10ped countries as fecenomic men."
The findings of Hopper in North Central Indiasz, Chennareddy

33, and Welsch in Eastern Nigeria34 indicate

in South India
that farmers in these areas allocate their resources
efficiently within their traditional technologies. Each of
these researchers calculated the marginal value product of
traditional inputs (using Cobb-Douglas functions) and
compared them with the marginal costs of these inputs. The
results were alike in each of these cases, with farmers
appearing to do a very adequate job of equilibrating

marginal costs and marginal returns.

Welsch found additional reason to believe that farmers
in Eastern Nigeria are rational in their response to new
technologies. With respect to lack of adoption of
fertilizers his assessment is as follows:

The Abakaliki rice farmers did not use fertilizer
in 1963, although they had tried it previously. Crop
yield was unpredictable, and lower yields often resulted

because of excessive straw growth and consequent
lodging and less of grain. Fertilizer trials on rice

C
0

at several locations in Eastern Nigeria over the past
several years have been inconclusive . . . In addition,

32W. David Hopper, "Allocation Efficiency in Traditional

hMimiAgriculture," Journal of Farm Economics (August 1965),
PP-6ll-624.

 

S 3Venkareddy Chennareddy, "Production Efficiency in
(fiMh Indian Agriculture," Journal of Farm Economics
member 1967), pp. 817~820. '

 

34 _ .
‘Wak .f3elane E. Welsch, "Responses to Economic Incentives by
a11.;1 Rice Farmers in Eastern Nigeria," qurnal'ef Farm

E .
“£fl22££§§ (November 1965), pp. 900‘913.

 

 

' — *._'w—-w .- -

25

the present variety of rice has been selected over the

years for its adaptability to the natural environment,

rather than for its response to fertilizer.35

This observation lends support to Schultz's contention
that laggardliness in adopting new technologies can be
satisfactorily explained by their profitability (see pages
21-22).

Eicher, in his discussion of Nigerian agricultural
development, cites evidence which indicates that the small-
holder played a very dynamic role in diffusion and expansion
of certain commercial crops. The growth in the production
of such crops as oil palm, cocoa, rubber and greundnuts, from
1900 to 1960, was the result of voluntary investments on the
part of Nige ian peasants in response to favorable world
prices. These significant increases in agricultural
productivity, notes Eicher, occurred without the assistance
of government directed agricultural programs and the aid of
an effective agricultural extension service (often assumed
as prerequisites for agricultural change).36

In another examination of economic man in Africa, Jones
pnmmnts evidence against what he believes to be a common

Insumception by Western observers of the need for a different

“Wtof economics to deal with the behavior of African

fa“Hers. His approach was to present historical evidence of
X

3
SIbid., p. 912.

in 8 Carl K. Eicher, "Transforming Traditional Agriculture
oftoutilern Nigeria," (paper presented at the Annual Meeting
e 1\1rican Studies Association. Bloomington, Indiana,

o
“”6? 26—29 , 1966).

 

 

26
trade among Africans, or responsiveness to price changes, of
rationality in production decisions, and of savings and
investment. One example from his study will serve to
illustrate how misinterpretation of the facts may lead Western
observers to deny the existence of economic man on this
largely underdeveloped continent:

In some parts of Africa . . . maniec farmers
present another sort of riddle to the uninformed
observer. In their study of food farms in the cocoa
belt of Western Nigeria, Galleti and his associates
commented on the large amount of maniec that was left
in the field unharvested. Harvests yielded only
about 45 percent as much as would have been expected
on the basis of crop harvesting experiments. . . .
Similar figures are reported from an agricultural
survey in Southeastern Ghana. . . . In the Ivory
Coast, too, an agricultural survey . . . revealed
maniec production to be only a small fraction of
what it would have been if all the crop were harvested.
. . The general explanation is again economic, and
derives from the distribution of costs. . . . It
costs little to plant a field of maniec and the crop
can be grown with a minimum of care. For maniec even
land cost is trivial when the crop is grown in fields
that would otherwise be abandoned to bush. By far
the largest cost . . . is that assumed in lifting the
mature roots from the field and transporting them to
market. Many African farmers grow maniec as
speculation, to be marketed if the price is high
enough . . . to yield a return greater than cost of
harvest, to be abandoned if it is net.37

Finally, Stern, and Bauer and Yamey provide still more
MWWHW for the concept of economic responsiveness among

fumensin less developed countries. Stern, using relative

Drums and crop acreages, found that Egyptian cotton producers38

\

7William 0. Jones, "Economic Man in Africa," Feed

£%Eizgh Institute Studies, (Volume I, 1960), pp. 107-134.
Uotatien, pp. 121-121 .

 

gy ._ Ilobert M. Stern, ”The Price Responsiveness of
'ptlan Cotton Producers," (vklos (Volume XII, 1959).

27
and Indian jute producers"9 shifted production patterns
with changes in producer prices.‘ Bauer and Yamey found

Nigerian primary producers responding to price incentives

for higher quality products.40

’The Behavior of Farmers in Less Developed Countries:‘ A

Summary

The question with which this discussion began «-
regarding the existence or non—existence of basic behavioral
differences between subsistence and modern commercial
farmers —~ is still unsettled. And universal agreement upon
the factors which determine adoption of new agricultural
practices has not yet been reached. While Schultz asserts
that:

Since the differences in profitability are a
strong explanatory variable, it is not necessary to
appeal to differences in personality, education, and
social environment.41

Rogers argues that:

Schultz does have a point, even if it is over-
stated. Econemic considerations a:g_one predictor
of innovative behavior by peasants, but they certainly
do not outweigh socio-cultural variables in all
instances.42

39Robert M. Stern, "Price Responsiveness of Primary
lhmhmers," Review of Economics and Statistics (Volume XLIV,
May 1962).

40F. T. Bauer and B. S. Yamey, "A Case Study of
.R“W0nsiveness to Price in an Underdeveleped Country,"
Eflmflflig Journal (Volume 69, December 1959).

4

 

 

1Schultz,ep."cit., p. 164.

 

Rogers, Modernization Among Peasants, p. 313.

 

28
The question has really emerged (as seen above) as

one of the relative weightings that should be given to
economic and non—economic variables (if such a line of
demarcation could be conveniently drawn) in the adoption
process. Economists have shown that in the aggregate
farmers in less developed countries are economically
responsive. This seems logical, for if one considers only
the socio-cultural barriers to adoption it is difficult to
conceive of how the farmer would be motivated to try any
new practices. On the other hand, if it is only necessary
to point out the economic advantages of a new technology to
secure its adoption, then why don't all farmers adopt new
practices at exactly the same time? Is it not reasonable
that both traditionally economic variables and other
variables -- such as differences in personality, education
and social environment -- could enter into the adoption
process? Assuming this to be the case, the approach used
in this study will be to examine the explanatory value of
arange of variables -- both economic and non‘economic --

whhfllmight be expected to influence the farmer's decisions

Umaninew agricultural practices.

 

Drflkrent Approaches to Securing Adoptioflufiflilfifli
Jaggggural Practices

 

An individual's conception of the nature of farmers in
1“3 developed nations will have substantial impact on the
prOCedure he would recommend for securing rapid adoption of
new agricultural practices.

135. profitability is a sufficient factor to secure the

29
adoption of modern practices, a logical strategy is to
ensure the availability of profitable new practices to the
agricultural sector. Schultz suggests the very direct
approach of investing in the production and distribution of
modern inputs in order to ensure that a supply of truly
profitable nontraditional production factors are provided
for the farmer to adopt. It is not adequate, cautions
Schultz, to simply transplant these practices used success‘
fully in modern agriculturos to less developed nations. At
a minimum, adaptive research must be conducted, and it is
likely that the development of new inputs should often
involve beginning with what is known in the basic agri-
cultural sciences and applying this knowledge to the
problems of the area. (Here Schultz alludes to the
Rockefeller programs to develop hybrid crop varieties in

43

Mexico.) In addition he notes the importance of investing

in the ”human resources" of the rural community.

In sum and substance, the man who is bound by
traditional agriculture cannot produce much food no
matter how rich the land. Thrift and work are not
enough to overcome the niggardliness of this type of
agriculture. To produce an abundance of farm products
requires that the farmer has access to and has the
skill and knowledge to use what science knows about
soils, plants, animals, and machines.44

But there are other ways to ensure the profitability of
mmkrn agricultural inputs and encourage their use. These
IMmIVe manipulation or restructuring of the agricultural

e O O O 0
WWW)" 1n such a way that it favors increased agricultural
\ I

Schultz, op‘. city pp. 145474.

411
Ilddq, p. 205.

ffi V ‘5'

 

 

30
VTOduction with the adoption of modern farming methods as
a cOnsequence. Agricultural price policy is one possibility
whiCh should be included under this heading, Raj Krishna
suggests the need for a "positive price policy" in order to
stimulate agricultural productivity. He believes that
agricultural policies in less developed nations have too
often been ruled by considerations of providing cheap food,
with the consequence of unfavorable terms of trade toward the
agricul tural sector ("negative price policy"); and that
s“Ch p0 licies have been patently unsuccessful in providing
Suffici ent food for these nations.45 Along with'the need for
general 1y favorable product prices, the importance of price
Stability to growth in agricultural output has been
suggeSted.46 Further, there may be justification, where
the C03 ts of producing agricultural inputs are high and a
consumer surplus is likely to result from their use, to

subsi * . . . -
d1 2e inputs in order to make them profitable tor the

 

farmer ‘ ,
to use. Changes 1n tenure arrangements also may
RTCSen . . . . . .
t some p0551b111t1es for 1ncreas1ng the use of
4 s

Dcvg]0 IRaj Krishna, "Agricultural Price Policy and Economic
(Ed5_)pment," Herman M. Seuthworth and Bruce E. Johnston,
New Yo; figricultural Development and Economic Growth, (Ithaca,
46ki—Cernell University Press, 1967), p. 540.
in deve S'ee the following for discussions on price stability
FA oping countries:
'1‘ The Stateaof FeedYandh.ggric111ture'(1915\9), pp. 137-14l._
’Volu. ‘ .Wauer an E. W. Pals 'Comment ' nylos '
L 3: XI, 1958). ' ‘ ’ “‘
in N13 I5‘a1d K. Helleiner, "The Fiscal Role of the Marketing Board
(Septemrian Economic Development, 1947-61,” Economic Journal
er 1964), pp. 582-610. ffifi
and De ‘ M. Singer, "Introductory Statement --~ Stabilization
5113103 elcapment of Primary Producing Countries, Symposium 11,"
\’ (voiume x11, 1959, Fasc. 3), pp. 269-276.

 

 

lip,

 

31
aglicultural inputs.

Fairly recently, the possibility of furthering economic
de\Ielepment through improvements in the marketing system has
been gaining some attention. This argument assumes that the

present marketing processes in less developed nations are not

generally efficient, and that changes in the marketing system

Could both reduce food prices to urban consumers and increase
I‘ul‘al incomes. Decreased consumer prices for food products
”Geld have a "real income effect," especially on the urban

Poor wlao spend a large proportion of their family income on

£0061. Since it is likely that these low income families

would Spend a considerable amount of any increment of income

on additional food, and their numbers are many in the. cities

or most less developed nations, one could expect a generally

inc: a ,. .
Ieabed level of demand for reed products. If this
increas;

ed demand is adequately reflected back through the
market

System to the farmer, he will be stimulated to

increa _ . . . .
5’0 his production. This may encourage him to purchase
imPI‘OV .
9:1 farm inputs (fertilizer, new seeds, machinery, etc.)
whose
D‘rices might also be reduced by more efficient

marke a
t 1 fig systems .47

10m the contrasting point of view, Rogers has suggested
a

"St

E) ategy for change" which concentrates on the interaction
etWe

change agents and their clientele. His strategy
fecusQ

11 upon the following principles: (1) a change program
s 0111
fl be tailored to fit cultural values and past

4 'v *'

PP. 1/ C. C. Slater, H. M. Riley,‘et.‘ al., Market Processes,
8~1/19. T '

 

 

 

32
exPeriences; (2) clients must perceive a need for the
Tvaation; (3) the change agent should be more concerned
withimpreving their client's competence in evaluating new
iQMS and less with simply promoting innovations; (4) change
qfimts should concentrate their attention upon opinion
lemhxrs in early stages of the diffusion process; and (5)
socnal consequences of innovations should be anticipated and
preVented if undesirable.
S;i;race Rogers believes that local beliefs and values have
a Critgi-(:al impact upon the success of an innovation, he
StIess;c3 s; that change agents should design their programs to
be compatible with the local culture. The consequences of
“0‘1 doing so are illustrated by his example of a U.S.-
trak“3<3 :irrigation engineer who found that farmers in his
Farlk15;.t> homeland were not using the wells he was building.

L0C“‘ j‘ziarmers believed that water from the irrigation wells

was "a - - . . . .
“]?‘1—J;fic1al" in contrast to ”natural” rainfall, and

refns . . -
€3C1 ‘te utilize them for fear their crops would be damaged.

1‘ . . . . .

I} (3 potential need for an innovation "must CXlSt in the

client

5SiYstem," according to Rogers, although the change
agent

C: E311 help to develop such a need. A useful tactic,

then -
’ 3L 53 for the change agent to select innovations on the

basis
<2) 4:? . .
‘re these needs perceived by the community.

It
Ilas seemed to Rogers that change agents have tended

E2mphasize shortprun, singlevinnovatien programs as

Oppose
61' 1Io longmrange programs to change the values and

evaluat .. .
l0n capabilities of their clientele. From the v1ew«
point t:. - ,

rlrlt not all innovations should be recommended to all

 

33
meInbers of a social system, a "more the better" philosophy
is not considered to be as acceptable as a program which seeks
to provide a more favorable attitude toward new ideas.
Research in diffusion has found that the opinions of
CeI'tain individuals in a community are highly valued by
other farmers, and that these opinion leaders are a strong
legitimizing force for new ideas. Consequently, if the
change agent can discover the community's opinion leaders
and convince them of the acceptability of an innovation, the
rate of adoption by other farmers will be considerably
enhanced.
Finally, Rogers notes that well-intentioned innovations
can r'ii‘Sult in unforeseen and undesired 'secial consequences.
It Should be the task of the change agent to anticipate these

1"." ‘ . . . . .
e fectb and initiate action to av01d them.48

A Brief Review of "Diffusion”

The elements of "diffusion of innovations," as given
by R0

gets, are: (1) a new idea; (2) which is communicated
thTOu

81" certain channels; (3) among members of a social
%. . 49 . . . . c

=- (4) over time. It is within this framework that

some

0 f the main concepts of "diffusion" will be briefly

consi
~ - deified, although not in the order suggested.

 

4 8
282. Everett M. Rogers, Diffusion of Innovations, pp. 278-

for P Also see Everett M. Rogers, "Redeveleping a Strategy
the AD anned Change," (a paper presented at the Symposium on
I)Zlicatien of Systems Analysis and Managerial Techniques

to Ed
June leational Planning in California, Orange, California,
2~~13, 1967, pp. 1-13).
Ibid., p. 12.

 

 

34

“SW Idea

An innovation is an idea perceived as new by the
individual. It really matters little, as far as human
behavior is concerned, whether or not an idea is
"objectively" new ....50

According to Rogers, an idea is new if it is perceived
“Jbe so by the social system. And, that new ideas are not
efidtqalent units, they have a number of characteristics to
gTeateerf or lesser degrees. Rogers suggests the following
as thc)s;e characteristics which innovations possess in varying
degre e s .
1... Relative Advantage«-the degree to which an innovation

is superior to ideas it supercedes (economic or
neneconomic advantage as perceived by the adopter).

 

:3 - 'Compatability~—degree to which an innovation is
consistent with existing values and past experiences
(which ensures greater security to the potential
adopter and makes the new idea more meaningful).

 

 

 

:3 - Complexity--the degree to which an innovation is
diificult to understand.

‘1 - Divisibility--degree to which an innovation can be
tried on a limited basis.

:> ‘ Communicability--degree to which the results of an
innovation can be diffused (i.e., obviousness pf
new haymaking equipment versus farm records).5

13‘3’ these Singh and Warlow added "rate of cost recovery,”

"Opera -
‘tlllng cost," "initial cost," "amount of labor savings,"

and "m Q}
<:hanical attraction." They tested the importance of

these
<::}1aracteristics on the rate of adeptien_of new corn

PTOduQ _
1t-3~on practices by Ontario farmers. The "rate of cost

recoveg

“ ' ~3r33’" seemed to be the most highly associated with rates

551()‘ »’- ' -

$1“ h1d., p. 13.
31316., pp. 121—134.

 

 

no

35
“f adoption, followed by "financial return" (relative
adVantage), complexity and divisibility in that order.
communicability, operating cost, and initial cost were also
significantly related to rate of adoption; while
COllzpatability, labor savings, and mechanical attraction were

"or. 52

“Guile Rogers alludes to the fact that innovations "carry

53 he has not given

a subj ective risk to the individual,"
attention to amount of perceived risk as a possible
explarlatory variable for varying rates of adoption. Johan
AI‘ndt , however, has studied the relationship between "perceived
TiSk" and innovativeness. He found that New York consumers

“ho perceived less risk involved with a. given product tended

to be those who had adopted it earliest. Unfortunately, due

to the Structure of his research, he was unable to determine

if C0118 umers had been early adopters because they perceived
1955 Pi 81: associated with the product, or whether they new
perceixaed less risk because they had adopted earlier and

thus had greater experience with the product by the time Of

the survey.“

\

5 2
Farm 1‘ Ram N. Singh and G. S. Warlow, ‘ Characteristics of
maratiens Associated with the R356 of Adaptation,
Report :30! 01 Guelph, Department of ExtenSion Education,

.14, October, 1966).
5 3

S I1c:gers, Diffusion, p. 84.
q __

lnnova {Johan Arndt, "New Product Diffusion: The Interplay of
and Pr :lveness, Learning, Opinion Leadership, Perceived Risk,
UniverS @Uct Characteristics," (Unpublished paper, Columbia
lty, New York).

 

 

 

 

36
113
Diffusion of innovation has come to be regarded as a
PNmess which occurs over time. Rogers found that if the
Mmmer of members of a community who had adopted an innovation
by certain dates were plotted over time it resembled an
S‘shaped (cumulative frequency) curve. He believes that
the .sliape of such a curve is caused by an "interaction effect"
amorig; the members of the social system. The "interaction
effcacz‘tf'results from the influence of those who already
adop>12<3d on the remainder of the community. At first the pace
0f idrlrlovation is sluggish, but as the number of adopters
incr‘c>:a¢;es it quickens until only the most recalcitrant
members of the community are left. Empirical evidence of

th. 3 go 0
IS S-shaped adoption curve, according to Rogers, is one

of . . .
tlll<3 reasons why adoption of innovations should be regarded

as a 55

Siocial process rather than a purely economic phenomenon.
'Tdie individual's adoption process may also be thought of
as . I u o o o o o 0
having a time dimensmn. Rogers initially believed this
PTOC . . . .
<31‘35 to be div1ded into five stages:

1.. Awareness--exposure to a new idea, but lacking
”complete" information about it. (Often thought
of as a random or non-purposeful occurrence.)

32. Interest--favors the innovation in a general way,
although he has not yet determined its utility
to himm-actively seeking information.

13. Evaluationa-mentally applies the innovation to his
present or future situation and decides whether to
try it or not«~seeks reinforcement at this stage.

4» ‘lrialw-use on a small scale to determine its utility
FF . o o - . o

\ ‘ _ . in the indiVidual's particular Situation.

Rogers, Diffusion, pp. 136-141.

 

 

\ 37

\
8 5. Ado tion--decision to continue use of the innovation
on a fuIl scale.56

Empirical evidence, however, did not seem to support
thiseonceptualization of the adoption process. Diffusion
re’Searchers found that stages were often misplaced and some-
Umes omitted altogether. This caused Rogers to reformulate

the acioption process in terms of four functions he believes

déscrfiitae the mental activities of the individual as he adopts

or rej ects a new idea:

liraowledge function--in this subprocess the individual
gains knowledge of an innovation but is not yet
motivated to seek additional information.

.Eigarsuasion function~~in this subprocess the individual
increases his information level to the extent that
he can form an opinion toward the idea.

IEAEjgision function--includes those activities an
individual goes through in choosing among
alternatives.

S25233firmation function--information seeking activities
following adoption or rejection of an innovation.
(This is explained by Rogers as an effort to reduce
post-decision cognitive dissonance.)

 

 

 

MEEEELE1§i_of'a Social System

‘9\SS noted before, not all members of a social system
mbpt: Ern innovation at the same time. Rogers has consequently.
dw1<3“53<1 the whole of the community into groups 0f individuals
cm t1)”€3 basis of their relative earliness or lateness of
mbp1t'ji<on. Rogers and others believe that different attributes

CMl‘t)
(3 assigned to these various adopter categories. The

£011
Q - . . o -
Wlilng chart describes some of the personal characteristics

 

6Rogers, Diffusion, pp. 76-93.

537
Of Ijnl Everett M. Rogers, with F. Floyd Shoemaker, Diffusion
.09w- IltD‘vations: A Cross-Cultural and Communication Approach

 

(ka: Free Fress of Glencoe, forthcoming).

 

 

 

38
assumed to be typical of these different categories as

applied to farm operators.

 

Adopter Category Personal Characteristics

 

Innovator Youngest age, highest social status,
largest and most specialized operation,
wealthy, willing to take risks.

Early Adopter High social status, large and
specialized operations.

Early Majority Above average social status, average-
sized operation.

Late Majority Below average social status, small
operation, little specialization, small
income, skeptical.

Laggard Little specialization, lowest social
status, smallest operation, smallest
income, oldest, tradition-oriented.

 

Source: Rogers, Diffusion, p. 185.

There are several concepts (in addition to those in

the above chart) which are basic to the diffusion researcher's

choice of variables. Below are listed some of the concepts
which the diffusion researcher frequently uses to explain an
individual's innovativeness -- willingness to accept new

ideas.

Literacv-—degree of mastery over symbols in a written
form

Education--usually highest grade attained in school.
Cosmopoliteness--degree to which an individual is
oriented outside his immediate social

system (opposite is localism).,

Extension Worker Contact--amount of interaction with
extension workers.

 

39

Interpersonal Trust-~willingness of individuals to
trust others.

 

Fatalism--degree to which an individual perceives a
lack of ability to control his future.

Source Creditability--degree of trust one places in

 

Achievement Motivation—-a desire for excellence in
order to achieve a feeling of
personal accomplishment.

 

Empathy-~3bility to gee oneself in another person's
situation.

Communication

 

The essence of the diffusion process is the human

interaction in which one person communicates a new idea

to another person. (Rogers: 1962).59

Information about a new idea may be communicated to an
individual from a variety of sources. Rogers has suggested
a dual classification of these sources on the basis of the
means used to communicate the message and the origin of the
message. The dichotomy based upon means involves the
division between mass media messages and interpersonal
("face-to-face" or "word-ofsmouth") communication. While
the mass media is an extremely rapid method of disseminating
information about new ideas, to large audiences, in an
unaltered form, it has certain limitations as an instrument
of social change. The one-way characteristic of mass media

often results in uncorrected misunderstandings of the

message, where the feedback possibilities in interpersonal

58Rogers, Peasants, pp. 302-303.

59Rogers, Diffusion, p. 13.

particular sources of information.

.. avian-n. Luv—i-mh -..

\—

 

 

40
communication permits the communicator to adjust his message
to the beliefs and level of understanding of the recipient.
Mass media also encounters the problem of "selectivity" to
a greater extent than interpersonal communications.
"Selectivity" is the psychological principle that individuals
tend to select and recall those messages which are congruent
with their beliefs and values, and ignore those messages
which are not. Mass media messages are subject to considerable
"selectivity" (such as turning off the radio), while a face-
to-face interaction between individuals has a greater
probability of overcoming this difficulty. The consequence
of these differences, according to Rogers, is that the mass
media can be an effective method of increasing knowledge of
innovations, bu: interpersonal communication is more success-
ful in ecuring changes in attitudes.

The second division is between "cosmopolite" and
"localite" sources of information. "Cosmopolite” messages
originate from outside the immediate social system, while
"localite" messages emanate from within the community.
"Localiteness" and "cosmopoliteness" may be combined with
mass media and interpersonal means of communication to form
a two—way classification of information sources as shown

below.

, ,fi..- -

 

41

 

 

 

Nature of Point of Origin
the Channel Localite Cosmpolite
Interpersonal Neighbor Extension Agent
Relative Wandering Storytellers
Mass Media Local Newspaper Radio
Wall Posters T.V.

Magazines

 

Source: Rogers, Peasants, p. 128.

Rogers has characterized the various adopter categories
on the basis of their communication behavior. The most
innovative are assumed to use mass media/cosmopolite sources
to a greater extent than other members of their community,

vhile the least innovative farmers tend to depend upon

 

 

interpersonal/localite sources.60
Adopter
Category Communications Characterists

Innovator Closest contact with scientific information
sources; interactions with other sources;
relatively greatest use of impersonal sources.

Early Adopters Greatest contact with local change agents.

Early Majority Considerable contact with change agents
and early adopters.

Late Majority Secure ideas from peers who are mainly late
majority and early majority; less use of
mass media.

Laggards. Neighbors, friends, and relatives with

similar values are main information source.

_‘

Source: Rogers, Diffusion, p. 185.

60Rogers, Peasants, pp. 124-133.

my run: w. . H

 

“N m. ./_I_ _ 1

 

42

Deutschmann and Fals Borda utilized a slightly different
classification of information sources in their study of
communication in an Andean village. Their classification is
based upon "a kind of psychological~sociological distance
dimension" hypothesizing that channels nearer to "self" would
produce later information than the channels farther from
"self" and that the more distant channels would result in
earlier adoption:

1. Egocentric or intra—personal (see‘forcmyself)

2. Intra-community personal (my neighbor said)

3. Extrascommunity personal (some stranger said)

4. Impersonal (radio, pamphlet, etc.)

These researchers found a greater dependence among
laggards upon interpersonal and intra-community channels
and less dependence upon extra-community and impersonal
channels as opposed to their more innovative colleagues.61

Although this review is, without doubt, rather incomplete,
it should be useful in outlining some of the notions used
by diffusion researchers which will be applicable to the
present study.

A Review of Risk, Uncertainty,
and Decision-Making

 

 

The variables affecting the decision-making process
of farm managers have only relatively lately become a
subject of serious interest in the study of agricultural
economics. One of the reasons for this development is
that traditionally pricesoriented economic theory of the
firm does not shed light on such questions as the

 

61Paul J. Deutschmann and Orlando Pals Borda,
‘Tmmmunication in an Andean Village" (a paper presented to
the Association for Education in Journalism Convention,
lhfiversity of North Carolina, Chapel Hill, August 27, 1962).

 

43
following: Why do some farmers adopt new technology

more rapidly than others? Why do some farmers prefer
risky undertakings . . . (Halter and Beringer: 1960).

62
However much they may disagree on other matters, most
researchers who are interested in agricultural development
have acknowledged that the adoption of new farming practices
involves an element of subjective risk to the farmer. As a
consequence, this section is devoted to examining several
views of decision-making under uncertainty in order to
identify concepts which might be applicable to the problem

of adoption of new agricultural practices in less developed

nations.

'Decision-Making With Incomplete Knowledge

 

Risk, Uncertainty, and Profit

 

Mention of Frank H. Knight's 1921 publication, Risk,

Uncertainty and Profit, usually brings to mind the definitions

 

expressed in the following paragraph:

The practical difference between the two
categories, risk and uncertainty, is that in the
former the distribution of the outcome in a group
of instances is known (either through calculation
a priori or from statistics of past experience),
while in the case of uncertainty that is not true,
the reason being in general that it is impossible
to form a group of instances, because gge situation
dealt with is in a high degree unique.

It is indeed unfortunate that such great emphasis has
been placed, in the usual interpretation of Knight's work,

upon this particular dichotomy. First of all, it contributes

 

V w

62A. N. Halter and ChristOpher Beringer, "Cardinal
Utility Functions and Managerial Behavior," Journal of Farm
W (February, 1960), p. 118.

63Frank H. Knight, Risk,;UnCertainty and Profit (Hart,
Sfimffner G Marx, 1921, Reprint: Harper Torchbooks), p. 233.

 

 

 

 

44
to the tendency to overlook Knight's efforts to understand
the process of decision-making under conditions of less than
perfect knowledge. Secondly, there is considerable evidence
that Knight intended his definitions of risk and uncertainty
to be illustrative rather than categorical.

Knight was concerned with the inadequacies of abstract
economic theory, especially the notion of omniscience among
the members of the economic system. In reference to this
assumption he comments that:

The facts of life in this regard are in a sense
obtrusively obvious and are a matter of common
observation. It is a world of change in which we
live, and a world of uncertainty. We live only by
knowing something about the future; while the
problems of life, or of conduct at least, arise
from the fact that we know so little. This is as
true of business as of other spheres of activity.

The essence of the situation is action according to

opinion, of greater or less foundation and value,

neither entire ignorance nor perfect information,

but partial knowledge.64

Clearly, Knight considered it very unlikely that
decision-makers either behave randomly in a vacuum of
complete ignorance or that they possess complete informa-
tion. Instead he suggests that they operate under
conditions of imperfect knowledge; choosing among
alternative actions on the basis of some "image of the
future state of affairs." For Knight, the mechanics of
formulating this image remained an enigma.

The ultimate logic, or psychology, of these
deliberations is obscure, a part of the scientifically

unfathomable mystery of life and the mind. We must
simply fall back upon a "capacity" in the intelligent

64Ibid., p. 199.

 

45

animal to form more or less correct judgements
about things, and intuitive sense of values. We
are so built that what seems to us reasonable is
likely to be confirmed by experience, or we could
not live in the world at all.65

He maintains, however, that there is at best a tenuous
connection between scientific inquiry and the process of
forming expectations and making decisions regarding everyday

affairs.

So when we try to decide what to expect in a
certain situation, and how to behave ourselves
accordingly, we are likely to do a lot of irrelevant
rambling, and first thing we know we have made up our
minds, that our course of action is settled. There
seems to be very little meaning in what has gone on
in our minds, and certainly little kinship with the
formal_processes oflogic which the scientist uses in

 

 

an investigation. he contrast the two processes by
recognizing that the former is not reasoned knowledge,
but "judgement," "common sense" or "intuition."

There is doubtless some analysis of a crude type
involved, but in the main it seems that we "infer"
largely from our experience of the past as a whole,
somewhat in the same way that we deal with intrinsically
simple (unanalyzable) problems like estimating distances,
weights, or other physical macnitudes, when measuring
instruments are not at hand.08
In the case of business decisions (such as expanding a
factory) and most other human conduct (choosing a career,
marriage, etc.), Knight suggests that the problems faced by
decision—makers are usually highly unique. It is therefore
cuiite unlikely that he will have experienced a sufficiently
ltxrge number of homogeneous (almost identical events from
nflrich to establish the probability of success or failure in

the present venture). As a consequence he is forced to

decide within the state of knowledge defined by Knight as

 

 

 

 

46
uncertainty (where it is not possible to calculate the
probability of various outcomes).

The essential and outstanding fact is that the
"instance" in question is so entirely unique that
there are no others or not a sufficient number to
make it possible to tabulate enough like it to form
a basis for any inference of the value about any real
probability in the case we are interested in.67

This does not mean that the decision-maker has no idea
of what to expect from various actions:

He "figures" more or less on the proposition,
taking account as well as possible of the various
factors more or less susceptible of measurement,
but the final result is an "estimate" of the probable
outcome of any proposed course of action.68

In addition to generating an "estimate” of the expected
outcome of a given action (through the "obscure process" of
consciousness), the individual decision-maker also assigns
some degree of belief (probability) to the correctness of
his "estimate." Knight insists that this procedure involves
"two separate exercises of judgement; the formation of an
estimate and an estimation of its value [probability]."

The businessman himself not merely forms the
best estimate he can of the outcome of his actions,
but he is likely also to estimate the probability
that his estimate is correct. The "degree" of
certainty or of confidence felt in the conclusion
after it is reached cannot be ignored, for it is of
the greatest practical significance. The action
which follows upon an opinion depends as much upon
the amount of confidence in that Opinion as it does
upon the favorableness of the opinion itself.6

Knight suggests an important relationship between the

 

67Ibid., p. 226.

68;pid., p. 226.

691bid., p. 227.

 

47
degree of certainty about the outcome and the gain necessary
to secure a positive response to a particular venture. In
his view the size of the possible rewards from any given
enterprise must increase in some "proportion" to the
uncertainty related to these rewards or the decision-maker
will be unwilling to engage in the venture.

We shall assume, then, that if a man is under-
going a sacrifice for the sake of a future benefit,
the reward must be larger to evoke the sacrifice if
it is viewed as contingent than if it is considered
certain, and that it will have to be larger in at
least some general proportion to the degree of felt
uncertainty in anticipation. It is clearly the
subjective uncertainty which is decisive in such
a case, what the man believes the chances to be,
whether his degree of confidence is based upon

objective probability in the situation itself 0;
in an estimate of his own powers of predictionuO

Knight suggests that there are at least three methods
by which the individual may form estimates, or degrees of
belief [a term not used by Knight] in his estimates, even
in seemingly quite unique cases: (1) by observing the
outcome of "similar" ventures undertaken by others;

(2) from the amount of success the given decision-maker has
had with dissimilar cases; and (3) on the basis of an
intuitive feeling or "hunch” regarding the perspective
outcome.71

Knight, unlike some of his interpreters, hedges his

distinction between risk and uncertainty by acknowledging

 

that there are considerable gradations of the states of

 

7°Ibid., p. 242.

7llbid., p. 237.

48
knowledge within which a decision-maker may find himself.

A11 gradations, we should say, except the ideal
extremes themselves [risk and uncertainty]; for as
we can never in practice secure completely homogeneous
classes in one case [identical events for which
probability distributions might Be calculated
a riori such as throws of a fair die], so in the other
it probably never happens that there is no basis of
comparison for determining probability .—T'. .72

 

Or:

. . . Nothing in the universe of experience is
absolutely unique any more than any two things are
absolutely alike.73
Finally, Knight believes that for various reasons

decision-makers are not a homogeneous group of individuals.
They vary in their capacity to form "correct judgements" due

to differences in the ability to perceive and infer the

course of future events. They differ in their capacity to

plan and execute ”adjustments necessary to meet the anticipated
future situation." They do not all hold the same degree of
confidence in their own judgements, nor are they presumed to

be equally willing to "take chances."

This brief review of Knight's work will provide a back-
ground for the remaining discussion of decision-making under
uncertain conditions. Additional attention will be given to
the mental process of decision—making, the state of knowledge
within which decision-making occurs, the effect of uncertainty

upon individual choices, and the differences among decision-

makers.

 

721bid., p. 227.

731bid., p. 226.

 

49
The Image
Some three and one-half decades after Knight's work,
another economist, Kenneth Boulding, set about the task of
elaborating the concept of an imagg_which would guide human
conduct. All of man's behavior, according to Boulding,
depends upon his subjective knowledge; his image of the

world.74

This image is a product of all the individual's

past experiences. It, perhaps, begins as an "undifferentiated
blur" and increases in differentiation and complexity as a
result of a constant stream of messages being admitted

through the senses. There is a necessity, argues Boulding,

to carefully distinguish between the image and the messages
that reach it:

The message consists of information in the sense

that it is structured experience. The meaning of a -
message is the change which it produceswin the image.

 

 

A message may have differing impacts upon the image. A large
proportion of messages are simply ignored (i.e., the ticking
of a clock). A second type of message may change the image
in a "regular or well-defined way that might be described as
simple addition" (i.e., learning more about price theory,
which would hardly reorganize the economist's image of the
economic system). Still another effect of messages may be

a "revolutionary change" [the term used by Boulding] in

which some portion of the image is completely reorganized.

 

74
Kenneth

Paperback, 19
75

B. Boulding, The [gage (Ann Arbor: Ann Arbor
6), pp. 1-49.

Ibid., p. 6.

...

 

50
(Perhaps an example would be Schultz's suggestion of
"efficient but poor" peasants which is said to have
reorganized a number of economists' ideas about peasant
agriculture.) One additional impact is possible. A
message may have the effect of increasing (or decreasing)
the certainty, probability, or clarity with which certain
portions of the image are held (i.e., the Cobb-Douglas
studies undertaken by several authors in less developed
countries -- as cited earlier -- may have increased the
certainty with which economists hold the concept of
"efficient but poor" peasants).

Not all messages, in Boulding's view, have equal access
to the image. Messages which are congruent with the existing
image are readily admitted, and result in either a simple
accretion in an image or no impact at all. Messages which
are contrary to the image tend to be rejected, unless they
are extremely forceful or are repeated too often to be
ignored. The acceptance of such a message may result in a
major reorganization of the image. Boulding suggests that
there are two reasons why messages receive unequal access to
the image. First, the image tends to have a built-in
stability or resistance to change. Boulding is rather vague
as to the basis of this stability, asserting only that it
is the result of an internal consistency or arrangement and
that "some kind of principle of minimization of internal
strain" is at work. Secondly, the image, in total, is said

to consist of both an image of fact (loosely defined) and an

51
image of value. The messages received by the image are,
according to Boulding, "mediated" through the individual's
value system (hence there are no "facts" as such). The
value system, therefore, has considerable impact upon the
interpretation of incoming messages. But still further, the
value system contributes to a resistance to such messages as
might be considered "bad" or "hostile" given the existing
image.

To describe the effect of the image on organizing
behavior (the reason for its existence), Boulding begins with
a discussion of the thermostat, which he believes expresses
the concept of an image in its most rudimentary form:

The thermostat has an image of the outside

world in the shape of the information regarding its

temperature. It also has a 'alue system in the sense

of the ideal temperature at which it is set. Its
behavior is directed toward the receipt of information
which will bring its image and value system together.

When its image of the outside world is ”right," that

is, conforms to its value system, it ceases to act.

As long as the image, as confirmed by messages received,

does not conform to its value system, it acts to bring

the other two together.76

This, in the simplest possible form, describes the
essence of the image's effect on behavior. Boulding believes
that all biological systems possess images, of lesser and
greater complexity, and, like the thermostat, attempt to
organize themselves in response to messages received from
outside the system. Man stands at the top of the biological

ladder, not because of improved sensory perception, but

Vbecause of his capacity to organize information into "large
7°Ibid., p. 22.

 

no b5
.‘ ob

.a,,.‘.
u" p-
....5

n._.

.g_

'v

 

 

52
and complex images." He has the ability to locate himself
in the world temporally and spatially (as a result of his
ability to communicate and record information). He is able
to observe various relationships and to infer cause and effect
from them. His image is characterized by greater degree of
self-consciousness and self-awareness than the lower animals.
There is also a reflective character to man's image: not
only does he know, he knows what he knows. Finally, man's
image has the capacity for internal growth quite independent
of external messages. His image, therefore, contains not only
what is, but what might be. This power of "imagination"
allows him to contemplate the world of potentialities,
evaluate them according to his value system, and choose the
"best" from among them.

The economist, criticizes Boulding, has not given
sufficient attention to man's image and its effect upon
economic behavior:

The economists have badly neglected the impact
of information and knowledge structures on economic
behavior and processes. There are good reasons, or
perhaps one should say excuses, for this neglect.

With deft analytical fingers the economist abstracts

from the untidy complexities of social life a neat

world of commodities. It is the behavior of
commodities, not the behavior of men which is the

prime focus of interest in economic studies.

He is aware, of course, at the back of his mind,
that prices, outputs, etc., are in fact the result of
human decisions. He likes to reduce these decisions,
however, to a form as abstract and manageable as 77

possible. Commodities are simpleminded creatures.

Even so, Boulding finds an implied concept of man's

77Ibid., p. 82.

53
image even in the simplest of economic theories.

Economic behavior is conceived as the process of
"maximization." Economic man is supposed to be capable
of at least three processes involving an image. In the
first place he is supposed to be conscious of the
alternatives which lie before him . . . Or in the
simplest form, we suppose his mind to be much like a
department store full of images of commodities, each
with a convenient price tag attached.

In addition to the image of alternatives economic
man is also supposed to be able to give a value ordering
to all relevant alternatives, that is all parts of the
image. Not only do the combinations of his mental
department store have price tags, they have utility
tags . . . .

His final task after imagination has performed
these labors of Hercules is a simple one. All he needs
to do is scan all possible combinations which are open
to him and all his alternative acts, rank them in order
on the parade ground of value, and pick out the top of
the class.78

When we introduce the difficulties of uncertain outcomes into
the problem, Boulding perceives the economists' usual
assumptions of the image on economic man as even more
incredulous:

The economists have tried to deal with the
problem of uncertainty by supposing that each of the
alternatives in our image presents to mind not only
with utility tags attached but also with whole
probability distributions. Economic man, clever
fellow that he is, now maximizes the expected value
of his acts, a feat of mathematical agility which
would take centuries of experience and enormous
electronic calculators to perfect.79
Having poked some good-natured fun at the traditional

constructs of economic man, Boulding submits his "revised
laws of economic behavior" (which as will be seen are not a

great improvement over the originals). According to Boulding,

 

781bid., p. 83.

791bid., p. 35.

 

 

 

 

a.
>4
o,‘\_

 

 

 

NW

54

man carries on his economic activities of today much as he
did yesterday, for in doing so he can be reasonably certain
of the outcome. (Although he doesn't indicate what made us
do what we did yesterday.) The further one diverges from
his habitual behavior the less certain he will be about the
outcome. The larger the negative value he attaches to
uncertainty "in his value orientation," the less likely that
he will abandon the familiar and known; the more likely that
he will do today as he did yesterday. At this point
Boulding suggests his first revised law of economic behavior:
"we will do today what we did yesterday unless there is very
good reason for doing otherwise."80

The second revised law is "that the good reasons which
are necessary if we do not do today what we did yesterday
are derived mainly from dissatisfaction with what we did
yesterday or what happened yesterday." Boulding contends
that changes can occur in the value structure during the
course of repetitions of habitual behavior (although he is
unclear how) until a point is reached where the misery of
contemplating present behavior overcomes the uncertainty
associated with a different behavior, and suddenly a
reorganization of economic activity is undertaken (i.e., a
new job, a new business, a new location, etc.). Another
possibility is that messages we receive (from the ticker

tape, the Wall Street Journal and, perhaps more importantly,

 

‘W’word-of-mouth) may cause us to revise our images of the

801bid., p. 86.

SS

alternatives and consequently select a different course of
behavior.81

This, unfortunately, is the extent of Boulding's
reformulation of economic behavior. It would seem that the
major contribution of his discussion is an insistence upon
the decision-making process as an interaction of the
individual's knowledge structure, value structure, and the
information entering this system. Thus, economic decision-
making is not simply a choice between alternatives -- of
which the outcomes are either perfectly or imperfectly known --
but instead is a choice of a new pattern of behavior over
the habitual pattern of the past as a result of an altered
perception of the alternatives.

Aspirations and Satisficing

 

In Simon's view, economic behavior can be conceptualized
in the same framework as used by psychologists to understand
other forms of human behavior: the motivating force
of certain drives and their ultimate satisfaction.

In most psychological theories the motive to act
stems from drives, and the action terminates when the
drive is satisfied. Moreover, the conditions for
satisfying a drive are not necessarily fixed, but may
be Specified by an aspiration level that itself adjusts
Upward and downward on the basis of experience.82

Applying these concepts to a firm, he suggests that a certain

level of profit might be set as a goal and that the firm

 

811bid., p. 93.

82Herbert R. Simon, "Theories of DecisioneMaking in
Econcmdcs and Behavioral Science," American Economic Review,
(June, 1959), p. 263.

nub

l V

.Hlb

“v

.»5

..-

.x»

 

56

would select an action, from among known alternatives,
which would fulfill this goal ("level of aspiration"). The
concept of profit "maximization" is replaced by a concept of
"satisficing" since the firm is not examining all of the
possible alternatives in order to choose the "most"
profitable action, but instead selects one from among known
alternatives providing it either meets or exceeds the minimum
profit criterion set as its goal. If none of the kpgwg
alternatives is "satisfactory" (meets or exceeds the firm's
goals), two possible actions may be taken: (1) a search is
initiated for a "satisfactory” alternative; or (2) the level
of aspiration is adjusted downward. If the process of locating
a "satisfactory" alternative or the downward adjustment of the
aspirational level to coincide with actual performance proceeds
too slowly, Simon believes that emotional behavior (apathy
or aggression) will very likely replace " ational behavior."83

Katona generalizes that an individual's level of
aspiration is the product of his past experiences and his
reference group. Past successes act as a stimulus to
aspiration levels, while "failure, disappointment, frustration
tend to lower the level of aspiration." Katona continues by
noting that not only our own successes and failure, but the
success or failures of our reference group impinge upon our
aspiration level:

If a person grows up in a small mining town in '

which his father, brothers, friends also grew up, he
will be strongly influenced to accept being a miner

83Ibid., pp. 253-283.

57

as his lot, and not aspire further. Thus his

expectations and aspirations will be limited, not

by any failure on his part, but by the limits of

his group.84

While accomplishment is assumed to breed striving toward
higher levels of accomplishment, Katona extends his argument
to hold that "apparent lack of motivation or absence of
striving toward a goal is usually the result of frustration."
Failure and disappointment may cause a person to be convinced
that his income and advancement is dependent upon others, or
upon luck, and not the result of his own effort and

activities [fatalism].

Knowledge Situations

 

Knight's definitions of risk and uncertainty are viewed
by Glenn L. Johnson as an unsatisfactory categorization of
the actual states of knowledge encountered by decision-

makers.85

Believing, as did Hardysé, that the conditions
of risk and uncertainty differ largely in the amount of

information at hand, Johnson proposes a classification of
"knowledge situations" which emphasizes the importance of

learning. He suggests an analogy between the process of

 

84George Katona, Psychological Analysis of Economic
Behavior (New York: McGraw—Hill Book Co., Inc., 1963),
p. 92.

85Glenn L. Johnson and Curtis F. Lard, "Knowledge
Situations," Manggerial Processes of Midwestern Farmers
(Ames: Iowa State University Press), pp. 41-51.

86Johnson and Lard cite Hardy in their discussion:
C. O. Hardy, Risk and Risk Bearing (University of Chicago
Press, 1923), p. 54.

 

 

 

 

S8

decision-making under uncertainty and the statistical
technique of sequential analysis. Using the sequential
analysis procedure the analyst draws a small sample and
from this sample determines whether to accept or reject
the hypothesis, or, as a third alternative, he may decide
that it is necessary (profitable) to gather additional
information before a decision is made. Johnson's scheme
of classification assumes that decision-makers behave in
much this same manner, deciding on the basis of the
subjective costs and returns of additional information
whether to commit themselves to acceptance or rejection of
a proposal or to seek still more information. The know-
ledge situations defined by Johnson are as follows:

I. Subjective Certainty-~"the manager regards his

information as so gdod that he need not take
precautions against being wrong.”

 

II. Subjective Uncertainty
A} Risk action, a situation in which a manager

regards present knowledge as adequate for
making a decision and in which the cost of
additional knowledge is exactly equal to its
value. Risk actions may be either:
(a) positive, or
(b) negative.

 

 

B. Learning, a situation in which a manager
conSiders his present knowledge inadequate
for action in the sense that he is subjectively
unwilling to decide and take the consequences
for his errors which he might make and in
which the costs of acquiring more knowledge

is less than its value.

C. Inaction, a situation in which a manager
regards his present knowledge as inadequate
for action and in which the cost of more
knowledge exceeds its value. In this situation,
no action is taken and no learning occurs.

 

59

D. Forced action, a situation in which a manager's
information is inadequate for him to be ready,
willing and able to make a decision subject
to the errors involved but in which some
outside force makes it necessary for him to
act. Forced action decisions were regarded
as either:

(a) positive, or
(b) negative.87

 

In the course of empirically examining these knowledge
situations, researchers "UHCOVGISL" what they believed to be
still another classification, involuntary learning.

Involuntary learning was preposed to cover those situations

 

in which farmers indicated that they continued to learn

even after a terminal decision had been made:

 

E. Involuntary learning, a situation wherein the
manager is subjbctively unwilling to learn more
since the costs of additional information equals
or exceeds its value to him, but in which some
outside force makes it necessary to learn or
for some learning to occur regardless of the
volition of the manager.88

Johnson's contribution is that it directs our attention
toward the importance of learning in altering the states of
knowledge within which the decision-makers select their

actions.
Cardinal Utility

While Boulding has been arguing that "maximization of
utility" is a much too formidable task for the decision-maker
to ever accomplish, other economists have been investigating
the usefulness of the concept of "cardinal utility" in

explaining economic behavior. The renewed interest in this

 

87Johnson‘and Lard, 9p. cit., pp. 44‘45.

 

881bid., p. 53.

60
concept is usually attributed to Von Neumann and Morgenstern's
development (in the mid-forties) of a method for numerically

measuring utility.89

The word "renewed" is appropriate
since, historically, the first expressions of utility were
in cardinal terms. Between then and the recent past, the
concept of utility as a number was largely displaced, within
the economist's theoretical constructs, by an elaborately
developed theory of ordinal measurement -- analysis of
indifference curves. The reason behind this change in
emphasis was the problems involved in quantifying a concept
which has at its basis the human sensation of preference --
desire for one item "more" than another. While it was
difficult to devise an intuitively satisfactory method of
assigning numbers to "how much" one item was preferred to
another, it was quite obvious that preferences could be
observed by directly ranking items (and groups of items).
Von Neumann and Morgenstern contend that ”under the
conditions on which indifference curve analysis is based,
very little extra effort is needed to reach numerical
utility."

Von Neumann and Morgenstern propose a situation in which
an individual has preferences among A, B, and C. By a
"very natural extension of this picture" they permit the
individual to not only compare events, but to compare
combinations of events with certain probabilities attached.

If the individual prefers A to B and also to C, he will

 

89John Van Neumann and Oskar Morgenstern, Theory of
Games and Economic Behavior (New York: John Wiley and Sons,
Inc., 1944).

 

 

61

clearly prefer A to a 50% probability of B and a 50%
probability of C. (A, B, and C must be considered
exhaustive in order to avoid the possibility of
complementarity between them.) But if he prefers C to A
and A to B, a preference of A over a 50-50 combination of
B and C contains some "fundamentally new information." This
would indicate that his preference for C over A is less than
his preference for A over B, or, more generally, the
differences between utilities have now become measurable.
It is from this modest intuitive argument that Von Neumann
and Morgenstern develop their method of numerically
measuring utility.

The Von Neumann-Morgenstern scheme can be operationalized
by asking an individual to state the probability level "u”
at which he would be indifferent between the uncertain
outcome qu + (l -u)X2 and a certain outcome X3 (where
X1>X3>X2). Arbitrarily setting numerical utility values
for X1 and X2 permits the computation of a numerical utility
for any given level of X3.

There are two common criticisms of this approach:
(1) a real-world individual may not understand the concepts
of probability; and (2) he may have an aversion to gambling
framework utilized in this line of questioning. These
problems, however, can be largely circumvented by slight

Immiifications in the Von Neumann-Morgenstern approach.90

 

90R. R. Officer and A. N. Halter, "Utility Analysis in
alPracfljcal Setting," American Journal of Agricultural
Economics, (May, 1968), pp. 257-277.

62

In a 1960 article, Halter and Beringer suggest "that
the Von Neumann-Morgenstern index for measuring cardinal
utility may explain certain aspects of managerial behavior
which are largely ignored by the traditional product-type
and factor-cost oriented theory of the firm." They derived,
using a method which did not require the respondent to have
a knowledge of probabilities, utility indexes for a number of
farmers. Halter and Beringer discovered a strong relationship
between the farmer's utility for increased income and the
type of farming he was engaged in, with higher utilities for
income being associated with those operations which are
commonly considered to be more risky (cash crops and fat-
stock feeding). Farmers with a high utility for increases in
income were also more likely to incur large debts than
farmers with a lower marginal utility for wealth. Low-risk
enterprises (dairying, general farming) were found to be
highly correlated with a large marginal disutility for losses.
It was also found that farmers with lower net worth positions
and lower gross incomes tended to have a greater disutility
for losses than their wealthier counterparts. Finally, it
was discovered that the size of gains sufficient to induce
the acceptance of unfair odds was at least 26 times the losses
necessary to induce acceptance of insurance schemes at unfair
odds.91

Somewhat later, Officer and Halter examined the

 

91Albert W. Halter and Christoph Beringer, "Utility .
Functions and Managerial Behavior," Journal of Farm Economics
(February, 1960), pp. 118-132.

63
contribution of cardinal utility functions in predicting the
decisions made by Australian wool producers. The main
hypothesis of their study was "that farmers' operational
decisions are more consistent with a criteria of minimizing
expected disutility than with a criteria of minimizing costs."
In addition they were interested in determining if useful
utility functions could be derived under field conditions and
whether or not these functions were stable over time. In
pursuit of their objectives, Officer and Halter conducted two
experiments, separated by a year, involving five farmers.
They derived utility functions for each of these farmers and
then asked them to choose alternatives (fodder reserve
levels) within several programmed situations. Predictions
of the farmer's choice, on the basis of his expressed utility
functions, were next compared with his actual choices.
Officer and Halter, using a rather crude method of analysis,
found that cardinal utility functions were somewhat helpful
in predicting farmer decisions. In addition, they found
that over a period of one year the farmers' decisions
(confronting them with the same situations) and utility
functions did not change radically.92
Friedman-Savage Utility Function
Friedman and Savage believe that decision~makers behave
"as if they were maximizing expected utility provided that

"a rather special shape is given to the total utility curve

 

V V v ‘ f v \ Vva f

92R. R. Officer and A. N. Halter,‘gp. cit., pp. 257-277,

 

64

for money."93 They consider the question of whether or not

individuals actually calculate and compare expected utilities

 

as largely irrelevant. What is important, according to these
authors, is whether or not a particular hypothesis is a
"sufficiently good approximation to reality for the purpose
at hand."

The "purpose at hand" in the present discussion is an
explanation of the apparent tendency for individuals to
both, and at the same time, insure against losses and engage
in games of chance. Gambling and insuring are the focal
points of their discussion because of the obvious influence
of risk on these forms of behavior. Friedman and Savage
believe that the model they have developed to understand
these phenomena may also be applicable to other forms of
decision-making such as occupational and business choices.

The "special shape" of the Friedman-Savage utility
function is based Upon a (graphical) proposal of the
necessary conditions underlying an individual's desire to
either avoid or accept risks. To examine these conditions
let us imagine the curve CDE (in Figure la and lb below) as
representing an individual's utility for all certain incomes
between 11 and 12. Next assume that our decision-maker is
presented with the chance A=[uI1 + (l-u)12] where the
probability u is set such that acturial value is II The

expected utility of A, U'= [uU{11) + (l-p) U(IZ)], is equal

 

93Milton Friedman and L. V. Savage, "The Utility Analysis
of Cfiuoices Involving Risk," Journal of Political Economy_
(Volume 56, 1948), pp. 279-304.

65

Utility (U)
Utility (U)

 

 

 

 

 

 

 

 

 

 

 

 

U(I1) U(I*)U U(I7) — U(I*) U(I2)
Il 1* I 12 I1 I 1* I2
Income (I) Income (I)
[a] [b]

Figure 1.--Illustration of Utility Analysis in Choices
Involving Risk. (Friedman and Savage, p. 290.)
to F [on the chord CFE drawn from U(Iz) to U(I1)]. 1* is
defined as the level of certain income which has the same
utility as the uncertain situation A, U(I*) = 5) It should
become obvious at this point that [a] describes a situation
in which the individual is willing to pay something for
certainty (buy insurance), since his utility for a lower
certain income (1*) is equal to his utility for the greater
expected (acturarial) income (I). In fact, he would be
willing to pay a maximum amount of T-I* in order to be
assured of the income 1* rather than take his chances on
either I1 or I2. The reverse is true of {b}. Here the
individual prefers the chance of either I1 or Iz'to any
Cmrtahiincome less than 1* (which is larger than the

expected income T). The premium he would be willing to pay

...—

r "thfi 4_ _ W—d”

66
for the gamble (A) is a maximum of I*-T.

In order to rationalize both gambling and insuring by
the same individual, Friedman and Savage suggest that the
shape of the utility function should be a combination of the
types discussed above (see Figure 2). Note that the lower
range of the utility function is convex from above while
the upper range is concave from above. Locating the
individual's present income at 10 (upper part of the convex
segment), we find that he would be inclined to insure
against losses since a chord drawn from U(IO)’ at D, to the
utility of a lower income would lie below his utility
function (for certain income). 0n the other hand, a chance
of II or I2 with an expected value of T is equally as
acceptable on a certainty of ID. If the expected value were
any amount greater than T, say T', the individual would be
willing to pay the difference between I0 and T' for the
privilege of gambling. Given a utility function of the form
presented, individual behavior involving both gambling and
insuring (at the same time) can be visualized as consistent

with utility maximization.

 

 

 

 

 

 

‘ E
E
>~
4.)
9H
I: F
:‘3 ('1)
11T10 I2 Income (1)

Fiwne 2.--Friedman-Savage Utility Function.(Friedman and
Savage, p. 295.)

 

 

67
Friedman and Savage have presented a seemingly useful

way of viewing utility functions for income. However, they
had little to say about the minimum income levels applicable
to the utility function, except that all values of U(I) will
not be described since the minimum income an individual can
have is a negative income equal to the amount he can lose in
any period. Other authors have had more to say about utility
and minimum incomes.

Focus-Loss

 

Focus-loss might be described as some minimum level of

 

outcome which constitutes the borderline between tolerable and
intolerable outcomes. The decision-maker, according to this
concept, will not willingly accept alternatives which might
result in outcomes falling below this minimal level. This
idea has been used to conceptualize farmer decision-making in
both modern and peasant agriculture. Schickele utilized just

such a premise in his discussion of the Great Plain's farmer's

94

adaptation to income uncertainty. He describes the

motivations of farmers in the following way:

The farmer's end is not simply to "maximize net
income" as it is usually assumed in analyses of the
firm. Let us break down the general goal of income
maximization into two more specific ones: (a) to
assure the farmer's survival in the case of heavy
risk losses (whenever they might hit) and (b) to
maximize income over time subject to (a). They are
not coordinate; the survival end has priority over
the maximization end.95

4Rainer Schickele, "Farmer Adaptations to Income
Uncertainty," Journal of Farm Economics, (August, 1950),
PP- 356-374.

9
51bid., p. 362.

 

 

 

 

 

 

68

In short, Schickele assumes that farmers will manage their
operations in such a way as to minimize the possibility that
their incomes would fall so low as to cause discontinuance of
the enterprise. He reasoned that farms with a low level of
assets were less capable of withstanding risk losses than
their larger counterparts. In support of this reasoning
Schickele found that small farmers organized their
operations in a way that did not maximize expected (long-run)
profits, but did provide for continued survival (diversifica-
tion, off-farm employment, etc.). Large farmers, by
contrast, more often specialized in the riskier high profit
activities in a manner consistent with maximization of
expected profits.

Wharton observed that the concept of focu5*loss could
be very useful in develOping a "neglected" explanation of
the peasant farmer's reluctance to accept new agricultural

96

practices. He believes that the low levels of income

associated with subsistence and semi-subsistence farmers
produce a much stronger "survival element" in decision-

making than would be found among their commercial counterparts.
Each community, according to Wharton, has a socially-

prescribed minimum standard of living -- Sms -- which

farmers are unwilling to fall below. (This minimum income

level is usually above the physiological minimum, but it may

‘

96Clifton R. Wharton, Jr., "Risk, Uncertainty and the
Subsistence Farmer: Technological Innovation and Resistance
t0<3pnge in the Context of Survival," (Paper presented at
thetmunt Session American Economic and Association for

ggmpzlarggive Economics, Chicago, Illinois, December 28, 1968),

69
nonetheless be insufficient for adequate diets in the poorer
communities.) The typical peasant family will strive to

reach some communal or societal achievement standard(pf living)--

Sas -- but more importantly it will struggle to avoid the
minimum subsistence standard -- Sms' Because actual levels
of living -- La -- are often in such close proximity to the

minimum acceptable standard (Sms), the peasant farmer is

 

particularly vulnerable to variations in income (due to
variations in costs, physical output, and prices).

The very survival of subsistence farmers, in Wharton's
view, has been the result of their ability to select
traditional technologies which have a small variability.

From considerable experience, sometimes handed down from
generation to generation, the subsistence farmer has learned
what to expect from traditional technologies. This is not

so with new technologies, implying that the peasant must base

his decisions upon subjective expectations. If the

 

subjectively evaluated variance associated with a new

 

technology is sufficiently large to result in incomes below

 

the minimum subsistence standard, the peasant will reject

the new method (as too risky) despite its superiority in
average returns. Wharton illustrates this situation with the
diagram in Figure 3. Note that while the expected outcome --
IKOT) -- for the traditional method is less than the

expected outcome -— E(O -- for the new technology, one

N)
standard deviation from mean outcome would still permit a
I; level of living above the minimum standard. At the same time,

i aStmwhrd deviation from the expected outcome of the new

70

 

 

Traditional New
‘ Technology Technology
i b--—-—------+O,
F
+O'I‘ >—-—--------""i
E(ON)
E(OT)

-OT _-----——---q

 

 

 

... / 7/14 f3;
/ //

Figure 3.-- Wharton's Diagram Representing the Differences
in "Expected" Outcomes from Traditional and
New Technologies. (Wharton, Risk, Uncertainty,
and the Subsistence Farmer, p. 40.)

 

 

practice (based upon the individual's subjective estimate of
variance) would result in the "intolerable" circumstance of
an income below the minimum standard. If somewhat less
variance were expected from the new technology, it would
stand a much better chance of being accepted, according to
Wharton.

The focus~loss concept has also been applied in empirical
research. In an attempt to study vegetable producrion in
Southern France, Boussard and Petit soon discovered that it
"was impossible to neglect farmers' reactions to extreme

price uncertainty." Confronted with this problem, they

71
developed a technique to include consideration of the
farmer's aversion to risk in their linear programming model.
At the foundation of their method was the assumption "that
farmers choose, among various actions, the one which will
nmximize the expected gain, provided that the possibility
of [212.15 so small that it can be neglected." Boussard
and Petit defined ruin in terms of income, at a level below
which the farmer would not risk having his income reduced.
The largest acceptable loss was, in turn, derived from the
nfinimum income level. Next the researchers asked extension
personnel to indicate the level of loss, for various crOps
that they would be "very surprised” to see occur. Having
Obtained a concensus of the amount of such losses (which is
an expression of the riskiness of various crops), Boussard
and Petit added the constraint to their linear programming
model that the maximum loss from the iii crop (the amount
of loss for a given crop which would be considered "very
surprising" times the acreage devoted to that crop) could
never exceed one-third of the largest acceptable loss. (in
other words, the "worst" conceivable outcome for a
particular crop could not amount to more than one-third of
the difference between the expected and minimum income levels.)
With this addition to their model the authors were able to
nearly duplicate the cropping patterns of farmers in the

area studied.97

 

7Jean Marc Boussard and Michel Petit, "Representation
oflmrmer's Behavior Under Uncertainty with a Focus-Loss
Unstraint," Journal of Farm Economics, (November, 1967),
PP. 869-880.

 

 

72

Sturt concluded from his study of 200 West Pakistani
farmers that "the capacity for risk-bearing appeared to be
a major factor affecting change." "Smaller cultivators,"
among those studied, failed to make as many changes in their
farm operations as "larger cultivators." (Recall Schickele's
explanation of this same phenomenon among Great Plains
farmers.) In addition, Sturt found that the amount of
change undertaken by these farmers was also related to the
availability of irrigated land, the lack of credit and the
98

unavailability of new inputs.

Sghematic Model of the Farm Manager

 

Nielson has prepared a schematic representation of the

farm manager.99

Nielson's model is almost self-explanatory.
The items subsumed by V symbolize the manager as an

individual with a certain configuration of background

 

experiences -- V1 -- directed by certain drives and

motivations ~- V -- which are monitored by the value

2
system (an effort to bring ”values" and "fact" together)
and, who possesses a particular endowment of talent and
capabilities (a critical item that Boulding omitted from
his model) -- V . P signifies the entirety of the complex

3
mental processes involved in decision—making. Past

 

98Daniel Sturt, "Response to Change in Pakistan,"
Journal of Farm Economics, (August, 1965), pp. 625-633.

99James Nielson, Aspects of Management of Concern to
Basic Researchers (Denver, Colorado: Farm Management
Research Committee of the Western Agricultural Economics
Research Council, 1962).

 

73
Antecedents Outcome

V P O

f V \r \

 

 

    
    

Drives and
Motivations

  
 

    
 

0
Outcome

Mgr'l
Success

P
Process

Mgr' 1
Behavior

    
 
   
 

  
 
 
 

     

 

     

  

Capa-
bilities

  

Figure 4.--Nielson's Model of the Farm Manager. (Nielson,
Aspects of Management of Concern to Basic Researchers)
experiences, motivations, and capabilities all have an
impact upon these processes, which -- as evidenced by
Knight, Boulding, and Nielson's reluctance to elaborate in
greater detail -- are apparently very difficult to
conceptualize. The outcome of economic decision-making, O,
is connected by "feedback" linkages to both capabilities
and drives and motivations. It is unclear as to why feed-
back is not also included between outcome and biography
(the set of past experiences). Still another omission, in
Nielson's schematic view of management, is provision for

messages from outside the system.

 

 

74
Summary

This chapter has examined various segments of the
literature for ideas which would be useful in conceptualizing
the adoption of new agricultural practices as decision-making
under conditions of uncertainty. A review of the controversy
over the importance of non-economic as opposed to economic
factors in determining peasant behavior has led to the
conclusion that both should be examined in a comprehensive
view of the adoption process. The question of greater
interest seems to be the relative weightings that should be
given to economic and non-economic variables as they interact
within the decision-making process.

Also included was a brief discussion of some of the main
concepts which have emerged from "diffusion research." Of
particular interest in this study were the distinguishing
characteristics of innovative individuals, the categorization
of channels of communication, and the characteristics of an
innovation. Variations of these concepts will provide part
of the complement of notions integrated into a conceptualiza-
tion of the adoption process as it evolves in later chapters.

Finally, a search of the literature on economic behavior
under uncertainty has provided several ideas which will serve
as a basis of a conceptualization of the adoption process.
Decision-making is viewed as taking place under varying
states of incomplete knowledge, the exact degree of
incompleteness depending upon the amount of knowledge the

decision-maker has been able to gather about the proposed

75

action. The views of several authors regarding the decision-
nmking process suggests that it involves an interaction of
the individual's knowledge from past experiences, his value
system, messages received from his environment, and his own
intellectual and managerial capabilities.

Knight has argued that a relationship exists between
the uncertainty perceived by the decision-maker and the
return necessary to induce him to accept the risks involved.
Friedman and Savage elaborated this relationship in terms
of a utility function which was designed to explain the
individual's willingness to accept certain risks and reject
others. Meanwhile, Schickele and Wharton have proposed
explanations of the unwillingness of poorer farmers to accept
the risks readily taken by relatively wealthier farmers.
These notions will be observed as the underlying arguments
in the conceptualization of the adOption process as presented

in the following chapters.

CHAPTER III
A CONCEPTUALIZATION OF FARMER ADOPTION

OF NEW TECHNOLOGIES UNDER
CONDITIONS OF UNCERTAINTY

A discussion of uncertainty is not "theoretical
, and impractical;" the major decisions of individual
1 persons and nations fall in this framework.
: [E. o. Heady1100

This chapter will attempt to evolve a conceptual frame-

work capable of bringing the study of adoption of agricultural
innovations under the umbrella of decision-making in an
uncertain environment. The forthcoming proposal is basically
an open and interactive model of an individual decision-
maker in an environment of uncertain outcomes. Within this
environment, the decision-maker is exposed to both
information which is useful in decision-making and
circumstances which impose constraints on the final form
that his decisions can take. The general framework of this

conceptualization is outlined in the succeeding sections,

while subsequent chapters examine its implications.

Sources of Uncertainty

 

Although it might appear sufficient to simply consider
the subjective risk involved in adopting new agricultural
practices as the product of uncertainty about their eventual
<nucomes, an interactive model would seem to suggest a
cfleser examination of the uncertain environment in which

‘flu>decision~maker operates. Since several factors are

__
1

100 ‘ . . o g o i .— , .- . o.
-‘ harl O. Heady, Economics of Agricultural Production

Mfllhxource Use (Englefibod Cliffs, New Jersey: Prentice
hil.1nc., 1952), p. 439.

76

 

 

l
‘i
A

 

 

77
presumably given consideration in the farmer's decision
Inocess, several sources of uncertainty are likely -- in
one form or another -- to have an impact upon his ultimate
decisions. The Interstate Managerial Study, conducted in
1956, suggested that farmers in a modern agriculture
required five types of information to engage in decision-
nmking: information about (1) production methods (current
technology), (2) new technology, (3) human relations,
(4) the institutional setting, and (5) prices.101 By
implication, these areas are sources of uncertainty in
United States agriculture. It seems reasonable that these
could also be sources of uncertainty in developing

agricultures.

Current Technology

 

102 103

It can be argued, following Schultz and Wharton ,
that farmers in a traditional agriculture have very well
developed expectations about the outcomes of their present
methods. Yet in the most traditional conditions conceivable,
uncertainty will surely exist; if for no other reason than
the vagary of nature. For even though farmers may possess
accurate expectations (as the result of years of experience),

the outcome of any given year remains uncertain due to

weather and biological factors such as insects and disease.

101Managerial Processes of Midwestern Farmers, op. Cit-:
pp. 26-27.

102

 

Schultz, op. cit., pp. 30-31.

 

103Whart0n: "Risk, Uncertainty and the Subsistence
Farmers" I) . 33 .

78
While the uncertainties imposed by weather variations and
other natural hazards vary from region to region, it would
be difficult to name an area where they are not an important
consideration. Within a changing, and hOpefully modernizing
agriculture, the uncertainties of the marketplace and new
technologies will likely be added to those of nature.

New Technologies

 

New technology is a much greater source of uncertainty
than traditional technology (current production methods).
Accurate expectations of outcomes from its use cannot be
immediately established. The farmer will question whether
the results obtained at experiment stations —- domestic or
foreign -- can be duplicated with his soils, climate, and
farm organization. Initially, he will lack sufficient
information to confidently evaluate a new technology. Heady
notes that extension of new technologies has usually
emphasized the improvement in the mean outcome over current

104 To asserts, as does Whartonlos, that the

methods.
variance of outcomes is also critical to decision-making.

Reduction of uncertainty about the performance of suggested
new technologies requires time for learning on the part of

the farmer.

Human Relations

 

Uncertainty about a new technology can result from a

quite different source, uncertainty in the area of human

104

P

Heady, op. cit., p. 443.

 

105Wharton, "Risk, Uncertainty and the Subsistence
Farmer," p. 36 .

79
relations. Byrnes isolates farmer uncertainty regarding
the competency and motives of change agents as an important

106 It is also

reason for failure of certain innovations.
conceivable that concern about peer group acceptance may cause
potential innovators to be reluctant to try "something
different." Rogers believes those less encumbered by concerns
of group acceptance have a tendency to accept new ideas more
rapidly.107

Institutional Arrangements

 

Institutional arrangements can be a source of uncertainty
whether the farmer uses traditional or new technologies. No
nation seems to be without some government program which
affects agriculture. Farmers may have incomplete knowledge
of these programs, and probably of greater importance, they
may be unsure of what programs to expect in the future. How
directly uncertainties regarding government programs --
present and future -- affect producer decision-making is a
matter of individual cases. Specific government programs
are not, however, the only institutional arrangements which
may be important sources of uncertainty. The institutional
framework governing land tenure, specifically landlord-
tenant relationships, is often identified as a source of
considerable uncertainty. The structure of the credit

system might be thought of as a possible source of

 

106Francis C. Byrnes, Some fissingVariables‘jnDiffusion
Research and Innovation Strategy, (New York, Agricultural
lmvelopment—Council, Inc.).

107

 

Rogers, Diffusion of Innovations, pp. 193~207.

 

80

uncertainty (”life and death" power of local money-lenders,
land as collateral on small loans, refinancing practices),
just as the risks of agricultural production are thought of
as limiting credit usage. Still another source of
uncertainty is the availability of new inputs. Untimely
availability and insufficient quantities of new inputs is
often given as a reason for their lack of use.
Prices

The uncertainty of producer prices is of considerable
importance to those farmers participating in the market
economy. In addition to causing difficulties in organizing
production, widely varying prices will likely have an
inhibiting effect on the innovativeness of farmers. Varia-
tions in prices can be classified into three categories:
(1) uncertainty about the price level of a given product for
particular years (inter-year variations in average prices),
(2) price fluctuations within a given season (intra-year
variability), and (3) price differentials between areas.
It seems almost too obvious to mention that the accuracy of
farmers' price expectations depend upon the functioning of
the relevant market system (market organization and "market

rules") and the variation in quantities supplied and demanded.

Expectations

 

By expectation I mean the act of creating imaginary
situations, of associating them with named future
dates, and of assigning to each of the hypotheses
thus formed a place on a scale measuring the degree

81
of our belief that a specified course of action on

our part will make this hypothesis come true.
(G. L. s. Shackle)108

As noted already, there are several sources of
uncertainty facing the farmer. Consequently, he often has
no way of knowing exactly some future outcome, and must
conjecture what will occur. The farmer is guided in
estimating future events by a set of expectations based upon
his judgment and his knowledge of the factors affecting given
outcomes. The formation and characteristics of such
expectations will be of major concern in this section. At
least momentarily, we will set aside the complexities of
dealing with several sources of uncertainty and concentrate
on a more abstract model.

For the purposes of this study, uncertainty will be
considered as any situation where the farmer does not know
for certain (probability equal to one) what outcome will

occur.109 It also seems reasonable to consider expectations

 

1086. L. S. Shackle, Expectations in Economics, (Cambridge:
Cambridge University Press, 1949), p. 10.

109It is the contention here that nothing is added to
our knowledge by attempting to separate uncertainty into
different categories such as those suggested by Knight and
elaborated by Heady. It will be recalled that Knight has
prOposed a conceptual division of uncertainty into "risk" --
where the probability distribution of outcomes is known --
and "uncertainty" -- where probability distributions do not
exist (Knight, p. 333). Heady bases his classification on
Knight's proposal, asserting that "risk" is the product of
a large number of independent observations such that the
probability function is empirically derived and that
"uncertainty refers to future events where the parameters
of the probability distribution cannot be determined
empirically" (Heady, pp. 441-449). Heady reaches the
conclusion that uncertainty "is a purely subjective

 

 

82
as being formed by an individual mental process (problem-
solving behavior) utilizing whatever information is at hand.

Such information can range from many observations on his

own farm -- akin to scientific investigation and statistical
analysis -- to the expectations of someone whose advice he
respects.

The information used by farmers could possibly be
classified into three types: (1) empirically derived
information, (2) a priori associative knowledge, and
(3) communicated knowledge. Since such a breakdown is prob-
ably unique to this study, some explanation of these terms
is necessary. Empirically derived information will be defined
as that info. mation which was obtained by observation of a
given phenomenon. It need not be thought of as a complete
empirical experiment. In addition, no claim can be made
regarding any one-to-one correspondence between the
information generated by these observations and some
"absolute" reality. The senses provide only data which must
be transformed into meaningful information within the context
of the second type of informe tion: a priori associative
knowledge.

A priori associative knowledge, in turn, can be defined
as that information, developed over time and residing with
the individual, which aids him in associating various relation-

ships and reasoning toward certain conclusions. Under this

 

phenomenon and is peculiar to the mind of the individual."

This view involves an unwarranted division of types of
information used in forming expectations, and suggests

argnmnns about the meaning of "objectivity” and "subjectivi ty"
as they apply to information sources.

83

heading are collected the individual's attitudes, beliefs,
and conceptual structures (i.e., concepts of plant nutrition,
weights and measures, mechanical concepts). They are the
product of his past learning experiences -- both formal and
informal -- and provide him with the capabilities for
interpreting situations and formulating expectations.

Finally, communicated information is the information
which passes between individuals through the various channels
of communication. It too is subject to the interpretation
provided for by a priori associative knowledge. The farmer’s
expectations are the product of some combination of these
types of information.

Let us take a typical example of the diffusion of an
agricultural innovation. Presune that a new variety of
corn has been developed and the intention is to extend this
innovation to farmers. The researchers at the experiment
station have evidence that the new variety is superior to
present varieties. This evidence might consist of field
trials where the new variety was observed to produce higher
yields than current varieties -- em irical information.
The extension agents learn about the variety from the
experiment station personnel -- communicated information.
They also may have a priori associative information which
permits them to see the researcher's evidence as reasonable.
And they have a priori information that the researchers are
usually correct in their evaluation of new technologies. The

eXpectations that they form are established on this basis.

84
When the extension agent presents the variety to
farmers, they are receiving communicated information based
initially on empirical information. They, too, have apriori

associative information with which to evaluate both the

 

source and the message. Presume that from experience the

farmer has a priori information that his extension agent is

 

competent and well intentioned. Suppose also he believes
that researchers can improve upon present practices. Consider
as well that he has some a priori information which allows
him to believe that better seeds will make better crops.
Finally presume that experience has shown him that on his
soils varieties produce about a third less than at the
experiment station. He now has information that can be used
to develop a set of expectations about the variety for his
own farm. In an intuitive way he assigns "degrees of
belief” for a set of outcomes which might be expected if he
were to adopt the new technology. As he observes the variety
growing on his own farm or his neighbor's farm he can revise
the probabilities of different outcomes on the basis of
additional empirical information. In retrospect it makes
very little sense to pursue arguments about the conditions
within which expectations are either "subjectively” or
"objectively" formed.

we have now reached the point of discussing the "degrees
of belief" that the farmer attaches to various outcomes.
Clearly we are suggesting that a Bayesian view of statistics

is applicable to expectations, that probabilities can be

85
established by other than measuring the frequency of given
outcomes over large numbers. The farmer contemplating buying
a farm has hardly enough observations to establish the
probability of his success or failure in a frequency sense.
Yet he must formulate some expectations in order to make a
reasonable decision. As a result. the information at hand is
processed by the individual and he is guided by his
expectations as to what action should be taken. The action
taken may, of course, be to postpone the decision and to
gather more information.

His expectations are essentially a set of associated
outcomes and "degrees of belief" that these outcomes will
occur. If one has reservations about using probabilities as
"degrees of belief,” they can possibly be dispelled by the
following representation. The scientist with his statis-
tically valid probability distributions developed from his
empirical observations undertakes the same mental
calculations as the farmer; ggly_his mix of information
sources is different. He expects a given outcome because

of empirical information from his experiments and apriori

 

associative information that is formalized in his knowledge

 

of statistics. The scientist's set of expectations are then
essentially "degrees of belief" based upon a different
emphasis in the information sources used.

If the farmer can establish a priori probabilities
regarxling forthcoming outcomes, then how can uncertainty be

taken irnx>consideration? This is somewhat of a conceptual

86
problem which deserves a careful examination. Since we have
downgraded the possibility of a "true" probability distribu-
tion that exists apart from human understanding we can hardly
say that as the farmer's probability distribution approaches
the true distribution uncertainty is reduced.

The fact of the matter is that we can theoretically
conceive of an analogous chain of events within the mental
processes of the individual. With little knowledge to the
contrary, the farmer may assign equal "degrees of belief"
to every plausible outcome -- nearly complete lack of informa-
tion and the highest degree of uncertainty possible. With
more and more information he begins to restructure his
"degrees of belief" in line with his increased understanding
of which outcomes appear most likely. Eventually, with all
of the information available that he is capable of processing
(or willing to accumulate) he will have reached some final
revision of his expectations. He is, nonetheless, uncertain
about what outcome will occur in a given year, but he has
been able to rule out these outcomes which at this present
level of knowledge are extremely unlikely.

Let us examine this sequence of events with a
hypothetical example. What we are expressing is a learning
process where additional information permits the decision-
maker to assign new "degrees of belief" that certain outcomes
will occur. Suppose that an extension agent is attempting to
introduce chemical fertilizer to a given group of farmers who

have had no eXperience with this practice. The range of

87
outcomes to which farmers might give consideration could vary
from fertilizer "burning up" their crops -- no yield —— to
more than a tripling of present yields. They may consider
decreases in yields for any given year equally as likely as
yields above the present levels. Assume for simplification
that the usual yield with present technology is ten units in
average years (six out of ten years), and eight units in poor
years (two out of ten years), and twelve units in good years
(also two out of ten years). The farmers' expectations for
traditional technology would probably closely reflect these
frequencies. If he has little information about the new
technology he might expect any outcome between zero and
thirty units as equally likely yields. Figures 5a and 5b
il1 ustrate these respective expectations at a given point in

time (t1).

 

 

WWW

8 10 12 Yield 10 15 20 25 30 Yield
-a- Present Technology -b- New Technology (t1)
Figure 5 -- Frequency Distributions Representing the "Degree

of Belief" (Probability) a Farmer Might Assign to Various
Crop Yields Under Present and New Technologies.

If the extension agent tells the farmer that the average
yield on the experimental farm with the use of chemical

fertilizer was tWenty units and that fertilization would not

88
"burn up" his crop, he might consider zero yield as unlikely
and expect an average (mean) yield somewhat below the
extension agent's claims (discounting for slight distrust in
the extension agent or differences in soils that he presumes
to exist between his farm and the research farm). At this
point in time (t2) he may have expectations approaching
figure 63. If his neighbors try the practice and even in a
bad year their crop yields ten units he will begin to attach
fewer degrees of belief in yields below this level. If, in a
very good year, the practice yields only twentyefive, this
information will influence his expectations. At this juncture

(t3) his expectations might be represented by figure 6b.

1'“th

 

 

 

~‘x' V ' (l
. i ' is
0 S 10 15 20 25 30 o 10 S O 25 30
-a- New Technology (t2) -b- New Technology (t3)
Figure 0 -- Frequency Distributions Representing a Farmer's

Expectations of Various Yields with Increased
Accumulation of Information.
After a considerable amount of information~gathering
(actively or passively), our hypothetical farmer still is not
certain what yield will occur in a given year, but he is
reasonably certain that zero and thirty unit yields will not
occur and he attaches a high "degree of belief" to the
possibility of yields at fifteen units.
Note that the example is based on discrete probabilities

at arbitrarily set levels. This is probably'a reasonable

89
approximation of the type of expectations that farmers have.
They certainly will not, in their expectations, approach a
continuous distribution of "degrees of belief." Our assurance
of this lies in the limitation of outcomes that it is humanly
possible to consider at one time. It is thus likely that
farmers focus their attention at certain points over the
range of possible outcomes and concentrate upon the "degrees
of belief" which should be associated with these points.
Other points hold less interest to the decision-maker or are

subsumed within a range centered about the points of interest.

Utility for Income

 

Now that we are equipped with some ideas of how expecta-
tions are formulated, to continue our conceptual framework
we will need some notions about the farmer's utility for
income. The concept of utility for income is notably
troublesome, yet it is crucial that we use some concept to
measure the value attached to various levels of income in
order that we may understand decision-making under uncertainty.
Our concern here is to conceptualize a "typical"
utility function for farmers in less developed countries.

10, which attempts to

The Friedman-Savage utility function1
explain gambling and insurance, can be a useful starting

point. The relevance of this type of function is that it
places the individual at some level of income and explains

his utility for gains and losses. The shortcoming of the

Friedman~$avage analysis is that they did not deal with a

 

110Friedman-Savage, op. cit., pp. 279-303.

 

‘ e

 

90
situation applicable to business decisions. They examined
situations with a small probability of a large gain and a
large probability of a small less, or the reverse in the case
of insuring. Most farm production decisions are hardly a
gamble in the Friedman-Savage sense. In short, the losses
associated with gains are not necessarily so small that
their disutility is negligible. We are, instead, interested
in the utility of a wider range of associated gains and losses.

To consider the case of the farmer in a less developed
country we will begin with his present income and speculate
on his utility for given levels of income other than present.
With respect to his utility for higher incomes, we can find
several positions represented in the literature. An upper
lll

limit on aspirations, as visualized by Mrs. Nair , would

suggest that the upper range of the (Friedman-Savage) utility
curve soon becomes parallel to the horizontal axis. Katona's112
assertion that the desire for additional income is a function
of the individual's past successes and the incomes of his peer
group would direct us toward a rather individualistic inter-
pretation of the shape of the utility curve for increased
incomes. A positively sloping utility curve is the least

13

that may be implied by Schultz' view of the peasant

farmer as an "economic man." Finally, we have Friedman and

 

111Nair, op. cit., pp. 192-193. Also see pals of this

 

text .

112Katona, op. cit., pp. 86~106. Also see pp. 56-57
of this text.

1135chultz, 9p. cit., pp. 24-52. Also see pp. 20-24
of this text.

 

91

114 contention that the upper range of a ”poor

Savage's
person's" utility function is concave from above. Perhaps
the best position to take, until better information is
available, is that the upper range will be assumed to have a
positive slope with the derivative of the curve depending
upon the individual and his particular situation (i.e.,

past successes and failures, the income levels of peer

groups and other members of the community).

At the Opposite end of the utility function, we have
somewhat sounder theoretical notions on which to postulate
the shape of the curve. Wharton115 has recently related the
concept of focus-loss to subsistence agriculture. In
Wharton's View, those farmers living on the edge of a
subsistence income, whether physiologically or socially
prescribed, are extemely fearful of taking actions which
have any likelihood of forcing them below the subsistence
level of income. (If money must be borrowed to take the
action under consideration it seems reasonable that the
critical level of income will be the subsistence level plus
the amount that must be paid back;) Any prospects of falling
below this critical level are intolerable to the farmer (as
starvation of his family would be conceived to be). The
closer a farmer's present income lies to this critical income

the more hesitant he is to accept losses. This implies that

 

114Friedman-Savage, op. cit., pp. 294-295. Also see
PP- 63-67 of this text.

11SWharton, "Risk, Uncertainty and the Subsistence
Fanmny" pp. 1~53. Also see pp. 63.70 of this text.

 

92
the utility function for income in the lower range becomes
discontinuous at the critical income level.
A representation of an income—utility function which
fits the specifications presented in this section could

serve as a useful summary.

 

 

 

 

Utility
A
. /,/” ' Boundary of
Sub51stence . Conceivable
Level "ZN *Incomes for
: the Farmer
fiC-p
l
l 1'»
0 I
V Present Income Income

-1

Figure 7.--Suggested Income-Utility Function.

The line labeled A represents an income-utility function
rising at a decreasing rate to the outer range of outcomes
conceivable to the farmer. Line B gives the function as
rising at an increasing rate. At the lower end utility

falls off more rapidly than it increases for gains in incomes,

116 who examined

as suggested by empirical work by Halter
farmers in the United States. The curve becomes discontinuous
at the subsistence level and goes to infinite disutility.

The segment "c" is smaller for farmers nearer the subsistence

level..

 

116Halter and Beringer, op. Cit-. PP- 113‘132°

 

93

Constraints Upon the Implementation'of
New Agricultural Practices

 

 

Before we examine the interaction of the farmer's
expectations and utility for income, it is necessary to
consider those factors which limit his capacity to employ
new practices. Since new methods usually involve purchased
inputs, the most obvious limitation on the farmer's
innovativeness would be the availability of these inputs.

Not only must inputs be available in the absolute sense, they
need to be provided at convenient locations. This is
necessitated by limitations on the distance that farmers in
less developed areas can travel to secure the items needed

on their farms. The supply of these products also should be
timely and in sufficient quantities so that farmers can be
certain that they can acquire inputs when they need them.

The availability of credit can also be a determining
factor in the implementation of new practices. Cash outlays
for purchased inputs may be beyond the immediate financial
capabilities of many of the farmers concerned. Without
credit they might be unable to cover both current family
living expenses and the cost of purchased inputs. The
availability of loans is not by itself a solution to this
problem. Loans must be large enough to adequately finance
new practices, and the terms of these loans must be such that
they encourage rather than discourage the use of credit.
Unduly stringent collateral and repayment requirements may
deter farmers from using credit. As an example, a relatively

small loan requiring land as security and insisting upon

94
repayment at harvest time -- when prices are lowest -- would
probably be unattractive to farmers, despite the benefits
they might derive from using new methods. 6

Another possible limitation on the farmer's capacity
to utilize new practices is land tenure. When the farmer is
a tenant rather than a landowner, he may be justifiably
unwilling to use new practices. If he must bear all the
costs and receive only a portion of the benefits, many other-
wise attractive innovations would become unattractive.
Secondly, he may be most unwilling to participate in improve-
nwnts in the farm -- irrigation facilities, building soil
fertility through fertilization, improved cr0p storage
facilities -- if he is uncertain that he will have continued
mmess to these resources.

Thus a favorable disposition toward a new technology is
not a sufficient condition for its adoption. The farmer must
also have available the resources -- inputs, credit, land
resources -- necessary for implementing his decision in a form

that does not discourage their use.

The Adoption Process: Decision-Making

InvolV1ng thE’intiraction of
Many Variables

 

 

 

flgationality"

A basic assumption of the present theory is that farmers --
peasants and commercial farmers -- behave "rationally." The
problem vdth this assumption is that opinions differ regarding
the definition that should be given "rationality." Economists

tend to have a very specialized meaning for this term which

...

 

 

95
usually involves the maximization of profits (in production)
or utility (in consumption). Criticism of the "maximization"
approach has given rise to the concept of "satisficing" --
choosing a satisfactory although not necessarily the "best"
alternative. Among non-economists "rationality" usually
receives a more general interpretation, meaning a choice of
alternatives which conforms to the individual's value and
belief systems. They normally emphasize the non-economic
variables which influence individuals' choices.

We need not argue for or against the need for including
"differences in personality, education, and social environment”
since the impact of a large number of variables can be
encompassed by our decision-making model. Differences in
attitudes, beliefs, and educational achievement can influence
dccision- aking through the impact of a priori associative
knowledge on the process of formulating expectations. As an
example, a fatalistic attitude might result in generally
negative evaluations of new methods, while a higher level of
education might permit a better understanding of the relation-
ship between chemical fertilizer and crOp yields through a
clearer concept of plant nutrition. Since we have allowed
for a flexible and individualistic interpretation of the
shape of the utility function for increased incomes,
variations in individual aspirations can be considered.
Attention has also been given to empirical and communicated
information which allows us to consider the impact of

different sources of information. Finally, we have given

96
consideration to the farmer's capacity to implement his
decisions, by including the availability of inputs, credit
and land tenure. With these possibilities for the interaction
of non-economic variables with economic variables, an assumption
of "rationality" must, by necessity, take on a rather general
character. Thus the "rationality" assumption is that farmers,
given (1) their state of knowledge about various alternatives
and the resulting expectations about possible outcomes and
(2) their individual aspirations for economic gains and desire
to avoid losses, choose the alternative actions that they
believe to be "best" and attempt to implement them within
the constraints of their particular situations.

Decision-Making: An Interactive System

 

The schema on the following page is a representation
of the decision-making process relating those concepts
which have been developed in the previous sections. This
conceptualization involves the interaction of various kinds
of information in the formulation of expectations, a
comparison of the expected utility from new and current methods,
and an implementation of the "best" alternative given the
constraints imposed by the availability of certain
resources -— credit, tenure, and inputs.

Let us examine this conceptualization in somewhat
greater detail. The composition of the farmer's expecta-
tions, at any point in time, is the product of the informa-
tion used in formulating these evaluations Initially he

may have little more than a priori infprmatiOn with which

 

97

.moofiuomam

Hmunpanowmw< 3oz op puommom new; mcwxe2-:oflm«oon mosaem mo 3ofl> owpeaonom--.w opzmmu

 

 

c o
o > c
w w o
u u no w
o m u
o a a
m c o
o o v
“LPIIIIK
oeouuoo

 

 

    

:Omequ
we cowpwu
-coEoHQEH

 

  

  

_ V
as

r.

 

 

 

 

.Haa><
munch
Hosanna fl
— unease .H

 

  

 

 

QEOUQM newt... HO>01~
womeomoom mom.flifit eEoocH anemone
>uaaausmaa .

 

l,

 

 

 

 

 

 

 

 

 

ad meowuewpmecsom
my \\a maumm .muogpc
“a \ Eafi@.=3b
m“ tux .omza seeaxaazm

 

       

:oEmmocwmzm Hmuoq

      

     

uzecwou vogue: 3oz «New: mmmz
.m> 302 now mcowu AM” a u:ow< :owmcouxm
mo saunas: -auuodxm mosauaaoa

 

Imam when:Mwmz
.omzH .2200

 

 

 

 

 

mmwouwup<
xuwcoseoo
.uooxm peed
.oopm Hmsuom
mmowaom
.lwem memoumup<

 

    

     

macho
poem we macaw
-mpmam< pew
m~o>o4 oeoucw _

urine-J

 

 

amt-w—
> ham...
‘5
\
$§

 

 

 

 

 

 

 

wall-ll. .c. .
osoucm _ta4c momofiwmm _ orzh HvonAlw
vomweaucH Aw .a wee
now Apmfiwu: a mmwwuzm amen

 

 

 

 

 

 

 

98
to evaluate the possibility of various outcomes from an
innovation, and consequently his expectations may reflect
a great deal of uncertainty. If they include the possibility
of outcomes consistent with returns below the minimum accept-
able income level, he will tentatively reject the innovation
and continue to use these practices which from experience
he has learned to expect at least the minimally acceptable
results.

Even expectations at this early stage will vary from
farmer to farmer due to differences in the attitides and
beliefs held by these individuals, variations in their past
experiences and formal education, and differences in their
perception of community attitudes. Farmers may hold
attitudes, such as a fatalistic outlook, which affect their
assessment of a new method in a negative manner. Or they may
hold attitudes, such as a generally favorable view toward
technological change, which would positively influence their
expectations toward new practices. Favorable experiences
with somewhat similar innovations in the past could stimulate
a rather positive view of a new technology, while the past
failures of poorly adapted innovations would justifiably
cause the farmer to regard a new practice with considerable
skepticism. The amount of formal education possessed by the
farmer may also influence his evaluations of new methods,
since especially in the case of technically complex
innovations the better educated individuals would be assumed

to reason toward a more technically accurate set of

99
expectations in a shorter period of time and with less
empirical information than their poorly educated colleagues.
Finally, a farmer who perceives his community's attitude as
favorable toward technological advance may be more favorably
disposed toward new methods than the farmer who believes his
neighborhood would resist change. Thus, a number of

hypothesized relationships between a priori information and

 

the adoption of new technologies would be consistent with
this conceptual framework.

As the farmers gather additional information (actively
or passively), they will differ in their choice of sources
of communicated ideas about new practices. Farmers who value
information from extension agents (agronomists) or the mass
media will conceivably be exposed to ”messages” about a new
technology at an earlier point in time than those who obtain
their ideas from discussions with relatives and neighbors.
The content of the respective “messages" will probably also
be quite different. We might call "messages" from the
extension agent and the mass media "direct" information, as
opposed to ”indirect" information from relatives and
neighbors. The reason for this particular terminology is
that "messages" from the extension agent and mass media
would seem to pass through fewer intermediaries in their
flow from the source of the innovation (perhaps the
experiment station) to the ultimate recipient. Since each
intermediary probably incorporates his own expectations

into the message he presents to others, a message from a

100
"direct" source of information will include fewer modifica—
tions than the ideas received from such "indirect" sources
as relatives and neighbors. In terms of the farmer's own
expectations, we would expect those who utilize "direct"
sources of information to formulate favorable expectations
toward new methods more rapidly than farmers who depend
heavily upon "indirect” information sources.

Still another form of information is observation of a
new method in use. Farmers who have an opportunity to
observe demonstrations of new practices may form favorable
expectations relatively early. But those farmers whose

desire for empirical information causes them to wait until

 

new practices can be observed on their neighbors' farms

will consequently be relatively late in formulating expecta-
tions favorable to the adoption of these methods. There is,
of course, the possibility of limited trials on the farmer‘s
own farm when an innovation lends itself to this practice

and the farmer is financially able to engage in such
experimentation. It should also be noted that the
reformulation of expectations is not discontinued with the
adoption of a new method as the farmer continues to assimilate
information about the practices currently being applied.

At the same time that the farmer is formulating and
reformulating his expectations about new methods he is
contrasting the expected utility from the new method with
that of the current practice. The relative attractiveneSs_of

the new method versus the current practice depends both upon

101
the expectations held by the farmer and his particular
utility function. Two farmers with the same set of expecta-
tions may respond quite differently to an innovation depending
upon their initial level of income. The larger and wealthier
farmer may be willing to accept risks in pursuit of higher
earnings that his smaller and poorer counterpart would find
intolerable. The possibility of a 50 percent reduction in
crop yields, as an example, might not reduce the wealthier
farmer's income below an acceptable level, but for the poorer
farmer such an outcome would represent considerable suffering
for his family. Thus the poorer farmer may need to be more
certain in his expectations and quite sure that untolerable
outcomes will not be forthcoming before he adopts a new
practice.

In addition to considering the disutility from possible
decreases in income, the farmer evaluates the utility from
possible increases in earnings. Differences in the motiva-
tion for additional earnings will vary the attractiveness of
new methods as perceived by different farmers. Past
successes or failures and the income levels and aspirations
of a farmer's peer group may influence his desire for
additional earnings. The income aspirations of the individual
farmer, therefore, may affect his willingness to adopt new
methods and accept the risks inherent in his expectations
about the innovation. Farmers who are highly motivated to
obtain increased earnings may consequently be willing to act
upon less certain expectations than their less motivated

colleagues.

102

A favorable disposition toward a new practice can be
conceptualized as involving the interaction of a farmer's
expectations and his utility for income, but this still does
not imply adoption. The farmer must be able to implement
what he believes to be the "best" alternative. He may find
that sufficient credit is not available or that the terms
of a needed loan involve too much "risk” to be acceptable
despite the benefits that might be derived from a new
practice. If the farmer is unsure that he will benefit
fully from a new practice, because of his status as a

tenant, he may be justifiably unwilling to employ a new

method. Finally, he may not find the needed inputs available

at a time and place that make it possible for him to utilize

a new practice.

In summarv the adoption irccess has been conce tualized
. 9 l

as the interaction of a number of variables which might be
expected to affect the formulation of expectations, the
shape of the farmer's utility function and his ability to
implement what he considers to be his "best" alternative.
This conceptualization of the adoption process will be
examined empirically in the following chapters utilizing
data from the MSU/SUDEHB survey of farmers in the Recife

area of Northeast Brazil.

 

 
 

CHAPTER IV
AGRICULTURE IN THE RECIPE AREA OF NORTHEAST

BRAZIL: THE SOURCE AND CHARACTERISTICS
OF THE DATA

'Introduction

 

The empirical examination of the adoption process
presented in the following chapters is based upon data
generated by a survey of farmers in the Recife area of
Northeast Brazil. This chapter provides an overview of
agriculture in Brazil's Northeast and a description of the
MSU/SUDENE Farmer Survey from which data used in the later
chapters was taken. In addition, selected characteristics
of the farms and farmers included in this survey will be
presented.

An Overview of the Agricultural
Economy in Northeast Brazil

 

 

The Northeast of Brazil is a large and populous area.
Robock describes it in the following way:

If the Brazilian Northeast were a separate nation,
it would rank second in population and third in area
within South America. With its present pepulation
[1963] of about 25 million, Brazil's "bulge" has as
many people as Thailand, more than Argentina, and.
almost three times the number of people in the New
England region of the United States. In area, the
Northeast is larger than Italy, Spain, and Portugal,
combined.117

' Climatic Sub-Regions of the Northeast
Four relatively distinct submregions are encompassed by

the Northeast (see Map 1, page 118). A narrow humid strip,

\‘\\\

 

117Stefan H. Robock, Brazil's Developing hertheastt A
“Study of Regional Planning and Foreign Aid (Washington, D.C.:
Brookings Institution, September l903), p. 2.

103

 

 

 

104
called the Zona da Mata or "Forest Zone," lies along the coast
between the city of Salvador to the upper edge of the state
of Pernambuco. This area, ranging from thirty to sixty miles
in width, has ample and dependable rainfall averaging over
fifty inches per year. The southern portion of the Forest
Zone in Pernambuco is very hilly, with an interspersion of
relatively flat, shallow valleys. The northern part of the
Forest Zone and the portion lying below Pernambuco consist of
extensive flatlands cut irregularly by deep stream beds.
Sugar cane production has been the principal economic
activity in this area since the initial colonization of the
region.

Lying to the west of the Forest Zone is the second
major sub-region of the Northeast, the Agreste. The Agreste
is a transitional zone separating the low-lying humid
coastlands from the vast semi-arid interior plateau. Its
width is approximately comparable to the Zona da Mata.

The Agreste has fairly reliable rainfall averaging
thirty to forty inches per year. It is a generally hilly
area, dotted with large rock outcroppings. The soil is of
poor quality, but their structure and depth, in conjunction
with a reasonably favorable climate, have permitted the
development of a diversified agriculture. The Agreste has
become the Northeast's primary internal source of foodstuffs.
Beef, dairy products and such staples as beans, maniec,
corn, and rice are produced in this area. Beef cattle are

generally raised on extensive holdings, while the staples

 

fo'i-r

—- zen-.1

105
are typically produced under share-cropping arrangements or
on the "minifundia" prevalent in the area.

Inland from the Agreste lies an extensive semi-arid
region called the Sertao. Normal rainfall in this area is
from twenty to thirty inches, but it is highly uncertain and
typically concentrated in the first five months of the year.
This leaves a dry season extending from June through
December. Soils are shallow and have a low water retention
capability. Drought is therefore a yearly occurrence, the
seriousness of the phenomenon depending upon the volume of
precipitation -- which may vary 30 percent from the average --
and the relative concentration or dispersion of the rainfall.
Beef cattle, tree cotton, and sisal are the major products
of the area.

The fourth sub-region consists of the tropical rain
forest areas of Maranhao and southern Bahia. The expansion
of slash-and-burn agriculture, especially in Maranhao, may
eventually result in the deforestation of these areas.118

Economic History of Brazil's Northeast

 

In the early 1500's the Northeast became the site of the
first Portuguese settlement in Brazil. With the ever-present
hope that the Brazilian territory would eventually bear untold
riches in precious metals, the colonizers sought an economic
base that would enable them to support the defense and

permanent occupation of the region. The solution emerged in

 

v

. 118Charles Slater, Harold M. Riley,'et‘al.,‘Market
Processes, pp. 2/3-2/7. Also, Robock, op. cit., pp. 70«74.

 

106

the form of large—scale production of sugar, a commodity
highly enough valued in European trade to offset the costs
and uncertainties of long distance sea transportation. In
addition to the suitable production conditions in the Zona
da Mata, the technical expertise of Portuguese in sugar
production and refining, the availability of an expanding
European market, and access to African slaves as a cheap
source of manpower, all contributed to the development of a
sugar colony that flourished for more than a century. As a
consequence, extensive sugar plantations emerged on the
better coastal lands of Northeast Brazil and instituted a
pattern of land use which has tenaciously resisted change
to this date.119

At the height of its prosperity the sugar economy
expanded rapidly. It is estimated that at one point in its
growth, retained earnings were sufficient to finance a doubling

120 The profitability of this

of production every two years.
monoculture, however, resulted in an economic structure that
had little impact on a broader economic development and
integration of the region. Income was concentrated in a few
hands, resulting in little incentive for the development of
locally produced goods. Further, both consumption and invest-

ment were largely oriented toward the external sector. In

fact, the basic organization of the sugar economy enabled it

 

119Celso Furtado, The Economic Growth of Brazil, (Berkeley
and Los Angeles: University of California Press, 1963),
pp. 1-12.

120

 

Ibidz, p. 48.

107

to withstand the ensuing decline in sugar prices without
major structural change. Since in a slave economy where the
major components of production are fixed capital it is in
the best interest of the entrepreneur to maintain high levels
of production, the sugar economy persisted even though by
the mid-1600's Caribbean competition had depressed prices to
one-half their former levels.121

The one economic outgrowth of the sugar economy was the
development of cattle raising inland from the Zena da Mata.
Both the demand for draft animals during the expansion of
the sugar industry and the unprofitability of using sugar
lands for cattle raising contributed to the growth of this
enterprise. With the decline of the sugar industry, cattle
raising inland assumed a new role, that of subsistence
production to feed a growing indigenous population as well
as emigrants from the depressed sugar growing area. Because
cattle raising for subsistence purposes does not depend upon
an external market for its maintenance, a second economic
pattern emerged in the Northeast which had a great capacity
to persist over time.122

The discovery of gold in Southern Brazil sparked the
first major influx of European immigrants into the colony.
This event was short-lived, however, reaching its peak in
1760 and declining to a minimal level of activity within the

next two decades. Despite the early atrophy of the mining

 

121Ibid., pp. 48-58.

1221bid., pp. 53-77.

“I

 

108
economy, there did exist for a time a rudimentary internal
market such as had never occurred in the Northeast. Its
existence is largely attributed to the broader distribution
of the economic rewards from mining and the greater cost of

. . . . . . 12
imports in the interior mining areas. 3

In 1762, the move-
ment of the capital of Brazil from Salvador to Rio de Janeiro
signaled the shift in the center of economic gravity from the
Northeast to the more richly endowed southern regions of
Brazil.

With the decline of mining another center of economic
growth developed in the Maranhao area of Northeast Brazil.
A struggling colony had for some time existed in this area,
progressing from capturing of indian slaves as the main
enterprise to the collection of cocoa, vanilla, cinnamon,
cloves and aromatic resins for export. A serious conflict
between the colonists and Jesuit priests over the use of
forced indian labor in the gathering of these tropical
products culminated in the Portuguese government taking the
side of the colonists and shortly thereafter establishing a
well financed trading company to develop the region. The
directors of the newly formed company seized upon the immediate
world market opportunities for cotton and rice, fostering
their production and subsequently avoiding the general
depression that was common to all other regions of Brazil in

the late 1700fls.124

 

123Ibid., pp. 77«92.

124Ibid., pp. 95~99.

 

‘n

.,'

.‘h‘

 

109
Following a brief revival of sugar and cotton prices -—
due to world political events in the early 1800's «- the
Brazilian economy fell into a period of general economic

decline by the mid—1800's.125

Since manufacturing had not
progressed rapidly, which can be attributed both to a lack

of technical expertise on the part of the Portuguese and a
weak internal market, Brazil was in desperate need of an
export commodity which could assist it in becoming reintegrated
with the world economy. The development of coffee production
in the Rio de Janeiro area met this need and provided the
stimulus for rapid development in the southern region. ’In its
early stages of development coffee production capitalized

upon pre-existing and under-utilized resources which had been
idle since the demise of the mining economy. Later, without
the alternative of importing slave labor, a second influx of
European immigrants arrived in the South to work on the coffee
farms. Further, the abundance of high-quality land permitted
a great expansion of subsistence food production which also
served as a stimulus for the most rapid population expansion
of any region in Brazil. Finally, the rapid deve10pment of

an internal market, spurred by coffee development, created
economic incentives which were previously nonexistent and

propelled the southern region into becoming the dynamic

'!
L

center of growth in the Brazilian economy.

In the final years of the nineteenth century, the economy

 

1933, pp. 107413.
1261b1d , p. 144.

—*

125

110
of the Northeast suffered another setback which was dealt
by the severe drought of 1877. The severity of the drought,
in which one-half million people were reported to have
perished, was exacerbated by the previous three decades of
ample rainfall which had facilitated migration from the sugar
producing area to the semiparid areas inland. The Government
of Brazil recognized the Northeast's drought as a problem,
but relief efforts were slow and disorganized. The relief
took the form of (1) emergency food and clothing, (2) make-
shift public works to employ drought emigrants, and

127 There

(3) relocation of emigrants to the Amazon basin.
were many claims of irregularities in handling the relief
funds, at that time entrusted to the state governments, which
became a permanent criticism of the works against the

droughts in the ensuing years.

Supported by the belief that the application of science

and engineering could avert such disasters in the future, the‘

government dispatched an Imperial Commission to make
recommendations for government action. The Commission
recommended building a series of dams and improvement of the
transportation facilities in the Northeast, an approach that
was to be followed without question for many years. The
Inspectoria of Works Against the Drought was-established to

128

implement this approach. The funds available to the

lnspectoria and its successor agencies have varied widely

 

127Albert 0. Hirschman,'g9urneys Toward Progress,
(New York: Twentieth Century Fund, 1963),p. 22.

1281bid., pp. 22-24.

 

 

‘i:"

111

over time, the highest levels of financing generally being
correlated with major drought years and the presence of
Northeasterners in high places in the federal government.129
The lower levels of financing were often the result of the
resumption of ample rainfall and fiscal problems faced by
the government. These fluctuations in financial support
suggest the theory that expenditures in poor areas generally
must take second place to those in the vigorously growing
economic areas, except in those times of national emergency.130

The highest level of funding was reached during the term
of the only President from the Northeast, Epitacio Pessao,
when 15% of the national budget was directed toward works
against the drought. President Pessao believed that
simultaneous development of reSeivoirs, roads, railroads,
and port facilities was imperative because they were all part
of a single system and to eliminate any one of them would
leave the program incomplete. He pursued this policy with
the accompanying detriment to Brazil's fiscal situation
throughout his term in office from 1919 to 1923. His
successor, in the face of fiscal difficulties, began
suspending public works and by 1925 a complete suspension of
such efforts had occurred.131

The construction of dams and public works received

another major stimulus during the Vargas Administration in

 

129;p1d., p. 18.

13°Ibid., p. 34.

1311bid., pp. 30-34.

”—

‘é‘ . .2:

 

V's-4s; “v

 

... .v .

 

112

132 However, the creation of irrigation

in the 1930's.
facilities to utilize the stored water resources were

continually delayed. Arguments for the value of the dams

in themselves (as sources of water for humans and livestock),

the fact that in normal years irrigation water is not a

prerequisite for agricultural production, and the reluctance

to disturb the landholding patterns downstream from the dams , ?E

133

probably all contributed to this delay. To date only

‘1‘

about 18,000 hectares of 250,000 hectares of potentially

 

E .
=4,
{4
V

irrigable land created by the dam projects are under
irrigation.134
Doubts began to arise in the 1940's as to the efficacy
of the "works against the droughts," especially dams, as a
solution to the problems of the Northeast. José Augusto
Trindade, first Director of Research and Extension in the
drought area, noted the need to change attitudes toward
water use and recommended expropriation of large landholdings
in order to make optimum use of irrigable land. ’He also put
forth the belief, to be reiterated by others later, that the
dams did not serve the small farmers who were affected most by
the droughts. Still another observation, by Trindade's
successor Guinares Duque, even cast a shadow on irrigation as

a solution to the Northeast's problems. From his calculations

132'1bid. , pp. 3849.

1331bid., pp. 42413.

“—

134Robock, opr'citr, p. 78.

 

113
of watershed-irrigation relationships it can be concluded
that of a total 8 million hectares in the Northeast's drought
area, a maximum of 800,000 hectares (10.0%) could be
successfully irrigated from available rainfall.

With the advent of interest in economic planning, an
alternative "economic solution" to the problems of the
Northeast began to take shape. An early step in this direction
was the creation of the Northeast Development Bank (BNB) in
the early 1950's, which was charged with utilizing a portion
of funds constitutionally set aside for the "fight against
the drought" to make long-term loans for the development of

the area's agriculture and industry.135

The 1958 drought, one
of the most severe ever in terms of displacement of the
Northeast's pepulatien, intensified the belief that previous
policies had not been adequate and further stimulated the
search for an "economic solution." In response to this

need, Celso Furtado, Director of BNB, prepared an economic
report on the region. He recast the problems of the

Northeast in terms of an increasing economic disparity

between this region and other parts of the nation. In fact,
he suggested that government policies on exports and imports

had transferred resources to the industrial South.
Purtado's view the main resource in the Northeast was cheap
labor for industrialization, but the high cost of food

limited the utilization of this labor for industrial

 

135;pid., p. 62.

136Robock, o . cit., p. 107.

 

1... .1 IA!
n-u om ...w

 

..-

 

114
purposes.137 His proposals for economic development of the
region contained three main thrusts: (l) intensification of
investments in industry, (2) reorganization of agriculture,
and (3) colonization of the Maranhao. Among his suggestions
for restructuring agriculture was greater use of irrigation
in the semi-arid regions for food production and a major
effort to achieve better utilization of the lands in the F‘

138 The Superintendency for Development ,

sugar producing area.
of the Northeast (SUDENE) came into being in 1959 to prepare

and implement overall development plans following the Furtado

 

suggestions and coordinate the activities of all those 9
agencies charged with development efforts in the Northea’st.1‘)9
With SUDENE'S emergence a new era had begun in the quest

for solution to economic ills of Brazil's oldest region.

The Data from Northeast Brazil

 

During 1966-67, a team of Michigan State University
researchers cooperated with SUDENE (the Northeast Brazil
Economic DeveIOpment Agency) in conducting a study of the
marketing processes in the Recife area of Northeast Brazil.140

Considerable emphasis was placed upon the marketing of farm

products, with the total scope of the project covering farm

 

137

Hirschman, op. cit., p. 75.

1331b1d., p. 76.

139Ibid.,.p.

140Charles Slater, Harold Riley, et al., Market Processes,
pp. 1/15p1/18.‘

 

 

,
unly‘

nuil§

 

'v .0.

Nu“

 

1""

w.“ .

.,'.
I

 

 

‘5

I.

 

115
production, assembly of farm products, wholesaling,
retailing, urban consumers, industrial production, and the
distribution of agricultural inputs. As a part of this study,
a survey of 781 farms -- representing rice, bean, manioc,
cotton and milk producers -- probed the production and

marketing activities of farmers in municipios identified as

 

important suppliers for the Recife urban market.

It should be noted that this survey was not a random
sampling of farmers in the Northeast, but instead provided a
representative sample of the farmers who in 1966 had sold at
least NCr $100.00 (United States $45.45) of the commodities
under study (rice, beans, cotton, maniec, and milk). Since
the main interests in the LAMP study centered about the
marketing of certain commodities, the observation of farms
which introduced less than NCr $100.00 into the marketing
system were believed to contribute relatively little toward
understanding the problem at hand. Although exclusion of
such farms is perhaps unfortunate for the purposes of the
present study -- decision-making with regard to agricultural
innovations -- it should be emphasized that annual sales at
or near NCr $100.00 still represents a very low level of
economic activity. Thus, even though the poorest of farmers
may be omitted, the data does provide a considerable range
of farm operations as given by size cf holdings and annual
sales.

The selection of the sample was accomplished in such a

way as to allow a greater probability of sampling within those

 

 

n1
.1
h

l
‘1»

 

“ n5
.

In,

' o

A

 

 

 

 

116

municipios which were the most important suppliers of given
141

 

commodities to the Recife urban market. At the same time

the sample size associated with the selected municipios was

 

adjusted so that all producers (those selling NCr $100.00 of
a specific commodity) had the same probability -- ex poste --
of entering the sample. In addition, each observation was
drawn by a method which assured that the sample reflected the
same distribution of farm size (over three general categories:
0-20 hectares, 21-120 ha., and 121 plus ha. ~- except for
cotton and milk producers) as was estimated for the population
of producers (with annual sales over NCr $100.00) in each
particular municipio. The primary sample was compiled from a
list of property owners assembled by the IBRA, the Brazilian
Land Reform Institure; while some non-owners were included

as substitute interviews when the intended interview could
not be completed. Table 1 below presents the location and
number of observations from which the MSU/SUDENE farmer data
was developed.

A broad range of questions were asked of the farmers in
this survey; covering their attitudes and personal character-
istics (age, size of family, educational background, etc.),
the availability of farm inputs including credit, their
marketing activities and the characteristics of their farm
Operations (including the use of specific practices pp
innovations). Several of these questions seemed to be
suitable as sources of data for the present study.

_+

141Ibid., pp. A/lO—A/lZ.

h.

YJT‘rf? :

 

.IPI
.LL

~
~1-

d‘V-

.
r”.
'o

"4..

117

Table l.--MSU/SUDENE Survey of Farmers in Northeast Brazil.

 

 

Commodity Location of Area No. of Observations

Rice Alagoas G Sergipe 127
Maranhao 27

Beans Alagoas G Pernambuco 130
Bahia 51

Manioc Pernambuco 155
Cotton Paraiba G Pernambuco 165
Milk Pernambuco "126
TOTAL . 781

 

Source: MSU/SUDENE Farm Survey (1967).

All of the products and areas encompassed in the MSU/
SUDENE survey were not, however, included in the data utilized
in the present study. Milk farmers and Maranhao rice farmers
are the portions of the survey which were omitted. Since
several of the questions asked of the milk farmers were
necessarily different from the questions used with the crop
farmers, milk farmers were omitted to allow treating the
remaining areas and products as a group. The Maranhio rice
area was omitted because of the very primitive nature of the
agriculture in this area -- slash and burn. The geographical
location of the areas which comprise the data in this study
is indicated by the map on the following page.

‘SeleCted Characteristics of the
Sample Farms and Farmers

 

 

The MSU/SUDENE survey data prevides a considerable

amount of information about the farms and farmers sampled.

118

Legend

Areas Sampled in Farm Surveys
1Rice Area
2Bean Areas
3Manioc Area
“Cotton Area

  

  

Belem ,9“ I
V‘ I
{3'36} I,
Syd? , Fortaleza
($.0
as V ’
a ’l
/
’/ Maranhao
I,’ ZONA
1’ PJaui' M2111
I‘ /’|
1,
(Q I
;- Pernambuc ., ,
55 (‘9‘ ,Q'Rec1fe
% ‘lagor,/
® Sergi ‘
t.’
O
Bahia 'alvador
. ATLANTIC
Minas Gerais OCEAN

Figure 9.--Map of Northeast Brazil Showing Sample Areas
for Commodities Studied.

119
, before we pursue the affects of certain
on adoption of new technologies, is to present
of the characteristics of the farmers studied.
g is a brief summary of the crop yields, farm
sales, farmers' attitudes, educational achieve-
ication behavior, and the agricultural practices

the farmers included in the sample.

elds were generally very low for the farms

the MSU/SUDENE survey, as may be seen by

ith the average yields in areas outside the

able 2). Although average rice yields in the

0 area were somewhat higher than the average

r areas of Brazil, they were less than 50 percent
a1 yields found in the United States. Bean

e survey farms were substantially less than the
other areas of Brazil and roughly one-third of
United States. Corn yields were even smaller
n, with the average yield of the survey farms
approximately one-fifth of average yields for»
tates. Finally, cotton yields for the survey
nly about 25 percent as large as those for other

zil.

 

120

Table 2.--Average Yields, By Commodities, for the MSU/SUDENE
Survey Farms, and Other Areas.

 

 

 

Production Area Rice Beans Corn Cotton
—~Founds Per Acre Bales Per
Acre

Northeasta

Séo Francisco-~Riee 1704 -- -- -—

(Alagoas and

Pernambuco)

Irecé--Bean (Bahia) -- 437 720b «-

Algoas-Pernambuco-- b

Bean -- 415 1204 --

Cotton -- -- -- .26
All Brazil-~Outside
NortheastC 1502 715 —- 1.48
United Statesd 4123 1319 $076 --

 

Sources: aAverage 1965 yields for major crops in the various
production areas sampled by the MSU/SUDENE Farm Survey (1967).
Most of these figures can be found in Market Processes in the

 

Recife Area of Northeast Brazil, p. 8/4.
b

 

Corn yields are for 1966 rather than 1965.

CAnnario Estatistico do Brazil (1965).
d1964 United States Census of Agriculture (1964
yields).

Size of Farms and Farm Sales

 

The size of farms included in the sample vary from as
small as two acres to as large as 7,472 acres. The bulk of
these farms, however, tend to be at the smaller end of the
continuum. Table 3 provides a summary of the distribution

of farm sizes.

121

Table 3.--Distribution of Farm Size Over the Sample Farms.

 

 

 

Average Size
Production (acres)
Area Mean Median 0-49 50-244 245-1234 1234 a above
acres acres acres acres
............. t"""""""
850 Francisco-r per cen
Rice 237 37 54 31 ll 5
Irecé--Bean 180 89 39 35 26 0
Al-Pe--Bean 200 74 38 50 10 2
Cotton 256 85 34 42 19 5
Manioc 47 17 84 8 1 1
Total 224 44 51 34 12 3

 

 

 

Source: MSU/SUDENE Farm Survey (1967).

A few large and still fewer very large farms result in
an extremely skewed distribution of farm size (as evidenced
by the wide dispersion between mean and median acreages --
especially in the 550 Francisco rice area). At the same time
many farms are quite small. Over 50 percent of $50 Francisco
rice farmers and a still higher percentage (84) of manioc
farmers (where 15 percent of the sample were on farms of less
than ten acres) operate properties of less than fifty acres.

The average gross farm sales computed for the MSU/SUDENE
study is $1,185 (crop farms), but averages tend to hide a
considerable amount of information. Table 4 presents a more
comprehensive view of gross farm sales. Over 40 percent of
the farmers sampled had gross farm sales of less than

one-half the overall average ($454). Yhile there were a few

 

...;

.5

.‘o,

 

 

122

Table 4.--Distribution of Gross Farm Sales for Farmers in
the Recife Area.

 

 

 

Mean Percentage Distributiona
Production (in U.S. Less than 3455- $2273- Above
Area dollars) $454 $2272 $4545 $4546
-------------- percent-----—------
850 Francisco--

Rice $2,843 40.3 34.8 11.3 13.7
Irecé--Bean 1,274 40.8 40.8 12.3 6.0
Al-Pe--Bean 834 59.4 35.9 1.6 3.2
Cotton 665 64.6 27.4 7.3 0.6
Manioc 684 73.5 21.1 4.1 2.7

 

 

 

Source: MSU/SUDENE Farm Survey (1967).

aConverted from NCr$ at the exchange rate of 2.2 NCr$=
1.0 U.S. dollar ($455 2 1000 NCrS, $2273 = 5000 NCr$,
$4546 = 10,000 Nets).
farmers who did have quite large gross sales relative to their
Recife area counterparts, still only 1.8 percent of the farms
sampled managed to equal or exceed the average gross farm
sales for the United States farmers ($11,176 according to
the 1964 Census). Thus again we find considerable dispersion
among respondents, with many farmers whose annual sales are
small and a few with much larger sales, yet still generally
less than the average for United States farmers.

Attitudes and Beliefs

 

A farmer's attitudes and beliefs are often thought to
have an affect on the decisions he reaches. The Recife Survey
asked a number of attitudinal questions. Those which seemed

to be most applicable to the present study are briefly

n
on”
-

we

n-A
-.

..vu

"r

”A.

 

123
this section. Later we will consider them in
111.
man is a social being, not only his own attitudes
.tudes that he believes his neighbors hold may
:ct on his behavior. The Recive survey asked
‘eveal how they believed their neighbors would
in innovative member of the community and how
Teel if he were able to progress more rapidly
-ves. Table 5 gives the percentage of farmers
that their neighbors would not look favorably
circumstance.
:0 Percentage of Farmers Who Believed That Their

)uld Look Unfavorabl} Upon Someone Who Tried
:w or Who Progressed More Rapidly Than Themselves.

 

 

"Local farmers think “Local faimers wouldn't
I it's odd when some- like to see me
body tries anything progress more
new." than themselves."
----------------- percent------------------
:o--
92.0 44.9
16.7 42.9
66.7 37 5
40.4 47.7
77.6 33.9

 

 

 

l/SUDENE Farm Survey (1967).

percentages are based upon the number of
:usually 60-80 percent of the total sub-sample)
,e interpreted cautiously.

 

124

The percentage of farmers who believed their neighbors
would respond unfavorably toward an innovative member of the
community varied widely by production areas, from as high as
92 percent for 550 Francisco rice farmers to as low as 17
percent for Irecé bean farmers. The farmers' assessment of
his neighbors' attitudes toward those who advance more rapidly
than themselves was considerably less variable between areas, 3!
from 34 percent to 48 percent believing that their neighbors {
would look unfavorably upon such a person.

Six questions probing the individual's own attitudes a

 

and beliefs were selected from the Recife Survey as measures 1'
of variables which might be expected to have an effect on the
adoption of new technologies. The responses to single
questions rather than indexes of responses to several some-
what similar questions are used as variables in this study.
The reason for this approach is two-fold. First, the
epistemological relationship between attitudinal questions
and individual attitudes is problematic enough without
blurring it still further by utilizing a set of questions
which may each measure a somewhat different attitude
(variable). Secondly, there is the difficult question of
how each of the supposed measures of an attitudinal variable
should be weighted as they are combined into indexes. As a
result, we will use a more cautious approach of choosing
single questions which seem to represent attitudes which
might have an effect on innovativeness.

The following discussion briefly reviews these six

questions and Table 6 indicates the response of farmers in

Table 6.-The Percentage of Recife Farmers Agreeing With Selected Attitudinal Statements.

Siaqio Kq

pasn are Aaqi uaqn
suoddeq aeqM eas
pue item 01 iaiiaq
st 11 paisafifins

ale seatiaeid {Blnl
-Inotifie nau uaqM

125

80.3

76.6

59.6

68.9

69.7

 

ardoad Jaqio
pue sentie[ai ut A1
-Ienbe asnai uea auo

44.0

51.0

H
/

45.3

69.7

 

ale Aeqi sa sBurqi
1531 sueratuqaai oqi
3t uorienits iaiiaq
e ut aq pInom am

24.4

23.4

12.2

24.5

17.6

 

481301 his law umn
mou moag ieeA ouo
'053 JON anreoai oi
iaaiaq 5,1: ‘asoouo
oi seq auoamos 31

44.0

22.0

27.0

38.0

44.8

 

[want

eq on new; ssau
-ran 0p 03 noq noun
0: aeiieq 5; it
‘Aeuom elem exam 01

55.0

49

46.0

41.8

65.1

 

agtt srq anomdur oi
qonm op 1,uaa OUOIB
Jamie; aqi sAepenoN

67.0

98.0

74.0

89.3

87.3

 

 

Production
Area

 

sao Francisco-
Rice

Irecé-Bean

Al-Pe-Bean

Cotton

anioc

M

 

MSU/SUDBNE Farm Survey (1967).

Source:

 

126
each of the production areas.
Farmers in less developed countries are often believed

to be fatalistic. This concept is defined by Rogers as

 

"the degree to which an individual perceives a lack of

142 The Recife farmers'

ability to control his future."
responses to the following statement was used as a measure
of his fatalism: "Nowadays the farmer alone can't do much
to improve his life."

Table 6 gives the farmers' responses for each of the

production areas. Since 75 percent to 98 percent of the

 

~— _ ._
“a. _.l_
~ i ."""

respondents -- depending upon the area -- agreed with the
above statement, one might suspect that farmers in the Recife
area are rather "traditional” in their views. However, such
premature generalizations should be avoided. When Michigan
farmers were confronted with nearly the same statement,

66 percent gave responses which indicated a similar

fatalistic view.143

It should be recognized that these
statements did not specify in what sense the farmer is

fatalistic; toward the possibility of improving things on

 

his own farm or in the agricultural sector in general.
Rogers also proposes an alternative view of fatalism.

He suggests that fatalism may be simply an ad hoc

 

14"Rogers, Modernization Among Peasants, p. 273.

 

143Dale Hathaway, at 31., ‘Michigan Farmers in the Mid-
Sixties, (Research Report No. 54, AgricuItural Experiment
Station, Michigan State University) includes the following

 

question: "Today farmers can't do much to determine the way
things turn out." Sixty-six percent of their sample agreed

with this statement.

127
rationalization for failure, a psychological mechanism for

reducing cognitive dissonance.144

This view will be kept
in mind during the later analysis of the variable.

With the fatalism variable the main concern was whether
the farmer thought much could be accomplished toward improving
his life situation. The following statement deals with the
farmer's view of which of two alternative components, luck
or knowledge, is most important in achieving a better
situation: "To make more money, it is better to know how to
do business than to be lucky."

From Table 6 we find that from 42 percent to 65 percent
of the Recife farmers believed that knowing how to do busi-
ness was of greater value in obtaining increased incomes than
simply being lucky. This may seem rather strange given the
high percentage of farmers who reported that they were
fatalistic. One rationalization might be that although
farmers believe in general that they cannot do much to improve
their situation, they also believe that if there are any
gains at all to be made, one should not depend upon luck to
achieve them. Still another argument is simply that farmers
hold inconsistent beliefs. This problem will be considered
again somewhat later.

The following question probes the Brazilian farmers'
willingness to defer economic gratification: "If someone

has to choose, it's better to receive NCr $90. one year from

 

144Rogers, Modernization Among Peasants, PP- 275'276°

 

128

now than NCr $30. today." (This amounts to roughly United
States $13.63 today as opposed to United States $40.90 in
one year.) Although an imperfect measure, this question may
give some crude estimate of the farmer's willingness to
defer current income in order to obtain a larger return in
the future. Table 6 reveals that most farmers -- roughly

60 percent -- prefer the smaller amount immediately to three
times as much a year from now. It should be remembered that
Brazil has experienced a considerable amount of inflation in
recent years and that the farmer would likely evaluate

90 NCr$ a year from now as somewhat less than three times

30 NCrS now (for a reason completely apart from time
preference).145

Much has been made recently about the unwillingness of

armors in less develOped countries, especially "peasant"

4%
1'rb The

farmers, to trust anyone outside their own family.
following question was used in the MSU/SUDENE survey as a
measure of an individual's willingness to trust others:
"One can trust equally in relatives and other people."
Interestingly enough, a large proportion -- 44 percent to
70 percent -- of Recife farmers agreed with this statement
(see Table 6).

The Recife farmers' general attitude toward new

technology was of particular interest in this study.

 

14SInflation rates of 20% or more per year have been
common since World War II. A level of 20% per month was
reached in early 1964.

14bRogers, Modernization Among Peasants, pp. 26-28.

 

 

_,.

129
Agreement or disagreement with the following statement was
utilized in attempting to assess this attitude: "We would
be in a better situation if the technicians left things as
they are." This question was intended to elicit the farmer's
overall attitude toward the change~generating activities of
people in these positions. Negative responses should indicate
a hopeful view of technical change. A positive response
would suggest that the farmer believes that the activities
of scientists and technicians -- introducing new methods --
can offer little to improve his individual well-being. In
general, Recife farmers exhibited a rather favorable attitude
toward new technology as only 12 percent to 25 percent
agreed with the above statement (see Table 6).

Also of considerable interest in this study was the
farmers' unwillingness to try new methods without first
observing them in use by others. The following statement
lends itself to this interpretation: "When new agricultural
products are offered it is better to wait and see what happens
when they are used by others." As might be expected, a
large proportion -- 60 percent to 80 percent -- of the Recife
farmers agreed with this statement.

Table 6 summarizes the attitudes discussed in this
section. Although a good deal of variation occurs between
production areas, a general profile of the attitudes of
Recife farmers can be developed. They are generally
fatalistic, even though it is fairly likely that they may

believe that knowing how to do business is a more dependable

130
way of getting ahead than being lucky. The chances are
greatest that they will prefer a smaller amount of money now
to a larger amount in even the near future. They are
surprisingly favorable toward new technology and about
equally split in their willingness to trust peeple who are
unrelated to them. Finally, over two-thirds of these farmers
believe that the best strategy is to let someone else try new
methods first.

Educational Achievement

 

The level of formal training received by the Recife
area farmers is generally low. In Table 7 the educational
accomplishment of Recife farmers is divided into the six
groups used throughout this study.

Table 7.--The Educational Achievement of Recife Farmers.

 

Years of Education templeted

Production 2-3 4-5 6-12 more than
Area none 1 year years years years 12 years

 

$50 Francisco--

Rice 31.7 12.7 23.8 19.0 11.1 1.6
Ireeé--Bean 30.6 28.6 24.5 6.1 8.1 2.0
Al-Pe--Bean 55.5 10.2 21.1 7 0 4.0 2 4
Manioc 45.3 19.3 22.0 8.0 5.4 0.0
Cotton 46.6 22.4 20.0 7.2 3.0 0.6

 

Source: MSU/SUDENE Farm Survey (1967).

From the table, we find that roughly one-third or more
had not attended school. Again, except for rice farmers, over

80 percent of the sample had completed less than four grades

131
of schooling. Even so, a few farmers had completed an
elementary level education or attended secondary school,
and a smaller number still had achieved more than a
secondary level of education. If formal educational achieve-
ment can be thought of as influencing the individual's ability
to process information about new methods -- literacy and
conceptual ability -- a wide range of such abilities should
be encountered among Recife farmers.

Communication Behavior

 

The Recife farmers were asked where they had learned
about the newest method they were using. Their responses
are given in Table 8.

Table 8. —-From Where or lhom Recife Iaimers Had Learned the
Latest Farming Method They h'ere Using.a

 

 

Information Source

 

 

Production Rela- Neigh- Agron- Busi- People Mass Another
Area tive bor omist ness- from a Media Source
man different -
place
- ------------------- pereent------------------4---
Sio Francisco--

Rice 10.0 53.7 12.7 3.6 7.2 2.8 10.0
Irecé--Bean 26.1 54.3 10.9 4.4 -- 2.1 --
Al-Pe--Bean 14.6 54.0 8.9 9.7 12.1 0.8 --
Manioc 20.0 50.9 10.5 3 7.9 -- 6.2
Cotton 25.0 49.3 11.8 6.3 3.4 ' .2.7 1.4

 

Source: MSU/SUDENE Farm Survey (1967).

3Computed on the basis of respondents who were using one
of several "innovations" selected by SUDENE/MSU researchers
and who indicated where they had learned about the one most
recently applied. Some farmers were either using none of
these methods or didn't respond to the question.

132

The importance of "indirect sources," neighbors and
relatives, is immediately obvious. Neighbors usually
accounted for about 50 percent of the responses, and with
the addition of relatives, two-thirds or more of the
responses are accounted for by indirect sources. "Direct
sources," such as the agronomist and mass media, account
for approximately 10 percent to 15 percent. Other sources,
which lie somewhere between these ”direct" and "indirect"
sources, account for the remaining responses.

The use of radio, farm magazines, and attendance at
extension meetings was also examined by the Recife survey.
Table 9 presents the percentage of farmers who had ever read
a magazine about agriculture, attended an extension meeting,
and those who have a functioning radio in their home.

Table 9.—-Pereentage of Recife Farmers Who Have Read 3 Far

iagazine, Attended an Extension Meeting, and Have a
Functioning Radio in Their Home.

 

Production Read a Farm Attended an Have a Func-
Area Magazine Extension tioning Radio
Meeting in the Home

 

Sao Francisco--

Rice 18.5 19.5 76.4
Irecé--Bean 32.7 26.5 65.3
Al-Pe--Bean 7.0 16.4 50.8
Manioe 12.6 10.7 ~ 56.7
Cotton 14.5 9.7 52.7

 

Source: MSU/SUDENE Farm Survey (1967).

133

While more than one-half of the Recife farmers had a
functioning radio in their home, a much smaller percentage
had read a magazine about agriculture. Attendance at
meetings where agronomists talked about production methods
was reported by an average of 14.3 percent of the sample
farmers with Irecé bean farmers indicating the highest level
of attendance. In all, the data suggests (deSpite the
rather frequent presence of radios in the home) that the
bulk of information about new farming practices comes from
an "indirect" source, flowing from farmer to farmer.

Use of Selected Agricultural Practices

 

Certain farm practices were selected by the MSU/SUDENE
research team as "innovations." The researchers believed
these practices to be useful improvements over present
methods and to represent a suitable base for measuring the
innovativeness of Recife area farmers. A complete assess-
ment of the economic contribution of each of these practices
is not available. Thus we have to trust the judgement of the
researchers who conducted the field work. The percentage of
Recife farmers using these selected practices is given in
Table 10.

The use of these practices vary considerably from area
to area. Fertilization of crop land was not a common
practice among the Recife farmers sampled. The application
of barnyard manure was reported by 24 percent of the 850
Francisco rice farmers, 35 percent of the manioc farmers,

and less than 10 percent of the farmers in the other areas.

134

Table 10. ~-Percentage of Farmers Using Selected Innovations
By Commodity, 1967

Vi

‘ . . n

 

 

 

 

 

 

 

 

 

 

 

 

 

0
Production 'o - ,fi
Area at o m ' H, .c
EU 0 d.) 3 d) U
0:4 H U) 'U . o H, . w
«40) <11 -H 01 1-1 ‘3 . 02
EN HN “c: U +4 o. :6 "d
o..—( gun-4 Q) 0H C‘. H H "450
.CH UH > +4 «3 "U On O_ U":
U-H weer-4 0 U 00 c: H °H°H
.4 u E+J s4 0 «4 m 'o u. 1944
CH Oh c. m E. c: 63 mm
00.) .230 E G :3 K N, H 03
Zn on e4 H m o, m .H.mn
, —w.
850 Francisco-« ' ' L ‘ I 1
Rice 24.4'11.8 58.3 70.1 26.8 23.6 10.2 51.2 18.9
Irecée-Beans 0.0L 0.0 65.3 44.9 89.8 14.3 49.0 65.3 24.5
Al-Pe--Beans 9.4 0.0 56.3 14.1 73.4 94.5 80.5 1.6 7.8
Manioc 35.3 0.7 50.0 18.7 0.0 8.7 0.0 2.0 20.0
Cotton L7.9l 0.6 52.1 64.8 0.6;34.5 0.6 4.2 91.5
Source: MSU/SUDENE Farm Survey (1967).

The use of chemical fertilizer was even less frequent, with
the $50 Francisco rice farmers being the only group to have
a significant number of its members (12 percent) employing
this practice.

The use of improved seeds was reported by over half the
respondents in each area, but this response is probably not
very meaningful as the data seems to reflect the selection
of the biggest and best looking seeds from the individual's
own production rather than the use of improved strains,
varieties, or hybrids. Insecticides were quite commonly used
by farmers in the $50 Francisco rice area (70 percent), the
cotton producing area {65 percent), and Irecé bean area

(45 percent); while less than 20 percent of the Alagoas-~

135
Pernambuco and manioc areas used this practice. The
application of fumigants to stored crops was mostly found
among the Ireeé (90 percent) and Alagoas-Pernambuco
(73 percent) bean farmers, although a few (24 percent) of
the $50 Francisco rice farmers were using this practice.
The ox and plow and the hand planter were practices most
frequently used by Alagoao-Pernambuco bean farmers
(95 percent and 81 percent respectively). The use of
tractors, on the other hand, seemed to be concentrated in
the $50 Francisco rice (51 percent) and Irecé bean
(65 percent) areas. Finally, the highest percentage of
farmers using a pesticide dusting machine was found in the

cotton area (32 percent).

CHAPTER V

ANALYSIS OF SELECTED VARIABLES AND THEIR
RELATIONSHIP TO THE ADOPTION OF AGRICULTURAL
INNOVATIONS BY FARMERS IN NORTHEAST BRAZIL

Introduction

 

The conceptual framework proposed in Chapter III describes
the adoption of new technology as an individual decision-
making process involving the interaction of a number of
variables. Using the data from the MSU/SUDENE survey of
farmers in Northeast Brazil, this chapter provides an
empirical examination of a portion of this conceptualization.
For the present, the emphasis will be upon those variables
which might be expected to affect the individual’s evaluation
of a new agricultural practice, whil- the next chapter
examines some possible constraints upon the implementation
stage of the adoption process. Attention is given to the
inter-relationships between the variables as well as their

relationship to the use of new agricultural methods.

Community_Attitudes

 

 

As suggested in the conceptual statement, the attitudes
a farmer perceives that his neighbors hold might be expected
to influence his decisions toward new agricultural practices.
This proposed relationship was examined in the context of the
Recife farmers' perception of how local farmers would react
to someone who tries something new or someone who progresses
more rapidly than his neighbors.

Combining the responses of sample farmers, it was

136

137

discovered that a direct relationship between perceived com-
munity attitudes and the farmer's use of certain modern
practices could not be supported statistically. Farmers who
believed their neighbors would respond negatively toward
those members of the community who tried something new or
who advanced more rapidly than their counterparts were no
less inclined to use barnyard manure, chemical fertilizer,
field insecticides or tractors than those individuals of the
Opposite persuasions. Given the possibilities for inter-
actions of numerous variables in the decision-making process,
the attitudes of the Recife farmers' neighbors may still be
of some influence in the adeptien process even though they
are not by themselves useful as predictors of innovativeness.

Setting aside this matter momentarily, it would be of
interest to know if those farmers who believed that their
neighbors would resent the more rapid advancement of a
community member also thought that their neighbors would
laugh at innovators. Although the two opinions are not
necessarily interconnected, there is a tendency for them to
be associated. Of those who believed local farmers would
disapprove of their more rapid progress, 54.3% reported that
their neighbors would think it was funny if someone tried
something new. Among farmers who held the opposite opinion,
41.3% believed that their neighbors would laugh at
innovators. (This association was significant at less than
the .005 level using chi-square analysis.) Thus, there seems

to be some evidence of an interconnection between these views

138
in the rural Recife communities, at least as the farmers
perceive them. Yet even farmers viewing their communities
as both Opposing individual advancement and regarding
innovators as foolish were not found to be significantly
less inclined to adopt certain modern practices (chemical
fertilizers, insecticides, and tractors) than those holding
other expectations about their neighbors.

We return to the question of what relationship, if any,
is there between perceived community views and individual
behavior toward new agricultural practices. One might
suggest that negative neighborhood attitudes toward
innovators or individual advancement would influence a
farmer's evaluation of new technologies. He might be
expected to discount the value of new technologies which
promise economic advance if he presumes that his community
will look Upon such progress unfavorably. Or, he may be
influenced by his neighbor's assessment of innovators as
foolish in developing his own general evaluation of new
ideas. There seems to be some justification for these
propositions, even though farmers in Northeast Brazil seem
to generally view the activities of technicians in a hopeful
light. (Recall that less than 25% of the Recife farmers
believed things would be better if technicians left things
alone.) By far the largest percentage of those holding a
negative view of new technology were farmers who believed
their neighbors would not like to see them get ahead -—

62.0%. The reverse is true of a positive attitude toward

139

new technology where only 47.9% of the group were fearful
that their neighbors would not like to see them advance more
rapidly than the rest of the community. (A chi-square
analysis of these relationships was significant at the .01
level.) The results are much the same for the farmer's
expectation of his neighbor's attitude toward innovators and
its relationship to attitudes toward the activities of
technicians. Here 63.0% of those holding a negative attitude
toward new technology believed their neighbors would regard
innovators as foolish, while 47.9% of those with a positive
attitude believed this to be the case. (A chi-square analysis
was significant at less than .005 for these relationships.)
This would seem to imply that negative neighborhood attitudes,
as perceived by the individual farmer, contribute to the low
esteem in which the activities of technicians are held by
certain producers.

A similar argument can be made for the significant
relationship between the farmer's view of neighbors'
attitude toward personal advancement and his knowledge of
how to improve crop yields. Recife farmers were presented
with the following question: "How would you try to increase
output without cultivating more land?"

While 62% of the Recife farmers were able to articulate
some method of increasing present yields -- using fertilizers,

insecticides, irrigation, closer spacing of crops, etc. --

 

NOTE: Unless otherwise stated statistical significance is
reported at the .05 level.

140
the remaining farmers seemed to have no idea how they could
increase present output without bringing more land under
cultivation. Of those who believed their neighbors would not
like to see them get ahead, 58% knew of some method to improve
yields. 3' comparison, 67% of those Recife farmers holding
the opposite view were able to suggest a method for increasing
yields. These relationships were significant at the .05 level.
(A similar difference was found between the farmers who held
opposing views about their neighbors' attitudes toward
innovators, but it was not of a statistically significant
magnitude.) An explanation for this finding might be that
individuals who believe their neighbors would look
unfavorably upon them getting ahead are less motivated to
learn ways to improve their situation.

There is some additional evidence which suggests that
farmers who believe their neighbors would respond unfav01ably
to their individual advancement were less aggressive in
gathering information about new agricultural practices.

While most of the Recife farmers reported that they had
learned the latest practice from neighbors or relatives --
72% -~, the importance of this "indirect" source of informa-
tion var'ed significantly with the individual's perception
of his neighbor's attitudes. By comparison, 66% of those
who believed their neighbors wouldn't Object to their
individual advancement depended upon neighbors and relatives
as a source of information as Opposed to 78% of the farmers

who perceived their neighbors to hold the Opposite view.

141
This suggests a tendency on the part of those who would
expect community disapproval to be less likely to use the
more direct sources of information -« local businessmen,
agronomist, and mass media -- in obtaining their ideas about
new agricultural practices. (See Table 11.)

Table ll.--Information Sources Used by Recife Farmers Holding
Opposing Expectations of Community Attitudes.

 

"Local farmers

 

 

o ' u ‘ I I a
would not line Percent Usrng lhese Information SerV1ces
to see me pro-
gress more Neighbors People From Local Agronomist
rapidly than and a Different Businessman and
themselves." Relatives Place Mass Media

Agree 78 7 3 12
Disagree 66 8 10 16

 

aRepresents 72.3% of the sample. The unspecified
category termed "another source” is omitted and a large
number of farmers did not give a response.

At this point we have some evidence that community
attitudes may have a certain amount of influence on the
decision-making process through their effect upon the
individual's general attitude toward new technology and upon
his motivation for learning about ways of improving his
situation. There is another interesting dimension to
Recife farmers' perception of community attitudes. Somewhat

surprisingly 58% of those farmers who believed business

knowledge to be superior to luck in earning more money also
thought their neighbors would resent seeing them get ahead.
At the same time, of those who were inclined to think of

higher earnings as a matter of luck, 46% believed that local

 

142

farmers would disapprove of their advancement. A chi-square
test of these relationships indicated that the knowledge-
oriented farmers were significantly (at the .01 level) more
likely to expect neighborhood resentment of their progress.

Farmers favoring business knowledge over luck were also
more likely to be among the ranks of those who believed local
farmers would regard innovators as foolish -- 57.3%, while ' a
less than half -- 46.7% -- of those counting on luck believed
that an innovator would be laughed at in their community. A

chi—square test found this association to be significant at

 

the .05 level. ' E
It would certainly be difficult to postulate an acceptable
causality between negative community attitudes toward individual
advancement and innovativeness and a choice of business know-
ledge over luck as a means of earning more money. The more
useful approach seems to be to reason in the opposite direction.
Perhaps the roughly 50% of the Recife farmers who chose busi-
ness knowledge over being lucky represent a somewhat more
progressive, business-oriented, segment of the sample. This
possibility seems to be supported by the fact that a choice of
business knowledge is associated with farmers with relatively
higher incomes and larger acreages. These farmers may have
done things that were "different" in the past and may have
progressed more rapidly than their neighbors, and thereby
personally experienced the resentment of local farmers. Since

expectations of community behavior (which is the basis of the

 

community variables under discussion) are the product of past

143

experiences, this group of farmers may be more sensitized
to community attitudes than their less progressive counter-
parts. In a somewhat similar vein, Rogers found that innova-
tive farmers often remarked that they could not win popularity
contests by their method of operation.147

In summary, the MSU/SUDENB Survey has provided us with
some useful data on the effects of perceived community
attitudes upon the decision-making of Recife farmers. Non-
progressive attitudes toward innovators and individual
advancement seems to be associated with low esteem for the
activities of technicians, less aggressive information
seeking, and less knowledge of how to improve yields. On
the other hand, slightly more progressive, "business-oriented”
members of a community may be somewhat more likely to assess
their neighborhood's response as negative toward innovators

and individual achievement.

Individual Attitudes

 

The logic behind the expected impact of the six attitudes
selected from the Recife survey will be discussed in this
section and an attempt will be made to determine if the MSU/
SUDENB data lends support to the arguments.

Fatalism

Our assumption in this study is that fatalism is an

attitude which might be expected to inhibit innovativeness.

(Note that it is not termed a "barrier" to innovativeness.)

 

147R . .g. .
ogers, QiiquIOn, p. 200.

 

144
The fatalistic farmer, it would seem, will be less willing --
than his non-fatalistic colleagues -- to entertain the notion
that adoption of new methods will improve his situation. His
fatalistic view of life may be presumed to negatively bias his
expectations toward innovations. This does not p:gvent_his
eventual adoption of new methods since additional information,

collected over time, can ultimately cause a revision of his

3. 'I'A --

expectations. It would, however, retard his progress toward
acceptance of innovations. §

The majority of Recife farmers agreed with a fatalistic

 

view of their situation, which is apparently not uncommon

(as indicated by Michigan data) to farmers elsewhere. Rogers
(see page 14} states that fatalism is a characteristic of the

148 . . .
The Rec1fe data, however, ndicates

peasant subculture.
that fatalism is not limited to small and poor farmers. In
fact, the poorest farmers in the sample were not significantly

more fatalistic than their wealthier counterparts. (See

Table 12 below.)

 

148 i o o . , '-
Rogers, Mouereizatien Among Peasants, p. 33.

r-
__

 

145

Table 12.--Percentage of Recife Farmers Who Expressed A
Fatalistic View By Income Levels.

 

Per Capita "Nowadays the farmer alone cannot do much to

 

 

 

Family improve his life."

Income .
(U.S. Dollars) Agree Dlsagree

---------------- percent-------------——------

Less than $50 82.8 17.2
$50 - $99 81.3 18.7 F
$100 - $199 82.4 17.6
$200 - $499 81.9 18.1
$500 and above 78.0 22.0 '

 

Vim .

Nor did fatalistic farmers operate smaller acreages.
Only the farmers on the very largest acreages were less
fatalistic (although not significantly so by a chiesquare
test) than the smallest farmers.

Table 13.--“ercentage of Recife Farmers Who Expressed A
Fatalistic View By Farm ize.

"Nowadays a farmer alone cannot Number
do much 39 improve his life." of

~.“-_ a-” —

 

 

3 Agree Disagree_ Observations
U - 49 84.6 15.3 322
50 - 244 82.2 17.8 214
2A3 - 1234 86.1 13.9 73
1235 and above 66.7 33.3 18

 

 

Education and age also did not seem to be associated with
fatalism. Fatalistic farmers were, however, significantly less

likely to defer income than their non-fatalistic neighbors.

146
Of the fatalistic farmers, 37.8% stated that they would
prefer 90 NCrS one year from now to 30 NCrS immediately.
At the same time, 49.5% of the non-fatalistic farmers would
be willing to wait a year to receive the larger sum.

Of perhaps greater interest is the effect of fatalism]
non-fatalism upon the adoption process. Fatalistic farmers
were not found to be significantly less likely to adept
tractors, chemical fertilizer, or insecticides than nen-
fatalistic farmers. But this relationship is beyond what
was postulated earlier. Recall that we thought it reasonable
to expect fatalistic farmers to be "less willing to entertain
the notion that adoption of new methods will improve his
situation." There is some evidence that a fatalistic view
did ”retard progress toward acceptance of innovations."
Hovever, it did not seem to manifest itself in a negative
attitude toward the activities of technicians (perhaps
because technicians offer a hope for improvements that
individual farmers feel incapable of accomplishing them-
selves). Yet the fatalistically inclined farmers exhibited
considerably less knowledge of new methods which might
improve the productivity of their farms and subsequently
their well-being. Table 14 indicates that the fatalistic
farmers were less inclined to know about fertilizer

specifically and about ways to improve yields in general.

 

W4 HW-

147

Table l4.--The Percentage of Fatalistic and Non-Fatalistic
Recife Farmers Who Knew About Fertilizer and Knew
How to Improve Yields.a - ' ,

 

 

Know What Know How to
Attitude Fertilizer Is Improve Yields
--------------- percent--------—-—---------
Fatalistic 54.8 59.4
Non~Fatalistic 73.5 74.5

 

aThese 1elationships were significant at the .005 level
using chi-square analysis.

This finding seems to indicate that non—fatalistic
farmers were more highly motivated to learn about new methods.
In addition, non-fatalistic farmers tended to use more direct
Sources of information (in learning about the newest method
they were using) than the fatalistic producers. Table 15
describes this situation.

Table lS.--Information Sources Used By -atalistic and Non-
hatalistic Farmers in the Recife Area.a

... _

Percent of Farmers Using These Information Sources

 

 

Relatives People From Agronomist
Attitude and a Different Business- and

Neighbors Place man Mass Media
Fatalistic 74.2 7.8 5.7 12.2
Non-Fatalistic 68.6 2.4 9.6 19.9

 

aRepresents 82.5% of the Recife sample of crop farmers.
An unspecified category termed "another source" was omitted
and several farmers did not respond to the question.

The percentage of fatalistic farmers who had attended

meetings where an agronomist had talked about new farming

148
methods was, however, not significantly different than for
non-fatalistic farmers. Nor were non-fatalistic farmers
more likely to have read a farm magazine. Radio listener-
ship, on the other hand, exhibits a unique pattern of
differences between fatalistic and non-fatalistic farmers.
The non-fatalistic farmers listen somewhat more frequently,
but listen fewer hours than their fatalistic colleagues.

In summary, "fatalism" seemed to be largely independent
of age, income level, farm size and the education level of
the respondent. The evidence suggests the notion that non-
fatalistic farmers are more aggressive in acquiring knowledge
about how to improve farm production. They tend to look
beyond relatives and neighbors more frequently in seeking
this information, although they don't seem to have read farm
magazines or attenied sessions with an agronomist more often
than fatalistic farmers. In general, the MSU/SUDENE data
indicates that fatalistic attitudes may have an inhibiting
effect upon the rate at which farmers form favorable
expectations toward new technologies. That fatalism does
not influence the decision—making process, but instead is
an ex pesto explanation for failure as suggested by Rogers,
tends to be negated by the previous analysis.

Luck versus Knowledge

 

The Recife farmer's choice between luck or knowledge as
the most important component of achieving higher earnings
seems to be related to past business success (due either to

\

good management or good fortune}. Those who believed in

business knowledge tended to earn higher incomes and operate

- a? a Lung“ _._._'

~_ 'l- A».

 

149
larger farms than farmers who favored being lucky. Tables 16
and 17 illustrate these relationships.
Table l6.--The Importance of Business Knowledge in Earning

Higher Incomes as Judged By Recife Farmers at
Various Income Levels.a '

 

 

 

Family Per "To make more money it is better to know

Capitq Income how to do business than to be lucky."

LevelB Agree Disagree
--------------- percent---------------------

Less than $50 42.3 57.7

550 - $99 45.2 54.8

$100 - $199 50.5 49.5

$200 - $499 62.9 37.1

$500 and up 67.3 32.6

 

I
d

Chi-square test for these relationships was significant
at .005.

b. . . . .
Family income from all soarces dixided by number of
persons in the household, a measure of level of living.
(U.S. dollars)

 

Table l7.--The Importance of Business Knowledge in Earning
ligher Incomes as Judged By Recife Farmers on
Various Sized Farms.d

 

 

 

 

Farm Size "To make more money it is better to know
in Acres how to do business than to be lucky."
Agree Disagree
---------------- percent--------—-v—------~-

0 - 49 45.1 54.9

50 - 244 47.8 . 52.2

245 - 1234 60.6 39.4

1235 and above 66.7 . 33.2

 

aChi-square tests gave a significance level of .05.

150

Apparently, relative business success reinforces the
belief that knowing how to do business is the best means of
improving one's situation, while less fortunate farmers tend
to attribute success or failure to luck. Educational achieve-
ment, which is also related to income and farm size, seems to
be associated with the farmer's choice between luck and
knowledge. Table 18 describes the greater likelihood of
Recife farmers believing in the advantages of knowing how
to do business as their level of education increases.

Table 18.--The Importance of Business Knowledge in Earning
Higher Incomes As Judged By Recife Farmers With

I

Various Levels of Education.d

 

 

 

 

 

Last Year of ”To make more money it is better to know
School Completed __how to do business than to be lucky."
Kgree Disagree
--------------- percent---------+-—-----~---

NOIIC 38.2 61.8

1 year 46.3 53.7

2 - 3 years 53.6 46.4

4 or more years 70.3 29.7

aChi-square table was significant at the .005 level.

It is well to know that farm size, income levels, and
educational achievement are the antecedents of this attitude,
but of more importance is some knowledge of its consequents.
Among poorly educated farmers, (roughly 60% of Recife farmers
were functionally illiterate) a belief that luck is the more
important component in business success may have a significant

effect on their decision-making process. It may be reasonable

151

to expect that farmers who would depend on luck are not as
motivated to seek out information as their knowledge-oriented
colleagues. Since the formation of realistic expectations
usually involves a certain amount of information gathering
and assimilation, the farmer's attitude toward luck versus
knowledge may be thought to affect the pace of innovation
adoption.

Like fatalism, there was no significant relationship to
be found between luck/knowledge and the adoption of non-
chemical fertilizer, chemical fertilizer and insecticides.
But, unlike fatalism, farmers who favored business knowledge
used tractors significantly more often than their associates
who regarded luck as the source of higher earnings. Tractor
usage was reported by 21.4% of those favoring knowledge,
while 13.7% of the remaining farmers used tractors. It
might seem reasonable to explain this in terms of the
relationship between luck/knowledge and income and farm
size, since tractor usage is significantly related to both
of these variables. The logic of a relationship between an
expensive tractor -- selling for about $6,000 in the Recife
area149 -- and larger farm size and higher incomes seems to
present an imposing argument for such reasoning. The
difficulty is that almost 30% of the Recife farmers reported
the availability of rental tractors in their area. In

addition, the use of chemical fertilizer is also significantly

 

9 , . , ' . ...f

14 Slater, Riley, 3: 31., Market Processes ...., reports
the prices of two tractors sold in the Recife area: Massey-
Ferguson 50X -- $6,395.45 and Valmet -1 $6,090.90.

 

152
related to both farm size and farm income, but not
significantly related to luck/knowledge. One can only
speculate, then, that the relationship between luck/
knowledge and tractor use mix reflect some of the impact
of farm size and income.

The information sources used by farmers expressing these
opposing views seems to support the contention that farmers
who depend upon luck are not as motivated to seek out informa-
tion as those who are inclined toward business knowledge as
the most important factor in financial success.

Table l9.--Information Sources Used by Recife Farmers Differing
In Their View of the Source of Higher Earnings.d

 

 

”To make more Percent of Farmers Using

money it is These Information Sources

better to know Relatives 'PLOple From Agronomi§t_—
how to do busi- and a Different Business- and

ness than to be Neighbors Place men Mass Media

lucky."

Agree 67.6 7.7 8.1 10.6

 

Disagree 78.2 6.5 4.6 10.3

 

aChi-square table was significant at the .05 level.

In gen-re , those farmers associating themselves with
business knowledge tended to learn about innovations from
relatives and neighbors less frequently and use more "direct"
sources with greater frequency than those farmers who
believed that luck controlled their fortunes. They were
significantly more likely to have attended meetings where
agronomists talked about new farming methods (18% attcnded

as opposed to 11.9%) and were more likely to have read a

..
mini»: :3?

 

nmm~

153
farm magazine (37.2% versus 26.2%). With respect to
listening to the radio, however, the behavior of farmers
who put their faith in business knowledge was not
significantly different than those who believed in luck.
Again a significantly (.005 level) greater percentage of
the business knowledge-oriented farmers knew about
fertilizer (65.6%) than their luck—oriented counterparts
(50.2%). Yet, they were not particularly more likely to
know how to improve yields. The evidence also suggests
that there is not a relationship between luck/knowledge and
the Recife farmers' attitude toward new technology. This
seems to indicate that a business—orientation does not
necessarily imply a favorable attitude toward new methods.
Such a finding might come as a surprise had we not
discovered earlier that fatalism and luck/knowledge were
apparently unrelated. While each of these attitudinal
variables should be recognized as representing a different
aSpect of the farmer's view of the world ("image"), it
would be intellectually more satisfying if all of his views
were completely consistent. This is undoubtedly too much
to expect of man. Certainly "rationality," as defined in
this study, is not negated by the apparent inconsistencies
in the belief patterns of (largely poorly educated) Recife

farmers.isq

M

1501f consistency would require that the farmers‘
beliefs be nonvfatalistic + business knowledge + positive
attitude toward technicians or the reverse, only 15.2% of
the Recife sample would hold consistent beliefs.

154

In general, the data from the Recife area indicates
that the business knowledge-orientation of the farmer, which
was strongly related to past success, appears to influence
the decision—maker's motivation to seek information. A
business knowledge—orientation does not, however, imply a
favorable attitude toward new technologies in general
(although the greater information-seeking efforts of these
farmers probably implies a more rapid formulation of a
favorable image toward specific new methods).

Deferred [atone

 

The survey question involving a choice between NCr $30.
immediately and NCr $90. a year from now was used here to
measure the relationship between the Recife farmers' time
preference for income and his adoption of new methods. It
was reasoned that the farmer whose time preference is
weighted heavily to the present would be relatively unwilling
to sacrifice present consumption in order to finance the use
of new methods which promised higher future earnings.

Many of the innovations available to traditional
farmers involve additional cash expenditures. In the absence
of abundant credit, the farmer must forego part of his
present income in order to earn the higher future returns
available with new methods. His willingness (or unwilling-
ness) to sacrifice current consumption on the expectation of
a higher level of consumption at a later pornt in time might
have an effect on how favorably he views innovations. Net
unexpectedly, the Recife farmers‘ time preference, as measured

by this question, was not significantly related to his use of

“”377?

155

tractors, insecticides, and fertilizer. Since there is no
reason to believe that his attitude toward deferring income
would influence the sources of information used or his
knowledge situation, the influence of this variable, within
the decision-making process, cannot be tested further until
Chapter VI (where the interaction of several variables will
be examined using an Automatic Interaction Detection program).

In the meantime there are certain other relationships
which are of interest. One in particular is the association
between deferred income and fatalism. Fatalistic farmers
were found to be significantly less inclined to wait a year
for larger returns than non'fatalistic farmers. By comparison,

37.8% of the fatalistic farmers chose the NCr $90. future

L

lternative, while 49.5% of the non-fatalistic farmers were

9)

willing to wait for the larger sum. Thus, it appears that
farmers who believe they can do little to influence the
future have a tendency to live for today, while farmers with
a more hopeful View of the future are more willing to wait
for future rewards. With respect to the luck/knowledge
dichotomy, we find that farmers who believed that higher
earnings are largely a matter of luck were significantly
(.01) less likely to choose the future return than their
business knowledge-oriented counterparts (35.4% as opposed
to 45.2%). Again an underlying confidence or lack of
confidence in the future seems to be at work in the associa-
tion of these variables. A somewhat surprising development

was the lack of association between ’illingness to defer

156

income and income levels. One would expect that the immediate
needs of the poorer farmers in the sample would severely limit
their willingness to defer income. Perhaps the wording of
the question reduces the possibility of such an outcome. The
specific wording gives the impression of a hypothetical gift
for which the respondent can choose the timing. If the farmer
were asked whether or not he would pay NCr $30. now for a
certain income of NCr $90. one year from now (more directly a
time preference investment question), one would certainly
expect the immediate needs of the poorer farmers would be
reflected in their choices.

There is little that may be concluded (at this point)
regarding the effect of a willingness to defer income upon
the adoption process. The effect of the variable was not
sufficiently strong to be independently related to the use of
certain innovations. Nor was there reason to expect it to
influence other variables within the decision-making system.

Attitude Toward New Technology

 

The majority of the Recife farmers -- according to the
MSU/SUDENE data, were inclined to favorably view the activities
of technicians, which has been interpreted to mean a rather
positive attitude toward new technologies in general. he
farmers who didn't share this view, as was noted earlier,
were somewhat more likely to believe that their neighbors
would disapprove of their individual advancement_or regard
innovators as foolish. Beyond this association, however, the_

data provides little by way of relationships with other

157
variables which could be utilized in characterizing the
farmers holding opposing views toward new technology. The
likelihood of a positive (or negative) attitude toward new
technology was not found to be a function of educational
achievement, income level or farm size. Nor was it at all
related to fatalism or luck/knowledge.

0f greater interest, however, is the influence of
attitudes toward new technology on the adoption of agricultural
innovations. The general perspective in which farmers View
new technology may be expected to bias their decision-making
with respect to innovations. The farmer rho perceives new
methods as a route to improvement of his situation will tend
to expect favorable results from innovations. This does not
imply that he immediately accepts every innovation disregard-
ing the uncertainties surrounding it. It does mean that he
engages in the process of information gathering and expecta-
tions formation under the assumption that the innovation might
be found useful. The farmer whose general attitude toward
new methods is negative may be expected to seek out little
in the way of additional inferration about innovations. He
would be expected to believe that any additional information
would simply support his present negative conclusions.

Since the majority of farmers seemed to be generally
favorable to new technologies, we will address our analysis
to examining the ways in which the minority «- those who view
new technologies negatively vs behave differently from the

rest. The Reci’e farmers' use of tractors and fertilizer

158

was not related to their overall attitude toward new
technologies, but insecticide usage did seem to be associated
with their view toward new methods. Forty-three percent of
all Recife farmers reported that they used insecticides. A
slightly higher percentage (45.7%) of farmers with positive
attitudes toward new technology applied insecticides, while a
significantly smaller proportion (35.8%) of the farmers with
negative attitudes used this practice. In addition we find
that farmers possessing negative attitudes toward new methods
are significantly less likely to know how to improve yields
than other Recife farmers f- 45.9%, as opposed to 68.1%. This
evidence begins to support the contention that farmers with a
negative view toward new technologies in general seem to
utilize this opinion in developing their expectations toward
specific innovations, and are consequently less likely to
seek out additional information about new practices.

Still more support for this hypothesis is provided by
an analysis of the information sources used by Recife
farmers divided along their views toward new methods. Local
(indirect) sources were reported significantly (.005) more
often by farmers with the negative attitudes as being the
origin of their knowledge about the newest practices they
were using.

On the other hand, the proportion of farmers utilizing
the agronomist or mass media as a source of information
was three times greater for producers with positive attitudes

toward new technology than for their counterparts with

159

Table 20.--Information Sources Used By Recife Farmers Divided
According to Their Attitude Toward New Technology.

 

Percent of Recife Farmers

 

 

Attitude ‘ ' ’ UsingiThese Information Sources
Toward New Relatives People from Agronomist
Technologies and Another Local and

. . . Neighbors Place .Businessmen4.Mass.Media
Positive 68.3 8.1 6.7 16.8
Negative 84.1 . 5.3_ .. 5.3__ ~ . 5.3

 

negative views.

Based upon the available data, the Recife farmers with
negative attitudes toward new technologies were also
significantly less likely to have read a farm magazine than
their positiIe-oriented colleagues —- 16.4% versus 36.8%.
They listened to the radio significantly less often -- 38.3%
as opposed to 29.6% seldom listen, while 20.3% as against
33.2% listen more than one hour per day. Yet, surprisingly
enough, farmers with negative attitudes toward new methods
were not any less likely to have attended a meeting where an
agronomist had talked about new production methods.

In all, the weight of the evidence seems to suggest
that Recife farmers who hold a generally unfavorable view of
new technologies were less inclined to seek out additional
information, upon which to base their expectations about
new methods, than those farmers who are predisposed in favor

of new technologies.

 

‘Interpersonal Trust
It was expected that a distrust of outsiders would be

a-sociated with a reduction in the sources of information

160
about new technologies that the individual is willing to
accept. If one can trust only relatives, it would seem
unlikely that the advice of government extension agencies
would be assigned much credibility. The individual might,
through his distrust of others, be excluding himself from
information that would help him to formulate favorable
expectations about new methods. The hypothesis that non-
trusting farmers would tend to concentrate upon their
relatives as a source of information about agricultural
innovations was not, however, supported by the data. Nor
was there any evidence of an association between trust and
other attitudinal variables, or trust and farm size, income
level, or educational achievement. In general, it was
concluded that a Recife farmer's willingness to trust others
had very little to contribute, at least as it was measured
in this study, to an understanding of the adoption process.

Attitude Toward Trying New Methods

 

Farmers who are willing to try new methods, without
first waiting to see what success others have had, would be
expected to use more modern techniques than their more
conservative counterparts. There are three possible reaons
for their willingness to experiment with locally untried
practices: (1) they may be very confident of their ability
to formulate accurate expectations; (2) more willing to
accept risks; or (3) better able to withstand the cost of
being wrong. The source of this venturesomeness will be

examined in the course of analyzing the variable.

161

A small percentage —- 26.4% .a of the Recife farmers
indicated that they would be willing to try a new method
without first observing it in use by others. One would
expect this small group, on the average, to be more innovative
than the remaining farmers. The evidence seems to be evenly
divided regarding this proposition. Over the four innovations
examined, these farmers who were willing to try things first
used barnyard manure and tractors significantly more
frequently, but did not seem to use chemical fertilizers and
insecticides much more often than their more conservative
counterparts. Table 21 presents this data in detail.
Table 21.--Use of Selected Innovations By Recife Farmers

Differing in Their Attitudes Toward Trying New
Practices.

 

c-v—w "...—.— .. —-—-.-——.-fi

3”-

Hhen new agricultural
products are offered
it is better to wait
and see what happens Barnyard Chemical Insec- a
when they are used by Manurca Fertilizer tiCidCS Tractors
others."

ent of Farmers Using

Pcrc
Selected Innovations

 

 

Agree 15.9 2.6 42.0 15.6
Disagree 23.3 ., ,3'1 g 45.4 .. 23.9

 

s was significant at the .05 level.

H.

a .
Chi—square analys

In general, these results seem to suggest that farmers

P40

who are w lling to try a new method without the guidance of
others' experiences may have been somewhat more innovative
than the remaining Recife farmers.

The second part of the analysis is to attempt to determine

what factors influence the farmer's willingness/unwillingness

162
to use locally untried practices. The best explanation
provided by the Recife data is that venturesomeness is
related to the farmer's ability to withstand losses. The
percentage of farmers who preferred to wait and observe the
experiences of others is greatest among the smaller and
poorer farmers. Table 22 describes the relationship between
farm size and the Recife farmer's strategy for trying new
methods.

Table 22.-~The Recife Farmer's Attitude Toward Trying New
Practices by Size of Farm.8

 

 

”When new agricultural proddEts are offered it

 

 

Size of is better to wait and see what happens when

Farm they are used by others.” ‘ '

(acres) Agree Disagree
---------------- percent---------------------~-

0 - 43 76.5 23.5

50 - 244 74.8 25.2

245 - 1234 63.0 37.0

1235 and above 50.0 50.0

 

a . . . . . . .
Chi-square analy51s 15 Significant at the .05 level.

The largest farmers were twice as likely as the smallest
farmers to assert their willingness to use (locally) untried
practices. They apparently recognize their better position
for risk—taking and are willing to act upon it. The small
farmers also seem to be cognizant of their situation and
typically prefer to observe some local experience with new
practices before they try them (to reduce their level of

uncertainty).

163
The association between the farmer's CXpressed willingness
to use untried practices and his level of income is much the
same as for farm size. Table 23 presents these relationships.

Table 23.-—Tho Recife Farmer's Attitude Toward Trying New
Practices By Level of Income.3

 

 

 

 

Per Capita ”When new agricultural products are offered
Family Income it is better to wait and see what happens
(U.S. Dollars) when thpy are used by others."
Agree Disagree

--------------- percent------—---------~-----
Less than $50 78.6 21.4
$50 - $99 70.0 30.0
$100 - $199 77.1 22.9
$200 - $499 63.3 36.6
$300 and above 64.0 36.0

 

a“. .... . --«,.
chi-square analySis Significant at tne .03 level.

Here it was found that as the farmer's level of living
decreases he has a greater tendency to be cautious in using
new practices. One might think of this finding in terms of
the theoretical framework where as the farmer's income level
approaches some "subsistence level of living" it becomes
more difficult to persuade him to accept the risks inherent
in untried practices.

Although capacity to absorb risks seems to be a fairly
adequate explanation of the Recife farmer's choice of strategies
in trying new practices, conclusive evidence does not exist
which would exclude such other possibilities as: (1) being

willing to accept risks (a gambler preference concept as

164
Opposed to capacity to absorb risks); or (2) degree of
confidence in one's ability to formulate accurate eXpecta-’
tions. The evidence that is available, however, seems to
cast doubt upon these alternative explanations. Higher
educational levels and a business-orientation might be thought
of as related in some way to an individual's degree of
confidence in his ability to formulate accurate expectations,
without locally generated empirical information. Yet neither
of these variables was significantly related to the Recife
farmer's strategy for trying new practices. A variable
exploring the individual's willingness to accept risks in
order to obtain certain gains in income (to be examined
somewhat later in this chapter) was also found to be unrelated
to the farmer's approach to trying new methods.

In summary, the individual's willingness/unwillingness
to try new practices without first observing the outcomes
experienced by others seems to be largely based upon his
capability to absorb the cost of making mistakes. Those who
indicated a willingness to use locally untried practices may
also be somewhat more innovative than the remaining farmers.
§EEBEIX

Reviewing the attitudinal variables examined in this
section, we find evidence in support of the conceptualization
of the decision-making process as suggested in Chapter 111.
Even though it was the unusual case where the attitudinal
dichotomies involving fatalism/nonmfatalism, luck/knowledge,

and positive/negative attitude toward new technology were

165
found to be significantly related to the use of certain
innovations; each of these variables seems to influence --
in a predictable manner -— the sources of information used
by the decision-maker, his use of certain communication
channels, and his knowledge of fertilizer or other ways to
improve yields. In addition it was found that the farmer's
attitude toward trying new practices seemed to be related to
his adoption of some new practices. These four variables
seem to enter into the decision-making process and to influence
this process in a fairly predictable manner. On the other
hand, it was impossible to determine if the Recife farmer's
time preference for income level had any influence upon his
decisions. And finally, interpersonal trust did not exhibit
the hypothesized influence upon his choice of information

SOUTCCS .

Educational Achievement

...... o“—

 

Educational level is usually assumed to be related to
individual innovativeness. Yet, from the standpoint of
this study, formal education is only a proxy measure of
certain skills which are useful in the formulation of

expectations about new technologies. Education involves a

certain mastery over symbols -- literacy and mathematical.
skills -- and a development of conceptual structures useful

in evaluating particular situations. As an example of the
impact of this second factor upon the adoption of new
technologies, consider the case of fertilizer. The farmer

who understands the relationship between soil type and

166
nutrient holding capacity is probably in better position to
form expectations about the effect of fertilizer on his
particular fields than the farmer who lacks this conceptual
framework. Presumably he will be better able to interpret
the recommendations of the agricultural specialist.

It is necessary, however, to be considerably less
specific in our analysis of the decision-making system
involving Recife farmers. Here we find that higher education-
al levels seem to be associated with knowledge about chemical
fertilizer, the difficulty perceived in using fertilizers, and
a knowledge of how to improve yields in general. Table 24-
describes these relationships.

Table 24. Tie Percentage of Recife Farmers Knowing About
Fertilizer, believing That It b’as Difficult to

Use, and l’nowing How to Improve Yields s- By
Education level. a

 

Number of Know What Believe That Know How To

Years of Fertilizer Is Fertilizer ls , Improve Yields

School Attended . Difficult To UseW
------------------ percent--~--~--~-----------

None 50.4 45.3 54.0

1 year 46.0 38.8 58.4

2 - 3 years 63.8 28.4 65.9

4 - 5 years 76.7 28.3 80.0

6 - 12 years 83.3 23.3 66.7

More than 12 years 100.0 .. - ‘14.3 g ,..L 7100°0...

 

aChi-square tables involving these variables were
significant at .05 or less.

b . .
Involves only those farmers who knew about fertilizer.

167

Despite certain interruptions in the trends, the data
suggests that the better educated farmers are more knowledge'
able about chemical fertilizer and other methods of improving
yields.

It would be an illusion, however, to attribute the
better educated farmer's greater likelihood of knowing
about fertilizer and other methods of imporiving yields
entirely to his superior conceptual abilities, if for no
other reason than the association between education and the
use of more direct sources of information. In Table 25 we
find that the better educated farmers were more inclined to
have read a farm magazine or attended an agronomist's meeting
than their less educated counterparts.
Table 25.--Percentage of Recife Farmers Who Had Read a Farm

Magazine or Attended a Meeting Where an Agronomist

Talked About New Production Methods -— Py
Educational Levels.d - - --

 

Number of Years

 

of School Read a Farm Attended an
Attended Magazineb Agronomist's Meeting
------------- ercent-----—---------«-—--
None 24.9 8.8
1 year 26.6 8.0
2 - 3 years 39.5 17.4
4 - 5 years 49.2 30.0
6 - 12 years 82.9 34.0
nMore_than 12 years 83.3 _. _ _.'. 71.4_~

 

#1
v—

aChissquare analysis was significant at .005 level.

b . ,
The survey question allows Ior someone to have read
"a" magazine to the farmer.

168

It should be noted that while Recife farmers with less
than a fourth grade education represent 83.7% of the sample,
they account for only 62.0% of those who had attended
meetings where an agronomist talked about new production
methods. In general, the better educated farmers were
either better prepared or more highly motivated to u.s e
these sources of agricultural information.

This same characteristic, involving better educated
farmers using more direct information sources, is exhibited
in the responses given by Recife farmers regarding the origin
of their knowledge about the latest practice that they had

employed. Table 26 summarizes these relationships.

 

 

 

 

Table 26.-~The Source of Information Used by Recife MT .ers
to Learn About Their Latest Practice Accordir ng
to Educational levels.a
Peicent of Recife farmers
Number of Using These Information Sources
Years of TEHBtives - _ Agfafiamigf
School and People From Local and

Attended Neighbors Another Place Businessmen Mass Media

 

None 76.0 9.8 6.7 7.5
1 year 74.2 9.3 6.2 10.3
2 - 3 years 72.3 6.3 6.3 15.2
4 - 5 years 68.6 0.0 7.4 24.1

6 years‘or
moreU . 62.8 . 4 . 0.0 . 2.9.. . . 34.3.

 

aChivsquare analysis was significant at .05 level.

b lore than 12 years category was omitted because it
include d only 4 observations. Total table was compiled on
the basis of 83. 3% of the sample either because of lack of
responses to the question of information sources or failure
to have adopted any ntw practices.

169

The table indicates that as the level of education
increases the degree of dependency upon indirect sources ~-
relatives, neighbors, and peeple from another place «- tends
to decrease; and farmers begin to look more frequently to
such direct sources as the agronomist and the mass media.
'Again we might speculate that the intellectual skills of
better educated farmers enable them to make greater use of
direct information sources, than less educated farmers, in
formulating their expectations about new technologies.

The relationship between education level and radio
listenership is so unique as to deserve special attention.
As the educational level of farmers increased they were more
likely to listen to the radio, but tended to devote less
time to this practice. Table 27 illustrates the relationship
between educational achievement and radio listening habits.
Table 27.--The Relationship Between Frequency of Listening

to Radio and the Educational Level of Recife
Farmers.a

 

 

 

Number of Years Listen to Radio
of School More than -
Attended Never Seldom l hr./day l hr./day

--------------- percent--------------------

None 12.7 36.3 28.8 22.1

1 year 8.1 30.6 26.1 33.1

2 - 3 years 7.4 33.1 23.5 36.0

4 - 5 years 0.0 16.9 40.7 42.4

6 years or more 0.0 g '30.0 .._ "42.5 . 27.5_

 

\

aChissquare analysis significant at .005 level.

17

0

Note that all Recife farmers with more than a fourth

grade education listen to radio, but while the preportion who

listen more than one hour per day increases through the fifth

grade it declines considerably for higher levels of education.

Education levels, in turn, are associated with size of

farm and level of income.

Since the flow of causality between

these variables is probably in both directions, the relation-

ship; will simply be reported.
relationship between educational level
that as the size of farm increases the

having higher levels of education also

Table 28.--Educational Level of Recife Fa

Table 28 indicates the
and farm size. Note
percentage of farmers

increases.

rmers By Size of

 

 

 

 

 

Farm.a

Size of Number of Years of School Attended
Farm None 1 2-3 4-5 6-12 More than
(acresl year years years years 12 years

---------------------- percent-------------------
0 - 49 51.9 19.6 20.2 5.0 3.4 ---
50 - 244 41.6 16.8 22.9 13.1 5.1 0.5
245 - 1234 23.3 16.4 24.6 13.1 13.7 6.8
1235 5.6 11.1 27.8 27.8 22.2 5.6

 

 

a . . . . . .
Chi-square analySis 15 Significant

at the .005 level.

The situation is much the same with respect to levels
A

of incomes (Table 29).

171

Table 29,--Educational Level of Recife Farmers By Level of
Income.

 

Per Capita Number of Years of SEhOOI Attended """
Family Income 2-3 4-5 6-12 More than
(U.S. dollars) None 1 year years .years years .12 years

 

 

 

Less than $50 57.9 17.9 19.7 3.6 0.9 ---
$50 - $99 37.9 18.6 31.0 6.2 4.7 1.6
$100 - $199 34.0 23.7 19.6 12.4 10.3 -»-
$200 - $499 30.5 18.1 22.2 20.8 6.9 1.4
$500 and above 16.0 6.0 16.0 30.0 24.0 A 8.0

 

aChi-square relationship was significant at .005 level.

Here again we find the level of education, as would be expected,
generally increasing with increasing levels of family income.
Neither of these relationships are particularly surprising,

but they do contribute somewhat to our understanding of

the data being studied.

In conclusion, there appears significant relationships
between the Recife farmer's educational achievement and his
knowledge of fertilizer and other methods of improving yields.
In addition, a higher level of education seems to be associated
with somewhat less dependency upon indirect sources of informa-
tion (particularly relatives and neighbors) about agricultural

innovations.

Communicated Information

 

In past sections we have examined the association of
certain variables with the information sources used by Recife

farmers. This has involved the assumption that "direct"

172
sources -- sources closest to the origin of the innovation -«
were related to a more rapid adoption of new agricultural
practices. The intention of this section is to examine that
assumption.

It will be recalled that one of the variables used to
identify the farmer's choice of information sources is based
upon a question probing where he had learned about the newest
method he was employing. If we are prepared to argue that
those sources used with respect to the most recent method
typify the information sources generally used by Recife
farmers, we can proceed with an analysis of the association
between information sources and certain innovations. Given
this assumption, we find the relationships described in
Table 30.

Table 30.--Perccnt of Recife Farmers Using Selected Innovations
By Source of Information.

 

 

Source of Percent Using Selected Innovations
information ~—~————.--

 

r

 

about latest Insecticide Tractor Chemical Non-chemical

method used Fertilizer. Fertilizer

Relative 44.0 18.0 1.0 17.0

Neighbor 42.3 17.5 2.8 18.2

People from

another place 31.6 15.8 0.0 18.4

Local business-

men 66.7 3.0 0.0 21.2

Agronomist 76.3 35.6 11.9 20.3
'Mass_Mediav ‘ 70.0 .... 30.0.} ’0.0

 

aChi-square analysis was significant at .001 level for
insecticides, tractors, chemical fertilizer and not
significant at .05 level for non—chemical fertilizer. Table
is based upon 83.6% of the sample who responded to the question.

173
The association between the Recife farmer's source of
information (about his last practice) and selected innovations
seems to support the notion that the farmers who make use of
"direct” information sources are more innovative than others.

Recife farmers who reported the agronomist as a source of

H.

nformation were more likely to use insecticides, tractors,
and chemical fertilizer than farmers reporting other sources.
(While the agronomist was reported by only 11.8% of the
respondents given in Table 30, these farmers accounted for
43.8% of the users of chemical fertilizer, 18.0% of
insecticide usage, and 21.2% of those who used tractors.)
Those who used the mass media were more likely than farmers
using other sources (except the agronomist) to have adopted
tractors and insecticides, but not chemical fertilizer.
Again, those farmers who learned of their latest practice
from a local businessman tended to use insecticides more
often than those who gathered their information from
relatives, neighbors, or people from another place. The

act that insecticides are sold by a large number of firms

in the rural Recife area, whereas tractors and fertilizer

are not, may help to account for this relationship.151
Finally, the use of non-chemical fertilizer (manure), however,
was not found to be significantly associated with the

farmer's source of information.
A similar case may be made from the relationship between

the farmer's use of certain forms of communication and his

 

151

 

Slater, Riley, et al.,'fiarket Processes, p. 3/27.

174
adoption of selected innovations. Table 31 describes these
associations.
Table 31.-ePercent of Recife Farmers Using Selected Innova~

tions by Use/Non-Use of Certain Information
Sources. '

...................

 

Use of Selected Percent U51ng Selected Innovations

 

Sources of Chemical Non-Chemical
Information Insecticide Tractor Fertilizer, Fertilizer

 

Had read a farm

magazine 68.6 37.3 7.8 31.4
Had not read a farm

magazine 37.4 13.7 1.7 15.0
Own a functioning

radio 49.5 25.1 4.1 22.9
Do not own a func-

tioning radio 32.6 6.6 0.8 10.9

Attended an agrono—

mist's meeting 63.0 35.8 ' 6.5 28.3
Not attended an

agronomist's

meeting 38.9 14.4 2.1 15.9

 

 

aA chi-square analysis cf each of these relationships
was significant at the .005 level.

The Recife farmers who had read farm magazines were
more likely to have used fertilizers, insecticides, and
tractors than those who had not. Those who had attended
meetings where an agronomist had talked about new farming
methods seemed to be more innovative than their counterparts
who had never attended such meetings. A difference in
innovativeness also seems to occur between farmers who had
a functioning radio in their home and those who did not.

In addition, it should be observed that a significant (.05)

association exists between eXposure to these channels and

175
the Recife farmer's latest farming practice. The percentage
of respondents who indicated that they had learned their
latest practice from an agronomist or from the mass media is
greater among those farmers who had read farm magazines,
attended an agronomist's meeting, or who had a radio in
their home.

In all, the evidence seems to support the assumption that
the farmers who make use of "direct” sources of information
tend to be more innovative than those who depend upon local
"indirect" sources of information. The use of "direct/
indirect" as opposed to "cosmopolite/localite” as a division
of information sources seems to be a conceptual improvement
since "people from a different place” are "cosmopolite" by
definition (see page 40), but such sources of information
(according to Table 30) do not appear to be any improvement
over one's neighbors and relatives.

Before concluding this section, it should be mentioned
that income and farm size, as well as certain attitudes and.
the farmer's education level, are related to sources of
information. The use of the agronomist and the mass media
as a source of information about new methods increases with
larger farm sizes and higher levels of income.

"Utility Functiczl

 

Until now, the emphasis has been upon the effect of
certain variables as they contribute to the farmer‘s knOW*
ledge of innovations, which.in turn shape his expectations of

these innovations. The theoretical framework of Chapter III

176

postulates an overall relationship between the farmer's
expectations, his utility function for income, and his
ultimate decision regarding particular innovations. Just
as we were unable to directly examine the farmer's expecta-
tions in past sections, we are unable to duplicate his utility
function in this section. Nonetheless, as in past sections,
there is certain data from the Recife survey which may provide
some insight into the interactive process of farmer decision-
making. Thus we turn our attention to a discussion of certain
prepositions related to the effect of the utility function
upon the adoption of new technologies.

As a beginning, consider the relationship between income
levels and the adoption of new practices. In past sections
we have found income levels to be associated with certain
attitudes -— luck/knowledge and attitude toward trying 16W
technologies --, educational levels, and the sources of
information used by farmers. All of these might interact in
the decision process causing innovativeness to be associated
with level of income. Adding to this the fact that new methods
often involve additional capital outlays, the financial
capabilities of higher income levels provides still another
reason for income to be related with innovativeness.
Finally, there is the hypothesized relationship between
nearness to a "subsistence level of income" and the farmer's
willingness to accept the risks inherent in new technologies.
Since each of these factors may be involved in a relationship

between income levels and innovativeness, as given in

177
Table 32, it is necessary to look for additional evidence
of the latter relationship. One possible source of such
evidence lies within the responses of Recife farmers to
the following question:
"Let's pretend you have tried a new method or
technique on your crop and the results were bad
causing you to lose half of your production. What

would happen to the crop and to the family?”

Table 3-.m’The Percentage of Recife Farmers Using Selected
Innovations By lncone_Lev;].“

 

 

 

Level of Percent Using Selected Innovations
Per Capita '
Family Chemical hon-Chemical

Income (U.S. Insecticides Tractors Fertilizer Fertilizer
dollars) ' .

 

Less than $50 29.4 7.3 0.0 13.2
$50 - $99 42.2 15.5 3.0 15.5
$100 - $199 57.8 20.6 2.1 19.6
$200 - $499 50.0 33.3 5.9 23.6
$500 and above .76.0 . . 60.0. 14.0 34.0

 

 

a C O I O I ‘-
Chi'square analySis was Significant at the .005 level
for each of these relationships.

Table 33 indicates the farmers' resnonses to this
1

question separated on basis of family income levels.

178

Table 33.--Recife Farmers' Indication of the Consequences of
the Failurg of an Innovation, According to Levels
of Income ........ . .............

 

 

Level of Consequences of the Failure

Per Capita . . of an Innovation..

Family

Income - Would Would Family Would get Would Would

1966 (U.S. have to have to would a parts try never

dollars) borrow sell suffer time job or again try
money assets become a again

rural worker

 

---------------------- percents---------—----««~-««
Less than $50 3.7 4.1 74.7 3.3 10.8 3.3
$50 - $99 4.7 3.9 71.3 1.0 14.7 4.7
$100 - $199 2.0 4.2 64.5 2.1 21.9 5 2
$200 - $499 4.2 12.5 51.4 1.4 25.0 5.6
$500 and
above 4.2 4.0 46.0 6.0 32.0 > 8.0

 

aChi-square table was significant at .005 level

Approximately three-quarters of the farmers in the
lowest income bracket -- less than $50 per family member --
believed their families would suffer from the failure of
an innovation. By comparison, less than half of those
farmers with $500 or more per capita family income believed
their families would bear the cost of such a catastrophe.
Again, the proportion of farmers in the highest income group
who would react to such a situation by "trying again“ was
three times greater than for the poorest farmers. Although
these empirical relationships do not provide a conclusive
argument for the hypothesized relationship between nearness

to a "subsistence level of income" and reluctance to accept

179
the risks of new technologies, they do seem to support his
proposal in a general way. If we add to this evidence the
association, as discovered earlier, between the willingness/
unwillingness to try locally untested practices and the
farmer's income level, the case favoring this particular
hypothesis is once again strengthened.

In turning to consideration of the Recife farmers'
utility for increased income, the MSU/SUDENE survey provides
very limited possibilities for analysis. From the stand-
point of the logic involved, the best of these possibilities
is embodied in the following question.

Now let's play a little game. Look at this

house. (Respondent is shown a drawing of a house

with three rooms.) As you can see, there are

three rooms and in each room there are four boxes.

In the first room, three of these boxes have NCr

$5.00 each; in the second room two boxes have

NCr $20.00 and in the third room only one box

has NCr $50.00. You don't know which boxes have

the money and to play the game you have to pay

NCr $.50. Then you may get into one room and

open only one box and keep the money you find. Do

you understand the game? (If not, the instructions

were repeated.) Would you pay NCr $.50 to join

the game? Which room would you prefer?

There is no pretense made that the data from this
question is, in fact, a measure of differences in utility for
increased income between the Recife farmers. There was
considerable concern among the MSU researchers (MSU/SUDENE
Study) that this question was too difficult for most of the
farmers to understand. It was included in this study upon

the chance that it might provide some crude grouping of

farmers on the basis of their desire for additional income.

180

If one analyzes the possible responses to this question

(as diagrammed in FigurelO), the following groupings could
be developed:

a. Very low utilitv--would not take any risk at more-
than7f5i7_odds_7would not play).

b. Low utility-—would require higher returns to accept
risks equivalent to rooms II and III.

c. figdipm_utilityf—would require higher returns to
accept risk; equivalent to room III. (Room II)

d. High utility--finds high returns sufficiently
attractive to accept the risk. (Room III)

 

 

 

 

 

’ Room III
75% -- ’3?”
I
Chance of ’1’
Losing ’,”,,.—--"” Room II
50% - :f’
,’ x’ W '— A Room I
/ '/
25% - ,’
I I
6’”
$1.... 1 1 ,
E 20 50 Gain NCr$

Figure]()--Iso-Utility Curves for Choices in the Risk Game.

The Recife farmers' risk cnoices were as follows: 42.7%

preferred not to play the game (very low utility); 15.5%
chose room I (low utility); 8.0% chose room II (medium
utility); and 33.7% chose room III. These choices seemed
to be independent of farm size, education level, and with

one excepti n, income. Farmers in the highest income

181
grouping( ’55 00 per capita family income) we1e .significantly
more willing than their colleagues to accept a small risk
for chance of a small return (36.0% of this group chose
room I as opposed to 15.5% for the sample as a whole), but
their acceptance of the other alternatives was a1prox \imately'
in the same proportion as the total sample. Contrary to
CXpectations, the Recife farmer's risk choice was not found
to be significantly related to his use of insecticides,
fertilizers, and tractors. This causes us to question the
variable (HSU researchers were doubtful that the risk game
was understood by many of the Recife farmers) and withheld
judgement on the importance of the farms r' 5 utility for
increased income as an interacting force in iiis dec ision-

making process.

Summarw'
... __J.

 

chapter 111 (p. 97) presents a schematic representation
of the adoption process as conceptualized in this study.
This presentation of tne adeition of ncw agricultural
practices will be enlarged upon in sum arising the findings
of this chapte1. Figure 13 should be of assistance in our
attempts to visualize the decisicn making system in terms of
the data that has been examined. Note that the implementa-
tion phase of the schema has been omitted since this is the
subject of the following chapter.

Although c mmu1.ity attitudes -- as perceived by the
individual farmer -- would have been a poor prediction of

the use of new agricultural practices, they did seem tc

182

 

 

“N’DUIUF'fiOZ
FNZF<FH>N
h
0

(DOD-PHD?-

HEOUHDO

 

    

 

 

£92 omauow 23 5 use“: we sounxm mad—$75333 23 we to; 9322.8 2: 9:3:

nomhmz
hzmmmuu
.m> zuz
-rhunukz.

 

.
I
..z 7.

llllllllllllJ
muzunuummm
Mm~m

 

mznuzh
auwauwux.
mom >h»;~ka

 

 

 

 

     

 

mmu~k0<ma
ouuxHZD >gg<uoa
rxk Oh mmwxuzuaqwz

 

 

 

 

       

nozhmz Zur
do» mzomk
-<kuunfi

      

 

. xm_4<h<m

 

 

 

 

 

 

 

   

 

 

 

 

 

.\ 3.3....
_ Emu 0:3
, fight”:
.. x M 28:. 52 N141].
WflLODOM A. , . .r\.—. / .
ZOnF<Z:OWZF ( D wmbkwbh
Ali“ olll:
.\ I: ../l

a g 35:55
a >83

.11-_ .55..
K 295 can:

«WV

 

 

 

 

 

\A .52:
. 58“.:
' I;"

.32...

 

 

 

”Nam 24¢“

 

183

have an influence on the formation of the farmer's own
attitude toward new technology. Specifically, it was found
that Recife farmers' attitudes toward new technology were
statistically associated with the attitudes they perceived
their community to hold toward innovators and individuals
who progressed economically at a more rapid pace than their
neighbors. These same community attitudes apparently have
some influence upon the individual's desire to learn about
new technologies, but this influence was rather weak relative
to the farmers' own attitudes and level of education. Note
that the position of community attitudes (to the far left)
in the schema reflects these relationships.

The Recife farmer's own beliefs and attitudes seem,
as the theory had suggested, to become involved in his exyecta-
tions formation about new agricultural methods. Fatalistic
farmers, farmers with negative attitudes toward new
technology, and farmers who believed that luck was the
controlling factor in business success are less inclined to
seek out information and depend upon less ”direct" sources
of information than the farmers holding the Opposing views.
In essence, individual attitudes become part of the informa-
tion used by Recife farmers in formulating their expectations.
Willingness to defer present income is not included in
Figurell because its effect, if any, upon the decision process
was not obvious from the available data. If this item were
to be included in the schema, it would be placed adjacent to

the other attitudes with an arrow connecting it with the

184
circle involving utility of the innovation, where the influence
of time preferences would be expected to affect the decision
outcome. Attitude toward trying locally untested practices
seems to be intertwined with the relationship between low
income levels and the reluctance to accept the risks inherent
in innovations. For this reason, it is located between income
and the evaluation of relative utilities in the schema.
Finally, it will be recalled that the farmer's business-
orientation (luck/knowledge variable) was associated with
past successes and is indicated as such in the schema.

In addition to the influence of farmer attitudes,
education achievement seemed to be associated with his choice
of information sources. Educational achievement, in turn,
was significantly related to farm size and income level,
which were also associated with the sources of information

1.1

U)
U)

6 by the farmer. Higher levels of education, larger farm
size, and higher incomes tended to be associated with the
more direct sources of information.

In turn, information sources were significantly associated
with the use of certain innovations, which leads us to conclude
that farmers who have the capacity and willingness to utilize
the more direct sources of information have greater knowledge
of innovations (less uncertainty about them) and formulate
favorable CXpectations more quickly than those who obtain
information from “indirect" sources such as relatives and
neighbors.

The Recife data provided some evidence that better

185
educated farmers were more knowledgeable about new agri-
cultural practices. This may be in SUpport of the theoretical
contention that the conceptual structures possessed by the
farmer influence his ability to interpret information about
new technologies and, consequently, formulate expectations.

Among other reasons for the significant association
found between income and innovativeness, evidence was uncovered
which seems to support the hypothesized relationship between
low incomes (nearness to a ”subsistence income level") and
the farmer's unwillingness to accept the risks of new methods.
On the other hand, the Recife data did not provide supporting
evidence for a relationship between the farmer's utility for
increased income and his adoption of innova ions. Here we
must leave a question mark in Figure 11.

Admittedly, net all reasonable variables have been
examined, non has the evidence been sufficient to conclusively
prove the thtoretically derived relationships. Yet, the
overall logic of an interactive decision-making process --
such as conceptualired in Chapter III -- does appear to provide
a framework which is useful in understanding the Recife

farmers' adOption of innovations.

CHAPTER VI

ANALYSIS OF SELECTED VARIABLES WHICH
MIGHT SERVE AS CONSTRAINTS
UPON INNOVATIVE BEHAVIOR
Even though farmers might hold favorable eXpectations

about certain new methods, they may find themselves unable to
adopt these practices because other factors constrain their
behavior. Three factors which might be thought of as having
an impact upon the farmer's ability to implement decisions --
credit availability, input availability, and tenure
arrangements -- are considered in this chapter., The extent
of discussion regarding these variables varies with the avai1~
ability of data provided in the MSU/SUDENE survey of farmers

in the Recife area.

Credit

 

The availability of credit is usually considered as
essential to the "transformation of agriculture." Improved
methods often involve the purchase of modern inputs, which
may be difficult for the farmer to finance from his present
earnings. In turn, limitations may be placed upon the
farmer's acquisition of external financing due to inadequacies
in the credit system or to self-imposed barriers on the amount
he is willing to borrow. Several questions about credit asked
during the Recive survey afford an opportunity to give rather
detailed consideration to credit usage in the Northeast of

Brazil.

186

3.5

.‘.;;e

(I?

 

187
Amount of Credit Used by Recife Area Farmers
An overview of the credit used by the farmers examined
in the Recife survey is provided in Table 34.

Table 34.--Percent of Farmers Using Credit and Amount of
Credit Used.

 

Percent of Mean Median Length of
Commodity Farmers with Amount Amount Loans in
Area . Loans Borrowed. Borrowed Months
(U.S. (U.S.
dollars) dollars)
Rice
550 Francisco 63.8 944.13 237.50 3 ~ 5
Bean
IrecG 53.1 752.27 511.36 3 - 10
Al-Pe 47.7 386.57 227.20 3 - 5
Manioc 36.7 349.77 114.55 1 - 4
Cotton 68.5 424.70 181.82 4 - 6

..-

Several characteristics of farm credit in the Recife
area are rather clearly portrayed by this table. First, there
is a great deal of variation in borrowing from area to area.
The percentage of farmers having loans varied from 36.7% in
the manioc area to 68.5% in the cotton area. The mean size
of loans was equally as varied with an average loan to
See Francisco rice farmers of nearly three times the mean
amount borrowed by manioc farmers. Similar variations exist
between the median amount of borrowings. The duration of
loans was very short, in all cases less than one year
(although there is the untested possibility that some farmers

may have been able to refinance their indebtedness).

188
Differences between the median and mean loans clearly
indicate the very skewed distribution of loan sizes among
those borrowing. Table 35 gives a more complete description
of the distribution of loan sizes.
Table 35.---The Percentage of Recife Farmers Who Borrowed

Amounts of Less than 3 Ci 1ven Anount and the Range
of Amounts Borrcved by Commodit.y Area

 

 

Range of Percent Percent Percent

Commodity Loans of Loans of Loans of Loans
(dollars) Under Under Under
$100 $500 . $2500

850 Francisco

Rice $ 7 - 18,182 24.4 65.4 93.6
Al-Pe Bean 23 - 2,273 24.2 75.8 100.0
l.ece Bean 18 - 4,545 18.5 48.1 96.3
Maniac 4 - 2,273 50.0 82.8 100.0
Cotton 22 - 7,727 33.0 78.6 .99.1

 

Except for Irece bean producers, approximately one quarter
or more of those farmers who be: rowed used less than $100 of
credit, from one-half to four-fifths borrowed less than $500,
and very few farmers had utilized more than $2500 in credit.
Together, Tables 34 and 35 indicate that manioc farmers tended
to receive the least amount of credit, followed by cotton
farmers (although they borrowed most frequently), and then
by Algoas- Pe rnubuco bean and see Francisco rice farmers. The
Irece bean farmers seemed to be able to acquire relatively
larger amounts of credit.

It is especially interesting to note that the amounts

borrowed per acre of crOps planted presents a very different

189
view of credit in the Recife area.
In Table 36 we find the manioc area «- the area with
the smallest average amount (mean or median) borrowed per
farm -- using the largest amount of credit on a per acre basis.

Table 36.s‘The Amount of Credit Used Per Acre Planted by
Commodity Area.

 

 

Loan Size Per Acre Planted

 

Commodity
Mean . ..Median

 

---------- U.S. dollars---------~----

Sao Francisco Rice 21.60 13.80
Al-Pe Bean 9.04 8.28
Irecé Bean 14.33 .9.63
Ianioc 26.55 13.47
Cotton 11.20 '_ , , ._7.36

 

 

Again the lrecé bean area, which had the largest median loan,
used considerably less credit per acre than either the manioc
area or the 550 Francisco rice area -- both of which had

much smaller median loans per farm. One possible explanation
of these findings is that the variation in credit per acre
simply reflects the difference in credit needs for the
production of various creps. Additional analysis indicates
that this is not an adequate explanation. First, the
duration and timing of loans suggests that credit used by

the Recife farmers may be misunderstood if it is viewed
entirely as production credit. According to the MSU/SUDENE
research it takes roughly 120 days from seeding to the

harvesting of rice and 18 to 24 months for maniec to mature,

t, y‘- .
win 3)

 

 

190

yet the length of loans was typically 3 to 5 months in the
550 Francisco rice area and l to 4 months in the manioc
area. The length of loans in the manioc area is inadequate
to cover a production cycle and loans in the rice area are
barely adequate. In addition, rice in the 850 Francisco area
is planted in seedbeds in either January or February and
transplanted to fields approximately 20 days later, yet

.9% of the loans were taken in March (as compared to 19.2%
in January and 12.7% in February), 20.5% in April, and 9.0%
in May. This suggests the possibility that loans were often
obtained to meet the family's living expenditures until the
next harvest.

It was also noted that the areas with the largest median
per acre loans were generally those with the smallest median
farm size. Table 37 contrasts median farm size with the
median loan sizes (on a per acre basis).

Table 37.-~Median Farm Size and Median Amounts Borrowed Per
Acre by Commodity Areas.a

 

Commodity Area Median Farm Size Median Loan Size

 

(Acres) (U. 8. dollars
a per Acre)
Manioc 17 13.47
sec Francisco Rice 37 13.80
A1 Fe Bean 74 8.28
Cotton 85 13.47
lrece Bean. 4 .. v. 89 ... ,4 . ... 9-63.

 

aIncludes median farm size for all farms and median size
of loan for onlv those who borrowed.

-fi

..e WC
1

"t'EI'

c

F» t...» «\u . . I. Ill l/H. aIL
Cu C To 2 1
ea TA , 0» e
O :4“ . D n l
p‘V “J. ‘L T. 1H,.
3 a: a...» rd 3 a:
n1.b 9.. .....15 1 a a

51

191
One would suspect from Table 37 that amounts borrowed per
acre were inversely associated with farm size. Table 38
verifies this hypothesis. Here we find that farms of less
than 50 acres borrowed an average of twice as much money per
acre as farms from 50 to 244 acres. Together, farms under
245 acres accounted for more than threevquarters of the loans
reported in the sample. It is not surprising, given this
data, that areas with the smallest median farm sizes had the
largest median per acre loans.

Table 38.*-Percentage of Loans Received and Mean Amount
Borrowed Per Acre by Size of Farm.

 

 

Size of Farm Percent of Loans Mean Loan
Represented 3y (U.S. dollars
Each Farm Sizea per Acre)

0-49 38.3 12.20

50 - 244 38.3 5.05

245 - 1234 18.8 3.23

1235 and above 4.6 1.49

All Farms 100.0 7.28

 

a - .
Based upon iarms that received one or more loans.

Further, this information adds weight to the contention that
a sizeable proportion of the Recife farmers borrowed in
order to survive until the next harvest.

Before concluding this section, it should be noted that
the likelihood of credit usage varies by size of farm. The
smallest and largest farmers were somewhat less likely to have

loans than the intermediate groups. Table 39 describes this

situati

Iable 3

C

 

 

 

 

192
situation.

Table 39. --Percentage of Recife Farmers Using Credit by Siz

 

 

Of Farm. ..........
Size of Farm Percentage of Farmers
(Acres) " ' . Using Credit IIIII
0 ~ 49 43.8
50 a 244 63.3
245 - 1234 60.0
1235 and above 52.8

 

In summary, it has been discovered that credit usage
varies corsiderably between and within commodity areas, with
each area represented by a skewed distribution of loan sizes.

A reasonably large percentage of the sample (generally 25% or
more) borrowed sums of less than $100. 00. Smaller farmers
borrow less frequently and tend to borrow larger amounts on

a per acre basis than their larger counterparts. The evidence
strongly suggests that small farmers (and perhaps some larger
farmers) oft en are borrowing to maintain their households until
the next harvest.

Sources of Credit

 

Not only does the percentage of farmers using credit
differ by farm size, but so do the sources from which they
obtain their loans. Table 40 describes the relative
importance of credit sources for farms of varying sizes.

Small farmers were much more dependent upon relatives and

neighbors; credit cooperatives; and landlords, local

p
\.
y'

9' . 14

1,:-

wh-‘V?
v. ig‘.
‘5

1‘q
"‘
5.15 aTL{

ab Ol‘fe

ifitp
\‘U') 8"
QLCH'C,

 

 

193

Table 40.--Percentage of Farmers Borrowing From Given Sources
By Size of Farm.

.........

 

Size of Relatives Credit Govern~ Landlord,
Farm and Coop~ ment Private Local Busi«
(Acres) Neighbors eratives Bank Bank nessmen,

. . a . ........ f.. .Buyer

 

First Loan

0 - 49 15.0 23.3 34.0 1.3 24.5
50 - 244 8.3 19.0 62.0 1.3 8.8
245-1234 6.4 11.5 74.4 3.8 2.6
1235 and

above 0.0 5.3 84.2 5.3 0.0

Second Loan

0 - 49 30.9 -« 14.3 2.4 45.2
50 - 244 20.0 -- 37.8 8.9 28.9
245—1234 19.1 -- 42.9 19.0 19.1
1235 and

above 0.0 -- 80.0 0.0 20.0

Third Loan

0 - 49 33.4 -- 0.0 0.0 61.1
50 - 244 21.4 -- 21.4 7.1 50.0
245-1234 33.3 -- 11.1 33.3 33.3
1235 and

above 0.0 -- 66.7 0.0 _‘ (33.3.

 

businessmen, and buyers than larger farmers. The larger
farmers, especially the largest farmers, seemed to use the
government bank more frequently (which suggests that larger
farmers had better access to this form of credit). Again,

small farmers depended very little upon private banks relative

to thei

Fix"
relativ
credit,

1
from tr

. \
Land.
10rd

.
:L'
.1“
”h‘nr
V \
g

up”;

icy .
V.ldn«_
‘ \-

 

194
to their larger colleagues.

Even though small farmers (51.4% of the sample) were
relatively less likely to use government banks as a source of
credit, they still accounted for 22.4% of the loans obtained
from that source. Mediummsized farms (SO-244 acres, which
represented 34.1% of the sample) accounted for 44.0% of the
government bank loans, followed by large farmss(245 to 1234
acres and 11.6% of the sample) which obtained 24.5% of these
loans. Finally, the relatively few very large farms (2.8% of
the sample) received 8.2% of the loans supplied by the govern-
ment banks.

Sources of credit also seem to be associated with the,
tenure status of the farm operator. Although the Recife
survey concentrated Upon landowners (owner-operators and
landlords), a sufficient number of tenants were included
to provide a comparison in terms of the sources of credit
used. Table 4] describes the differences in credit sources
Table 4l.--Percentage of Farmers borrowing From Given Sources

By Tenure.

Local

 

Tenure Relatives Credit Govern- Business-
and Coep- ment Private Land- man and
Neighbors erative Bank Bank owner Buyer
Land~
lord 10.9 18.6 59.0 4.2 —- 6.7
Owner-
Operator 12.3 20.0 50.1 1.5 5.0 10.2

..Tenanr. “.8-0 “24:0 -.-. .4.8~b_ ..... 2.0.-.0.._

 

V

utilized

with ten;

 

195

utilized by tenants, owner-operators, and landlords (owners
with tenants . There are notable differences between tenants
and landowners. Nearly half of the loans obtained by tenants
came from their landowners. Local businessmen were also a much
more important source of credit to tenants than landowners.
At the same time, the proportion of tenants securing loans
from the government banks was less than one—half the percentage
of owners who were able to obtain credit from this same source.
Finally, none of the tenants in the survey had obtained credit
from cooperatives or private banks. In general, one can
conclude that tenants found it more difficult than the property
owners to obtain credit from the commercial sources (govern-
ment banks, private banks, and credit cooperatives) and were
consequently more dependent upon landowners for their loans.

Since we have determined that the use of particular
sources of credit varies with the size of farm and farm tenure,
it would be useful to understand somewhat more about the
sources themselves. The following table indicates the
percentage of total loans attributed to each source and their
mean and median size of loans.

The largest share of the loans received by Recife farmers

C"

was supplied y the government banks. Credit from this source
should be considered as subsidized since the average rate of
inflation (roughly 3.0% per month) exceeded the interest rate
on these loans. Perhaps one of the reasons for the small

percentage of credit sepplied by the private banks (3.3%) is

due to competition of the subsidized government credit program.

Table 4

Source

 

196

Table 42.--The Percentage of Total Loans Accounted for by
Various Sources of Credit and Their Mean and
Median Size of Loans. ___

 

 

Source of Credit Percent of Total
Loans from this Mean Loan Median Loan

Source ...... Size. . . Size
Government Bank 51.7 $829 $414
Credit Cooperatives 16.8 266 182
Neighbors 7.2 138 45
Local Businessmen 5.7 329 73
Landowners 5.2 83 34
Buyers 4.1 335 273
Private Banks 3.3 1763 273
Relatives 2.9 ' 199 . 91

Credit cooperatives were the second most important supplier
of credit. Although the size of loans from the credit
cooperative were smaller than from the banks, they tended to
provide somewhat larger loans than were usually obtained from
other non-bank sources. Neighbors and relatives combined
provided 11.1% of the loans, generally smaller in size than the
credit cooperatives. Local businessmen and buyers accounted
for 10.0% of the loans, with the amount obtained from buyers
tending to be somewhat higher than those supplied from local
businessmen. The smallest loans, upon which tenants are
largely dependent, came from the landowners. It is worth
:noting that the 550 Francisco rice farmers seemed to be able
to obtain somewhat larger average loans from private banks

(mean loan equal to $3377), businessmen ($1035), and local

buyers I

 

 

197
buyers ($491) than other Recife area producers.

The evidence developed in this section suggests that
smaller farmers tend to depend heavily upon nonmbank sources
of credit which loan relatively small amounts of money,
except for the credit cooperatives which appear to serve
small farmers by providing somewhat larger loans. Buyers
and businessmen provided some relatively large loans,
although they were apparently concentrated in the Sfio
Francisco rice area. Landowners, whom tenants are particularly
dependent Upon, provided the very smallest of loans. The
larger loans were typically obtained from government and
private banks. These sources were used proportionately more
often by the larger farmers, although 22.4% of the government
bank loans were granted to small farmers (0 - 49 acres).

Reasons for Not Borrowing

 

The Recife survey asked farmers who had not borrowed
why they hadn't; and asked the farmers who did borrow why
they didn't borrow more. Their responses were quite

1nterestin-, especially as viewed in terms of farm size and

0'4

level of income. Table 43 describes the reSponses given by
Recife farmers as reasons for not borrowing.

The percentage of farmers who did not believe that they
needed credit increased with the level of family income.
Yet, in terms of the total sample of Recife farmers, 43.0%
of those who ”didn't need credit" were to be found among
farmers who earned less than $50 per capita family income.

The meaning of the response "didn't need credit” may well

:12-

 

198

Table 43.«-Reasons Given by Non-Borrowing Recife Farmers,
.For Not Using Credit by Levels of FamilyIncome.

w V v w

 

 

Family Per Capita Couldn‘t
Income Didn't Need Get a Afraid to Other
(U.S. Dollars) . ' Credit.. Loan C ...Borrew '. ..
Less than $50 25.6 38.2 34.0 2.1
$50 e $99 28.3 26.4 41.5 3.7
$100 - $199 51.4 14.2 28.5 5.7
$200 and above 73.3 30.3 15.1 6.1
.All Incomes , _ 32.5__ .. __3l.7 ,. . 32.5.. _-_3.4_

 

8This table represents crop farmers only and the
differences were significant at the .01 level by the
chi-square analysis.
have been different for farmers at different income levels.

To the relatively poorer farmers it may have meant that they
were able to survive from harvest to harvest without needing

a loan to meet family living expenses. For larger and
relatively higher income farmers the phrase may have reflected
an ability to finance the costs of production —- hired labor,
purchased inputs, and perhaps loans to tenants -- from their
own resources.

A large percentage -- 38.2% -- of non-borrowers in the
lowest income grouping (less than $50 per capita) indicated
that they "couldn't get a loan." This reason was relatively
less important among the intervening grOUps, but then increased
to 30.3% for farmers with family incomes of more than $200
per capita. Such a result could be rather perplexing without

giving consideration to some additional information. It

each ix

grcap i1

1.‘
‘- 1
‘5 a W:

5., -
‘C-ill 1‘"
J

'Tr
dildO‘nfac

 

 

 

 

199
should be kept clearly in mind that these percentages
represent the relative importance of the reasons given for
each income category. While only 28.5% of the largest income
grouping failed to use credit, 53.9% of the lowest income
group did not obtain a loan. Consequently, those not able
to acquire credit represent 20.6% and 8.6% for the low and
high income groupings respectively. Again, low income
farmers who were unable to obtain a lean represent 9.1% of
the entire sample, as compared to 1.6% for high income
farmers who found themselves in the same situation. Taken
as a whole, it seems that low income farmers had greater
difficulties in securing a loan than their somewhat wealthier
counterparts.

Fear of borrowing appears to be associated with lower
family incomes (as was the fear of accepting the risks of
innovations). Farmers with under $200 per capita family
income were much more likely to indicate that being "afraid
to borrow" was the main reason that had kept them from
obtaining a loan. Other reasons for not borrowing represented
a small percentage within each income grouping.

Much the same situation is discovered when the Recife
farmer's reasons for not borrowing are compared on the basis
of farm size. (In fact, the data for farms of less than
fifty acres is nearly identical to the responses given by
farmers with less than $50 per capita family income.) The
relative proportion of farmers who believed they "didn‘t

need" credit increases with the size of farm. At the same

tixe

3
7- AP.
611:;
'1
did 5'.

:C'TC' :

 

 

200

Table 44. --Reasons Given by Recife Farmers For Not Using
Credit By Size of Farm.a

.................

 

 

Size of Farm Didn't Need Couldn‘t Get Afraid to

(in acres)... Credit ..... a Loan. “ i Borrow... .Other
0 - 49 24.1 38.2 35.2 2.4
50 - 244 42.5 23.3 31.5 2.7
245 and above 55.6 , 14.8 " ..18‘S_ . 11.1

 

 

aChi-square analysis significant at the .005 level.

time, the larger farmers seemed to have relatively less
difficulty obtaining loans than their smaller counterparts,
and were generally less fearful of borrowing.

Turning to the Recife farmer's reasons for not using
more credit, we again find important differences on the
basis of income and farm size. Table 45 describes the
responses of borrowers divided by levels of income.

The percentage of farmers who indicated that they
"didn't need more credit” tends to increase with higher
levels of income. In total, 13.2% of the borrowers (7.1%
of the total sample) believed that they had obtained
sufficient credit to meet their needs.

Risk was an important consideration to those who
borrowed, as well as those who didn't. A large percentage —-
30.5% ~~ of Recife farmers who had obtained loans were
unwilling to borrow more because of the risks involved.

The proportion was, of course, larger among the lower income

farmers ~s 39.3% for those withe ings under $50 per caplta

201

Table 45. --Reasons Given by Recife Farmers For Not Using
More Credit by Income Levels.a

.................

 

 

Per Capita Family Bank Does Don‘t Need
Income Not Give Lack c _ d Moree

(U.S. dollars) Larger Assets RISK Credit Other
. . Loans . .. . . .. .. . . . ..... ‘

Less than $50 19.6 28.7 39.3 9.0 3.3

$50 - $99 27.1 15.7 34.3 ‘15.7 7.1

$100 - $199 27.8 21.3 31.1 13.1 6.6

$200 and above 46.2 13.7 13.7 17.5 6.2

All Incomes 29.7 21.0 30.6 13.2 5.4

 

“Chi-square analysis was significant at the .005 level.

’Responses ceded under: "Banks don't make larger loans
availghle to farmers." '

L» . . . . n - . 4.

lesponses coded under. I have no assets to gua1ancee
a la: 'c1 loan."

espenses coded under: "I would run the risk of be cing
unablg to pay bag} or even lose my p1operty. .
)lerpopsrs coded u.1der: ”I don t have anything on wh1ch

to use a la rgcr amount of money.

as cozyared to 13. 9% for farmers with per c apita family
incomes in excess of $200.

A 1ack of sufficient assets to secure a larger loan also
seems to be associated with lower income levels. Proportionally
twice as many of the lowest income farmers gave this reason
as compared to their counterparts within the highest income
grouping. In total, 21.0% of those farmers utilizing credit
found their assets inadequate to obtain larger loans

Finally, we find that higher income farmers complained
most frequently that “banks did not give larger loans."

This reason for not using additional credit is also

importantly related to the size of farm, as seen in Table 46.

202

Table 46.--Reasons Given By Recife Farmers For Not Using
More Credit By Size of Farm.a f . u

 

 

 

Size of Farm Banks Don't Don‘t Need
(in acres) Give Larger Lack More
Loans _ _ Assets Risk 1' Credit _ Other
0 - 49 20.1 31.3 30.6 11.8 6.3
50 « 244 36.1 13.6 35.3 11.3 3.9
245 and above. 41.4 12.1 g 20.7 19.0 6.9

 

 

aChi—square analysis was significant at the .05 level.

The percentage of borrowers on farms of 245 acres or more who
indicated the unwillingness of banks to give farmers larger
loans as primary constraint on the amount they borrowed was
roughly double the prOportion of small farmers (0-49 acres)
giving this same response. This fact, in conjunction with
the observation that large farmers use government banks
relatively more frequently than smaller farms, suggests the.
following as a possible interpretation of their situation.
(See Figure 12 on page 203.)

Note that the supply curve for credit is drawn completely
elastic to the extent of the government bank's resources.

ta

Q)

Beyond this point the supply curve is upward sloping
positive real rate of interest. Large farmers would like to
obtain more credit at the subsidized government bank rates
(or even at a low real rate of interest), but credit from
this source must be rationed since the supply does not equal
the demand by larger farmers. Further, they-may not consider

the positive rate lenders as important sources. Thus, large

"‘Lrnrf

‘
nub. v.4 .

Figgre

Sial

 

 

 

203

SS

    
 
 

Supply From
Other Sources

Real \\\\\
Interest +

O

 

- Subsidized Gov't DD (Large Farms)

Bank Credit

 

 

Figure 12.--Supply and Demand For Credit By Large Recife Farmers.

farmers often explained that they did not borrow more

because "banks do not give farmers larger loans.” The

smaller farmer, on the other hand, may find that the upward
sleping section of the curve is relevant, since he seems to

be relatively less able to obtain government Sponsored credit
(at least he tends to use other sources more frequently than
his larger counterpart). As a consequence, "risk" and asset
limitations would be more important among the reasons he gives'
for not acquiring additional credit.

Returning to the examination of Table 46, we find the
small farmers (0-49 acres) were considerably more encumbered
by lack of assets than their larger counterparts. It should
be noted that the rerage loan to farmers of this size (0—49
acres) was approximately $132. Consequently, a tempting
conclusion is that the small farmers in the Recife area must

meet with extremely stringent collateral requirements when

204
they attempt to borrow money.

Finally, farmers with less than 245 acres were considerably
more likely to cite the ”risk” of borrowing as a reason for
not obtaining larger loans.' At the same time they were, of
course, less likely to indicate that they had obtained
sufficient credit to meet their (perceived) needs.

In summary, larger and relatively wealthier farmers tend
to give different reasons than smaller and poorer farmers for
either not borrowing or not borrowing more if they are
currently using credit. Risk as a deterrent to credit usage
seems to be associated with smaller farm sizes and smaller
farm incomes. Lack of a need for credit or lack of a need for
additional credit (representing 7.1% of the sample) is a
response found relatively more frequently among larger
farmers and farmers with higher incomes. At the same time,
difficulty in obtaining credit or additional amounts of credit
seems to be related to small farm size and lower farm incomes.
Finally, the complaint that "banks do not give larger loans,”
as an explanation for not using more credit, was more often
voiced by larger farmers and the farmers with higher incomes.

Amounts of Credit Used by Farmersn;;_Regression Analysis

 

 

 

An initial regression model of the amounts of credit
used by Recife farmers indicated that property values and
the amount of crops planted, both related to size of farm,
were the only significant explanatory variables when the
population involved all farms in the sample. Further, these

ass than 5.0% of the

‘\

variables were capable of explaining l

nrian
would
naly:

varieb

 

 

205
variance. This suggested that a more fruitful approach
would be to divide the sample according to farm size and
analyze the more homogeneous subsamples. A number of
variables were preposed152 and a least squares add regression
technique was applied to the problem. This approach utilizes
the computer in selecting variables in order of their
importance in explaining the variance within the sample.
Small Farms (0.49 acres)

Among the proposed variables the computer selected
reasons for not borrowing as the first two variables to be
included in the equation computed for small farms. Fear of
borrowing resulted in an R2 of .0646 (explained 6.5% of the

2 .1318.

variance) and ”no need to borrow” increased the R
Since the sample included both borrowers and non—borrowers

this finding is simply a correlation between not having a

 

— -—.-.-—_

152.. .
LlSt of variables:

y--Amount borrowed (dependent V'riable)

x1-~Assets (property value)

xze-lncome level (per capita family)

xK—-Percent of acreage farmed

x;~-1ncome per hectare planted

x5--Days of illness, 1966

x6--Crop failure in 1966 (dummy variable)

xq-—N mber of crop failures (1960-1966)

xé--Risk choice l-low risk and low return (dummy variable)

xQ--Risk choice II-mnderate risk and moderate return (dummy
“ variable)

x10~—Risk choice Ill-high risk and high return (dummy

variable)

x11-~lrrigarion on farm (dummy variable)

x ~~Used tractor (dummy variable)
12“ g_.- 1 a . . 'd .,.- bl

x13 Used chemica- fert1112?r ( ummy lel& e)

xla-wflired labor (dummy variable)

Xis~~Number of tenants

x16--Afraid to borrow

117-«do need to borrow

206
loan and one of the reasons for not borrowing. The value
of the farmer's assets (preperty value) was the next variable
to be selected by the computer, and together with reasons
for not borrowing explained 20.8% of the variance (R2=.2077);
The last signficant variable was the farmer's level of
income (larger amounts of credit associated with larger per
capita family incomes), which increased the level of the R2
statistic to .2336. Each of these variables were significant
at less than the .05 level.
Medium-Sized Farms (SO-244 acres)

The number of tenants on farms was the initial variable
chosen by the computer to explain the variance in amounts
borrowed by farmers Operating 50 to 214 acres. Perhaps the
landlords in this group commonly borrowed in order to relend
to their tenants. The amount per tenant. as indicated by the
equation, was $60.80 (133.78 hCrS). An R2 statistic of .1914
was accomplished by this variable alone. The second variable
chosen was the farmer's income level, increasing the R2
measure to .2950 (roughly 30% of the variance explained).

"No need for credit, as a variable explaining why some

farmers in this group had not borrowed, was the third
variable selected by the computer, which resulted in an R3 of
.3569. A fourth variable chosen, income per hectare of crops
planted, seems to indicate that the farmers earning higher
returns per hectare utilized more credit than farmers obtain-

ing lower levels of productivity from their land (R2=.4191).

Finally, the percentage of the farmer's acreage used in crOps

 

seems t
Large F

T};

*0 act

nT‘

 

inal‘l

207
seems to be positively related to the amounts of credit used
2 .
(R =.4346).
Large Farms (245~1234 acres)
The most important variable, according to the computer
program, in explaining the amount of credit used by large
farmers was the value of their assets. .Alone, this variable

2=.7538).

explained 75% of the variation in amounts borrowed (R
Three additional variables «- risk choice III (high risk-
high return), risk choice II (medium risk-medium return), and

. 2
number of tenants -- increased the R

statistic to .8077, but
their interpretation is uncertain. Farmers who were willing
to accept higher risks for higher returns seemed to be
willing to borrow substantially larger amounts than other
farmers, but risk choice 11 and the number of tenants per
farm were negatively associated with the sums borrowed.
Finally, the variable indicating those farmers who perceived
"no need to borrow" was included in the equation (R2=.8178).
Very Large Farms (1235 or more acres)

Only one variable, percent of the farm devoted to crops,
was significantly related to the amount borrowed by very
large farms -- R2=.2586. On the average, these farmers
planted only 12.5% of their total acreage (including land
planted by tenants). By comparison, small farmers utilized
an average of 61.1% of their acreages for crop production,
mediumwsized farmers planted 33.1% of their farms, and large

farmers raised crops on 20.0% of their land. The maximum

percentage of land devoted to crOps by any-of the very large

farmer

notior

Recife

Fm.

.1. (‘1‘
L, LA:
“‘05
"w. k'.
y:
0,;Ch
!!: ‘v
L'x 1.1:]
"H 1241
I" N.‘
p
yrgns

 

208
farmers was 66. 7 .

In all, the regression equations seem to reinforce the
notion that an explanation of the amount of credit used by‘
Recife farmers involves different variables for farms of
different sizes. With respect to small farms, those variables
indicating the reasons farmers did not borrow and variables
reflecting their ability to obtain loans ~~ assets and income
level -- were variables of primary imnortance. In all
probability, much of the borrowing done by farmers of this
size represented loans for the purpose of covering family
expenditures until the coming harvest. The variables selected
by the conputer in explaining the amounts borrowed by medium-
sized farms (averaging $332) seem to reflect a more "commercial"
orientation in credit usage. Such items as the number of
tenants per farm, income per hectare (a proxy variable for
productivity per hectare) nd percent of the farm devoted to
crops indicate a relationship between the production process
and the use of credit. Nonetheless, there were still farmers

in th7.s group who w re afraid to borrow, even though this

 

——.- -"—-

C1e<lit kecrcssvor Equations

w—J... --.. .—-..-.----— o

 

--—-

Smalquarms (0-40 acres): R2=.2336 . ‘
i = 114.066 - 195.592x16 ~ 2131511x17 + o.olox1 + 0.191x2
bkmiium~8ized Farms, (50 244 acres): Rz=.4346
? = 68. 382 + 133. 7: ox + 0.2oaxz n 857.366X
~ 686.689316 + 0.7 9x4 + 579.241x5

 

17

Farms (24S«1234 acres): R2=.8178

57.830 + 0.043X + 1564.464X ‘ 1849.671X9

Y a

« 104.963X10 ~ 443.992X15

I7
Very Large Farms (1235 acr eS'or more): R =.258o
Y = 673 + 12450.704X3

van

The '

red

’1.-

"‘
Lh’t‘S

l

:0 \’
..CA

1.1"

‘ltr‘

209
variable seemed to be less important than for small farms.
The lack of interpretable relationships between large farms
(245-1234 acres) and variables which reflect the use of
credit for production activities may indicate that many of
these farmers were either not interested or unable to make.
maximum use of their farms for agricultural production.
This notion is reinforced by the small percentage of their
total acreage that was devoted to crop production. On the
other hand, the relationship between amounts borrowed and
asset position may be of such importance that it simply
outweighs other variables. Finally, the association between
the percent of very large farms (over 1235 acres) devoted to
crop production and amounts of credit used suggests that some
of the very large farmers were not intensively utilizing
their properties for agricultural production.
52:11:11,:

Somewhat less than half (45.6%) of the Recife farmers
had not borrowed money in 1966. The largest share of these
farmers could he found on small farms and among those with
relatively low incomes. Approximately one-third of the
farmers who didn't use credit indicated that they were
afraid to borrow. Another third were unable to obtain loans.
lost of those remaining indicated that they "did not need
credit." "No need for credit" was the reason given most
frequently by the larger and relatively wealthier
non~borrowers. Yet, in an absolute sense, the poorer and

smaller farmers were more important among all the farmers

giving

small 1

that ti

their <

f‘"

I] r?

 

210
giving this response. It is reasonable to assume that the
small farmers believed that they "didn't need credit" if
they had managed to meet their family's consumption needs
from one harvest to the next without a loan. The possibility
that they were able to finance new agricultural inputs from
their own meager resources seems quite remote.

When the smaller and poorer farmers did borrow, they
were relatively more likely than their larger and wealthier
counterparts to have obtained loans from neighbors, relatives
and other non-commercial sources. These sources typically
loaned fairly small amounts. Credit cooperatives, however,
seemed to provide somewhat larger loans to the small and
mediumnsized farmers.

The largest loans, however, were usually obtained from
the government banks. Larger and relatively wealthier farmers
seemed to have less difficulty obtaining credit in general,
and they were the ones relatively more likely to receive
their loans from the government bank (and private banks).
Nonetheless, a large preportion of the government bank loans
(probably the smaller ones) went to the relatively smaller
and poorer farmers simply because they constituted such a
great preportion of the sample.

In all, 7.1% of the Recife farmers indicated that they
had obtained sufficient credit to meet their needs. A large
number of these responses were from the larger and relatively
wealthier farmers. Although some Small and mediumssized

farmers also gave this reason for not obtaining larger loans.

a fer:

0‘
A

repo'

1 o

551;)

rub
PI ....“
Al L a
.~b¢ inlau
.JM ‘1»

1H {d
tl. Db
a: VI“
...¢ .1

..ll

.1;

Tin
ulu

. .. 1L
hon
LL — In
. ‘11

 

.. \ .
. s .- a 3 n .
ha 1?... .w \
.s k (C, .
VI: 9 \ \ae
Vii ...; ....Oi

211
"Tenure

Although the Recife survey involved largely landowners,

a few renters were interviewed and a number of the owners
reported tenants on their land. With respect to the relation-
ship between tenure and innovativeness, one would generally
eXpect owner-operators to be the most innovative and both
landlords and renters to be less innovative, but for quite
different reasons.

The tenant is usually considered to be less progressive
than the owner~operator. One of the common reasons suggested
is that he is unwilling to make the long-range decisions and
investments necessary to improve his farm operation. His
unwillingness results from uncertainty about whether he will
remain on the land long enough to reap any of the benefits.

It nay he found that renters are at the mercy of unsympathetic
landlords. This appears to be the case with Recife sharecroppers.
The most common sharecropper arrangement is
equal shares. In other words, the sharecropper
gives half his production to the landowner and keeps

the remaining half for himself. The usual agreement
also requires that the sharecropper sell to the
landowner any part of his share which is not
consumed. The sharecrOpper must sell this part at
harvest time rhea pri as are lowest. If the
sharecronper does not sell to the landlord, he will
be removed from the land the following year. “3

It is hardly reasonable to expect the renter to be particularly
innovative under the circumstances cited above.

A second reason that might be given is that the tenant
(sharecropper) receives only a share of the increased product

-~153Charles

*Processes, pp.

.—-—-—D~.

ter, Harold M. Ri1e*,'et a]., Market_

\
I

 

5313
7sll

from

 

 

who! .1. 6.

*IS ‘Arnfi
‘véfl I 8
JR p hl Ml ..

L ‘Iu‘ halo r:

212

from a new method. If he must bear all of its cost, the new
input may be unprofitable to him even though it is quite
profitable in the aggregate sense. A third reason for
expecting the tenant to be less innovative is that he may
find it difficult to finance new methods of production,
especially where credit sources demand land as collateral for
loans. Finally, we might expect tenants to be very backward
in their beliefs and attitudes, and to lack the educationa .1
background of other members of the community.

The Latin American landlord has at times received
considerable attention because of his lack of progressiveness.
Arguments over the reasons for his laggardliness have

54 Typical of the reasons suggested

apparently not ceased.
are that his income is adequate without giving his holdings
the attention they would require for maximum profits, that he
is more politically oriented than economically motivated, and
that he holds land for purposes other than agricultural
profit. For the moment, we will assume that the Recife-area

landlord 15 enough similar to the general concept of Latin

 

l,
1Mlirnost Peder, "The Latifundia Puzzle of Professor

Schultz: Comment,” Journal of Farm Economics, (May 1967),

pp. 537'510; Theodore h. ”Schultz. ”Theilatifundia Puzzle

of Professor Schultz: Reply," Journal of Farm Economics,

(May 1967), pp. 511-513.

Dale W. Adams, "Resource Allocation in Traditional Agri-
culture; Comment," Journal of Farm Economics) (November 1967),
pp. 930~932; Theodore w. Schultz,filesource Allocation in
Traditional Agriculture: Reply," Journal of Farm Economics,
(Noveviler 1967), pp. 933~935. -

E. Grunig,'lnformati0n, Entrepreneurship, and

 

 

 

 

 

 

 

James ___ 2--.“...—
Economic Devflenmcnt: '3. Stanly oi toe L‘sc is—Iin—li kinv.
Procc§ses of (o olembian ratlfuncrstps. *L7 ”n ublished Lh.D.
dissertation, University 01 hisconsln, l96 68), pp 9—12.

Americ

innova

farmer

 

TE

 

 

 

213
American landlords that we would expect him to be less
innovative than the owner~operators.

Many of our preconceptions about the differences between
farmers with different tenure statuses are not, however,
supported by the data from the Recife survey. First, we find
that tenants, except for fatalism, held attitudes and beliefs
which were not significantly different from landlords and
owner—operators. Again, tenants seemed to be nearly as
likely to know how to improve yields as landlords and owner-
Operators (53.5% as compared to 64.1% and 60.6% respectively).
Also, the educational achievement of tenants was not
significantly different from owner-operators, while landlords
tended to have completed somewhat more schooling than the
others.

Table 47.-~The Percentage of Tenants, Landlords and Owner-

Operators Who Had Achieved Given Levels of Formal
Education. -

 

Years of School Completed

 

 

Tenure None 1 yr. 2-3 1:5 6 or
yrs. yrs. more

yrs.
---------------- percent-------------------

Tenant 48.8 14.0 20.9 16.3 --
Owner-Operator 47.7 20.3 20.3 6.8 4.9
Landlord ..34.8. 14.6 25.8 13.1 11.6

 

.Although none of the tenants had completed six or more years

of schooling, they were somewhat more likely than owner~

Operators to have completed from two to five years of formal

education. Landlords were considerably less likely to not

214
have attended school at all, and were much more likely to
have completed more than six years of education. In fact,
2.5% of the landlords had attended school beyond the twelfth
year (as compared to 0.5% of the owner~operators and none of
the tenants).

At least one of our contentions about tenants seems to
be correct. As may be recalled from the previous section,
they seem to find it more difficult to obtain credit than
their landowner counterparts. The greatest percentage of
tenants depended upon loans from landowners, which were
generally for very small amounts. At the same time, this
source of credit was relatively unimportant to the Recife
farmers who owned their farms.

Of greatest interest, of course, is the relationship
between tenure and the farmer's use of particular inputs.
From examination of the use of four selec;ed innovations,
some general impressions about the relative innovativeness
of tenants, owner—operators, and landlords can be obtained.

Table 48.-~Thc Percentace of Tenants Owner‘O erators and
(z, 9 . 9
Landlords Uglng Selected Innovations.a

 

Tenure Non—chemical Chemical insecticides Tractors
Fertilizer Fertilizer

 

Tenant 6.9 2.3' 58.3

25.6
Owner-Operator 16.6 1.8 32.6 14.3
Landlord . 21.7 II..7 '4.0 . 58.1 21.8

 

f

aA chi-square analysis of insecticides and tractors was
significant at the .05 level.

    

Landlci
J

mm .

m 5'1;

 

 

215
Landlords seem to be more likely to use insecticides and
tractors than owner-operators, although the differences were
not significant for barnyard manure and chemical fertilizers.
This result is probably a byproduct of the landlords generally
higher level of education and more favorable financial
position. At the same time, the percentage of tenants using
tractors and insecticides seems to more closely resemble his
landlord rathtr than his owner-operator counterparts. The
likely conclusion is that landlords either require their
tenants to use these innovations or provide them with tractors
and insecticides. In general, evidence does not support the
contention that Recife area owner-operators are more innovative
than tenants and landlords. On the contrary, the evidence
seems to suggest that landlords are the most innovative and
that tenants ray use insecticides and tractors at the landlord's

stance.

*4.

ins
In summary, we have found that many of our initial

snag-stions about differences between tenants, owner-operators,

.3

Lb

.n landlords are not supported by the Recife data. Except

A

C
a

for “talism, the attitudes and beliefs of farmers with
different tenure statuses were not found to be statistically
significant. In addition, the tenants seemed to be as well
educated as the ow1.rtoperators, while landlords tended to
have somewhat more schooling than either of the other two
groups. However, tenants were found to be at a disadvantage
in obtaining credit. Finally, the hypothesis that owners
Operators are more innovative than landlords and tenants seems

to be unsupported. If any generalizations may be made, it

216
appears tlat landlords and tenants are more innovative than
owner-operators. The most reasonable explanation of the
tenant's behavior is that decisions regardinU farm inputs

D

are made by the tenant's landlord.

gum: Availability

 

A farmer can hardly be expected to be iDDOs ative if
modern agricultural inputs are difficult to obtain.
Consequently, it seems reasonable to consider the availability
of certain inputs in an examination of the Recife farmer's
innovativeness. Fortunately, the Recife survey addressed
certain questions to this matter. In a separate survey, the
MSU/SUJZNE researche s interviewed a sa .ple of thirty rural
suppliers located in the same areas from which farmers' survey
was conducted. Table 49 gives the percentage of respondents
who stocked selected inputs. In addition, the table int Mi ates
the percentage of farmers from the Recife sample who were
using these inputs

A general impression gained from Table 49 is that a
relatioflnsiip seems to exist between the local level avail-
ability of agricultural inputs and their use by Recife

farmers. The availability and corresponding use of selected

seeds, alone, fails to fit this overall pattern. An

explanation of this particular outcome 15 probably best given

re from the Recife research report:

5.4

by the following passa

The high response on improved seeds, however, is
probably due to tne wor din; of the question, which
led most farmers to repl affirzatirely even if they
were doing nething .nore thrn selecting the biggest
and best looking seeds from their own production for

Tab}

 

{-a-

’ 1’7
9...;
‘ )

 

217

Table 49.-~Percent of Rural Distributors Supplying Selected
Inputs and the Percentage of Recife Farmers Using
These Inputs.a .-. ,.. . ............ ... ’.

 

 

Percent of Percent of
Inputs Rural Suppliers Recife Farmers
Carrying - Using Selected
.Selected Inputs . Inputs

Chemical Fertilizer ' ~- ' 3
Tractor ~— Rental 15 17
Selected Seed 18 53
Insecticides 78 43
Fungicides 48 27
Hand Planters . 52 i - . -22-

...........

 

“Inputs correspond with practices considered innovations
in the Recife Farmer Survey.

planting. "ad the question specified use of

special i pioxed variet1“ cs, strains, or hybrids, the

response V.OU1d have been much lower, since the

supply of such improved seeds that have been,

developed for the Northeast is very limited. 155
With exclusion of “selected seeds," we find that our general
impression is supported statistically. The correlation
co-efficient between input availability (as given by the

\

lnput Sur‘:y) and the use of selected practices is r=.9481
(which is significant at the .05 level).

In more specific terms, insecticides seemed to be the
most readily available input, and (aside from "selected seeds")

the most commonly used innovation. At the other extreme,

the survey of rural suppliers failed to find a firm which

 

15 a . .. _N .. q. “.1. .
- charles Slater, darold M, Riley, et al., Market;
Process s,

p. 8/21.

9

 

218
handled chemical fertilizer, an innovation which only a very
few Recife farmers were found to be using. The availability
of tractors for rent is found to be very closely related to
the percentage of farmers using them. In addition, 7.0%
of the rural suppliers had tractors to sell.

The Recife Farmer Survey also asked about the avail-
ability of certain inputs. Farmers were asked if they
believed that they could find chemical fertilizers, insecti-
cides, and rental ractors in the stores in their own
municipie (county). Table 50 indicates the responses of
the farmers interviewed.

Table SO.--The Percentage of Recife Farmers Who Believed
Chemical Fertilizer, Insecticides and Rental

Tractors Were Available in Their Municipio, By
Commodity Area.

 

 

 

Percent Who Believed These Inputs

were fiyailable In Their Municipio

Chenical fihental

Fertilizer Insecticides Tractors
sao Francisco--Kice 7.9 50.4 43.3
Al-Pe--Bean 11.7 i 69.5 28.9
Irecé--Bean 8.2 93.9 98.0
Maniac 4.0 42.4 14.7
Cotton 25.4 87.9 15.3

 

It is interesting to note that nearly all of the Irecé bean
farmers thought that they could rent a tractor locally. It
was also found that 65.3% of the Irece farmers used tractors.
In fact, the relationship between the availability of rental

tractors -- as nerceired by the Iarmeis -- and their use in

the

hora

am

 

 

219
the different areas was highly correlated (r=.8873‘.

The use of insecticides and chemical fertilizers,
however, does not seem to vary directly with their local
availability. The bean farmers sampled did not use chemical
fertilizer at all, even though a few farmers in each area
believed they could buy this input from local storekeepers.
At the same time, the percentage of Sao Francisco rice
farmers who used fertilizer was greater than the proportion
who believed it was available locally (11.8% as opposed to
7.9%). Apparently, some of these farmers were capable of
reaching beyond their own community to obtain inputs for
use on their farms. The percentage of manioc and cotton
farmers using fertilizer was minimal (0.7% and 0.6% respective-
ly), despite the fact that 25.4% of the cotton farmers thought
that they could buy it locally.

The percentage of farmers who believed insecticides
were avai able locally and the percentage who utilized this
input seems to be somewhat more related than was found with
fertilizers. Still, they were not significantly correlated.
tith rCSpoct to insecticides, again we find the percentage
of She Francisco rice farmers using this input exceeding the
percentage who believed that it was available locally. The
same was true of manioc farmers. Yet only about onewfifth
of A1~Pe bean farmers who believed that they could obtain
insecticides locally were utilizing this input.

The most reasonable conclusion from the Recife data is

that a general relationship between input availability and.

220
'nput use seems to exist, yet other factors involved in the
decision system can inhibit many farmers from adopting an
innovation and a few farmers may use new inputs even when they

do not find them available locally.

EETTEQIC

In Chapter V a schematic representation was develOped
relating several variables to the decision-asking process
involved in the adoption of new technologies. Figure l3 adds
to this schema three variables which can affect the ability
of the farmer to implement his decision to utilize a new
technology.

The indicated relationship between credit and either farm
size or income is supported by a number of findings which were
forthcoming in this chapter. First, there is an association
between farm size and the percentage of farmers receiving
loans, with the small farmers being less likely to have
obtained a loan than their larger counterparts. Next, farm
size was found to be related to the source of credit utilized;
small farmers tending to make greater use of credit cooperatives
and non-commercial sources while larger farmers depend more
heavily Upon government banks. The percentage of farmers who
were afraid to borrow or indicated that they were unable to get
a loan decreased as farm size and farm income increased.

Among borrowers, lack of sufficient assets and the additional
risk involved were more often cited by small farmers than by
larger farmers as reasons for not using more credit.

Tenure also affects the source of credit and size of loans.

221

.HH.nu h< 0. wood can a“ uqu—um «o wouvxm ucfixez.

 

>PH4Hw<AH<><
Faun“

 

 

 

    

 

 

~1ZO><I~MDZW

   

a

o

o

H m. ochcwury; ,
.F Kr? (AWIEOI
a

o

e

...

 

 

MEN:

 

 

 

mzuuz~
omm<mmux
“an rkquha

 

 

       
 

 

.0855." .
Safiza 5,203 .
5:. 2 $967.53: “I!

  

 

 

 

 

 

 

 

: . .1 mafia.“ IA;
mo smut. ., :35 f
4 cz<
axed hunm¢u
£5:

 

 

 

 

 

  

 
  

{refit

reuuuuon osu uo 30m>

 

)7fl9:.

{its

umumaom
zomr<2uow2~

 

 

 

. \7.

up”.-
1...;

 

Jupam
zo—h<u;am

I

 

 

 

 

 

 

mu~m «é .<.

 

 

 

uwuucusum -.

ma ouaufim

 

 

 

 

 

 

 

 

 

nmaa<~<a
=uuh zuz
zxa
maakapn< .
E 29.28

 

 

 

 

 

 

222
Tenants seem to be heavily dependent upon landowners for their
credit and resulting loans are generally small by comparison
to those from other sources. This preceding relationship is
depicted by the arrow connecting tenure patterns to credit
in the schema.

The CXpCCth relationships between tenure and attitudes
was not supported by the data. Except for a somewhat greater
tendency to be fatalistic, the attitudes hele by tenants were
not significantly different than landowners. Again, contrary
to the proposed hypotheses, owner-operators Jere discovered
to be less likely to have used particular innovations than
either landlords or tenants. Even more surprising, the use
of innovations by tenants more closely resembles the pattern
of the landlord than that of the owner—operator.

Finally, the analyses in this chapter suggest that the
local availability of inputs is likely to have an impact upon
their adeptien and use. Thus, the potential effect of input

availability is also indicated in Figure 11.

AL.

VI 1'!

ts

("P

 

 

CHAPTER VII

THE INTERACTION 0" SELECTED “KRTAPIIF IN THF
ADOPTION P' tOChSS: lfi AUTUMAFI C l\thuL1IU
DETTCTlON APP? \OACH TO THU ANnLYSIS
OF INN VATIVENESS

Centre 1 to this discussion of 1PQQV'tlx ness has been
a concept of the adoption of new methods as involving the
interaction of several variables witiin the decision-making
processes of individual Recife area farmers. Within the
context of the theoretical framework developed earlier,
Chzn tors V and VI have attempted to outline the sys stema.tic
inltrac tion of certain variables as they appear to be related
to the ace tion process. The t.sk of this chapter is to
examine this interact: V1 process, once again, in an effort

\
.
o

to rev-clog: a greater tn‘dcr<‘...:r..1-ng of innovativeness among

Recife farmers and to provide some estimate of the relative

importance (prediCtive value) 0 those variable s which seemed
to be involved in the adoption p1. oc es:. A computer program
termed ivtozuatic interaction getection (A.1.D.) will he used
in this analysis. This appr ash permits the s mult aneou 5

analysis of certain variables which have been constructed

'itb the benefit cf those aml alvses undertaken in the two

‘4

previous chapters.

criction Dete t1on Program

'gfiflfi—-l n”...-

                 

1

Automatic interaction Detection is a coInputer program

7‘?

th University of

('0

developed by SonquiSt and Mcrgal a

ill Chi

the 11

 

 

 

 

224
Michigan.156 They explain the purpose of its development in
the following excerpt:

It [A.I.D.] is focused on a particular kind of
data-analysis problem, characteristic of many social
science research situations, in which the purpose of
the analysis involves more than the reporting of
descriptive statistics, but may not necessarily
involve the exact testing of specific hypotheses.

In this type of situation the problem is often one of
determining which of the variables, for which data
have been collected, are related to the phenomenon in
question, under what conditions, and through what
intervening processes, with appropriate controls for
spuriousness . 1 57

In essence, this program is specifically designed for the
analysis of data where the interaction of the independent
variables is eXpected to have an effect on the value of the
lependent variable. Sonquist and Morgan, in their description
of A.1.D., elaborate on this particular point:

The objective is to explain the variance of the
dependent variable Y. Where the number of predictors
is small, the problems of isolating the relationships
between X- and Y are manageable, but when tne number
of prediciors is large, which is typical or many
survey data analysis problems, then an analysis of the
joint effects of X1 or Y presents serious problems.158

From this brief description, the purpose of the A.1.D.
computer program seems to closely coincide with the analytical
objectives of this study. From the beginning, an interactive

system of decision-making has been proposed; and secondly,

the objective of developing and testing of an overall

 

156John A. Soncuist and James V. Morgan, The Detection
of Interaction Effects, (Ann Arbor, Institute for Soc1aI
Research, Report No. 35, University of Michigan).

157

 

121$;1 p. 2'

158Ibid., p. 2.

225
conceptual framework of the adoption process has been at the
forefront of this research. Thus, the A.1.D. program
promises to be a useful tool in examining the interaction of
certain independent variables as they affect the innovative-
ness of Recife area farmers.

Advantages and Disadvantages of A.1.D.

 

In practice, the A.1.D. program seems to have certain
advantages over multiple correlation for the type of analysis
involved in this study. First, one need not specify in
advance the interaction terms (xixk) that will be needed in
the analysis. It is obvious that preparing and interpreting
a regression model with a large number of interaction
possibilities would soon become rather burdensome. At the
same time, the A.1.D. program allows considerable flexibility
for the interaction of variables and, in effect, specifies
these interactions in the course of the computations.
Secondly, the A.1.D. program can be used to analyze a
dichotomous (O-l) dependent variable. Finally, the A.1.D.
technique lends itself to a very comprehensible visual
display of the program results (in the form of a branching
tree with specific variables related to each branch).

A.1.D. does have certain limitations. From the
experiences of this study, it appears that small sample
sizes (less than N=400 or 500) pose a problem to the analysist.
Operationally, the program divides and subdivides the sample
on the basis of the variables which explain the greatest

amount of the total sums of squares (maximizes the between

sums
the r
as t}

sampf

 

c]:

11.1

 

'

 

226
sums of squares at each split). Small sample sizes mean that
the number of useful "splits" is limited. One reason is that
as the program divides the sample into subgroups, the
sample size of the subgroups rapidly diminishes to sizes from
which additional partitions are of questionable value.
(Sonquist and Morgan set N=25 as the minimum group size from
which further partitions should be considered legitimate.)
Secondly, the number of observations for each classification
within a variable (predictor) also can quickly become too
small to trust the significance of their mean values (which
are important to the operation of the A.1.D. program). For
example, in the A.1.D. I model which follows, the "tenant"
classification within the predictor variable "tenure"
initially included 16 observations from a total sample of
N=119. Five steps later this classification included only
six observations in this particular subgroup. If there had
initially been five times as many observations (N=80) for
this classification, we could be much more comfortable with
the mean value (percentage who had adopted fertilizer)
calculated from a probable "tenant" subgroup size of N=3O
after five computational steps. In short, small sample
sizes limit our ability to take advantage of the possible
Predictive value of certain variables. The difficulties of
dealing with small sample sizes is, of course, not unique to
the A.1.D. program, but it nonetheless should be recognized
35 a problem in its usage.

Another difficulty is that the A.1.D. program does not

227
supply us with the coefficients we are accustomed to dealing
with in multiple correlation analyses. The output does not
provide a beta coefficient (and accompanying statistics)
relating numerical values of the independent variables to
the value of the dependent variable. For some analytical
purposes this may be objectionable. The program does, however,
provide an estimate of the value of the dependent variable
(and accompanying standard deviation) for groups which include
certain classifications from within the various independent
variables (say a mean innovation index of 58.1% for farmers
with more than one year of education, per capita family
incomes of over $200, and access to sufficient credit). For
the purposes of this study, an identification of the inter-
active relationships between the dependent and independent
variables, and their relative importance (amount of variance

explained) is probably sufficient.

The Dependent Variables

 

Innovativeness is defined as the degree to
which an indtviduai adopts new ideas relatively
earlier than others in his social system.
(Rogers: 1969)159

 

aneral Approaches to Indices of lrnevation

 

 

Typically, there have been two approaches to measuring
an individual's innovativeness: (1) those which use an

adOption/non-adoption dichotomy and (2) those which use the

 

 

159Rogers, Peasants, p. 294.

228

160 The first approach involves computing

time of adoption.
the percentage of a list of "applicable practices" adopted
by an individual and considering this to be a measure of
his innovativeness relative to other members of his
community. The second approach involves determining the
date of the individual's first use of an innovation and
measuring his innovativeness in terms of the earliness or
lateness of his adoption relative to his colleagues. An
implicit assumption included in this second approach is that
an individual's response to some particular innovation is
sufficient to measure his general level of innovativeness.
The first approach also involves an implicit assumption,
that all innovations are equivalent units. It may be
recalled, from Chapter II, that Rogers and others have sug-
gested that innovations differ with respect to various
characteristics (see pp. 34-35). Yet, it is usual practice
to construct innovation indexes ignoring these differences.
Intuitively it would seem that an individual who has
adopted chemical fertilizer should be considered more
innovative than the farmer who has adepted non«chemical
fertilizer (application of barnyard manure), but the impact
0f either of these practices (differing in relative
advantage, complexity, and compatability with previous
experiences) upon the typical index of innovation (where

both were considered as "innovations") is implicitly
160Tom W. Carroll, "Diffusion Research: Application to

National Surveys in Developing Countries," (Working Paper
#20, Department of Communications, February 1968), p. 9.

r-v-

 

 

229
assumed to be equivalent. It should be recognized, then,
that an innovation scale of this type is in fact a weighted
index, with all "innovations" given exactly the same weight.

An Improved Innovation Index

 

It would seem reasonable to argue that adoption of an
innovation wnich is both technically complex and a
considerable departure from past practices would be a greater
test of an individual's innovativeness than adoption of an
innovation which is simple and not greatly different from
present practices. Perhaps the same argument could be made
for differences in other characteristics such as divisibility
(technical characteristic which permits small scale trials)
and conmunicability (observability of the results of an
innovation), although the reasoning is not as clear for
these characteristics. Suppose, for example, that both
tractors and the ex and plow are included within a list of
practices to be used as an index of innovativeness for
farmers in a less developed country. It would be the unusual
peasant who would have considerable experience with mechanical
apparatus and internal combustion engines but little knowledge
of farm animals. The tractor, barring such extreme cases,
would clearly be a more complex and less compatible
innovation than the ex and plow. It should then carry a
greater weight within the innovation index. Assigning
Weights on this basis presents some problems, however, since
we as yet have no objective method of scaling the degree of

Clifferences found in these characteristics. Nonetheless,

230

it seems that innovations could be placed into groups that
are somewhat alike in terms of their complexity and
compatibility on the basis of the judgement of a competent
observer. Since neglect in doing so is to assume them equal
in these characteristics, a carefully considered subjective
ranking -- recognizing these differences in the innovations --
would probably provide an improvement in the index.

Consideration of the relative advantage of innovations
is the point of great disagreement between economists and
diffusion researchers (emphasized by the first and de-emphasized
by the latter). Both argue that the farmer does what is to
his advantage, but diffusion researchers seem to believe it
is the perceived advantage (involving both economic and non-
economic factors) that contributes to adoption of an
innovation while economists prefer to regard the actual
profitability of the technology as the motivating factor.161
The way in which one regards the relative advantage of an
innovation has considerable impact on the method the
researcher uses to weight this factor in an index of
innovation. The approach followed in this chapter is to
regard the innovative individual as the one who is first to
formulate realistic eXpectations about the profitability of

various innovations and adopt those which would be profitable

161Diffusion, p. 126 for the “perceived relative
advantagE“ View.

See Schultz, Transforming, p. 164 for the "objective"
profitability view.

See Wharton, §ybsistence, p. 459 for an interesting
discussion of peasant ”rationality" as viewed by economists
and non~econemists. ‘

 

 

231

(with some discounting for uncertainty) in his particular
situation. This view places the emphasis on the profit-
ability of the innovation, with innovations which have less
impact on agricultural output and efficiency being considered
as less important than more productive innovations. Such a
view is consistent with adoption as a decision-making process
proposed in this study (expectations about innovations being
formulated and acted upon by the individual) and consistent
with the developmental goal of increasing agricultural
production through the diffusion of output-increasing
technologies. Establishing a relationship between innovative-
ness and the profitability of the applicable innovations adds
an extra dimension to the traditional definition of
innovativeness. The innoxative individual is thus not only
relatively early in his adoption of new ideas, but also more
inclined to adopt those ideas which are most profitable (and
generally have the most impact on increasing productivity).

An improved innovation index, then, would include
relative weightings for various innovations on the basis of
their characteristics. While determining the weightings
for such characteristics as complexity and compatibility
must probably still rely on the judgement of the
researchers, the relative profitability of innovations could
be derived empirically. One method would be to use
experiment station comparisons of old and new practices
and rank the various innovations on the basis of their

contribution to increased farm incomes. Non-comparability

232
of experiment station and farm level results from innovations,
however, is a recognized problem. It would also be possible
to measure differences in output at the farm level and from
a budgeting study determine the relative profitability of
various innovations. Still another method would be to use
Cobb-Douglas or other production functions to determine the
marginal productivity of certain innovations at common levels
of usage and rank the innovations on this basis. Utilizing
some of these approaches, which might involve only a marginal
increase in information gathered in a diffusion study, would
certainly serve to improve the adequacy of innovation
indexes and bring economists and non-economists to a
greater level of agreement on procedures for studying the
diffusion of innovations.

The Index of Inppvativeness Used in This Study

 

Unfortunately, the data derived from the LAMP/SUDBNE
Farm Survey does not provide the information needed to
calculate the relative profitability of the technologies
considered as innovations. Since the LAMP study was
primarily concerned with market processes, the researchers
involved did not gather this type of information. The
choice thus becomes one of following traditional practices
and assuming all innovations as equivalent units, or
attempting, although perhaps quite subjectively, to provide
SOHKBITflaIiVC weighting of the innovations considered.
EHche there seems to be no theoretical loss in weighting

the? innovations differently (as all innovation indexes

ix

SC

Z33
involve a weighting) and there may be some gains in even a
somewhat subjective weighting, a simple weighted index will
be used in this study.

There were nine "modern farm inputs" studied by the
LAM? research team in Northeast Brazil: (1) non—chemical
fertilizer (barnyard manure), (2) chemical fertilizer,
(3) improved seeds, (4) insecticides, (5) fumigants used
in grain storage, (6) ox and plow, (7) hand planter,

(8) tractor, and (9) pesticide dusting equipment. The
use of these inputs is given in Table 31-

These innovations will be divided into two groups on
the basis of their expected potential for increasing
productivity, compatibility with past practices, and
complexity. This is, of course, some distance from an

ideal index, but it does seem to be a movement in the

'fi

correct direction. Of the nine innovations, insecticide ,
fumigants, chemical fertilizer, and tractors are judged to
be the most complex, have the least compatibility with past
experiences, and offer the most potential for increasing
productivity. Improved seeds would normally fit within
this category, but in this case the data seems to reflect
the selection of biggest and best looking seeds from the
individual's own production rather than the use of improved

162

strains, varieties, or hybrids. It might also seem

that tractors would not be a suitable innovation to include

 

lfizCharles Slater, Harold Riley, et :—_~.}_._, market_
Frocesses, p. 8-21.

 

 

a:

SI

Dr:

234
as a more profitable input because of the large number of
small farms included in the sample. The practice of
renting tractors is reasonably widespread in the bean and
rice areas, however, and although large farmers were more
likely to use tractors,one-third of the respondents who had
used tractors in 1966 farmed fifty acres or less. These
four practices (insecticides, fumigants, chemical fertilizer,
and tractors) are therefore weighted more heavily than the
remainder of the inputs. For lack of a better system of
establishing relative weights, which will hopefully be
included in future studies, the use of each of these four
innovations is weighted twice as heavily as the use of any
of the other innovations in calculating individual innovative-
ness scores. The inadequacy of this method of weighting
innovations is clearly recognized, but it is believed to be
at least somewhat more adequate than the typical system of

preparing indexes of innovativeness.

 

 

The Dichotomous Dependen£_Variables Used in This Study

In addition to the innovation index proposed above,
the adeptien/non-adoption of three innovations will be
considered individually (dichotomous dependent variables).
As a counterpart to the usual index of innovations it would
also seem useful to attempt to identify (from the available
data) the variables with the greatest impact upon the
farmer's decision-making process with respect to the use
of particular innovations. Fortunately, the A.1.D. program

provides a statistical technique for examining the

YET
ana

ini

235
relationship between independent variables and the
adoption/non-adoption of selected new practices. Three
such practices were chosen for individual attention:
chemical fertilizer, tractors, and insecticides. The
program subdivides the sample on the basis of explaining

the percentage of farmers adopting a particular practice.

Selected Independent Variables

 

The A.1.D. analysis reported in this chapter involves
variables which have been selected with the benefit of the
analyses in previous chapters and the results of an
initial computer run. A number of possible variables,
both those which were utilized and those which were not,
are discussed in this section. (See pages 245-247 for
a summary of the independent variables.)

Farm Size

 

Size of farm was found to be associated with income
levels, credit usage, educational achievement and informa-
tion sources used by the Recife farmer. Although it was
not reported earlier, the size of farms was significantly
associated with the use of certain innovations —- chemical
fertilizer, tractors, and insecticides. One of the reasons
for not emphasizing size of farm in earlier chapters and not
including it in the A.1.D. program is its close identifica-
tion with other variables which seem to have a more direct
relationship to farmer decision-making with respect to
innovations. Again, since it is difficult to hypothesize

a "uni-dimensional" (direct) relationship between size of

far
pra
inc
jus

Ten

USed

236
farm and innovativeness, it would seem to be a better
practice to deal with variables (such as education level,
income, credit usage) which have a clearer theoretical
justification.

Ienure

 

The evidence of Chapter V indicates that, contrary to
our original hypotheses, tenants are more akin to landlords
than owner-operators in their likelihood of using new
technologies. And that the proportion of tenants and land-
lords using insecticides is higher than with owner-operators.
The rationale proposed for this finding is that landlords
are more innovative because of their larger farm size and
greater financial capability, while the innovativeness of
tenants reflects their relationships with the landlord.
Specifically, landlords may either furnish certain inputs
to tenants or insist that they use them. Some verification
of these arguments is possible by including tenure as a
variable in the A.1.D. programs.

Income level

 

Income level as a variable must be recognized as having
a "multi-dimensional" character. They reflect both the
financial ability to purchase new inputs and the capacity
to withstand the risk inherent in new technology. Income
is also interrelated with a number of other variables:
educational level, credit usage, and information sources
Used by the Recife farmers. Consequently, a program

division ("split”) based upon income should be carefully

exm
of '
As ‘
as ‘
the
ten
"ri

CIR
R

inc

|
A

on

v11
to
111;

One

237

examined in an attempt to identify which of the dimensions
of this variable are affecting the decision-making process.
As before, this variable is included in the A.1.D. program
as per capita family income, a family "well-being" form of
the income variable. It is hoped that in this form it will
tend to reflect the willingness of individuals to accept the
"risks" involved in new methods.
Credit

There are several ways in which credit could be
included in the A.1.D. analysis. One could use the amount
borrowed or the amount borrowed per acre as the variable.
In either case the variable would strongly reflect the size
of farm (as large farmers borrow more in total, but small
farmers tend to borrow more per acre). A second approach
is to simply divide farmers on the basis of those who borrowe
and those who didn't. This would provide a very minimum of
information about the relationship between credit and use
of new agricultural methods. Another alternative is to
construct the variable on the basis of the sources of
credit used by the farmers, remembering that the amounts
advanced and the typical clientele differed by sources. It
will be recalled that banks (government and private) tended
to loan the larger amounts and the larger farmers were more
likely to borrow from this source. Nonetheless, approximately
one-quarter of the bank loans were obtained by farmers with
less than fifty acres. Credit cooperatives were frequently

used sources of credit among small farmers and tended to

prox
comm
busi

wit}

tota

had

thre.

238
provide their clientele with larger loans than non-
commercial sources -- neighbors and relatives, local
businessmen, and landowners. Non-commercial sources,
with the notable exception of buyers (who loaned relatively
large amounts to a few Sao Francisco rice farmers and whose
total credit activities accounted for only 3.4% of the loans
received by Recife farmers), tended to loan very small
amounts (especially landowners) and were an important
source to smaller farmers. It might be argued, then,
that the ability of farmers to finance new agricultural
methods is related to the sources from which they were
able to obtain credit. We would expect farmers who had
obtained credit from banks to use new methods more
frequently than those who had borrowed from credit
cooperatives, and farmers who had borrowed from either
of these sources to be more innovative than producers who
had obtained their loans from non-commercial sources.

Still another approach would be to divide farmers into
three groups depending upon whether or not they had
borrowed, and if those borrowing believed they had obtained
sufficient credit to meet their needs (recalling that 7.1%
of the Recife farmers didn't borrow more because they had
"no need" for additional credit). The resultant groups
would consist of (l) farmers who had not borrowed for
various reasons, (2) borrowers who did not obtain larger
loans because of internal ("risk") or external (couldn't

get larger loans) credit rationing, and (3) those who

bel
nee
wou
(su

who

Some
BDOU
tion
than

\J

to H

239
believed they had obtained sufficient credit to meet their
needs. Intuitively, it would seem that the latter group
would be in the best position to adopt new methods
(sufficient credit), and that the second group (borrowers
who might have wished to borrow larger amounts) would be in
a somewhat less satisfactory position to purchase (or rent)
new inputs, but still better able to adept new practices
than those who did not borrow at all. This generalization
is, of course, affected by the individual's ability to
finance new practices from his own resources.

The credit variable used in the A.1.D. analysis combines
some of these possibilities in order to glean the greatest
amount of information from the program. Five classifica-
tions are involved in this variable. The first includes
those farmers who didn't use credit, allowing the program
to "split” the sample on the basis of use/non-use of credit.
Classifications two through four relate to the sources of
credit used by farmers -- (2) non-commercial sources,

(3) credit cooperatives, and (4) banks. This permits the
program to seek out those sources of credit which are used
most by the more innovative farmers. The final classifica-
tion involves those few farmers who "didn’t need" additional
credit, allowing the program to "split" off this group if
sufficient credit (as perceived by the farmer) is found to
be related to innovativeness. In essence, an attempt has
been made to construct one credit variable capable of

providing information equivalent to two or three variables.

net
WOU

inn

Var:
on i
uric

far:

Z40

Attitude Toward Tryipg New Practices

 

Slightly over oneoquarter (26.4%) of the Recife
farmers indicated that they would be willing to try a new
method without first observing it in use by others. One
would expect these individuals, on the average, to be more
innovative than the remaining farmers. As such, a dichotomous
variable involving the farmer's attitude toward trying new
methods may be one method of measuring the importance of the
"risk factor," as separated from the financial capacity to
purchase new inputs, to decision-making with regard to the
adoption process. It should be remembered that the willing-
ness of Recife farmers to use untried methods was found to
be associated with their capacity to absorb losses (in terms
of income levels and farm size).

I1} 5.1.1:}; £212.19;

The "risk-choice" question (p. 179) was designed to
determine the farmer's willingness to accept risks in order
to obtain additional income. It was hoped that this
variable would provide some ranking of the Recife farmers
on the basis of their utility for increased income. The
evidence given in Chapter IV indicated, however, that the
farmer's "risk-choice" was not significantly associated with
the use of insecticides, chemical fertilizers, and tractors.
The A.I.D. program provides an additional opportunity to
examine the usefulness of this variable.

Brice Responsiveness

 

The farmer's response to increases in prices is often

used
reas
devc
crop

be a

241
used as a measure of his economic motivation. For this
reason price responsiveness (expected increase in acreage
devoted to crop A e the total acreage currently devoted to
crops, given a hypothetical price increase) was expected to
be associated with the farmer's responsiveness to the profit
opportunities represented by new practices. The decision
was made to eliminate this variable from the A.1.D. program
following the initial computer runs. In several cases it
was selected by the computer as a statistically important
variable, but these results were all incomprehensible.
Nearly without fail, the program would select the less
price responsive individuals as the most likely to use
innovations. Since a suitable interpretation of this out-
come could not be developed, the variable was simply
omitted in the final programs.

.Tnp_1.1_t___.§_‘:'_ai labi lity

 

The local availability of modern inputs is thought to
be a prerequisite for progress in backward rural communities.
The Recife survey asked farmers to indicate whether or not
they believed rental tractors, insecticides, and chemical
fertilizer were available in their municipio (county). The
responses to these questions were used as a measure of input
availability in the A.1.D. program. For the analyses of
single innovations (dichotomous dependent variables) the
farmer's response regarding the availability of that
particular input was used as a variable. An index of input

availability was prepared for use with the programs

242
involving an index of innovativeness as the dependent
variable. This index was simply a summation of the
affirmative responses given by farmers to questions about
the local availability of each of the following inputs:
chemical fertilizers, insecticides, and rental tractors.

Educational Level

 

Educational achievement is usually assumed to be
related to individual innovativeness. A rationale for this
relationship, as hypothesized in this study, is that the
skills concomittant with increasing levels of formal
education enable the better educated individuals to formulate
realistic expectations about new methods more rapidly than
uneducated and poorly educated farmers. Evidence presented
in Chapter IV seems to support this hypothesis. It was found
that the better educated farmers were more likely to know
about fertilizer and how to improve crop yields, and were
more inclined to use "direct" sources of information than
their less educated colleagues. In order to provide
continuity with the earlier analysis, the same classifications
of educational achievement are utilized in the A.1.D. program.

Information Sources-I

“a“...

 

Information_§ourge§;l_is a variable based upon the

 

farmer's re-ponse to the survey question asking him to
indicate the source from which he had learned about the
latest method he was employing (relative, neighbor,
agronomist, mass media, etc.). Since the question ignores

the possibility that information might have been received

by t
the

most
aidi
usur.
Give
RHOR

Infc

243
by the farmer from several sources, it must be assumed that
the respondent's choice represents the source that had the
most important impact upon his decision—making process. In
addition, it must be assumed that the source indicated is
usually of considerable importance to the respondent.

Given these assumptions, information sources-I was included

 

among the variables used in the A.1.D. program,

Information Sources-II

 

Information sources~II involves the individual farmer‘s

 

exposure to extension meetings (meetings where agronomists
had talked about new methods), farm magazines, and radio.
Exposure to each of these sources of information were found
to be associated with the use of new agricultural practices
in Chapter IV. For the purposes of the A.1.D. program, a
variable was constructed which identifies the use of each of
these sources and combinations of their usage. The variable
was composed of the following classifications:

Extension meeting, farm magazine, radio

Extension meeting, farm magazine

Extension meeting, radio

Extension meeting

Farm magazine, radio

Farm magazine

Radio
None

OOVO‘Ul-fihLNNl-l
0

As initially conceived, this variable would permit the
program to "split" on the basis of the most important of
these three sources. Alternatively, it could "split" on
the basis of the combinations of information sources which
are most associated with new practices. In the final

computer run, however, this variable was constrained in

244
such a manner as to consider only the second alternative.
Small sample numbers related to some classifications and
little indication of any one source being independently
superior to others resulted in the decision to constrain
the variable in the final computations.

Attitude Variables

 

The farmers' responses to certain attitudinal
questions (fatalism/non-fatalism, luck/business, attitude
toward technicians, and willingness to defer income) were
included in the A.1.D. analysis. The variables were not
expected to become involved in the early "splits" performed
by the program, but it was believed that they could be
involved after other variables had been taken into
consideration. The farmer's interpretation of his neighbor's
views were included as variables in the initial computer
run, but omitted in the final computations. The reason for
omitting these variables was to provide for their
substitution by more informative variables in the computer
output. (After two intermediate "splits," Sfio Francisco
farmers who believed their neighbors wouldn't like to see
them get ahead were found to use tractors more often than
those who believed their neighbors would not be disturbed

by their individual progress.)

Summary of the Variables Included

o“----*

in the A.1.D. frogram

 

 

The following listing summarizes the variables and the

classifications within variables which were utilized in the

 

Mi).

Te

245
A.1.D. program:

Tenure: (Free)
1. Landlord
2. Owner-Operator
3. Tenant

Income (annually): (Monotonic)
l. < $50 (per capita family)
2.550-599
3. $100 - $199
4. $200 - $499
5. $500 and above

Credit (Free)
1. Didn't use credit
2 Obtained credit from non-commercial sources
3. Obtained credit from credit cooperative
4 Obtained credit from banks
5 Didn't need additional credit (borrowed from
one of the above)

Attitude Toward Trying New Practices ("Try Second"):
(Free)

"When new agricultural products are offered it is
better to wait and see what happens when they are
used by others.” (Question from MSU/SUDEHE Farm
Survey.)

1. Agree

2. Disagree

 

Risk Preference (see p. 179): (Monotonic)
1. Very low utility
2. Low utility
3. Medium utility
4. High utility

Education

1. None
2. One year
3. 2 - 3 years
4. 4 - 5 years
5. 6 - 12 years
6. More than 12 years
Input Availability: (Monotonic)

A.1.D.-I (Fertilizer--Sio Francisco Rice Area)
1. Fertilizer available locally
2. Fertilizer not available locally
.I.D.-II (Insecticides--Sao Francisco Rice Area)
1. Insecticides available locally
2. Insecticides not available locally

A

246

A.1.D.-III (Tractors-~Sao Francisco Rice Area)

1. Rental tractors available locally

2. Rental tractors not available locally
A.1.D.-IV G A.1.D.-V (Innovation Index--Sao
Francisco Rice Area and Combined Bean Areas)

1. Fertilizer, insecticides, rental tractors ggt_

available locally

2. One of the above available locally

3. Two of the above available locally

4. All of the above available locally

Information Source-I (Free)
1. Relative
2. Neighbor
3. Agronomist
4. Businessman
5. People from a different place
6. Mass media
7. Another source

Information Source-II (Monotonic)
1. Extension meeting, farm magazine, radio
2. Extension meeting, farm magazine
3. Extension meeting, radio
4. Extension meeting
5. Farm magazine, radio
6. Farm magazine
7. Radio
8. None
Fatalism: (Monotonic)
"Nowadays the farmer alone cannot do much to improve
his life." (Question from MSU/SUDENE Farm Survey.)

1. Agree
2. Don't know
3. Disagree

Luck/Knowledge: (Monotonic)
"To make more money it is better to know how to do
business than be lucky.” (Question from MSU/SUDENE
Farm Survey.)

1. Agree

2. Don't know

3. Disagree

Attitude Toward New Technology: (Monotonic)
"We would be in a better situation if the technicians
left things as they are." (Question from MSU/SUDENE
Farm Survey.)

1. Agree

2. Don't know

3. Disagree

indic
varia
catic
thei:
giver

247

Deferred Income: (Monotonic)
"If someone has to choose, it's better to receive
NCr $90.00 one year from now than NCr $30.00 today."
(Question from MSU/SUDENE Farm Survey.)

1. Agree

2. Don't know

3. Disagree

Note that either the word “free" or "monotonic" is
indicated in parentheses across from the variable. A "free"
variable means that the computer can rearrange the classifi-
cations in descending order on the basis of the value of
their means. ”Honotonic" instructs the computer to perform
the A.1.D. operations with the classifications ordered as
given by the researcher.

A.1.D.-I: Fertilizer Use bysao Francisco
Rice Farmers

 

 

Nearly all of the farms using chemical fertilizer were
from within the 830 Francisco Rice Area. Still, even
within this area the percentage of farmers who adopted this
practice was very small -- 11.8%. The use of fertilizer
by these farmers was examined with the aid of the Automatic
Interaction Detection program and the results are presented
by Figure 14.

According to the program, the most important predictor
of fertilizer usage was the farmer's identification of the
agronomist as a source of information. 0f the 14 She
Francisco farmers who had learned their latest method from
the agronomist, 42.9% were using chemical fertilizer. This
figure is contrasted with the 7.6% of those farmers indicating
other sources of information. In all, 12.4% of the variance
in the sample was explained by this initial "split." The
second choice predictor, which would have explained 10.0%

of the total variance, was education. Income and credit

2L48

acofluu>uomno mo honenx n z
muoawflwuuom
mama: unannonun u H

 

 

 

 

 

 

 

 

 

a~ . z
1 .~.a . w .
aN..o 2
Has Human emu“ xx
x
// om I 2
I, «o.- g n
A ac.“ agsouov
\\ “my "panama
\
\
- . 2 \\
.~.n~ . w \
A..nv
Ace "amen xmax

 

 

 

o a z

«n.mn u h
an nacho“

new "sneeze

 

 

 

.mhoflh

a» ouflm oumwucaam 0mm up on: nouflauuuom "

 

a” . z
”0.0 . H
:1.:

"zofihtwnau

   

 

 

 

 

“my

om . z
”m.a~ . H
“as H Au

”20_h4unn,

 

 

H-.a.~.< -. a” auauaa

 

     

 

mo" 3 z
no.5 n H

Physio.

oi»

“mg u~-cazu

 

ad" . z
.a.- u w
"as u<~op

 

 

 

 

 

 

 

«a a z
,gm.~w I P
numwnocouu<v
mug “H.0nZM

 

 

249

variables were considerably less helpful at this early
stage, in explaining the adoption of fertilizer (see
Appendix A).

Educational achievement was assigned primary importance
in the second "branching" of the "tree" (dividing Group #3
into Groups #4 and #5). The "split" occurs between
farmers who had less than two years formal education
(n=49) and farmers with at least two years of schooling
(n=56); and adds an additional 4.3% to the explained
variance (RZ=.1672). All too clearly, the most poorly
educated segment of the Recife farmers were not among the
innovators with respect to fertilizer usage. One might
expect that low levels of income would be a close
substitute for educational disadvantage at this juncture,
yet its influence appears to be somewhat limited. While
7.2% of the variance within Group #3 is explained by
education, only 3.1% would have been explained by a division
of farmers into groups with less than $50 per capita family
income and those with more. Instead, the individual's
exposure to extension meetings, farm magazines, and radio
("information sources-II") seems to be nearly as important
as his educational achievement in explaining the adoption
of fertilizer (explaining 6.4% of the parent group variance
as compared with 7.2% for education). Eleven farmers who
were exposed to the combinations of extension meetings, farm
magazines, and radio or extension meetings and farm

magazines would be separated from the remaining 94 farmers.

250
Among the better informed farmers, 27.3% were using fertilizer
as compared to 5.3% of the larger group.

Among the Recife farmers who had completed more than one
year of education (except for those who had indicated the
agronomist as an information source), those few (n=6) who
had obtained "sufficient credit to meet their needs" were
more likely to use chemical fertilizers than the remaining
farmers -- 33.3% as opposed to 12.0%. These six farmers
represented two-thirds of the farmers who believed that
they "didn't need” additional credit, the remaining third
was included among those who indicated the agronomist as an
important information source.

Turning attention to those farmers who were apparently
less fortunate in obtaining credit, we find the ordering of
the credit classifications as would be expected, but with
mean values that are not greatly different. Among the non-
borrowers, 9.5% were found to be using fertilizer as compared
to 10.0% for those borrowing from non-commercial sources
(neighbors, relatives, businessmen, buyers and landowners).

A somewhat greater percentage -- 15.8% -- of those borrowing
from banks were employing this innovation. Despite the fact
that these findings are rather satisfying in terms of the
proposed theory, they should be interpreted cautiously.

It should be kept in mind that the parent group (Group #5)
consisted of only 56 farmers and that Group #6 -« those
‘With "sufficient" credit contains only six observations. In

general, we need to be more cautious about our findings as

251
we move outward through the "tree."

Differences in the willingness to try new methods ("Try
second") and variations in exposure to extension meetings,
farm magazines and radio ("information SOUTCCS‘II") were
nearly as useful as credit in explaining the variance within
Group #5. Of those 11 farmers willing to try new methods
without benefit of observing other peoples‘ experiences,
27.3% were using fertilizers. This is compared to 11.1% of
the farmers (n=45) who preferred the alternative ~- "wait
and see" -- strategy. Once again, we find those farmers
exposed to both extension meetings and farm magazines (n=ll)
more inclined to use fertilizer than the remaining farmers --
27.3% as contrasted with 11.1%. Finally, income level was
almost as important a variable and would have separated
out the 9 farmers who had per capita family incomes of less
than $50 and did not use fertilizer at all.

In all, the first three "splits” are capable of
"explaining" 18.7% of the variance (R2=.1869) in fertilizer
use among 350 Francisco rice farmers. Two additional
”splits" were made by the A.I.D. program, but they were of
doubtful value and were ”pruned" from the tree. An
alternative ”split” of Group #7 is indicated by dashed
lines, dividing these farmers into groups with "high" or
"moderate" utility for increased income (as measured by the
"risk preference” variable) and "low" or "very low" utility
for additional income. Such a "split" would have provided

an additional explained variance of 1.2% (R2=.1991), but

give:
”pruw

inter

factt
fert.
thou,

they

like

amen

 

252
given the fact the irrational "splits" were already
"pruned" at this point it should be given a very cautious
interpretation.

In interpreting the A.1.D.—l output, it appears that
factors related to the development of expectations about
fertilizer are critically important to its adoption. Even
though 88.2% of the 850 Francisco farmers indicated that
they knew what fertilizer was, those who use the agronomist
as an important source of information were the farmers most
likely to adopt the technology. The most poorly educated
among the Recife farmers (one year or less of schooling)
failed to have any of its members using fertilizer. Finally,
the farmer's exposure to extension meetings and farm
magazines repeatedly emerged as a "second best" explanation
of fertilizer usage.

Secondly, the availability of "sufficient" credit
resources seemed to improve the farmer's chances of using
fertilizer. In addition, there is some reason to believe
that the willingness to take risks and higher utilities for
increased income may be related to innovativeness with
respect to fertilizer. These conclusions, however. should
not be strongly supported on the basis of the present

information.

'A.I.D.‘II: Insecticide Use Among

--‘e.’

Sao Franfiisce Rice Farmers

 

 

Unlike fertilizer, the use of insecticides is reasonably

widespread among 850 Francisco rice farmers. And, as

253
exhibited in Figure 15, a configuration of variables
different from those associated with fertilizer usage seems
to be related to the use of this innovation. The avail-
ability of adequate credit resources appears to be particularly
important to insecticide usage. Of those farmers who reported
bank loans or indicated they had obtained "sufficient"
amounts of credit, 88.8% were using fertilizer. Only 55.4%
of their counterparts who either had not borrowed or had
obtained loans from non—commercial sources were utilizing
this practice.

The division of the 550 Francisco farmers on the basis
of credit usage resulted in the explanation of 13.4% of the
variance in insecticide use. A partition of these farmers
on the basis of tenure would have been nearly as useful.
Landlords and tenants would have composed one group (with
84.4% using insecticides) and the second would have consisted
of owner-Operators (of whom 52.7% were using insecticides).
This relationship, between tenure and insecticide use, was
encountered earlier in Chapter V. At that time it was
suggested that landlords were generally better educated and
had greater financial resources than owner-operators, and
that tenants were probably either furnished this input by
their landlords or required to use insecticides.

Even among those farmers who had not obtained bank
loans or didn't use credit at all, the better educated
individuals were most likely to use insecticides. While

86.4% of these farmers who had completed at least four

.2594

uncauu>uomno «o unsanz r z
moemowuuencw
mafia: omeucnouem a H

 

 

 

 

 

 

 

 

.mueauaw ouwm oumwucaum 0mm kn van avawuuomcM

 

 

 

 

 

 

 

 

as. "egazuw

 

 

 

n

 

o a 2
an.»» . w
an» M vV

Nag "zoah<u:ce

 

 

 

 

mN I z
poo” o H

. Aux a av
pnau "ze~h<u:om

 

 

 

 

ad . z
am.Oa . »
”5.”.n.a ma ates
has "H.oazu on . z # ma . z
no. on . am.om . »
“waldo, hats m we
a" u z adv z:e new "zoua4onnm
.o.~m . »
an.~ unnouov
any “maouwzm b I 7. a Nu .- ...
. ah.mw . » as.om . w
nuancosv Am»; m AV

nay nzoew<swcu

 

V

 

 

 

Ch

~ I 2

f0

\
ii

on. k

hood“ vV
“Nay season”

 

 

 

 

. z
.H.wa . »

Rona» .u
“Had "wzouzn

VI
I

 

 

 

 

 

so . 2
...mm . »
nu. a nose»: ov
“me .puntae

 

 

 

 

em . z

«~.aw . H
hm.v museum.
flue ":Hommo

 

 

“an-.o.~.< .-ma ouauum

 

mg“ n z
.o.oa . »
sakes

 

 

255

years of education were using insecticides, the figure is
only 39.5% for farmers with three years or less of schooling.
In an identical comparison among the 850 Francisco rice
farmers who had either bank loans or credit "sufficient” for
their needs, it was found that 100.0% of those with more than
three years of education were using insecticides as
contrasted to 83.3% for farmers with less than three years of
schooling. This seems to suggest that improved credit
arrangements might have considerably increased the number
of farmers in Groups #4 and #5 who were using fertilizer.

Focusing again upon education, the partition of Group
#3 between farmers with three or less years of schooling and
those with more explained 12.9% of the total variance in
insecticide use. The next best predictor would have been
the sources of information given by the respondents,
explaining 9.5% of sample variance. Interestingly enough,
the farmers of Group #3 (non-borrowers and loans from non-
commercial souices) would have been subdivided into a group
containing those who had learned about their latest method
from neighbors and businessmen (n=34) and those who had
utilized other sources (n=31). Only 4 members of the second
group had indicated the agronomist as a source of informa-
tion and twc had given the mass media as a response. Over
two-thirds of the latter group had indicated either relatives
or another (undefined) source of information. In all
probability, such a split would have indicated that neighbors

were a better source of information than relatives and

256

another source. In fact, neighbors may have been a
reasonably good source of information about insecticides
given the relatively widespread use of this practice. This
might be contrasted with the considerable importance placed
upon the agronomist -- a ”direct" source of information --
by A.1.D.-I as the program attempted to explain the use of
fertilizer -- a practice which was not widely used. One
might generalize that as the number of farmers using a
particular practice increases, the importance of "direct
sources" of information in the adoption process decreases.

The third partition created by A.1.D.-II indicates that
the poorly educated tenant is more likely to adopt insecticides
than the landowners who are in the same circumstance with
respect to education and credit. Again, this is probably
because the decision of whether or not to use insectidies
was "handed down” from their landlords. Note that 85.7%
of those few tenants remaining in the parent group (Group
#4) were using insecticides, while only 30.6% of their
landowner counterparts had adopted this practice. We should
remember, however, that we are reaching the point in terms of
sample numbers that partitions should be interpreted more
cautiously.

This particular "split" explained approximately 7.2%
of the variance in the use of insecticides. A division on
the basis of information sources, similar to the one
discussed above, would have been nearly as useful. A third

alternative, based upon the farmer's choice between luck

257
or knowledge as the best means of attaining larger incomes,
would have explained a considerable amount of the variance
in Group #4 (< 3 years of education). According to the
program, 54.5% of those who favored knowledge were using
insecticides as opposed to 25.0% of the farmers favoring
luck.

The final ”split” on the lower branch was based upon
information sources, dividing farmers into a group who
indicated neighbors or businessmen as important sources and
one composed of the remaining sources. This is not a very
informative partition since it largely reflects the
difference between neighbors and another (unidentified)
source of information. "Pruning" this partition from the
”tree" would reduce the explained variance by 5.5%. An
alternative would be to "split” this same parent group on
the basis of the farmer‘s reaction to luck as opposed to
knowledge as a means of earning higher incomes. The
percentage of adopters among those favoring knowledge was
44.4% as compared to 16.6% for farmers favoring luck. Such
a partition would have explained 5.8% of the variance in
insecticide use. It is interesting to observe that after
adjustments have been made for the interaction of other
variables, an attitudinal variable emerges as potentially
useful in explaining the adoption of insecticides.

Turning our attention to those among the 850
Francisco rice farmers who had either received loans from

banks or obtained "sufficient” credit to meet their needs,

258
we find that the lower income individuals were considerably
less likely to use insecticides. While 73.7% of the farmers
with less than $100 per capita income used this practice,
nearly all -- 97.1% -- of those with higher incomes had
adopted insecticides. There is some indication, given the
reasonably favorable credit situation of these farmers, that
the "willingness to accept the risks of new methods" is
involved in this partition. The fact that these farmers had
either obtained bank loans (generally larger amounts than
from other sources and with government-subsidized interest
rates associated with most of them) or indicated that they
"didn't have any use” for additional credit suggests that
the ability to finance inputs from their own resources is
of diminished importance in interpreting this particular
"Split.” In addition, those variables which were found to
be associated with income levels -- educational achievement
and information sources -- were not nearly as useful in
explaining the variance within the parent group (see

BS ./TSS. for Group #2 in Appendix B). In all, the evidence

1 1

(I)

seems to suggest that the "dimension" of the income variable
involving the lower income farmer's reluctance to "accept
the risk" of new methods may have been prominent in this
partition.

This particular "split," based upon income levels,
explained 2.7% of the variance in the use of insecticides.
A partition on the basis of tenure would have separated

the 6 tenants and 27 landlords from 21 owner~0perators, and

259
explained 2.2% of the variance. Tenure, in this case,
appears to be a reasonably good substitute for the income
variable (as landlords tend to own larger farms and have
higher incomes than owner-operators).

Continuing with the larger income farmers (above $100
per capita), we find that only among those without any
formal education (n=6) were there farmers who were not
using insecticides. Unfortunately, the size of this group
and the amount of variance explained by this partition -—
0.5% of the total variance -- make one hesitate to draw
strong conclusions. Nonetheless, it is interesting that,
among farmers who were apparently able to withstand the
“risk" of innovations and probably able to finance these
inputs with credit, the few who had not adopted
insecticides had also never attended school.

In summarizing A.I.D.-II, it was found that four
interacting variables -- credit, education, income, and
tenure -— were capable of explaining 36.8% of the variance
in insecticide usage. (The total "tree," as presented in
Figure 15, resulted in a R2 of .4336.) Credit seemed to be
the most important predictor of insecticide use, accounting
for one-third of the explained variance. The likelihood of
a 850 Francisco rice farmer using insecticides was enhanced
by access to either bank loans or "sufficient" amounts of
credit (from whatever source) and inhibited by failure to
use credit or a dependency upon non-commercial sources for

loans. Which suggests that an improvement in the availability

260
and use of credit would have increased the use of insecticides
in this area. Better educated farmers were inclined to use
insecticides more frequently than their poorer educated
colleagues. All of those farmers possessing a combination
of ”sufficient" credit or bank loans, income levels sufficient
to "accept the risk" of new methods and some formal education
were found to be using insecticides. Farmers with more than
three years of education, even though they had either not
borrowed or had obtained their loans from non-commercial
sources, were twice as likely to use insecticides (86.4%) as
those in the same circumstance but with less education (39.5%).
Among those poorly educated farmers who had either not
obtained credit or had received loans from non-commercial
sources, the tenants were much more likely to use insecticides
than their landowning counterparts. his, once again,
suggests that the landlord had influenced the tenant's
decision.

A.1.D.—III: The Use of Tractors by
850 Francisco Rice Farmers

 

 

Slightly over half (51.3%) of the Sao Francisco rice
farmers were using tractors. The availability of rental
tractors makes a considerable contribution to the large
proportion of farmers using this practice. A.1.D.-III
(given Figure 16) identifies the local availability of
tractors for rent (as perceived by farmers) as the most
important variable explaining their use. This variable

alone accounts for 12.3% of the variation in tractor usage.

263]

 

acoflun>uomn0 mo noneaz n z
mucuuenu
«can: nwnucouuoa n 9
NM 3 z
fio.o u H
Rammed
Any "song as mung

 

 

.nuoeuam oowu

 

 

 

ow
vv.fl~

 

I z
n H

Rousonnmcdc
Au“ "zOZx a; zone.

 

 

 

ouawucaum 0mm xn nuouoanh we «no one

 

 

 

 

 

an -

«o.m~ u

 

2
H

ancoxocnoa,
an". ”panngo

 

 

 

 

 

 

~H u
“p.00 u

 

Is
o!

'

Amuorouuom.cozw
«do

a . .
pH“ .FHCV

 

 

 

 

 

so u z

am.mn . »
.Hua>< uozv

HA~<>< puaza

 

 

 

 

 

 

 

 

 

 

 

 

"may-.a.a.< --oH «sauna

 

 

maa a z
om.mn a H
A<h0h

 

 

 

 

 

 

,
I
a
a
a
./ ea . 2 an . 2
EL 3.: . H 3.2 .. H
n5.m...fiv “can” vv
. Fwy H.0cra may "urcozn
a
mm . a" “mg
. m . 2
.Q N” . H am.am . »
A~.n.ow Room“ V
m u 2 on a a
3.3 . w nu 3:9. . H
mom“ V“ \ hmuk a V“
an“ “grooza s and "zocs4He.u
c
a
a
c
an . z .x N" . z
ah.mo a.» V «o.co~ n.»
flown av "was a do
has "deadening “a” "zocp<oaau

 

 

 

 

  

    

Nm . z
.»~.~a . n
nonnaaaa><c
. < saga”

 

 
 

  
     

262
Among those who found rental tractors available locally,
71.2% were utilizing this method of tillage. This proportion
is contrasted with 35.5% who did not believe they could rent
a tractor in their municipio. The importance of the avail-
ability of rental tractors is underscored by the fact that
31.0% of the poorest farmers in the 850 Francisco sample were
using tractors.

Nearly as large an explanation of the variance in
tractor usage (in the initial partition) could have been
achieved using income levels or educational accomplishment.
Individuals earning per capita family incomes of $500 or
more (83.3% using tractors) would have been separated from
the remainder of the farmers (43.2% using tractors),
explaining 10.4% of the variance. In terms of education,
those with no formal education (27.8% using tractors) would
have been set apart from the rest (61.4% using tractors).

It is interesting to note that a division of income,
in the form suggested above, is the basis of the next most
important ”split." Even when they didn't find rental
tractors available locally, the wealthiest farmers ($500 or
more per capita family income) were still very inclined to
be using tractors (87.5%). Either they owned tractors
'themselves or rented them from sources outside the immediate
nounicipio. (It should be noted that the question upon
Which this variable is based asks if tractors can be rented
"from suppliers in the municipio where you live?" It

might have been possible to rent tractors from more distant

263
sources.) Farmers with incomes below this amount, and

without locally available tractors to rent, were largely

 

inclined to use other means for tilling their land «~ as
only 28.8% were using tractors. Thus, once we have adjusted
for the availability of rental tractors, the financial
capability of the farmer is very much related to whether or
not he used this means of power. An interesting sidelight
is that income levels and the belief that rental tractors
were available locally were found to be significantly
associated at the .005 level (by chi-square analysis).
The reason for this relationship is open to speculation.
Perhaps the relatively wealthier farmers were less isolated
and more aware of the availability of such inputs throughout
the municipio. Or perhaps the wealthier and poorer farmers
were clustered with rental tractors being more readily
available in the vicinity of the wealthier farmers.

This partition, dividing between the highest level of
income and all lower levels, explained an additional 8.2%
of the variance in tractor use. (Approximately 20.5% of the
total variance is explained by input availability and income
levels alone.) The second most important variable at this
point was the farmer's exposure to extension meetings, farm
Inagazines, and radio. Those farmers who were exposed to at
1_east two of these channels of communication adopted tractors
more often (68.8%) than the other members of the group
(25.5%). Educational achievement was also a reasonably good

predictor at this juncture, with a division suggested between

('1‘

P—Ij

FF)

264
these with more than three years of education (42.9% using
tractors) and farmers with three or less years of schooling
(32.7% using tractors). This alternative bears mentioning
since exactly this ”split” was found in A.1.D.-II in the
same position in the tree, although it followed credit usage
rather than input availability.

According to the A.1.D. program, sources of information
best explain the variance in tractor usage among the lower
income farmers. Although the group whose members were more
likely to be using tractors includes farmers who looked to
the mass media (n=1) and agronomists (n=4) as sOurces of
information, the partition is laroely a division between
neighbors as an information source and "another source"
(which is left unidentified in the survey data). As such,
it adds little to the information available about the
farmer's decision-making regarding tractors. An alternative
variable would he the farmer's exposure to extension meetings,
farm magazines, and radio. Again, we find that the
individuals exposed to two or more of these channels of
communication were more likely to be using tractors (54.5%
with n=1l) than their colleagues with less exposure (22.9%
‘with n=48). While the partition selected by the computer
accounted for 4.6% of the variation, the alternative
:suggested above would have explained 3.1% of the total
variance in tractor usage. Although less important
Statistically, the alternative partition would seem to be

more informative. It suggests that the farmers with a

26S

greater exposure to modern (or perhaps more "direct")
channels of information were more innovative than others.

Returning to the farmers who indicated that they could
rent tractors locally, we find that all of those individuals
with six or more years of education were using tractors.
Among the corresponding group, with less than six years of
schooling, only 62.5% of the farmers had employed tractors
in tilling their land. This represents a partition of the
educational variable at a somewhat higher level than had
been observed in A.1.D.-I and A.1.D.-II. It may be that
even the most rcorly educated farmers are aware of the
obvious technical advantages of tractors, while the economic
advantage is best understood by the better educated

0

individuals. Whatever the reasons are, 4.4% of the total
variance is explained by this “split" and there were no
substitutes suggested by the program.

An additional 0.8% of the variation in tractor usage
is accomplished by dividing the farmers with less than six
years of education on the basis of income levels. This
partition would group together those farmers with $50
or more per capita income (n=35) and those with smaller
earnings. Only 40.0% of the members of the latter group
”were found to be using tractors, while 65.7% of those
:farmers with at least a $50 per capita family income had
adopted this innovation. Since there is probably very
little "risk" associated with trying tractors, a difference

on the basis of income levels would undoubtedly reflect the

Cd

01'

IO

ha

St

SID

Fi

of

ha

aC

9X

We

266

capacity of the farmer to pay the rental fee. Yet, whether
or not a real difference is given by this partition is open
to question. Other "splits" at this location in the "tree"
have already been "pruned" because they were unsatisfactory
statistically, and this partition itself involves a very
small number of observations in the lower income groups.
(Even though it represents five-sixths of the farmers at
the less than $50 level who also believed tractors were
available locally.)

The remaining partitions (yet to be discussed) in
Figure 16 are perhaps even more tenuous. Although a division
of Group #6 into sub-groups of non-borrowers and borrowers
has some claim to statistical validity, it makes little
sense theoretically. It is difficult to conceive of a
reason why those farmers who hadn't used credit would be more
likely to either own or rent tractors than those who had
obtained loans, especially when the wealthiest of the
farmers had been separated by an earlier "split."
Statistically, this partition explained 3.7% of the total
variance. The best alternative -— based upon exposure to
extension meetings, farm magazines, and radio -- would have
accounted for 2.5% of the variance. Farmers who were
exposed to two or more of these channels of information
Wwould have been placed in one group (n=7, with 71.4%
using tractors), with a second group b‘ing composed of
those with less exposure (n=26, with 34.6% using tractors).

IJnfortunately, this alternative is extremely questionable

267
since the three classifications composing the group with
greatest exposure to modern channels of information contain
a total of only seven observations.

A partition of GrOUp #7, which would have been
satisfactory from the theoretical standpoint, was "pruned"
from the "tree" for statistical reasons. This particular
"split," also based upon credit, would have separated one
individual with "sufficient" credit, and who also had used
a tractor, from the others. A more acceptable alternative,
based upon the farmer's evaluation of luck as opposed to
business knowledge as the best means of "getting ahead,"
has been substituted into Figure 16. None of the farmers
who favored luck were using tractors, while 21.4% of those
favoring business knowledge had adopted this practice. The
original "split" would have accounted for 2.7% of the
variance, as contrasted with the 1.0% increase in explained
variance contributed by the substitute partition.

Considering only the first three branches of the "tree"
(Groups #1,2,3,4,5,8,9), A.1.D.-IlI accounts for 24.8% of
the variance in the 850 Francisco farmers' use of tractors.
This portion of the "tree" largely reflects the farmer's
ability to secure the services of a tractor. Where rental
tractors were not available locally, it was quite unlikely
that farmers would be using this source of power unless
they were among the few with fairly substantial income (over
$500 per capita). Only 28.8% of those farmers with less

than $500 per capita family income, and who believed rental

268
tractors were not available locally, were using this
innovation. Yet, 66.6% of this same income grouping who
indicated that they could rent tractors locally were found
to be employing this practice. This seems to indicate that
an increase in the availability of rental tractors would
have expanded their use.

Even though access to the services of a tractor seems
to be the most critical factor in tractor usage, certain
variables related to expectation formation were included
in the "tree." Even among farmers who found tractors
available locally, those with higher levels of education
(6 or more years) used them more frequently (Groups #8 and
#9). And when we substitute exposure to extension meetings,
farm magazines, and radio for the uninformative "split" of
"information sources-I," we find that the individuals with
greater exposure to these channels of communication
(utilizing two or more of them) were more likely to be
using tractors. Including this alternative partition,

A.1.D.—III accounted for 27.9% of the variance in tractor

usage.
Finally, if we accept the possibility -- as suggested
in Figure 16 -- that farmers with per capita family incomes

of less than $50 are relatively less able to acquire the
:services of tractors for hire, A.I.D.«III would account for
28.8% of the variance in tractor usage among the 550

Francisco rice farmers.

A.1.D.«IV: An .\nalvsis of Innovativeness
Kmong Sao lranc1sco RlC larmers

 

 

 

A.1.D.*IV outlines the relationship of certain
variables to the general innovativeness of Sao Francisco
rice farmers. The range of the innovation index, used as
a measure of the farmer's innovativeness, varied from O to
100. The average index number was 35. %. For farmers with
more than one year of schooling the average index value
increased to 42.6%, and dropped to 25.0% for those with
less education. This particular partition, which accounted
for 17.5% of the variance in innovativeness, suggests that
the better educated farmers were at a considerable advantage
in formulating their expectations toward new technologies.

Other variables would also have been capable of explain-
ing sizeable amounts of the variation in innovativeness. A
"split" on the basis of credit, separating those who had
"sufficient" credit or had received bank loans from farmers
who either didn' t borrow or obtained loans from non-
commercial sources, would have explained approximately
14.2% of the variance. Exposure to extension meetings,
farm magazines, and radio would have accounted for 13.2%
of the variance; separating those who were exposed to
extension meetings and one or more of the other sources
:irom the rcma ining farmers. A division of the sample into
groups of farmers with $50 or more per capita income and
farmers with lower levels of income would have explained
12.8% of the total variance.

The second partition di v1d ed the better educated group

270

mnoHuu>uonno we nunuaz I z

ceauuu>ov vuuwcaum I u
once» nova“ :502 I H

 

a“ I z
a.o~ I a
an.aa I »
ha.H.
fi«~u "aaea xm_a

 

.muoE»am ovum ounwoauum 0mm unas< unoco>muu>ona~ ”>H-.n.~.< -- an ounuwm

 

 

Mm I z

 

 

 

 

MN I z
..NH I o
.H.m~ I »
h¢.m.~.mv
fi-c H.oaza

 

 

 

 

 

 

 

 

a I z
0.x” I a

mafia ~.0ezH

 

 

 

.
0Q!“
Pr'sMr-lo-d

paw

c..,
VI If

asz II

L.)

m F'022

L)

r-s
W

 

 

 

 

 

 

 

 

 

 

 

 

an I z

w.afl I o

n..mm I H
noon” vv

Rev ”aroozH

 

 

 

 

 

 

,
“may

 

n.
v.

uuwxm x

mmnH

23“.!

Q.

Will

ca Dataz

v

Hm

 

 

 

ma I z

 

~.~H I o

<

 

.o.mm I
Pu.ac
haw ”Facaao

 

an I z

~.c~ I o

.o.nq I H
803 .3

“my "mzoozH

am I 2
..0a I o
.o.mN I w
E a ma
“NV "zo~»<unam
. ad" I z
m.o~ I o
.~.mn I »
q<aop
we . z
a h.o~ I o
.o.~a I H
at» g Ac
any rzoah<onnu

 

 

 

 

 

 

a“ I z
a.a~ . o
.~.am . H
Am.vU
new "hanwau

 

 

271
on the basis of income, resulting in a mean index number
of 49.6% for those with over $200 per capita family income
and 35.7% for farmers with incomes below this level. An
additio.al 6.2% of the total variance was explained by this
"split." It would appear, given the importance of credit
to the innovative behavior of the higher income group (as
indicated in the next partition), that the reluctance of
lower income farmers to accept the "risks" inherent in new
methods was an important factor in this partition. If the
ability to independently finance new inputs had been the
critical advantage of higher incomes, it would seem unlikely
that credit usage would have had such a significant impact
upon the farmer's innovativeness.

It should be indicated, however, that a partition of
the better educated farmers on the basis of income was
somewhat equivalent to dividing the group into those who
had used the agronomist as a source of information and those
who had not. Ten of the fourteen Sao Francisco rice farmers
‘who had indicated that the agronomist was important to them
as a source of information about new methods are to be
found within the thirty-four members of Group #5. Thus,
tht: advantages possessed by this group are not only related
to their level of income, but probably also to their access

to "direct" sources of information. In fact, a,partition on
the basis of use/nonsuse of the agronomist as a source of
infOrmation was one of the better alternatives of income

leV€>ls (exceeded only by a division on the basis of credit

equ

thc
im
f0:
fa:

(1)

272

equivalent to that given in Groups #6 and #7).

Among the better educated, higher income farmers;
those with either "sufficient" credit or bank loans were more
innovative than others. The average index of innovativeness
for these farmers was 58.1%, as compared to 35.6% for
farmers who either didn't borrow or secured their credit
from non-commercial sources. This particular "split"
contributed 7.8% to the variance explained by the total
"tree."

An additional 6.3% was explained by the credit
variable in its application to the poorly educated farmers
(for a total of 14.1% of the explained variance attributed
to this variable). The sec Francisco farmers who had com-
pleted less than two years of schooling were divided into
those who had received bank loans (n=17) -~ with mean index
of 36.3% -- and those who either didn't borrow (n=22),
borrowed from non-commercial sources (n=9), or believed
they had obtained "sufficient" credit (n=3) -— with an
average index of 19.3%. (The latter classification contains
too few observations to be of much practical importance.)
(Dne might well conclude from this evidence, that even among
the poorly educated farmers a more adequate source of credit.
alilows individuals to be more innovative.

The best alternative to credit, among the poorly
edHeated farmers, would be a partition on the basis of the
avadilability of inputs -- specifically insecticides, chemical

faritilizers and rental tractors. Such a "split" would have

273
placed in one group the farmers who believed that one or
two of the important inputs were available locally and in
another those farmers who did not believe any of these
inputs could be acquired locally. It should be noted that

none of the farmers with less than one year of education

thought that all three inputs -- tractors for rent,
insecticides and chemical fertilizer -- could be obtained

locally (while 8.8% of the better educated farmers believed
this to be the case).
Two of the remaining partitions, based upon information

sources, largely reflect the difference between neighbors

as a source of information and the unidentified classification
called "another source.” Although these partitions are
statistically valid, they are not at all informative.

Before we examine some acceptable alternatives to these
partitions, it is interesting to examine the remaining
partition based upon risk preference. This partition may
have some claim to validity, despite the fact that it is
preceded by an uninformative partition. Among those farmers
vvith more than a year of schooling and less than $200 per
Cardta family income, one might postulate that a few
inaiividuals would be more willing than others to accept
celftain risks in order to secure a larger income. Setting
aside those with the poorest sources of information, this
aPIVCars to be the case. Eleven farmers with somewhat more
”ti7lity for increased incomes, as measured by the risk

p1°~iaerence variable, had -— on the average -- adopted a

274
greater percentage of the new methods (mean innovation
index of 48.1%) than their less venturesome counterparts
(mean index of 36.3%). The problem with accepting this as
strong evidence of the importance of variations in the
motivation for economic gain, is that we have reached the
very minimum acceptable size of a subsample to allow another
partition (n=25). The most that may be said is that there

is some vidence that differences in the utility for

 

increased income has an effect on the individual's
innovativeness.

Returning to the partitions based upon the ”information
sources-I" variable, we find that among the poorly educated
farmers the availability of inputs is a reasonably good
alternative. While the ”split” selected by the computer
would have explained 3.6% of the variance in innovatives,
the availability of inputs could account for 3.1% of this
variation. A partition of this type would have grouped
together those farmers who believed one or two of the
hnportant inputs -- rental tractors, insecticides, fertilizer --
were available locally and those individuals who did not
think they could acquire any of these inputs in their
immecliate areas. The associated innovation indexes would
have been 23.0% and 11.5% respectively.

IRn equally good alternative does not exist for the
subsarnple including the better educated farmers with less
than 3:200 ppr capita family income. The best that is

avanhilile is a partition between those iarmers who were

27S
willing to try a new practice before others in their
community had used it and those who believed the best
strategy was to "wait and see” what happens to those
farmers who are first to use an innovation. Unfortunately,
only three observations were included in the first group.
Thus, even though such a "split" would have explained 4.2%
of the variance, as compared to 6.3% for the rariable
selected by the computer, this partition is of questionable
usefulness because of the very small size of one of the
resultant groups. A second alternative, explaining only
1.5% of the variance in innovativeness, would have
separated those farmers who believed that being lucky was
the best way to "get ahead" (n=lS) from those who would
have placed their faith in knowing how to do business
(n=19). The mean index values for such a "split" would
have been 30.3% and 40.0% respectively.

The A.1.D.-IV program, as given in Figure 17, provides
an explanation of 49.5% of the variance in innovativeness
among 850 Francisco rice farmers. if we consider only the
effects of education, income, and credit, an explanation of
37.9% of the variance would be provided. Adding the
alternative partitions suggested as replacements for the
"information sources-I" partitions (input availability and
luck/knowledge), we can account for as much as 42.5% of the
variation in innovativeness.

Setting aside these gross statistics, it is interesting

to note the important role -~ in the decicion-making process --

276

which seems to have been assigned to the intellectual
skills acquired with at least some amount of formal education
(more than one year). A careful analysis of data indicates
that a relationship between educational achievement and the
use of "direct” sources of information is clearly indicated,
since the farmers with more than one year of school nearly
monopolized the use of agronomists and the mass media. Again,
we find that the better educated group had a much greater
exposure to extension meetings and farm magazines than their
relatively poorly educated colleagues. In fact, exposure
to these sources was an important alternative to education
as a partition. Yet there seems to be somewhat more to the
impact of education than its association with a greater use
of the "direct" sources of information (which, of course,
is important in itself). Recalling the "tree" related to
A.1.D.-I, we found that even when we had eliminated the
critical importance of the agronomist as a source of
information, educational achievement (greater than one
year of schooling) was still an important predictor of the
use of fertilizer. Thus, we probably cannot ignore the
importance of the conceptual structures, ohtained with even
a fairly short exposure to formal education, in their impact
upon the adoption process.

A.1.D.mIV suggests that the lower income farmers are
inhibited in their innovativeness by the "risk" involved
in new agricultural practices. This is supported by the

findings of A.1.D.-II, which suggested that lower income

277
farmers (less than $100 per capita in this case) were more
reluctant to adopt insecticides, despite the fact that they
might well have had sufficient credit to finance these inputs.

The use of credit was important to the innovativeness
of both the better educated and the poorer educated Sao
Francisco rice farmer. Since education and income are
strongly associated (20 of the 29 lowest income farmers had
less than two years of schooling), it can probably be said
that credit is important to the innovativeness of both
poorer and wealthier farmers. Individuals who were able to
obtain "sufficient” credit to meet their needs and farmers
who had received bank loans were found to be generally more
innovative than non-borrowers or those who obtained loans
from non-commercial sources (relatives, neighbors, local
businessmen, landowners, and buyers).

The local availability of inputs seems to be most
critical to the innovativeness of farmers who were dis-
advantaged in terms of educational achievement, credit, and
probably income. The innovativeness of poorly educated and
poorly financed farmers were greatly enhanced by the local
availability of inputs, 3 variable which did not emerge as
important among the remaining farmers.

Again, there was some evidence that the farmer's utility
for increased income affects his innovativeness, Finally,
it should be noted that attitudinal variables, as measured
in this study, seemed to be of much less importance than

other factors in their effect on innovativeness. Perhaps

278
with increased sample sizes it would have been possible to
identify an interactive effect of attitudinal items (as
suggested by the use of luck/knowledge as an alternative
partition of Group #4).

A.1.D.-V: Innovativeness Among
Recife Bean farmers

 

 

Among the Recife bean farmers, merging the Alagoas—
Pernambuco and Irecé samples, a range of 0 to 76 is found
in the index measure of innovativeness. To explain this
variation the A.1.D. program generated a "tree" which bears
a striking resemblance to the one proposed for the 850
Francisco rice farmers (A.I.D.-IV).

The first partition is quite unlike A.1.D.‘IV separating
12 farmers who were exposed to extension meetings, farm
magazines, and radio from those farmers who were exposed to
less than all three channels of information. Since this
partition is somewhat surprising, it is useful to examine
the other characteristics of this very innovative group.
Upon closer inspection we find that all of Group #2 had
completed more than one year of schooling, as Opposed to
31.8% of G‘oup #4. Approximately 83.2% of these farmers had
either received loans from the credit c00perative and banks
or indicated that they had obtained "sufficient" credit to
meet their needs. This is compared to a figure of 44.2%
among the remaining members of the sample. Three quarters
of these more innovative individuals had income levels of

$100 or more, as contrasted to 31.0% of the other farmers

279

 

 

nu . a
v.» u o
«o.c~ . H
AmOXV
.HNU aaooum ssh

 

 

 

m I

m.m .

.o.»n .

. mos“

RONV nzooew

%<0'.

y

«k

owoum wove“ cue:

anewua>uemno
mo nonnaz
duzvcuum

II I
Lani:

 

we . z
» ~.o . o
o.a~ . »
n~.~v
“as "q~<>< wamzu

 

 

 

 

 

 

 

aw I z
N.~H I u
«h.vn I h

any "4~¢>< Benz”

 

 

.uuofiuum neon omaoom unos< neoco>uuu>occu

 

 

L<021

hey

ES

 

 

 

 

 

 

 

0H .

”.9 .

am.~v 1
nooNn Jo
Anne "arson"

L‘U:z

«v I a

o a” I b

 

 

 

am” . z

o.n~ . a

...»n . p
ha.~v

any u_~.oaz~

 

 

 

 

. a z
m.uH I o
u h

and "zouw<uaom

 

 

an . z
u..« I b
.~.am . H

has
n~v ”HH.oaz~

 

 

 

 

 

 

 

 

an I 2
Q.V¢ I 0

“may ”F.3egu

 

 

 

V
V
m

 

 

 

 

 

">-.a.m.< -- mu ouauwm

 

can . z
”.0“ . a
no.0n . p
q<pop

d

 

280
under study. In addition, it appears that.they are more
likely to try new methods without the benefit of first
observing the experiences of others and somewhat more
inclined to believe that knowing how to do business (as
Opposed to being lucky) is the best means of earning higher
incomes. In effect, the 12 innovative farmers selected by
this partition are a rather elite group -- better informed,
better educated, better financed and holding generally more
modern attitudes than was typical of the remaining
individuals in the sample.

This initial ”split" accounted for lf.8% of the variance
in innovativeness among the Recife bean farmers. As should
be expected from the preceding discussion, education, credit
and income were important alternatives to the suggested
partition. Education, separating the bean farmers into one
group with less than one year of formal education and
another with at least that many years of schooling, would
have explained 12.1% of the variation. A partition between
those who received loans from banks and credit cooperatives
or believed they had obtained "adequate” amounts of credit
and those who either did not borrow or received their loans
from non-commercial sources could have contributed 11.6%
to the explained variance. And finally, a division on the
basis of farmers with $100 per capita income or,more and
those with smaller incomes, would have accounted for 7.7%
of the total variance.

With the second partition, on the basis of education,

281

the "tree" begins to resem le the explanation of innovative-
ness suggested for the $30 Francisco rice farmers. Bean
farmers with at least one year of schooling had an average
index number of 38.3%, while those without any formal
education avaraged 30.4% on the innovativeness scale. (The
equivalent partition in A.1.D.~IV involved average scores
of 42.6% for farmers with more than one year of formal
education and 25.0% for those with a year or less of formal
education.) This particular "split" contributed an
additional 6.4% to the explained variance. The best
alternative to education would have been a partition between
those farmers who were not exposed to extension meetings,
farm magazines, or radio and those who were exposed to at
least one of these channels of information. (An important
alternative to education in A.1.D.-IV was a partition between
those who were exposed to extension meetings and those who
were not.) A division of the sample into groups composed of
farmers with bank loans, credit cooperative loans and
"sufficient” credit or non-borrowers and farmers with loans
from non-commercial sources would also have been an
important alternative. (Such a partition —- omitting credit
cooperative loans which did not exist in the rice area --
was indicated as the "best” alternative to education in
A.1.D.-IV.)

Among the better educated farmers, those who had
learned their latest method from relatives (n=17), agronomists

(n=6), peeple from a different place (n=6), and businessmen

282
(n=8) seemed to be more innovative than those who indicated
that neighbors (n=17) and the mass media (n=1) were as
important to them as a source of information. This particular
partition is somewhat difficult to interpret, especially
given the small number of observations in some of the
classifications (an unfortunate consequence of small sample
sizes). What can probably be said, with some assurance of
validity, is that farmers who used agronomists as a source of
information tend to be more innovative than those who depended
upon their neighbors for information about agricultural
innovations. This statement gains support from the fact
that the mean innovativeness score, across all farmers, in
both the bean and rice areas was highest for those learning
new ideas from the agronomist. It may also be the case
that relatives were a better source of information than
neighbors, although there is no particularly useful
rationalization for such a contrast. (A.1.D.“IV does not
provide support for the opposite view because of the small
number of observations included within the "relatives"

classification whenever this variable was involved in a

()9

partition.) Statistically, this "split" accounts for 3.9
of the variance in innovativeness among bean farmers. An
alternative partition would have involved credit divided in
the same manner as discussed earlier.

The availability of locally supplied inputs seems to
be particularly important to the innovativeness of those

farmers who did not have any formal education. Individuals

283
who believed that at least two of the three important
inputs -« rental tractors, fertilizer and insecticides --
were available locally tended to be more innovative
(T534.7%) than farmers who thought that local suppliers
handled, at most, only one of these items (T528.O%). (It
may be remembered that this same partition would have been
the best alternative to credit -~ Groups #8 and #9 ~- among
the poorly educated rice farmers.)

An explanation of 2.2% of the total variance was
provided by this "split." The best alternative partition
would have been to separate those farmers who had not been
exposed to extension meetings, farm magazines, and radio

(T=27.4%) from those who were exposed to one or more of

H

hese channels of conmunicatien (1&33.3%). Because of the
small number of observations in some of the classifications,
such a partition would largely reflect differences between
those who were exposed to radio and those who were not.
Income levels and credit usage, which were found to be
important predictors of innovativeness among rice farmers,
were equally useful in explaining the innovativeness among
those bean farmers who had completed at least one year of
schooling. As with their counterparts among the rice
farmers, the bean farmers who had obtained loans from the
better credit sources tended to be more innovative than
other farmers with similar backgrounds. Figure 18
illustrates the partition of Group #6 between those farmers

who had obtained bank loans, credit c00perative loans, or

(1.)

PH

 

 

284
"sufficient" amounts of credit and those who had either not
borrowered or received their loans from non~commercial
sources. The innovativeness scores associated with this
partition were 49.9% and 36.0% respectively. (Index
numbers in the equivalent A.1.D.~IV partition equaled 58.1%
and 35.6% in the same order.) Partitioning the sample on
this basis contributed an additional 4.9% to the variance
explained by the "tree."

Farmers with per capita family incomes of $200 or more
were found to be more innovative than individuals with
incomes of less than that amount (based upon the partition
of Group #7). The relatively wealthier farmers had an
average innovativeness score of 42.5% as opposed to 32.3%
for their poorer colleagues. Since there is no evidence to
indicate differently, we must assume that the effect of
income is a mixture of differences in the ability to
finance new methods and willingness to accept the "risks"
involved in applying them. For whatever reasons, this
partition contributed 2.5% to the explained variance.

Credit was also used to subdivide those bean farmers
who lacked formal education, yet found inputs readily
available. The A.1.D. program separated the farmers who
had obtained loans from the credit cooperatives (n=2) from
those who either did not borrow or used other sources of
credit (n=26). Although this partition is very questionable
statistically, there may be some validity in setting the

credit cooperatives apart from other sources. In earlier

285
chapters, it was found that credit cooperatives were a very
important source of loans to small and medium sized farmers.
Loans given by this source also tended to be larger than
other non-bank sources, which also supplied a very large
proportion of the small farmer's credit. It was clear that
credit c00peratives were superior to other non—bank sources,
and it is possible that they may even do a better job of
serving smaller farmers than the banks. The mean innovation
index number, across all bean farmers, was somewhat higher
for those using credit cooperatives (T=44.3%) than for those
using banks (T=4l.2%). Thus, since size of farm was found
to be closely associated with educational achievement, a
partition setting credit cooperatives apart may have had
some real meaning.

If instead we assumed that this "split" was simply
spurious, it would be possible to provide an alternative
partition based upon the f rmer's willingness to forego
present income for a larger amount in the future (deferred
income). The group who would be willing to defer income
were more innovative (T=44.3%) than those whose preference
was for present income (T=32.l%). While the original
"split" (based upon credit) would have explained 2.1% of
the variance, the alternative partition could account for
1.9% of the variation in farmer innovativeness.

Among those farmers who found inputs less available
(Group #8), those few individuals (n=S) who were willing to

try new methods without observing other people's experiences

286
were discovered to be more innovative (T=33.0%) than those
who would prefer a "wait and see" approach to innovations
(T323.5%). Interestingly enough, the more innovative
individuals tended to be from the lower income classifications
of the parent group. To the extent that this partition
involves a "real" difference, the five more innovative
people were apparently simply more willing (as Opposed to
being better able) to accept "risks." Finally, this
partition contributed 1.5% to the variance explained by
the "tree."

In total, the "tree" illustrated by Figure 13 explained
29.5% of the variation in innovativeness among bean farmers.
(It should be noted that additional partitions were
suggested by the A.1.D. program, but have been "prunec"
from the ”tree” because of their questionable statistical
value.) The upper portion of A.1.D.-V was found to contain
the same variables as A.1.D.-IV, although their configura—
tion was different. Again we have found farmers with at
least some education to be more innovative than others; the
farmers with $200 or more per capita income to be more
innovative than those with less, and farmers with bank
loans, credit cooperative loans, and ”sufficient" credit
to be more innovative than their counterparts.

There is evidence, especially in terms of the initial
partition involving 12 of the better informed farmers,
that access to "direct" sources of information hastens the

adoption process. Once again, the positioning of a

287
partition utilizing differences in the local availability
of certain inputs suggests that the innovativeness of the
most poorly educated farmers is apparently enhanced by
supplies of modern inputs at convenient locations. Finally,
as in the case of rice farmers, there is some evidence which
suggests that farmers who are more willing, or better able,
to accept the "risks" involved in new methods tend to be

more innovative.

Mair

Considering all of the A.1.D. programs presented,
certain generalizations seem to emerge. At the forefront
is support for the basic contention that the adOptien of
innovations involves a decision-making process which can
be described in terms of the interaction of certain
variables hypothesized to (1) have an effect upon the
formulation of expectations, (2) indicate the farmer's
willingness to accept "risks,” and (3) suggest an
individual's ability to implement his decisions. In a
sense the A.1.D. could do little else than support the
notion of an interactive system, yet it need not have
paralleled the theoretical and analytical developments of
earlier chapters. We might have found, for instance, that
only those variables which relate to the individual's
ability to put new methods into practice -- input avail-
ability, credit, and income levels -~ were important.
Sueh a finding would have indicated an argument for

'Reconomic determinism" of innovativeness. 0n the other

288
hand, the A.1.D. programs might have indicated that individual
and community attitudes largely governed innovativeness,
suggesting a "socio-cultural determinism" of the adoption
process (although such an outcome would have been extremely
surprising given the analysis of Chapter IV).

Yet, neither of these outcomes were suggested by the
evidence. The variables describing the adoption of various
new practices, and innovativeness in general, are a
combination of those expected to affect the decision-making
process and those affecting likelihood of implementing the
resultant decisions.

The Effect of Attitudes Upon Innovativeness

 

A prediction of the Recife farmer's innovativeness on
the basis of his attitudes, given the evidence of this
chapter, would undoubtedly be very inaccurate. To the
extent that farmer attitudes may affect innovativeness,
their influence was clearly overshadowed by other variables.
This is not to say that attitudes have no effect upon
innovativeness, as Chapter IV suggests that attitudes may
influence the information-gathering process, but at leas
as measured in this study they were not among the more
important predictors of the farmer's innovative behavior.
There was, however, some indication that with larger sample
sizes it might have been possible to identify an interactive
effect of farmer attitudes, especially since attitudinal
variables were occasionally important alternative partitions

late in the development of the "trees." Looking ahead to

289
possible policy conclusions, this finding would suggest that
programs designed to hasten the adoption of more productive
technologies, at least in Northeast Brazil, need not be
directly oriented toward changing the attitudes of farmers
(such as those examined in this study).

Education, Information Sources, and InnovativeneSs

II..-

 

Educational achievement, which is not easily altered
in the short run, was continually found to be among the more
important predictors of adoption, especially in the case of
describing variations in the general innovativeness of both
rice and bean farmers. Although the general educational
level of Recife farmers is very low, 68.3% of the rice
farmers and 86.0% of the bean farmers had not completed
more than three years of school; even a very few years of
schooling was often found to enhance the individual‘s
innovativeness (see A.1.D.-I, IV, V). There was, of course,
a relationship between the farmer‘s educational achievement
and the sources of information that he tended to use, but
there seems to be more to the impact of education than this
"dimension" alone. The best assessment of the "other
dimension" of education seems to be the development,
concommitant with educational achievement, of mental
processes which enhance the farmer's ability to process
information and develop expectations regarding new agri«
cultural practices.

The advantage of utilizing "direct" sources of

information (these emanating from nearest to the origin of

290
the innovation «- agronomist and certain mass media channels)
is fairly evident, even though it is not always readily
apparent from the A.1.D. presentations. The clearest example
is with fertilizer usage in the 850 Francisco rice area
where 42.9% of the farmers who indicated the agronomist as
an important source of information were using this innovation
as compared to only 7.6% of the remaining farmers. The
initial partition of A.1.D.-V also suggests the importance of
"direct" sources of information. Here, it may be recalled,
a rather elite group of farmers who had been exposed to
extension meetings, farm magazines, and radio (and at the
same time constituted a large proportion of those indicating
the agronomist as source from which they had learned their
latest innovation) were found to be considerably more
innovative than their counterparts with less exposure to these
channels of information. In addition, the more innovative
among the rice farmers (although separated on the basis of
other variables) included nearly all of those individuals
who had used the agronomist as an information source (as
GrCUp #5 contains 10 of the 14 farmers who had used this
source). Finally, exposure to extension meetings, farm
magazines, and radio provided an important alternative to
many of the partitions selected by the program. One might
therefore conclude that providing greater access to more
"direct" sources of information, and encouraging their use,
might substantially improve the overall innovativeness of

the Recife farmers.

291

Innovativeness and the Willingness to Accept "Risks"

 

One of the important general hypotheses of this study
was that the adoption of new agricultural practices involves
"risks" and that those farmers who were more willing or able
to accept these "risks" would be the more innovative
members of the sample. While approaching this matter of
"risk” directly would be quite difficult, it is possible
to examine certain indicators of the farmer's willingness
to accept the uncertainty of new methods. One approach
was to examine the effect of a farmer's willingness to try
something new before others in his community had tried it,
as opposed to a more widely acknowledged strategy of "wait
and see." A positive relationship between innovativeness
and the farmer's willingness to try locally untried practices
received some support among bean farmers (GrOUps #20 and
#21 of A.1.D.-V). Again, the identification of this
characteristic was the most important alternative to an
existing partition (of Group #4 in A.1.D.-IV) among rice
farmers. Although the number of farmers who exhibited this
trait was quite small in each case, it does provide at
least tentative evidence that farmers who are more willing
to accept the uncertainty inherent in new methods adopt
them more rapidly.

An alternative approach is to assume that one of the
main reasons why lower income farmers are less innovative
is that they are reluctant to jeopardize the "well«being"

Of their family by risking the failure of an unfamiliar

292
practice. The difficulty with this approach is that lower
incomes not only mean less buffer against unfortunate out-
comes but represent a decreased capacity to finance the
inputs usually associated with new methods. In each of the
A.1.D. programs, except A.1.D.«I, income was an important
predictor of innovativeness. (Recall that income was.
represented in terms of family per capita figures in an
attempt to make the variable reflect family "well-being.")
The relative importance of the two effects of income, as
suggested above, is fairly clear with respect to the
adoption of tractors (A.1.D.—III). Here we found that when
rental tractors were not available locally, the relatively
wealthy farmers (income 1 $500 per capita) were still very
likely (87.5%) to use them. Undoubtedly, they had been
able to purchase their own tractors. The outcome is not
nearly as clear for A.1.D.-II (use of insecticides) and
A.1.D.~IV (general innovativeness among rice farmers) but
the surrounding evidence seems to support the contention
that the poorer farmers' desire to avoid risks was important
in these partitions. In these particular cases the
relationship of credit to income levels suggests that self-
financing of inputs was probably not the critical
advantage of higher incomes. Finally, in the case of
A.1.D.-V there is no basis for speculation about the
reasons for the importance of incomes, so it must simply be
assumed that the partitiongiven represents a combination

of the factors suggested above.

293

While we have argued that the disutility of losses
causes poorer farmers to be reluctant to try new practices,
it can also be argued that differences in the utility for
increased incomes can affect the farmer's willingness to
accept the "risks” of new methods. Although we must remain
skeptical about our measure of utility for increased income --
responses to a risk game proposed in the Recife survey «-,
it did appear in A.1.D.-IV and A.1.D.-I following the manner
that was hypothesized. Those with ”moderate" and "high"
utility for increased incomes tended to be more innovative
than those with "low" or "very low" utility for gains in
income. Adding to the uncertain nature of this outcome is
the position of the partitioning related to this variable.
In each case the partition occurs at the end of the "tree,"
where it must be interpreted with considerable care, and
in A.1.D.-I it represents an alternative to an unacceptable
"split." The most that may be said on behalf of the
relationship between innovativeness and utility for income
is that there is tentative evidence suggesting that farmers
with greater utility for increased income are more
innovative than others.

In total, there seems to be sufficient evidence to
argue that a farmer's willingness to accept "risks,"
which is in turn related to his utility for gains in income
and especially his disutility for losses, is related to his
innovativeness. Thus, the contention that those who are

Willing or able to bear the uncertainties of new

294
technologies tend to adopt them more rapidly seems to be
supported empirically.

Innovativeness and the Ability to Implement Decisions

 

Two variables were presented in Chapter V -- credit
usage and input availability -- which were expected to have
an effect on the farmer's ability to implement his decisions.
According to the A.1.D. programs, both of these variables
were important predictors of innovativeness. Farmers who
either had not borrowed at all or had borrowed from non-
commercial sources (neighbors, relatives, landowners,
businessmen and buyers) were found to be less likely to
adopt new methods than those who had either obtained loans
from credit c00peratives (not vailable in the rice area)
and banks (government and private), or indicated that they
had received ”sufficient” credit to meet their needs.
Credit, divided in this manner, was the most important
variable in explaining the use of insecticides among the
850 Francisco rice farmers (explaining 13.4% of the
variance). There were exceptions, however, to this usual
division of the credit variable.. Only those farmers who
had obtained "sufficient" amounts of credit were found to
be greatly different from others in their use of fertilizer.
Also, there was ev1dence, among bean farmers, that the credit
c00perativos may have done a better job than the banks of
serving the needs of the smaller farmers. Finally, in
A.1.D.-III we find a partition of credit which is exactly

the reverse of what one would postulate, non—borrowers being

295
more innovative than others. Given its position within
the "tree," one would be rather reluctant to assert that
this particular "split" negated other findings with respect
to credit. On balance, it seems that those farmers who
were able to obtain ”sufficient” credit or had received bank
and credit cooperative loans were more innovative than
individuals who either had not borrowed or had received
their loans from non-commercial sources of credit. This
outcome suggests that improving the credit system in the
Northeast of Brazil might well contribute to the overall
innovativeness of the area's farmers.

The local availability of inputs also seems to be
important to the adoption process. In the case of tractor
usage among the 850 Francisco rice farmers, the local avail-
ability of rental tractors was the most important variable
in explaining the adoption of this innovation (accounting
for 12.3% of the variance among farmers). An index of
input availability (including tractors, insecticides,
and fertilizers) emerged as a useful predictor of
innovativeness of rice farmers. The location of this
variable within the "tree" suggests that local availability
is most important to the innovativeness of the smaller
poorly educated farmers. Perhaps their situation imposes
special limitations on their ability to acquire inputs
from other than the nearest trading center. In general,
it appears that an increase in the local availability of

certain inputs would widen their use among the farmers in

b)
'O
0‘

the Recife area.

Tenure and Innovativeness

 

In Chapter V it was found that most of our preconceptions
about the relationship between tenure and the use of certain
inputs were incorrect. Landlords and tenants seemed to be
about equally likely to use insecticides and tractors,
while owner-Operators were discovered to be less likely to
have adopted these practices. The rationale suggested then
was that landlords were generally better educated and better
financed than owner-operators, and that the landlords
either furnished modern inputs to the tenant or required
that he use them. This contention seems to be supported
by A.1.D.—II, where poorly educated and poorly financed
tenants were found to be considerably more likely to use
insecticides than their landowner counterparts in the same

situation.

CHAPTER VIII
SUMMARY .ND CONCLUSIONS

Review of the_Besearch Project

 

 

There seems to be considerable agreement that one of
the pressing problems of agricultural development is to
secure a more rapid adoption of improved practices among
the farmers in developing countries. There is less agree-
ment, however, about the appropriate strategies for
accomplishing this task. Some believe that emphasis should
be placed upon the social interactions involved in the
adoption process, while others prefer to focus their attention
upon the economic considerations related to technological
change. These differing points of view stem from contrasting
opinions about the behavior of farmers in less developed
countries. One view is that they behave quite differently
from their counterparts in a modern agriculture. The
reason given for these differences is that the traditional
farmer is primarily a "social being” and that he lives by
attitudes, beliefs, values and cultural norms which widely
diverge from those of the modern farmer. An alternative
view is that farmers in a traditional agriculture are no
less "economic men" than those in deve10ped countries.

The frame of reference used in this study was to view
the adoption of new practices as an individual decision-
Hulking process under conditions of uncertainty and involving
tlle interaction of social, economic and personality

297

298
variables. Although many researchers have recognized the
adoption process as a matter of individual decision~making
and have acknowledged the element of "risk" in adopting
unfamiliar practices, few have approached the problem with
this particular frame of reference. Such an approach provides
sufficient flexibility for examining a wide variety of vari—
ables which might be hypothesized to have an effect upon the
farmer's decisions toward innovations. The objectives of
this study were to:

1) Develop a conceptual framework of decision-making
under uncertain conditions which can be applied to
examining the adoption of new agricultural
practices in less developed countries.

2) Within the context of this conceptual framework,
to postulate relationships between certain
variables and the adoption of innovations, and
test these relationships using data provided by
a survey of farmers in Northeast Brazil.

3) Examine the relative importance of the postulated
relationships as explanations of the variance in
the adoption of innovations among the Recife
area farmers.

4) Suggest, on the basis of this research, some
elements of a strategy to hasten the adoption of
more productive agricultural practices by the
farmers in Northeast Brazil.

5) Provide an assegsment of decision-making under
uncertainty as a framework for studying adoption
of innovations, and suggest further research
which might be conducted utilizing this approach.

In pursuing these objectives the study began with a

review of the literature on the behavior of farmers in less
developed countries, the diffusion of innovations, and

decision«making under uncertain conditions (Chapter II).

In turn, some of the useful theoretical notions and

299
insights contained in this review were incorporated into a
conceptual framework of the farmer decisionvmaking process
under conditions of uncertainty (Chapter III).

Under the guidance of the conceptual framework, relation-
ships were hypothesized between innovativeness and certain
variables which were amenable to analysis using the MSU/
SUDENE farmer survey data from Northeast Brazil. As the
study progressed, these relationships and the inter-
relationships between the variables themselves were subjected
to statistical analysis (Chapters V and VI). Each of the
relationships examined can be placed within one of seven
hypotheses.

1) Farmers' attitudes have an influence on their
innovativeness.

2) Farmer innovativeness is positively related to
his educational achievement.

3) innovativeness is related to the sources used by
farmers to learn about new agricultural practices.

4) Farmer innovativeness is positively related to the
financial capacity to accept "risks." (Poorer
farmers are less innovative because they are
reluctant to subject their families' well—being
to a crisis involving the failure of an innovation.
Stated differently, they are unable to accept the
”risk" perceived in new methods.)

5) Farmer innovativeness is positively related to
his anticipated utility from increased income.

6) Farmer innovativeness is positively related to
the availability of inputs and credlt.

7) Farmer innovativeness is related to tenure
arrangements.

A more general hypothesis mw to which these seven are

subservient ~- is that: an interactive system_of social,

 

 

300

economic and personality variables as deve10ped in the
conceptual framework can be verified empirically, and that
the relative importance of these variables to the adoption
of an innovation can be established. Identification of such
a systgm would be useful to planners as they deve10p
strategies for increasing the pace of technological change
within the agricultural sectors of deve10ping nations. Thus,
the analysis of the seven hypotheses was conducted in such
a manner as to provide an assessment of the more general
hypothesis as well.

The final phase of the project involved an examination
of the relative impact and interaction of variables which,
on the basis of earlier analyses, appear to have an effect

Upon farmer innovativeness. Automatic'lnteractign

 

   

Detection -- a relatively new technique —* was utilized in

-- -0

this analysis (Chapter VII), providing a further assessment

of the theoretical and statistical developments in earlier

 

----..-

tO—uetd\lng the Kuoptien or lnnova tions

Analysis of he cision- -Mal:ing Under Uncertainty As
‘l

 

From the author's point of view, there appear to be
some important advantages of studying the adoption of
innovations within the framework of individual decision“
making under uncertainty. First, such an approach explicitly
rxnzognizes what many researchers have already acknowledged,
that the adoption of innovations is a decisionvmaking process

and that decisionemakers have less than perfect knowledge

301
upon which to base their expectations about new and
unfamiliar methods. Explicit recognition of the "risks"
inherent in new technology, by itself, suggests an added
dimension in our attempts to understand differences in
innovativeness among farmers.

This approach has caused the author to view the
adoption process as an ”interactive system" of variables
affecting the farmer's decisions with respect to
innovations. It would seem to be useful to encourage this
view among other researchers interested in the process of
technological change. One reason is that it stimulates
additional consideration of ”why" certain variables are
found to be related to innovativeness and under what
conditions (interactions with other variables) these
variables have an impact upon the adaption of new practices.
Another argument in support of this view is that it tran-
scends the disciplinary disagreement over the emphasis which
should be given to the economic versus non-economic factors
believed to be influencing the farmer's decisions toward
innovations, and might serve to expedite interdisciplinary
cooperation in this area of research.

Finally, it is the author's opinion that the findings
of this study support the usefulness of viewing the
adoption of agricultural innovations in developing countries
as an individual decision-making process involving the
interaction of social, economic and personality variables,

:nd that approaching the problem with this frame of reference

302
will provide an improved conceptual basis for developing
strategies to increase the rate of technological change in

the rural sectors of the developing countries.

'Qpnclusions

 

Perhaps the most concise and meaningful way to
summarize this study is to examine the research findings
within the context of the seven generalized hypotheses
presented earlier. This provides an Opportunity to review
both the specific and more generalized relationships
concurrently.

1) Farmers with "modern" attitudes are more
innovative than farmers with ”traditional"
attitudes.

Both the individual farmer's own attitudes and the
attitudes he perceived that his neighbors held were examined
in this study to the extent permitted hy the available data.
Among the Recife farmers surveyed, 46.7% indicated that
thei‘ neighbors would regard early adopters of new agri-
cultural practices as foolish and 43.2% believed that local
farmers would not like to see individuals progress more
rapidly than the community as a whole. One might expect
that these community attitules could be interpreted as
social pressures inhibiting individual innovativeness.
There is, in fact, some evidence that this is true, since
the farmers indicating the presence of such views were more
likely to hold negative attitudes toward new technology in

general and also seemed to be less aggressive in gathering

knowledge about new agricultural practices. Iowever, the

303
impact of community attitudes was not found to be'
statistically significant in terms of explaining variations
in individual innovativeness itself.

A large prOportion -~ 82.0% -- of the Recife farmers
expressed a fatalistic view of the possibility of improving
their situation through their own efforts. This expression
of fatalism was found to be unrelated to the farmer's
educational achievement, income level, and the size of his
farm. Although their less fatalistic counterparts were more
likely to know about fertilizer and other methods of
improving yields, they did not seem to be significantly more
innovative.

Even though most of the farmers appeared to be fatalistic,
nearly half -- 46.7% -- of them believed that "to make more
money it is better to know how to do business than to be
lucky." Relative business success seems to reinforce this
belief, since the larger and wealthier farmers were more
inclined to be of this persuasion. Affirmation of this
belief was also more commonly found among the better
educated members of the sample. The ”knowledge-oriented"
farmers were expected to be somewhat more motivated to seek
out information than their "luck-oriented” colleagues.

There seems to be evidence in SUpport of this proposition,
as "knowledge«oriented” farmers were found to use the
agronomist, mass media, and businessmen more frequently as
sources of information about new methods. In addition, they

were more likely to have attended extension meetings, read

304
farm magazines, and were more often aware of chemical
fertilizer than those who believed that luck was in control
of their earnings. Finally, this variable was found to be
of some value in explaining the adoption of insecticides
and tractors (see A.1.D.-II and A.1.D.*III), although its
contribution was relatively small.

The majority of the Recife farmers -- 69.4% «v seemed
to hold favorable attitudes toward new technology in general.
It is interesting to note that while most of these farmers
believed there was little they could do individually to
improve their situation, they were nonetheless hopeful that
technological advance would bring them a better life. The
central argument related to this variable was that attitudes
help to shape the farmer's expectations of new methods, and
that farmers with a generally negative attitude toward
technological change would be less responsive to specific
new practices. Evidence was found to support the contention
that such farmers were less aggressive in gathering informa-
tion about new methods. Farmers with negative attitudes
toward new technology were also found to be less likely to be
using insecticides, but they were not any less likely to
use tractors or fertilizer. Finally, the Automatic Inter-
action Detection programs bypassed the farmer‘s attitude
toward technology as an explanation of farmer innovativeness.
Thus, again, we must conclude that other factors were more
critical to the farmer's decisionvmaking process than his

general attitude toward new technology.

305

Among the Recife farmers, only 57.3% would have preferred
NCr $90 one year from now to NCr 330 today. It was reasoned
that an unwillingness to defer income might represent a
reluctance, on the part of the farmer, to sacrifice present
consumption in order to finance new methods which promised a
higher return in the future. The farmer‘s time preference
for income, however, did not seem to be significantly related
to his innovativeness. Still, there was a logical relation«
ship found between other attitudes and the willingness to
defer income. Fatalistic farmers were less willing to defer
income than nonvfatalistic farmers, and "knowledge-oriented"
farmers were more willing to defer income than their "luck-
oriented” counterparts.

Interpersonal trust was also eramined in this study.
Over half -- 57.0% -- of the Recife sample indicated a
willingness to trust outsiders (non~relatives). The
hypothesis that non-trusting farmers would concentrate
upon their relatives as a source of information about
agricultural innovations was not supported by the data.
Since there was no alternative evidence suggesting that
differences in interpersonal trust affected the innovative-
ness of Recife farmers, the variable was assumed to be
unimportant to the adoption process.

From the information available in this study, it would
appear that certain attitudinal variables *— fatalism, luck/
knowledge, and attitude toward technology -~ may have an

effect upon Recife farmer's decision~making process (especially

306
with respect to the information gathering), but the
magnitude of this effect was very small by comparison to
other variables. At least for the Recife area of Northeast
Brazil, this would suggest that altering other variables
would provide a greater stimulus for increasing the adoption
of new practi es than efforts to directly change the farmer‘s
basic attitudes.

2) Farmer innovativeness is positively related to
his educational achievement.

Farmers in the Recife area were generally not well
educated. A large preportion of the sample (44.0%) had not
attended school, and only a few (7.0%) had completed more
than six years of formal education. Educational achievement,
among these farmers, was significantly associated with farm
size and income levels. The larger and relatively wealthier
farmers tended to be better educated than their smaller and
poorer counterparts.

Greater educational accomplishment was assumed to improve
the farmer's capacity to utilize "direct" sources of informa-
tion (the agronomist, farm magazines, radio messages, etc.)
and to enhance his ability to evaluate particular situations
(conceptual ability). The better educated farmers in the
Recife sample were, in fact, found to be making greater use
(If the agronomist and mass media as sources of information
about.agricultural innovations. They also seemed to be more
launeledgeahle about chemical fertilizers and other methods
(If improving yields. More importantly, the farmer's level

of talucatienal achievement seems to provide an important

307
explanation of his innovativeness. Educational achievement,
as a variable in the A.I.D. programs, was the most important
predictor of innovativeness among the 550 Francisco rice
farmers (A.1.D.*IV) and the second most useful explanation
of variations in innovativeness among the Recife bean
farmers (A.1.D.-V). In addition, it appears that even a very
few years of formal education (1 to 3 years) had a substantial
impact upon the farmer's level of innovativeness. This
suggests that the pace of technological change would have
been considerably hastened if even an elementary level of
education had been widespread among the Recife area farmers.

3) Innovativeness is related to the sources of

information used by farmers to learn about new
agricultural practices.

Very few of the Recife farmers reported that they had
learned about a new method from an agronomist or the mass
media (12.9%). Still fewer indicated that they had learned
about new practices from a local busine-sman (5.7%) or peeple
from another place (6.6%). Instead, most of them seemed to
iave depended upon either neighbors or relatives as a source
of information about new agricultural technologies (71.0%)163
Slightly over half (58.7%) of the farmers owned a radio, while
less than one-fifth had read a farm magazine (16.5%) or
attended an extension meeting (14.2%). In general, the
communication of ideas about agricultural innovations flowed

from farmer to farmer by word of mouth.

 

-..—— r_'_

‘03. .
‘ an undefined category termed "another source"

accounts for the remaining 3.6% of the information sources
used by Recife farmers.

 

...—...

308

One would expect that such a great dependency upon
farmer to farmer communication would serve to slow the pace
of technological advance, and this seems to have been the
case among the Recife farmers. Considering the sample as a
whole, those farmers who reported the agronomist as a source
of information were more likely to be using insecticides,
tractors, and chemical fertilizers than farmers reporting
other sources. Those who used the mass media were more
likely than farmers using other sources (except the agronomist)
to have adopted tractors and insecticides. Again, farmers
who indicated local businessmen as a source of information
about new methods were more likely to use insecticides than
those who had gathered their information from friends,
relatives and people from another place. A similar associa-
tion was found between the farmer's exposure to radio, farm
magazines, and extension meetings. The Recife farmers who
had read farm magazines were more likely to have used
fertilizers, insecticides, and tractors than those who had
not. Those who had attended extension meetings used these
practices more frequently than their counterparts who had
never attended such meetings. And even the presence of a
'radio in the home seemed to be associated with a greater
degree of innovativeness.

The Automatic Interaction Detection programs computed
for rice and bean farmers also support a positive relation-
ship {mummen innovativeness and the use of the agronomist and

masm=1nedia as sources of information about new agricultural

309
practices. The clearest example involved fertilizer usage
in the 850 Francisco rice area, where 42.9% of the farmers
who indicated the agronomist as a source of information were
using this innovation as compared to only 7.6% of the‘
remaining farmers (A.1.D.-I). Among the Recife bean farmers,
exyosure to extension meetings, farm magazines, and radio
seemed to be strongly related to farmer innovativeness
(A.1.D.~V). Finally, the more innovative of the rice
farmers (although separated on the basis of other variables)
included nearly all the individuals who had used the
agronomist as a source of information. Consequently, it
appC?TS that techniques which could be employed to encourage
and extend the use of the agronomist, the mass media, and
perhaps even the businessman as sources of information
about new agricultural practices would serve to spur the
pace of technological advance in the Recife area of
Northeast hrazil.

4) Foorer farmers are less innovative because they
are reluctant to subject their family's well-being
to a crisis involving the failure of an innovation.

Although there was a wide variation in income levels

among the farmers in the Recife sample, most of them were
‘very poor considering the fact that nearly tweethirds of
the sample had annual incomes of less than $100 per family
lnembeq; The niggardliness of the present incomes may well
ccntrfilnnmrto their continued poverty. In addition to the
(fiyvious problen of financing new inputs from their own

zneager resources, the results of this study suggest that

310
poorer farmers are more reluctant to accept the "risks”
inherent in trying unfamiliar practices. The use of
chemical fertilizer, insecticides, tractors, and non~chemical
fertilizer was found to be significantly related to
increasing levels of income. As was the farmer‘s evaluation
of the personal consequences of the failure of an innovation,
where the relatively wealthier farmers were less inclined to
indicate that their "family would suffe*" and significantly
more willing, than their poorer colleagues, to "try again.”
It was also discovered that the willingness’unwillingness to
use locally untried methods was significantly associated
with the farmer's income level, with poorer farmers being
more reluctant to try new practices without first observing
the orperiences of other members of their community. Finally,
it was observed that lower income farme‘s were less innovative
than their wealthier counterparts (A.1.D.-II, A.I.D.-IV, and
A.l.D.-V), with sufficient reason to believe that the
reluctance to accept "risks” was importantly involved in
this relationship.

5) Farmer innovativeness is positively related to
his utility for increased income.

Utility theory would suggest that the farmers who are
:most highly motivated to increase their incomes would tend
to be the most innovative. The findings of this study are
far less than conclusive on this point (due to probable
iJuudequacies of the method used to measure utility for
incrnte) although they suggest, in the author's riew, that

sucli a hypothesis deserves further attention. The variable

311
used to indicate utility for increased incomes was found to
be of some value in explaining fertilizer usage and general
innovativeness among the $50 Francisco rice farmers. In each
case farmers designated as having "moderate" and "high"
utility for increased income were more innovative than those
designated as having ”low” and "very low" utility for
additional earnings.

6) Farmer innovativeness is positively related to
the availability of inputs and credit.

Favorable expectations toward new practices are of
little practical importance if farmers lack the necessary
resources for implementing their decisions. Two such
resources are credit and the availability of modern inputs.

The aVailability of modern inputs was found to be an
inportant constraint on their use by Recife farmers. The
relative availability of certain inputs -- chemical
fertilizer, rental tractors, insecticides, and hand
planters -- as indicated by the percentage of rural
suppliers handling these products, was highly correlated
with the percentage of the sample farmers using these
innovations. Somewhat more specifically, the local avail-
ability of rental units was discovered to be the most

important explanation of tractor use/non-use among the

$50 Francisco rice farmers. An index of locally available

iJuNJtS was found to contribute to an explanation of

innovatixencss among tecife bean armors (A.1.D.~V). In

addititni, it would have provided a useful alternative to

leLLLt and sources of information in explaining the variation

312
in innovativeness among the 850 Francisco rice farmers
(A.1.D.-IV). Finally, there was evidence to suggest that
the local availability of inputs was most critically
important to the innovativeness of the poorly educated and
poorly financed segment of the farmers in the Recife sample.
Apparently these farmers were unable to travel beyond their
local communities to acquire the inputs used on their farms.

Farmers who had been able to obtain "sufficient"
credit (by their own assessment) or had secured loans from
banks and credit cooperatives tended to be more innovative
than farmers who either had not borrowed or had obtained
loans from such sources as relatives, neighbors, land-
owners, huyers and local businessmen. Smaller and poorer
farmers were more likely than larger and relatively
wealthier farmers to have either not borrowed or have
received loans from non-commercial sources (relatives,
neighbors, landowners, etc.). Loans from these sources
were typically of smaller amounts than offered by banks and
credit cooperatives, while loans in general tended to be
small (median of all Recife farmers was $247.73) and of
short maturities (less than one yearl.

About one-third of Recife farmers who hadn't used
credit indicated that they were afraid to borrow.
Approximately the same proportion of those who had borrowed
indicated that they were unwilling to accept the "risk"
involved in a larger loan. In either case, the smaller and

poorer farmers more frequently gave these reasons than

313
their larger and relatively wealthier counterparts. Thus,
not only are poorer farmers more reluctant to accept the
"risks" inherent in new technologies, they tend to be less
willing to accept the "risks" involved in loans which would
enable them to finance improved practices.

In addition, about one'third of the non~borrowers
indicated they didn't need credit, while a slightly smaller
proportion indicated that they couldn't get a loan. The
larger nd relatively wealthier farmers were more likely to
give the first re5ponse, while smaller and poorer farmers
were more likely to give the second. There is also evidence
to suggest that when smaller farmers indicated that they
”didn't need credit” what they really meant was that they
were able to survive from harvest to harvest without
requiring a loan to meet their family's living expenses.
Finally, credit c00peratives appear to be providing small
and medium-sized farmers, who had difficulty obtaining
bank loans, with larger amounts of credit than they could
have obtained from such sources as neighbors, relatives,
landowners, buyers, and local businessmen.

7) Farmer innovativeness is related to tenure
aTIYUlgLFKHItS.

Initially, the owner-operator was expected to be more
innovative than either Recife landlords or tenants. The
insecurity of tenancy and the anticipated disinterest of
large landlords were assumed to give owner~operators the
edge in terms of innovative behavior. These preconceptions,

however, proved to be incorrect. In general, landlords and

314
tenants seemed to he more innovative than their owner-
operator counterparts. The most reasonable eXplanation of
this finding was that the landlords in the sample tended to
be better educated and have greater financial resources than
owner-operators, while tenants utilized modern practices
either because landlords furnished the necessary inputs or
required their use.

A Strategy For Increasing the
r ‘u‘ o ——.—.-—é-‘_—-‘—O
Adoption offlnew Egricultural Iractlces

 

It is generally recognized that the adoption of new
agricultural practices is essential for expanded food and
fiber production. At the same time there is a growing
awareness that change in the rural sector must serve other
developmental goals as well. Such goals as improving the
relative well-being of small farmers and reducing the
disparity between urban and rural incomes are becoming an
important part of the deve10pment policies of a growing
number of less developed nations. Dissatisfaction among
the lower income groups is viewed as a present and potential
source of social and political unrest. Failure to respond
to this segment of the population may be expected to lead
to social and political upheaval. As a consequence of
these conCerns the strategy suggestions given in this
section, while not intending to neglect the larger farming
units, give particular attention to the problems of securing
the adoption of new technologies among the smaller and poorer

farmers within the rural population.

‘ ‘illllll
li‘lllll

 

 

315

Although the strategy suggestions given in this section
are based upon research findings from Northeast Brazil, they
may well have wider applicability among developing countries
(especially within Latin America). The following are some
of the elements which would appear to be important
components of a strategy to increase the use of new agriv
cultural methods, at least among farmers in the Recife area
of Northeast Brazil.

Informational Input

 

 

In order for farmers to accept innovations they must
first learn a sufficient amount about these new practices
to formulate the expectation that they will benefit from
altering their present farming methods. For the most part
the communication of such information in the Recife area
was from farmer to farmer by word of mouth. However, those
few farmers who did utilize the agronomist and the mass
media as sources of information were found to be
significantly more innovative. This suggests the need to
devise schemes to extend contact with the agronomist and
the mass media to a greater number of farmers in order to
increase the pace of technological change. Since small,
poorly educated farmers are least likely to use these
sources of information, it is necessary to place special
eaphasis on designing approaches which will reach this
grOUp. On the other hand, due to their generally higher
level of’education and greater capacity to accept “risks"

the larger farmers may assume an important role in these

316

SS'Cducated

(I)

approaches. These farmers could assist 1
farmers to understand mass media messages about new

practices in terms of local farming conditions. In

addition, they might provide the early field experience

with new practices necessary for their wider acceptance
within the community.

A version of the radio farm forum technique would
seem to be a very useful approach for the following reasons:

(1) Radios are readily available in rural

communities (over 50% of the Recife area
farmers had functioning radios).

(2) This form of mass communication does not
depend upon a literate audience (roughly
one third of the Recife sample had not
attended school at all).

(3) More importantly, this technique could
help expand the number of farmers reached
by an extension program and multiply the
effectiveness of the agronomist.

Properly structured the radio farm forum grOUps could
become permanent units which the agronomist would meet with
at regular intervals. This could serve to extend the
number of farmers contacted by the agronomist, especially
if sufficient effort is made to obtain the participation
of small farmers who would otherwise be unlikely to receive
attention from the extension service.

The radio programming would serve to increase the
awareness and interest of the group in new farming practices.
1A relatively better educated member of the community

(probably a somewhat larger farmer) could act as monitor of

the group and lead discussions following the program. These

317
discussions could not only reinforce the message presented,
but would provide an opportunity for farmers to raise
questions and doubts which seemed important to them. In
some cases these doubts and questions could be answered
during the group discussion. In other cases the monitor
could record the groups' concerns and supply them to the
broadcaster or the agronomist who could attempt to deal
with them on the next program or at the next scheduled
meeting. Finally, when the agronomist would meet with the
groups they would already be aware and at least partially
informed about the topics of discussion.

Local demonstration plots, which could be the outgrowth
of the radio farm forum groups, would be an additional means
of assisting the farmers to develop realistic expectations
about new practices and thus reduce the risk they perceive
in trying new methods. The agronomist should encourage
those farmers whose opinions and judgement are respected in
the community ("opinion leaders") to plant the demonstration
plots. To multiply the usefulness of the demonstration
plots, the extension service could link several of them
together in the form of a tour. A truck or bus could be
rented to provide transportation for farmers who would
otherwise be unable to tour the plots. At each stop the
farmers who planted the plot should be encouraged to discuss
the new practice with his visitors. This approach would have
the advantage of closely paralleling the usual pattern of

communication of new practices between farmers.' (It will

313
be revealed that over 50% of the farmers in the Recife'
survey indicated that their neighbors were the most
important source of information about new practices.)

The gathering of farmers at the trading center on
market day could be another focal point of extension
activity enabling the agronomists to reach a larger number
of farmers. Simple displays illustrating the advantages of
new practices over present practices could be placed near
the center of market activity. Slide—film presentations,
movies, demonstrations (such as proper use of pesticides
and fertilizers) and discussions about new practices could
be presented by the extension service. In the evening the
extension service might provide a program, arranging for the
use of a local theater or even outdoors, which would mix
entertainment with an introduction to improved farming
methods. A mobile team, bringing their displays and equip-
ment with them, could assist agronomists serving several
different locations to implement these programs.

There is also evidence that local input suppliers may
provide a useful source of information about new practices
and could support the extension service's program. Since
suppliers of farm inputs could provide considerable transfer
of information in the course of merchandising their wares,
the agronomist should assist them in becoming well informed
abcnrt the proper use of their products. Profitable use of
run» farm inputs is in the best interest of both the agronomist

arui the businessman. Further, the agronomist could

319

encourage the input supplier to develop displays in his
store and on—going demonstrations (such as feeding trials)
contrasting the use of his inputs with existing practices.

In summary, the extension service should attempt to
increase its impact on the rural sector through programs of
greater use of mass communication methods and greater
emphasis on group or mass meetings as contrasted with
individual contact with farmers. Special emphasis should
be given to making contact between the agronomist and
smaller farmers more frequent, at locations and under
conditions which are convenient for this segment of the
rural population. Finally, the agronomist may find the
farm input supplier as a helpful ally in extending
information about new practices to farmers.

Feducing the__"Risk" of New Methods

 

 

Poorer farmers seem to be reluctant to accept the
"risks" involved in trying new methods. As a consequence,
institutional arrangements which would reduce the "risk"
perceived in new practices, and encourage their use among
the poorer segment of the rural population, should be
given consideration in rural development strategies.
Pregatuns to lessen the instability of producer prices would
Certainly contribute to reducing some of the "risk"
‘aSsociated with new methods. Improving the farmer‘s know«
ledge oi’new methods through greater contact with the
aBITnumnist and the mass media as sources of information

Emnrhd also be of value in alleviating his fears, but the

320

major ingredient of such arrangements undoubtedly involves
some method of insuring the farmer against undue losses from
trying a new practice. Insurance schemes which would
guarantee a certain minimum yield in association with
approved new practices would eliminate the possibility of
disaster and make these practices more attractive to farmers
who can ill afford a crop failure. Where new practices are
divisible, the guarantee might cover only a limited acreage
encouraging the farmer to employ the innovation on a "trial"
basis. Such schemes might be financed through government
funding, slight additions to the cost of farm inputs, or
possibly through insurance premiums paid by the farmers
themselves. The extension service (and researchers) would
need to be particularly sensitive to the profitability of
the practices they recommended. if the approved practices
were indced more profitable, it would seem that an insurance
arrangement of this type would greatly increase the security
of farmers desiring to use new practices at a very low cost
to those financing the scheme. The biggest difficulty with
an insurance scheme in Northeast Brazil would seem to be
the provision of the administrative capabilities for its
initiation and operation, and financial provision for the
effects of periodic droughts.

Alternatively, an insurance scheme could be integrated
with a supervised credit program. Such an approach would
reduce the farmer's fear of borrowing to finance new

practices by assuring his ability to repay his loan. This

321
might encourage the use of credit to purchase the non~farm
inputs involved in new agricultural practices.
92.2.4.5;

Since only a small preportion of the Recife farmers
believed they had obtained sufficient credit to meet their
needs at the present level of technology being applied in
the area, increased use of modern inputs will require a
considerable expansion of credit resources. Expansion of
loanable funds, by itself, would probably be an inadequate
response to the credit problem. Changes in the credit
system appear to be necessary if the productivity of small
(less than 50 acres) and moderate-sized (SO~244 acres)
farms is to advance concurrently with larger farms. A large
preportion of the small and moderate-sized farmers, which
in turn represent the majority of the Recife farmers, were
fearful of using credit under the present system. The use
of land as collateral and the unpredictability of rains over
much of the area surrounding Recife are probably important
factors contributing to their fearfulness. Alternative
approaches to securing loans and provisions for refinancing
(on reasonaole terms) or insuring loans which are defaulted
due to unfavorable crop years would undoubtedly serve to
zilleviate much of the "risk” perceived by potential
luarrowers. In general, expansion of loanable funds and

lr-nges ii the credit system seem to be necessary for

n
E)

fi

rurreasing the productivity of agriculture in the Recife

p.

i D
m

I‘

a)

Input Availability

There is strong evidence to suggest that increasing the
local availability of inputs would stimulate the use of new
agricultural practices, especially among the poorly educated
and poorly financed segment of the Recife farmers. Some of
the Recife farmers seemed to be able to reach beyond local
markets to secure certain inputs, but most of the farmers
are undoubtedly dependent upon local markets to provide
needed supplies. Consequently, schemes to improve the local
availability of modern inputs should be given considerable
attention by the agencies responsible for the development
of the area.

The type of schemes used would depend upon the existing
marketing system for farm inputs. Where private outlets for
farm inputs are present in most rural communities, the most
practical approach might be to encourage the input
manufacturing companies to train the local suppliers in
modern management practices and in the proper use of various
inputs so they could more efficiently serve their clientele.
Either private or public credit could be made available to
local input SUppliers enabling them to stock adequate
quantities of the inputs needed in their communities. Where
this is not a viable alternative, cooperatives or even
gyyvernnent agencies might be formed to supply inputs to
farmuars. In any case, it appears to be necessary to provide
trmrining and financial support to the suppliers of farm

irunats who are usually overlooked in agricultural development

323

plans.

Education

While the previous suggestions represent possibilities
for increasing the pace of technological change in the short
run, improving the education level of farmers seems to offer
an Opportunity for increasing agricultural productivity in
the longer run. Since even a few years of education seems
to contribute considerably to innovativeness, efforts to
provide at least some training to greater numbers of the
rural pepulation might be expected to have a sizeable impact
upon the modernization of the rural sector. This would
suggest that scarce resources for education should be
largely directed toward mass educaticn cf rural children
through a limited number of years (perhaps three to four
years) and some form of an adult education campaign. This
would be in contrast to devoting efforts to what is
probably an unobtainable goal of a complete elementary
and secondary education, and providing such an education to
only the fortunate few. It is undoubtedly possible, through
careful planning of curriculums, to provide both rural
chilriren and adults with the basic skills and concepts
necessJUj'iku'their economic viability in a more technical
socjrzty within the constraints of a limited term of schooling.
fmuflrasis would be placed upon developing needed.basic skills
and making the training relevant to the rural environment.
As an] example, reading could be taught from bulletins on how

to IVrise corn and mathematical concepts could be approached

324

in terms of farm budgeting and accounting.

Suggestions For Furthethesearch

From the author's point of view, the most productive
type of continuing research on the adeption of new farming
practices would be a thorough evaluation of actual efforts
to introduce new farming practices at the community level
in less developed countries. Further refinement on develop-
ment of concepts is likely to make a small contribution by
comparison. The point has been reached where available
concepts are probably adequate, but the question of how to
best utilize what is known is still unanswered. The next
logical step is to examine ongoing programs intended to
stimulate the use of new methods, especially among small
farmers, in order to determine the organizational,
institutional and managerial requirements for increasing
the pace of technological change in the rural sector. The
goal of systematic studies of such programs would be to
determine the best approaches to providing information about
new practices, extending credit, marketing farm inputs, and
reducing the risk of trying new practices in order to secure
a rapid rate of adeptien of more productive farm practices
raith the limited economic, technical, and managerial resources
that less developed countries can devote to this purpose.

Perhaps further investigation of farmers‘ expectations
<mf returns from alternative practices would also be of
'VaIULu A practical research design would be to periodically

exznnine the eXpectations held by farmers from their first

325

awareness of an innovation until they either adopted or
definitely rejected the practice. The researcher would

then be in a position to trace the changes in the farmer's
expectations and the impact of various information sources

in altering their perceptions of the innovation. A better
understanding of the process of expectations formation, by
concentrating on this element of the decision-making process,
might provide some useful insight on how the informational
input of development pregrams could most efficiently be

presented.

BIBiIC'fiPAF’lIY

BIBLIOGRAPHY

Adams, Dale W., "Resource Allocation in Traditional
Agriculture: Comment,“ Journal of Farm Economics,
November, 1967), pp. 930—932.

 

Alchian, Armen A., ”Uncertainty, Evolution, and Economic
Theory," Journal of Political Ecgnpmy, (June, 1950),
pp. 211-221.

 

Arndt, Johan, "New Product Diffusion: The Interplay of
Innovativeness, Opinion, Leadership, Learning,
Perceived Risk and Product Characteristics,"
unpublished paper, School of Business, Columbia
University, 1967.

Bauer, P. T. and Yamey, B. S., ”A Case Study of Responsive-
ness to Price in an Underdeveloped Country,"
Economic Journal, Vol. 69, (December 1959).

 

Real, George M. and Sibley, Donald N., "Adoption of
Agricultural TechnOIOgy by the Indians of Guatemala,"

Rural Sociology Report No. 62, Iowa State
University, 1967.

Bean, J. A., "Reaction to Uncertain Price Expectations,"
Journal of Farm Pconomics, (February, 1955), pp. 90 95.

—‘———- ...-

 

Bond, A. Dewey, "Producer Responses to and Attitudes
Toward Federal Price Support Programs for Potatoes,”
unpublished Ph.D. Thesis, Michigan State University,
1953.

Boulding, Kenneth E., Thgmlmagef Ann Arbor Paperbacks, The
University of Michigan Press, 1966.

Boussard, Jean Marc and Petit, Michel, "Representation of
Farmer's Behavior Under Uncertainty with a Focus—Loss
Constraint," Journal of Farm Economics, (November,
1967), pp. 8697330. ‘

 

Braybrooke, David and Lindblem, Charles E., A Strategy of
Decision:' Policy Evaluption as a Social Process,

1.‘*'"7"""r--1T"‘——H _.—-'._' -" - .° - ‘ ,‘ -" ‘
hew lora, the free Press of ulencoe, 1963.

 

 

328

Brownlee, O. H. and Gainer, Walter, "Anticipation and
Uncertainty in larm ilanning,” Journal of Farm Economics,
(May, 1949), pp. 266 -275.

 

Carroll, Tom W., "Diffusion Research: Application to
National Surveys in Developing Countries, " unpublished
paper, Department of Communications, Michigan State
University, 1968.

Castillo, Celia, "Commeumt A Critical View of a Subculture
of Ieasantry" in Clifton R. Wharton, Jr., Ed.,
Sunristente iletU1tUTC and Economic DevelOpment,

d-fl‘F

thizagéj Ald1ne.ublish1ng Cempany, T569:_pp. 1§6u14l.

 

Coward, F. Walter, Jr., "Dimen sions of Adoption in the
Commerci.a lization of Agriculture in a Developing
Nation," paper presented at Rural Sociological Society
Meeting, San Francisco, California, August 26-28,
1967.

Davidson, Jack R. and Mighell, Ronald L., "Tracing Farmers'
Reaction to Uncertainty," Journal of Farm Economics,
(August, 1963), pp. 577-586.

 

Deutschmann, Paul J. and Pals Borda, Orlando, "Communication
in an Andean Village,” paper presented to the Associa-
tion for Education in Journalism Convention, University
of North Carolina, Chapel Hill, August 27, 1962.

Dillon, lohn L. and Heady, Earl 0., "Free Competition,
Uncertainty and Farmer Decisions, " Journal of Farm
Iconon1cs (August, 1961), pp. 643-651.

 

Richer, Carl K., "Transforming Traditional Agriculture in
Southern Nigeria: A Contemporary Experience,"
paper presented at the Annual Meeting of the African
Studies Association ,Bloomingtcn, Indiana, October
26-29, 1966.

.1- , "Comment: The Subsistence Farmer in Traditional
liconomics" in Clifton R. Wharton, 11., Ed., Subsistence
Agricu ture and Lee nemic De.clopment, Chicago, Aldl ne

-0—o_

ubllfih1ng Compan), 1969, pp. 22/~228.

 

Bidman, Vern E., "An Application of Statistical Decision
Theory to Commercial Turkey Production," Journal of
FUTJ‘IA 11C DOCEICS, (November, 1967), pp. 852’868.

 

 

FAQ, The Stilte of Food and Aoricultuie, Rome, (1959),
pp.13/vldi.

 

329

Feder, Ernest, ”The Latifundista Puzzle of Professor Schultz,
Comment,” Journal of_£arm Economics, (May, 1967),
pp. 507-510.

 

 

Friedman, Milton and Se vage, L. J., "The Utility Analysis of
Choices Involving Risk," J01:rnal of Political Economy,
(August, 1948), pp. 279- 393.

 

_. , "The Expected Utility Hypothesis and Measurability
of*Utility, " gougnnl of Political Economy, (December,
1952), pp. 463 474.

 

Furtado, Celso, The_:£cnorlic G1owth of Brazil, Berkeley and
Los Angeles, Unixe1sit) of Califolnia {3053,1963.

 

 

Grunig, James E., Information, lntrenreneu1sh1p, and Economic
DeveIOpnent: K Stnd7'01L tne Decision 1111194Processes
"thodofihian Kat1-aTdistas unpubIXsTgT—ph D. disserta-
tion, University oi 'nisconsin, 1968, pp. 9-12.

 

 

 

 

 

Halter, Albert N., "Measuring Utility of Wealth Among Farm
M1n3ge15," unpublished Ph.D. Thesis, Michigan State
Univelsity, 1956.

H"- ____, anl Beringer, Christoph, ”Utility Functions and
la1u1gtrial Beh:1vior, ” dgurnalkgf Farm Economics,
(lehruary, 1960), pp. 118-132.

 

, and Dean, C. w., “Use of Simulation in Evaluating
Managerial Practices Under Uncertainty Application to a
Large Scale Ranch," Journal of Iarn Economics, (August
1965), pp. 557-573.

 

Harrison, Kelly Max, ”Agricultural Market Coordination in the
Economic Development of Puerto Rico," unpublished Ph.D.
Thesis, Michigan State University, 1066.

Hathaway, Dale, Michi qnn Fa rmers in the Mid- Sixti.es, Research
Report Lo. 54, Mi 1:111t1ra1 111011111111. Statlon, Michigan
State University.

 

, Earl 0., Economies of \ingultur l IIEQBCt1°n and

RTSOELSC Use,11g1cne d Cliiis, Pientice-Hall, Inc.,
’32 , pp. W439 461.

Helleiner, Gerald, "The Fiscal Role of the Marketing Board

in Nigeri 1an Economic Development, 1947*61," Economic
'Jou1nal, (September, 1964), pp. 582~610.

 

Hicks, J. R., value and Capital, Oxford, Oxford University
Press, 1939.

330

Hildreth, C., "Problems of Uncertainty in Farm Planning,"
Journal of Farm Economics, (December, 1957),

pp. 1430-1441.

 

Hirschman,.A1bert 0., Journeys Toward Progress, New York,
l'wentieth Century lund 1963.

 

 

Hopper, W. David, "Allocation Efficiency in Traditional
Indian Agriculture," Journal of Farm Economics,
(November, 1965), pp. 611—624.

 

Johnson, D. Cale, F071? rd Prices for Ag11rnltu1e Chicago
Uni1eisity of Chicago Press, 1917.

Johnson, Glenn L., "The Friedman-Savage Utility Hypothesis
in the Inter-state Managerial Study," Journal of Farm
Economics, (December, 1955), pp. 1110-1114.

 

 

, and Lard, Curtis F., "Knowledge Situations" in
"- Glenn L. Johnson, et al., Managerial Processes of
Midkestern Farmers, Ames, The Iowa State UBYEETEEty
tress, 1961, pp.-H1- 53.

 

 

Johnson, Paul R., "Do Farmers Hold a Preference for Risk,‘
Journal of Farm Economics, (February, 1962), pp. 200‘
297.

 

Jones, 1. 0., ”Economic Man in Africa,‘ Food Research
Epstitute Studies, Vol. 1, (113), 1960), pp.107~134.

 

 

(atona, George, Psychological Analysis of Economic Behavior,

.__../-»—— .. -.f—-~-— --

New York, McClaw-Hlll Booli Company, Inc., 19“63.

 

Knight Frank H., Risk, Uncertainty, a.1d Profit, New York,
Haiper Torch Bool;;, Harper and'Row, Publishers, 1921.

 

Krishna, Raj, "Agricultural Price Policy and Economic
Development” in Herme .n M. Southworth and Bruce F.
Johnston, Fds., Agricultur2l Devel_opment nd _Economic_

Grow1h, Ithaca, Cornell University Press, 1967,
pp. 1“7 540.

 

Langham, Max R., "Game Theory Applied to Rice Farmers,"
Journal of Farm Economics) (February, 1963), pp. 151-

..o- ~—.

162.

 

Larson, Donald Walter, "A Diagnosis of Product and Factor
Aarket Coordination in the Bean Industry of Northeast
Brazil," unpublished Ph.D. Thesis, Michigan State
University, 1968.

CC; York, JOhR Wiley and Sons, Inc., 1967.

331

Mellor, John W., "The Subsistence Farmer in Traditional
Econoxnies” in Clifton R. Wharton, Jr., Ed.,
Subsistence Agriculture and Economic Develogment

Chicago, .\ldine Publishing“ Comipany, 1969, pp.1209*
226.

 

 

 

Morgan, Bruce W., An Introduction to DayCSian’StatiStical
Decision Processes, anlewooJClitfs, Prentice«Ha11,

1nc., 1968.

 

 

Morrison, T. E. and Judge, G. 6., “Impact of Price
Expectation by Poultry Farms, " Joutnal_pf Farm
Economics, (November, 1955), pl 63 2063.

 

       

Nair, Kuse.m,813§sgms in the Dust, New York, Frederick A.
Piaeger,”19ol.

 

 

Niehoff, Arthur H., A Casebook 01 Socia 1 %(h nge, Chicago,
Aldine Publishing (onyaux 1900.

Nielson, James, "Aspects of Mana agement of Concern to the
Basic Researchers, " Desc1iling and Measuring Managerial
Abilities and Qermce‘sfl'iff‘lolt No. 4’, [arm Management
pG%(dlCl_C6—Hlb$10n ol host eln Agricultural Economics
Res*arch (ouncil, Denver, Colorado, 1962.

 

 

Officer, R. R. and Halter, A. N., "Utility Analysis in a
Practical Setting,” Aggrican Jougnal of Agricultural
Economics, (Hay, 1968), pp. 257-277.

 

 

A Report of the President's Science Advisory Committee,
The World Food Problem, Report of the Panel on World
F663'365§177“v01. l, The White House, May, 1967,
pp. 1-6.

 

Robock, Ste fan U. Brazil's Develojing Northeast, Washington,
D. C., lle Broolings 121t1t1:tloz1, 1963.

 

Rogers, Fverett M., iffus ion of Innovations, New York, The

_.

Free Press f GiCHCO\;”1962.

 

_, Modernization Amongfifeasants, New York, Holt,
Rinehart and hinston, Inc., 1909.

 

, "Develoning a Strategy for Planne Ad Change,"
paper presented at the Symposium on the Application
of Systems Analysis and Managemen Techniques to
£111 c..tional. Planning in California, Orange,
California, June 12~13, 1967.

Schickele, Rainer, "Farmers' Alaptation s to Income Uncertainty.
Journal of Farm nooru~1c~, (August, 1930), pp. 356*374.

w—gr-D-.. “h--O_-

 

332

Schultz, Theodore W., T1ansforming Traditional AgriCulture,
New Haven, Yale Univeisity Press, 1964.

 

 

, "The Latifundia Puzzle of Professor Schultz:
Reply," Journal of Farm EConomics, (May, 1967), pp. 511~
513. ‘

 

 

, "Resource Allocation in Traditional Agriculture:
Reply," Jou1nal of Farm Economics, pp. 9331935,

 

 

Shackle, G. L. 8., Igpectations in Economics, Cambridge,
Cambridge University'P1ess,11949.

 

Simoz1, Herbert R., "Theories of Decision-Making in Economics
and behavioral Science," American Iconomic Review,
(June, 1959), pp. 251- 283.

 

 

Simon, Julian, "The Effect of Economic Status and Traditional
Culture Upon Economic Behavior: An Optomistic Comment,"
American Journal of_§gripultural Beonomics, (May, 1968),

---—.—.- -———---—.—_—-.—-—-‘

pp. 438~4Sl.

 

Singer, H. M., "Introductory Statementw-Stabilization and
Drvelopment of Primary Prodz1cing Countries, Symposium
]I, “ _Xlzlo.:, Vol. XII, (1959, Pasc. 3), pp. 269~276.

Singh, Ram V. and Waz‘low, G. 8., Characteristics'of Fa.rm
Innovations Assofl :11ed hit the Rate of 1\Joi_t_1on,

Unive1s1t) oi Guelph, uepa1tmen_51 Ixten510n
Education, Report #14, October 1966.

 

Slater, Charles C.; Riley, Harold M.; Farace, Vincent;
Harrison, Kelly; Neves, Ferdinando; Bogatay, Alan;
Doctoroff, Mark; Larsen, Donald; Nason, Robert and
Webb, Thomas. Market Processes in the Recife Area of
Rplthca st BraziI, Past Lansing,1esear(:H—Report No. 2,

market ting in Developing Communities Series, Michigan
State UniVCISity, 1969.

 

 

 

Soncu s

 

, Jol1n A. and Morgan, James N., [be Detection of
1111tion Fffec1s Ann .\1bor, Institute for Social

1 t
.13.? 5:1-
Re1sca1cn, bnixers1ty of Michigan, 1964.

Stern, Robert M., "The Price Responsiveness of Egyptian
Cotton Producers," Kvklos, V01. 12, 1959.

Sturt, Daniel, ”Response to Change in Pakistan, 'JoUrnalv
of Farm Economics, (August, 1965), pp. 625 633.

 

Tendler, Judith, Agricultural Ciedit in Brazil, U. 8.
Agent) 101 -dtCLRdLlUndL Development, Washington, D.C.,
October 1969.

 

 

333

Thomas, D. Woods, "Sociological \spects of the Decision
Process, " Jou1na1 of Farm Economics, (December, 1955),
pp. 1115 1118.

 

Tisdell, Clement Allan, The T11eory of Pri Ce Uncertaintv,
Productionja and Profit, Princetcn, Princeton University
Press, 1968.

 

 

Welsch, Delane, "Response to Economic Incentive by Abakaliki
Rice Farmers in Eastern Nigeria," igurnal of Farm
Economies, (November, 1965), pp. 9001913.

 

 

 

 

Wharton, Clifton R., Jr., Ed., oulsistence Aericulture and F
Economic. 1101':le ment, Chicago, Alcfine hfolisning Company,
1969.

_ , "Malay Rice and Rubber Smallholders: Economic
Man Revisited,” Agricultural Development Council, Inc.,

 

_____._m, "Risk, Uncertainty and the Subsistence Farmer: ;
Technological Innovation and Resistance to Change in

the Context of Survival,“ paper presented at the Joint

Session American Economic Association and Association

for Comparative Economics, Chicago, Illinois, December

28, 1968.

Williams, D. B., "Price Expectations and Reactions to
Unce“tainty by Farmers in illinois,” Journal of Farm
F<01.o:nics, (February, 1951), pp. 20-37.

0— ~-———...-.-—'———-

 

Venkareddy, Chennareddy, "Production Efficiency in South
Indian Agriculture," Journal of Farm Economics,
(November, 1967), pp. 816- 820.

 

m nd nWergenster1 Oskar, Theory of Games

Von Neumann, John .
Behavio1, New York, John hiley and Son-.

and Peonomic
194$.

 

 

AI’I‘E‘JD 1X

APPENDIX A

AID I»‘«Use of Fertilizer Among
850 Francisco Rice Farmers

 

Variables . Group #1.. Group #3 ..Group #5.. Group #7
Tenure .0157 .0001 .00549 .00420
Income .04296 .0315 .03191 .02597
Credit .03651 .02840 .03:56 .00833
Try 2nd .0090 .0182 {.03367} .01515
Risk Prof. .0191 .0094 .02147 ‘ .02843
Education {.1000} .0321 .02151 '{.03406}
Chem. Fort. .02939 .0050 .01634 .01515
Info 1 .12317 .0215 .02848 .02634
Info 11 .0309 {.0642} {.03367} .01515
Fatalism .0009 .0011 .00104 .00617
Luck vs. Know. .01111 .0119 .00835 .01473
Att. t/w N.T. .00677 .0174 .01734 .00869
Def. Income .00858 .0117 .01362 .05:64

 

 

 

Proportion of variation in each group explainable for each
predictor (BSS/TSS)i
V0 ' 1 1' ~-° 1

opllt mace on tJlS vaigaale.

{ }Next-bcst 838.

04
LN
'J“

APPENDIX B

AID II**~Use of Insecticides Among
850 Francisco Rice Farmers

 

a :3 C) (D C) an
>3 >1 *1 *3 H H
o o o o g g
.5 ~S >5 -5 'o w:
Variables y: c: i: i: :3 f:
. .. H
-- 'V ~.
Tenure {.13205} .13054 .17346 .01617 {.10390} .0849?
V
Income .07755 .05858 .05186 .06494 .12707 .02477
V

Credit .13402 .00383 .03713 .00649 .00000 .00380
Try 2nd .00050 .00213 .02101 .05500 .00223 .07353

Risk
Pref. .00873 .0052 .01303 .07059 .04058 .07353
v v
Educ. .11709 .19871 .01773 .00649 .06250 .14216
Insec-
ticides .00057 .00029 .00240 .00509 .00017 .03992
v

Info I .06619 {.14613} {.152351 .21130 .07353 .0563?
Info 11 .05529 .07608 .000426 .03030 .06250 .02477
Fatalism .00741 .00669 .04267 .06481 .00444 .02477

Luck vs.

Know. .03472 .06166 .09874 {.09091} .02500 .01535
Art. t/W

N.T. .03641 .05855 .01536 .00260 .00586 .01018
Def.

Income .00648 .01457 ..00440 .04538 _ .00009...0277a

 

Preportion of variation in each group explainable for each
predictor (BSS/TSS)i

V _. . .
Spllt made on this variable.

{ INextsbest BSS.

Jill]:

71

”(a

.msopw mafia cw pneumaoo mw ofinnfium>

.mmm pmmp-uxoe.
r

.ofinmwhm> mflgu do owme ufiamm

w

>

H . .
.mmmh\mmmc nepuflwoum some new oamezflmfimxo about come a« coflpmflhw> we cofiuFOQOMA

 

 

mammo._ mefimo. cocoa amouo. nmmeo» . ma.~:. enemas. maoucH .«mq
momma. mmmfio. mmmoo. newec. mNHNo. cmewo. mcomo. .9.2 3\w .uui
omHHH. Hmnao. mmfioo. meson. ovooo. mmHHo. ommca. .Bocm .m> Mum;
mvaoo. w “No. mmmao. eamoo. mwmflo. owvoo. vumoo. Emwmwuam
wmmmo. Amnmmo.w cmono. mwooc. fiwmmno.w mummvfl.w chemo. HM one“
mNmmo. flames. waved. mmcmco.w mOVHH. NHMOH. onmmo. H omzH
p D

us. all ...v ...0 In... ...... vaNH. .nu'OwUnuHh.

oncoo. HHNNQ. Hamo. momma. omvvo. mmmoa. mummo. defluJuzwm
D
vwmvo. Hammo. fiemmmo.w. nommo. naom . ooooo. nmmoo. .mogg Mafia
onoco. Noomo. namoo. oocce. momma. mmvoo. mmmwo. USN xhk
nooom. mvnmfi. «moao. Noemc. mnemo. ncmmo. omNHo. wwweyu
p
mmmmo. mnemo. owono. evwmo. o.omo. thma. ANOVOH.V oEoucH
>

05000. Hemvo. mmcoo. “memo. NmmHo. mnmoo. mmono. ohncoh
n» asouo o» macho we asepu ma amoeu an macho ma muonu Ha msouc mofinmfiam>

whoeumm oofim oomwocwum 0mm acoe< mpouomph mo own--:HHH QH<

U xHozmmp<

338

H .
.Amme\wmmv Ho

poadokm some how

.oHpmflww> was“ no mums uwHam

.mmm ummn-uxozfi D

D

ofipmcfimfimxm msowm some ca :efiumwuw> mo :ofiuHomopg

 

 

Naeoo. .amooo. cameo. Omega. Howmo. mecca _ wwNoO. .xeom
mflmoo. enmec. nmmoo. mmomo. mmnmo. ommoo. mwmmo. .pu<
Hmomo. memfio. mmmoo. mwwoo. Nacoo. Nmmmo. emamo. .zocx .m> :u.:4
noofio. moaeo. gmmoo. ommmo. mfiooo. moeoc. mvaoo. Emmfispme
moemo. ommmc. cwmao. <5. mo. Newman mmnmo. woflmfi. HM owcH
ommoo. Hemp“ Nuanm oemmfi. Amnmmo.w mmmwo. meemc. H cmch
D D
monmo. omHeH. ecmoc. moflmw. mnmmo. oevmo. Nmoeo. xovcH u "NH
Aonmwo.w mwoao. \mooo. unmoo. mflooo. mmnao. muqnfi. . use m
D
Nwwmfi. Noamo. omNNc. mo co. Nommo. mNHeo. HNOHO. .weeu xmwm
D
ommno. me\oo. Aomfimo.w a mo. momma. ooooo. memfio. wan xpk
Namvo. :0 No. oooeo. e mm. «meow. AmmmCH.w mwmmwa.~ ufiwehu
D
mmmoo. mmomo. mmfimc. n ma. mhmoo. DVHHH. mmmmfi. osoecw
D
ofiu asopo m» macho c» msouu Na scum mm macho. ma usouc Nu msouu moHanHmD
mpaEpnw «x oomw mew omw "xvvmw :oaum>o::H--->» cH<
a Mmmzmmm<

.QSOHm m

0H

.oflnmflhm> mwnu co owns wflqaw

nu cw uzwumeoo ma oflawwhmD

.mmm umop-uxozfi D

 

 

D
flammw\w 0 avefiwopm come Hem omnwcwwmmxm ozowm auwe an cowpmflum> mo moflwasmonm
wummo. mmmoc. a mac. ommmo. mmmoc. Hmfioeo. esoocw .wom
emsmo. HMfioo. ammoo. magma. Nmafio. mmomo. .~.z z\p .ou<
mmwwo. nwoqo. mcaao. HeOHo. omfioo. wmooa. .aocx .m> xosq
heave. “Heme. mmwfio. maeu.. eooao. womao. Equmumm
Hmmho. mmmmma.w mmmu . owowo. Ammeoo.v OmeH. HM own”
D
9
”w mmoOHH.w Howmc. whomo. mmvmo. “ammo. emmmc. H emcm
ommoo. emmmo. wfimmo. mNmHo. chemo. Neomo. xmwcw unmcH
D
cameo. mmwco. --- ommfio. mmmmo. fiamomfi.w coaumonem
D
mHHoo. momma. vuoao. «some. HHNHO. eomoo. .woum xmflm
Hmmoo. Hmfioo. momeo. ooooo. mnmfio. “Nome. new >99
mfimmo. ommcm. emmmo. AQHeco.V nmooo. HHOHH. uflcmpu
>
mHeMH. HHooo. mmmoo. mfimmo. mmemo. oenuo. mEooaH
D
mmwmo. ommmo. momoo. Heemc. owmoo. ecows. opncoh
ma msouu an macho e» DDOHD mu asouu ma mnouu Ha nacho monnwwudD

whoeaam :wom

”xomzH

m XHQmea<

newpw>occH---> mam

inued

E-Cont

.LHD IX

1‘ “0

AP?

{42‘

--Innovation

‘Y
\

ID

L

:3

Grouo #21

up #19

Gro

#3

‘ .
I\A"L)

I

p #14

‘
Q

Croz

Variables

 

“54

k

.01

.02524

.02716

r-1
C.)
C)

r4

1’)
CO

6.70

.00

\D
P!

CDC)

'.\
\-

L‘JT

f.‘
I.
\

C"
CD

I":

20

35

N

-C\!

CO
r4

.07147

9

l“

.07

J-)
0".

lulu—u;

f'

g)

.00066
.00821

‘7
C.
bx
{’7

.00583

3
.02538

\O

\C;

8
.00345

.02040
.00442
.04005

sk Pref.

i.

Try 2nd
R.

.0054

\C)

Ln
(\1
C:

‘
5

catiox
Input Index

\0

.00007

.00422

7

t")
(\l

.00

If)

C:
r"!

'K

\L;
00
(‘0

LJ
('3

.02409 .

1

.0966

C)
74
f) m

r4

40

‘1?
[x

’-
\

7017

.00

.(

(\I

.01612

I")

.01747

.07279

II

0
L; 4
'4
04
hi

5?

.086

..,
.3

1
212

.022

I’)
t’)

.-
k

(4-

V

(\1'

Li“;

C.;

('\I
0."
Pi

\"\
)‘v

r'fi

(‘J
C)

P‘
\o

\O

CO

801

\ o

.r‘
oUu

_n

Know.

VS .

I“
s—I.

A3

I

6

,
|\

.01442

.01169

.06962

232

0

Pi

.00881

{\
t0
0;

P4

.09210

(\
v
Pd
"'1'

 

'w—y