31293010199696

 

 

 

'9 llflWlWflflWiflflfliWWW L ,,.,

ABSTRACT
METHODOLOGICAL PROBLEMS IN THE MEASUREMENT OF VALUES

by Henry James Watts

This thesis deals with some of the problems in the
development and use of a quantitative instrument designed
to measure values in different cultural settings. A frame-
work for the development of such an instrument is presented.
This framework is based on some of the literature dealing
with concept formation, scientific method. and the definition
of values. Throughout the thesis, the relationship between
theory and method is emphasized. Eventually. a conceptual
framework is developed which attempts to state the logical
implications of an instrument which assumes that samples of
varying degrees of industrialization will respond in a
predicted manner depending on where they fall along a
'hypothetical continuum of industrialization.

The hypothetical predictions were tested by a
secondary analysis of data Obtained by the repeated
administration of the same instrument to three different

samples. The instrument used was A.O. Haller's MSU work

Henry James Watts

Beliefs Check List. This instrument consists of 44 items
which are sub-divided into six sub-scales. The six sub-
scales focus on orientations toward work, structured time.
physical mobility, change, internal or external determin-
ation of events and deferred gratification. All of the
samples consisted of high school boys. The locations of the
three samples were Lenawee County. Michigan. Turrialba.

Costa Rica. and Lansing, Michigan, and the sizes were

439. 112 and 87 respectively. The boys of the Lansing sample
were Mexican-Americans.

The data was analyzed by using means, variances:
product moment correlations, and principle axes factor
analyses with varimax rotations.

The findings indicated that only the Lenawee sample
yielded scores consistent with the hypothetical predictions
on all sub—scales of the instrument. Both the Turrialba and
the Lansing samples showed considerable variation of responses.
Possible reasons for the variations were discussed both in
terms of shortcomings of the instrument and in terms of the
possibility that the value orientations are not distributed
as predicted. Finally, some of the implications for

future research were discussed.

METHODOLOGICAL PROBLEMS IN THE MEASUREMENT OF VALUES

BY

Henry James Watts

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Sociology and Anthropology

1962

5.25'944

<57?! NJ

ACKNOWLEDGEMENTS

For making their data available to me for secondary
analysis. I am indebted to A. 0. Haller. Arturo De Hbyos.
and Charles P. Loomis. Support for the original collection
of data was provided by the United States Office of Education.
Department of Health. Education and Welfare; Michigan State
Agricultural Experiment Station; the Carnegie Corporation;
and the Research Grants Branch of the Division of Hospital
and Medical Facilities of the United States Public Health
Service. I wish to thank International Programs for direct
financial support in the analysis of the Turrialba data.
through a grant to Dr. A. O. Haller.

Thanks are also due to the Michigan State University
Computer Laboratory and the IBM Tabulating Department of
NHchigan State University. Mrs. Dorothy Tervo provided
great assistance in the arrangement and typing of tables.

The members of my committee who aided in the planning
and the completion of graduate work as well as this dissertation.
were Charles P. Loomis. J. Allen Beegle. Jay W. Artis. Iwao
Ishino and Charles F. Wrigley. My greatest indebtedness is
to A. 0. Haller who served as chairman of my committee.

major professor. friend. wailing wall and a continual source

of encouragement.

ii

TABLE OF CONTENTS

Chapter Page
I. INTRODUCTION . . . . . . . . . . . . . . . 1
Statement of the Problem 1
Importance of the Problem 3
Organization of the Thesis 5
II. CONCEPT FORMATION AND SCIENTIFIC METHODOLOGY 7
Purpose of this Chapter 7
Explanation 7
Laws. Causality and Prediction 23

Empiricism. Operationism. Verification
and Meaning 28
Theory 37
Scientific Method 39
III. TOWARD A WORKING DEFINITION OF VALUES . . . 42
Purpose of this Chapter 42
Problems in the Definition of Values 42
Attitudes 46
Beliefs. Opinions. and Attitudes 49
Values and Value Orientations 53
Toward a WOrking Definition 68
IV. METHOD . . . . . . . . . . . . . . . . . . . 73
Purpose 73
The Instrument 73

The Typological Basis of the Work

Beliefs Check List 81
The Samples 85
Predictions and Method of Analysis 89
Techniques for Data Analysis 94
V. RESULTS . . . . . . . . . . . . . . . . . . 98

iii

Chapter
VI. SUMMARY'AND CONCLUSIONS . . . . . .

General-Summary

Discussion of the Alternative
Conclusions

Limitations of the Study

Implications for Future Research

BIBLIOGRAPHY . . . . . . . . . . . . . . . .

iv

Page
147
147
151
152
155

158

Table

10.

11.

12.

13.

14.

LIST OF TABLES

Factor matrix for WBCL Sub-Scale l.
Lenawee sample (N-439) . . . . .

Factor matrix for WBCL. Sub-Scale l.
Turrialba sample (N—112) . . . .

Factor matrix for WBCL Sub-Scale l.
Lansing sample (N-87) . . . . . .

Means and variances for all samples for

Sub-Scale l . . . . . . . . . .

Factor matrix for WBCL Sub-Scale 2.
Lenawee sample (N—439) . . . . . .

Factor matrix for WBCL Sub-Scale 2.
Turrialba sample (N-112) . . . .

Factor matrix for WBCL Sub-Scale 2.
Lansing sample (N-87) . . . . . .

Mean and variance for all samples for
Sub-Scale 2 . . . . . . . . . . .

Factor matrix for WBCL Sub-Scale 3.
Lenawee sample (N—439) . . . . .

Factor matrix for WBCL Sub-Scale 3.
Turrialba sample (N-112) . . . .

Factor matrix for WBCL Sub—Scale 3.
Lansing sample (N-87) . . . . . .

Means and variances for all samples
for Sub-Scale 3 . . . . . . . .

Factor matrix for WBCL Sub-Scale 4.
Lenawee sample (N—439) . .'. . . .

Factor matrix for WBCL Sub-Scale 4.
Turrialba sample (N-112) . . . . .

Page

100

100

101

105

107

107

109

112

113

113

114

115

118

119

Table Page

15. Factor matrix for WBCL Sub-Sample 4.

Lansing sample (N—87) . . . . . . . . 121
16. Means and variances for all samples

for Sub-Scale 4 . . . . . . . . . . . 122
17. Factor matrix for WBCL Sub-Scale 5.

Lenawee sample (N—439) . . . . . . . . 123
18. Factor matrix for WBCL Sub-Scale 5.

Turrialba sample (N-112) . . . . . . . 124
19. Factor matrix for WBCL Sub-Scale 5.

Lansing sample (N-87) . . . . . . . . 125
20. Means and variances for all samples

for Sub-Scale 5 . . . . . . . . . . . 127
21. Factor matrix for WBCL Sub—Scale 6.

Lenawee sample (N—439) . . . . . . . . 128
22. Factor matrix for'WBCL Sub-Scale 6.

Turrialba sample (N—112) . . . . . . . 129
23. Factor matrix for WBCL Sub-Scale 6.

Lansing sample (N-87) . . . . . . . . 130
24. Means and variances for all samples for

Sub-Scale 6 . . . . . . . . . . . . . 131
25. Factor matrix for WBCL. all items.

Lenawee sample (N-439) . . . . . . . . 136
26. Factor matrix for WBCL. all items.

Turrialba sample (N-112) . . . . . . . 139
27. Factor matrix for WBCL. all items.

Lansing sample (N—87) . . . . . . . . 141
28. Factor matrix for WBCL Sub-Scale. total

scores. Lenawee sample (N-439) . . . . 144

vi

Page

Table

29. Factor matrix for WBCL Sub-Scale total
scores. Lansing sample (N-87) . . . . . 145

30. Factor matrix for WBCL Sub-Scale total
scores. Turrialba sample (N-112) . . . . 146

vii

LIST OF APPENDICES

APPENDIX Page
A. CORRELATION MATRICES FOR ALL SAMPLES . . . . 163
B. FACTOR MATRICES FOR POOLED SAMPLES . . . . . 192

C. THE MSU WORK BELIEFS CHECK LIST AND
SCORING KEY . . . . . . . . . . . . . . . 197

viii

CHAPTER I
INTRODUCTION

Statement of the Problem

Current attempts to measure values indicate an
increasing trend toward the development of quantitative
techniques. The implication of this trend is that the use
of more sophisticated techniques will improve the precision
of the measurement of values. This appears to be a legiti—
mate position when the limits and boundaries are kept in
mind. In the extreme case. some researchers allow the
technique to become an end rather than a means to an end.
When this occurs. numerical data is sometimes believed to
have actual meaning rather than symbolic meaning. Within
this context. any attempt to interpret the data is ruled
out on the grounds that such interpretations deviate from
the scientific canon of objectivity.

However. the scientific method is usually stated in
a way which includes both measurement and theory. An under—
lying theme of this theSis is that the controversy over

”quantitative versus qualitative.’ is not only unnecessary.

but is in fact. a false dichotomy in the domain of the

measurement of values.

Regardless of the quantitative procedure employed. it
is impossible to measure a non-observable variable without
specifying or defining the variable. Thus. any attempt to
measure values must begin with an attempt to define values.
This turns out to be a difficult task since there is a wide
range of definitions in the literature. Since values are
inaccessible to direct measurement the researcher must con—
cern himself with operational definitions and theoretical
constructs in developing a quantitative instrument. Once
the quantitative data has been gathered and analyzed by any
of a variety of techniques. the meaning of the data must be
imposed by relating it to other bodies of data. Meaning
does not emerge from the data in a final unambiguous form no
matter how sophisticated the method appears to be.

The problems of measurement and interpretation are
increased when we are concerned with the measurement of
values in cross-cultural samples. Following the terminology
of Sears: PTranscultural variables are variables that can
be measured in all cultures,"1 It is the present contention
that transcultural variables in the general area of values

(hereafter referred to as transcultural value variables)

 

1Robert Sears. VTranscultural Variables and Conceptual

Equivalence.9 in Bert Kaplan. (ed.). Studyinngersonality
Cross-Culturally (Evanston. 111.: Row Peterson and Co).l961,

p. 45.

have an important place in sociology. As yet. however. we
know very little about how to measure them in such a way as
to account for variation among cultures and among individual
participants in these cultures. Yet many of our hypotheses
require tests based on valid instruments measuring such
variables. The present thesis attempts to state the logical
implications of a theory of transcultural value variables
for the relations among responses to questionnaires designed
to measure them. If the theory which predicts these relation-
ships is valid and if an instrument designed to measure the
component transcultural value variables is valid. then
certain relationships among means. variances. correlation
coefficients and factor structures of the responses to the
questions should be found to exist. If the predicted
relationships obtain. then the theory which posits the
existence of the variables must be accurate. and the instru-
ment designed to measure them is also valid (in the sense
that it satisfied these--but not necessarily a11--criteria
of validity). If the predicted relationships do not obtain

either the theory or the instrument or both are invalid.

Importance of the Problem

The problem is important. then. for several reasons.

(1) The method will permit the researcher to reject as

invalid transcultural value instruments which fail to meet
the criteria. thus opening the way for development of new

and more precise instruments. (2) Because the method can
detect items which are invalid. its use should help in
refining instruments proposed to measure transcultural values.
(3) When repeated attempts to measure a certain proposed
transcultural value variable fail to meet the criteria. the
proposed transcultural value variable itself (or the theory
of which it is a part) may be called into question.

At present it is almost impossible to do any of the
above. The usual external criteria of validity. prediction
of variation in overt behavior or correlation with instruments
of known validity. are unavailable to us. In essence the
proposed criteria are an extension of internal techniques
for testing validity.

The method investigated here is limited in that it
does not provide a coefficient of validity. that it does
not provide for external tests of validity. and that it
does not test for scalability. Moreover. inasmuch as it is
concerned only with validity it does not provide ways of
testing other important characteristics of instruments such
as reliability. It is not. therefore. a blueprint for
completely evaluating instruments designed to measure trans-

cultural value variables.

Organization of the Thesis

As pointed out above. one of the first concerns in
the measurement of values is the problem of definition.
However. before dealing with this problem directly. the
literature on concept formation and scientific method is
examined in more general terms. The purpose of this is to
focus on some of the general principles which are advocated
and to state the present writer's opinions in regard to
some of these principles. The problem of defining values is
dealt with directly in chapter three. by examining some of
the literature in this area and attempting to formulate a
working definition.

Chapter four examines the instrument and the samples.
In addition. a series of hypothetical predictions are developed
for the purpose of testing some of the underlying assumptions
of the instrument. These assumptions involve the uni-
dimensionality of scales as well as the relationship of the
three samples (Lenawee. Turrialba and Lansing) to a
hypothetical continuum of industrialization. Therefore.
the chapter includes a section on the factor analysis used
in the study as well as a brief discussion of the typological
basis of the instrument (i.e.. A. O. Haller's. The MSU

Work Beliefs Check List).

The results of the data analysis are presented in
chapter five. Some effort is made to assess the instrument
in terms of the specific items as well as the unidimensionality
of the six sub-scales. The final chapter continues the
discussion of the general results. In addition. the limitations
of the study are presented and the thesis concludes with a

brief summary and some implications for future research.

CHAPTER II
CONCEPT FORMATION AND SCIENTIFIC METHODOLOGY

Purpose of this Chapter

The purpose of this chapter is to provide background
material for the remainder of the thesis. This is accomplished
by examining some of the contemporary literature on concept
formation and scientific methodology. At times this discussion
may appear to be somewhat general and therefore tangential
to the present problem. Hewever. as subsequent chapters will
show. there are numerous unresolved problems of definition.
measurement and meaning in the nebulous domain of values.
Inevitably. many of these prdblems are amplified and new
ones are added when we expand the measurement of values
to more than one culture. For these reasons. this back-
ground material is viewed as essential in explicating the
boundaries and limits of the present study. With this in

mind. let us begin with the problem of explanation.

Explanation

Despite some lack of agreement on the aims of

science. there seems to be considerable agreement that

explanation is the chief objective.l However. no such
agreement exists on the crucial characteristics and the function
of scientific explanation.2 It is the contention of Hempel
and Oppenheim that: ”The decisive requirement for every
sound explanation remains that it subsume the explanandum
under general laws.f'3 Before commenting on this statement.
a more detailed look at how these authors define an
explanation will be useful.

Briefly. there are two major parts to an explanation.
One part is the explanandum or sentence describing the
phenomena to be explained. The other part is called the
explanans and consists of the class of sentences which
account for the péghomena. The explanans is further sub-
divided into one subclass containing those sentences which
state the antecedent conditions. and a second subclass
containing sentences about general laws. Within this

framework. an event is explained when it is shown to occur

 

lCarl G. Hempel. FThe Logic of Functional Analysis.”
Llewellyn Gross. (ed.) Symposium on Sociologicgl_Theory
(Evanston. 111.: Row. Peterson and Company. 1959); see also
Ernest Nagel. The Sppucture of Science (New York: Harcourt
Brace and World. Inc.. 1961).

2Carl G. Hempel and Paul Oppenheim. "The Logic of
Explanation.9 in Herbert Feigl and May Brodbeck (eds.).
Readingsin the Philosophy of Science (New Ybrk: Appleton-
Century—Crofts. Inc.. 1953). p./412.
EM
3Ibid.. p. 331.

in accordance with the general laws when the specified
antecedent conditions are present.

It is important to note that Hempel and Oppenheim
reject the frequently cited argument that human behavior
does not have the characteristic of repeatability and is
thus not amenable to causal explanation. They point out
that insofar as we are dealing with a single event. it is a
unique event in the physical as well as the psychological
and social sciences. This does not prevent such individual
events from occurring in accordance with general laws and
thereby being explained by such laws. Moreover. questions
about the general laws may be explained by subsuming them
under more general laws.5 This point seems particularly
important when we get into the question of levels of
explanation for it pinpoints a frequent but misleading.

P. . . insistence that explanation means the reduction

of something unfamiliar to ideas or experiences already
familiar to us.f'6 As Hempel and Oppenheim point out.

the free fall of an object is much closer to experience than
the law of gravitation which explains the phenomena. In

short. as we move toward the more general laws. we

frequently move further away from the familiar rather than

 

41big.. p. 331. 51bid.. p. 326. 61bid.. p. 330.

'1

10

closer to it.7

After this brief discussion of the view of Hempel and
Oppenheim on the nature of explanation. let us take a closer
look at their conclusion that: ”The decisive requirement
for every sound explanation remains that it subsume the
explanandum under general laws._"8 While sociologists are
reluctant to use such terms as filaws? or ”explanation.”
they seem quite willing to accept the implications of the
statement. Consider such concepts as dissonance. imbalance.
assymetry. incongruity. and status inconsistency. Each of
these concepts is invoked in order to explain certain kinds
of behavior under certain conditions. Whatever the phenomena
which we are trying to explain (i.e.. the explanandum).
it is in essence subsumed under a general law. modified to
fit a probabilistic model. Basically. the general principle
states that dissonance. imbalance. assymetry. incongruity. and
status inconsistency. all cause discomfort to the organism
thereby forcing the organism to initiate some kind of behavior
which will tend to reduce the discomfort. The exact nature
of the behavior will vary depending on the specific antecedent
conditions and the specific theory. Some causes of dis-

comfort can be easily removed while others cannot. Some

 

71bid. 81bid.. p. 331.

11

are physical while others are psychological. To attempt to
accurately predict the behavior under such conditions would
require a more detailed spelling out of the particular theory.
On the other hand. we could go to a still more general level
and subsume each of the general concepts under a more

general principle such as homeostasis or the equilibrium
model.

A similar example could be given using the notion
of functionalism as it is used in sociology and contemporary
anthropology. However. the above example should suffice to
illustrate the main point that social psychology and socio-
logy appear to be in agreement with the Hempel and Oppenheim
view of explanation. even though the specific terms are not
congruent. Before leaving the concept of explanation. let
us first examine the views of Nagel.

The reader is cautioned at the outset that the
following treatment of explanation is a highly abbreviated
treatment of selected aspects of Nagel's exhaustive treatment
of the subject. For a true picture of Nagel's view. the
original work should be consulted.

According to Nagel. there are four major patterns of

explanation. These types are:

 

9Nagel. op. cit.. pp. 20-26; while all four types
are given for comparative purposes. the present thesis shall
be concerned primarily with the probabilistic type of explan-
ation. It should also be noted that Nagel assumes that all
types strive for the deductive pattern.

12

1. The deductivepmgdel has the formal structure of a
deductive argument. in which the explicandum is a
logically necessary consequence of the explanatory
premise. This type includes situations where the
explicandum is either a law or a historical fact.
and where both the explicandum and the premises are
necessary truths.

2. Probabilistic explanations are those in which the
explicandum is made more probable by the premises
without asserting the absolute truth of the
explicandum.

3. Functional or teleological explanations use such
expressions as "in order that” or ”for the sake of.”
to account for the performance of organs. units.
individuals or institutions as a means of maintaining
a system. This type is particularly prevalent in
biology and in the study of social behavior.

4. Genetic explangpions are essentially historical and
evolutionary. They attempt to explain how certain
objects. traits or characteristics have evolved
from earlier forms.

According to Nagel:10

 

loIbidol pp. 503-04.

 

13

. . . most if not all the generalizations empirical

social research has succeeded in establishing are

formulated in terms of familiar "common-sense”

distinctions. and possess a comparatively narrow

scope of valid application (or low order generality).

Moreover. most if not all of these generalizations

assert relations of dependence that hold between

stated phenomena only in a (more or less precisely

specified) fraction of the instance of those

phenomena. rather than invariably or with strict

universality. . . .
The two most common reasons given for the lack of universal
laws in social science are that the subject matter is so
complex that it is impossible to identify all of the
relevant variables. or that the Ffree will” of humans makes
perfetct predictability an unattainable ideal. Where the
notion of free will is rejected. the general idea is fre-
quently restated in terms of the view F. . . that men's
actions are governed by their interpretations of external
stimuli. rather than by such stimuli directly."11 Neither
of these actually provides a satisfactory reason for the
lack of universals in the social sciences. As Nagel points
out. there are numerous examples in mathematics and physics
where highly complex subject matter has been reduced to a
simple level. once more effective means of dealing with them

were invented. Moreover. while the question of individual

volition is involved in responses to a given social

 

11Ibid.. p. 504.

14

situation. ,. . . this fact by itself does not explain why
there are no universal laws relating each of the several
interpretations placed upon a given type of social stimulus
to a particular form of human response.,"12 Instead. Nagel
contends that there are two methodological points which
answer the question in a more satisfactory manner.

The first point has to do with the lack of precision
in the terms which are used in the social sciences. In
other branches of scientific inquiry. the terms used in the
formulation of universal laws are precise enough to
designate a highly homogeneous class of objects. The logical
relations of propositions can be more readily seen and
inconsistencies have less chance of survival when the
language employed is precise. The language of social
science is derived mainly from everyday discussions of
social issues. These concepts are frequently vague and
lacking in specificity. In other words. there is no clear
demarcation of classes of things. and those distinctions
which are made are too broad to deal with narrower differences
which might be very important.13

The second point which Nagel makes to account for

the development of universal laws in the natural sciences

 

12231.. p. 505. 131mm. pp. 505-06.

15

is the use of the notion of the idealized state. By stating
the universal laws in terms of ”ideal” conditions. the
scientist is able to systematically account for the empirical
findings in terms of differences between the ideal conditions
and those under which the observations occur. The use of

the ideal case is much less frequent in the social sciences
than in the natural sciences. Actually. Nagel tends to
restrict his examples from social sciences to the field of
economics. and he attributes the failure to be more success-
ful to inadequately developed theoretical notions and assumptions
for a successful bridging of the gap between the ideal case

and actual.14 Thus. Nagel concludes that: _. . . correlations
between empirical data are rarely perfect. and generalizations
based exclusively on such correlations are almost inevitably
bound to be statistical."15

In the opinion of the present writer. it appears that
Nagel has slightly overstated his case. While it seems
undeniable that no branches of the social sciences have
enjoyed any success comparable to the natural sciences with
the use of the ideal case. it is not accurate to imply that

only economics has attempted to use the method. The notion

of the VIdeal Type? has been used in both sociology and

 

l4Ibid.. pp. 503—09. lsIbid.

l6

anthropology for some time. Since this subject shall be
treated in greater detail in a later chapter. further comments
on Nagel's view shall be postponed until that time.

In an earlier section. it was pointed out that what-
ever the phenomena to be explained. it is in essence. sub—
sumed under a general law modified to fit a probabilistic
model. In the light of the above discussion of Nagel
and the previous discussion of Hempel and Oppenheim. this
seems to be a fair synthesis of their views at a very general
level. The heavy emphasis on statistics in contemporary
social science research tends to indicate that the workers
in this field readily accept the statistical nature of their
explanations. Whether the majority of the workers in the
field accept Nagel's rejection of the ”complexity of social
science subject matter? and ”free will? as reasons for the
lack of universal laws is another question. More important
in the present context are Nagel's methodological points on
the lack of precision of concepts. and the use of the ideal
state.

To begin with Nagel's own statement about the most
effective way of making concepts more precise and thereby
reducing indeterminacy:

Several devices reduce the vagueness and increase

the specificity of linguistic expressions. Counting
and measuring are for many purposes the most effective

17

of these tecnhiques. and are perhaps the most

familiar ones. . . . The artisan in metals may

be content with knowing that iron is harder than

lead. but the physicist who wishes to explain this

fact will require a precise measure of the difference

in hardness.l6
Thus. it appears that measurement becomes a crucial factor
in the development of precise concepts. The most pressing
question one might ask at this point could be. ”How do we
use high level measurement to improve our concepts in a
discipline where the ggtio apple is non-existent?" Siegel
contends that the most common types of measurement achieved
in the behavioral sciences are nominal and ordinal.17
thereby implying that even an interval scale is not commonly
used in this area. Without quibbling over Siegel's comment.
the point seems established that the level of measurement
currently achieved in the social sciences is far short of
the kind necessary to increase the precision of concepts.
Quite frequently. the social scientist cannot even make a
statement as precise as ”iron is harder than lead.” At
least the artisan knows that 311 iron is harder than 211

lead. It has already been pointed out that the social

scientist is seldom if ever. in a position to state invariate

 

léIbid.. pp. 8—9.

l7Sidney Siegel. Nonpgrametric Statistics (New Yerk:
McGraw—Hill Book Co.. Inc.. 1956). p. 29.

18

laws or generalizations. Thus. when the sociologist talks
about differences between urban and rural people. he must
state the differences in terms of rates. trends. proportions
or means. and he must be careful to point out that he is

not implying that ”all” or any single case of the class
urban or the class rural would demonstrate the particular
qualities attributed to the class in his analysis. There-
fore. the sociologist may be operating at a level of measure-
ment no higher than ordinal measurement when he expresses

a relationship. Moreover. the classes of things which he is
talking about are frequently so heterogeneous that even this
low order relationship cannot be expressed in terms of
distinct and all inclusive classes of persons. Obviously.
in this example terms such as urban and rural fall into the
category of concepts which Nagel referred to as being too
broad to pick up subtle but important differences. At the
same time. one might wonder about the utility of attempting
to improve our concepts by emulating those disciplines where
measurement is highly developed. until we can at least
approach the types of scales common in those disciplines.

It is sufficient at this time to point out that the present
writer is of the opinion that measurement in the social

sciences is insufficiently developed to provide a panacea

for our language problems. At the same time. it is felt

19

that greater care in spelling out exactly ”what and how”
measurement is used in the social sciences would greatly
improve our communication and ultimately lead to improvement
in the precision of concepts. This view should not be
misconstrued as a rejection of measurement in social science.
The writer considers measurement to be a very important
aspect of social science. when used by careful workers and
kept in proper perspective. However. it is felt that
failure to recognize the gap between ordinal or interval
measurement and ratio measurement leads some workers in
sociology to assume that because they have numbers they have
precision. The present writer does not share this view.
Nagel's second point deals with the use of the ideal
case. This subject shall be treated in a later section
dealing with ideal types. Since the present position will
be developed at that time. it would be repetitious to do
any more than remind the reader to keep the subject in mind.
at this point.
At the beginning of this chapter. it was pointed out
that there is frequently a time dimension to the ”truth”
of a proposition. The discussion on explanation has
pointed out some of the problems and difficulties which

arise when we attempt to improve the level of explanation

in the social sciences. For the most part. the discussion

20

has led to frustrating dead ends. The concepts we use are
vague and it was argued that one of the most useful methods
of increasing precision in the ”exact” sciences (i.e..
measurement). is insufficiently developed to have the same
utility in the social sciences. It was argued that
theoretical constructs are useful and necessary. despite
the explicit and implicit indications that theoretical
notions in the social sciences are not highly developed.

Yet. the present writer is not too disturbed by this
apparent confusion. One fact seems to emerge from the
confusion with increasing clarity. While certain sciences
may be exact. the philosophers of science are not. They
tell us to improve our language partly by using a kind of
measurement which we do not have. but they do not tell us
how to acquire the measurement. They tell us to utilize
the logical device of stating the ideal case but then they
remind us that our theoretical notions are too primitive.
at the present time. to bridge the gap between the ideal
and the actual. There is no doubt that the philosophers of
science recognize the needs of social science. There may be
some question as to whether they recognize the solutions.
It is not the intention here to suggest that they do not
make constructive suggestions as to how to improve the

situation. for they do make such suggestions. waever.

21

the examples. analogies and models held up to the social
sciences are seldom taken from the subject matter of
social science. Mere frequently. the model is physics or
chemistry.18 For example. after presenting an analogy from
physics. Nagel writes: ”The obvious moral of this analogy
is that social scientists should likewise develop more
discriminating classifications of social phenomena. if
strictly universal social laws are to be established."19
Nagel then proceeds to demonstrate why it is unlikely that
social sciences will refine their concepts beyond a certain
point. Following this he illustrates how. if such refinement
was done. we might no longer be dealing with social phenomena
or we might be formulating universal laws in terms of dis-
tinctions more subtle than is necessary. Finally. he con-
cludes that:

. . . social scientists may find it more advantageous

to establish statistical generalizations rather than

strictly universal ones. if the former are more

effective means than the latter for answering the

sort of questions we normally ask about social

phenomena. Accordingly. if the essentially
”practical” nature of our current interests in

 

18Philosophers of science such as Nagel and Hempel do
use examples from the social science but this is usually
as an adjunct to probabilistic and teleological explanation.
rather than in their discussions of the establishment of
universal laws.

19Nagel. op. cit.. pp. 506-07.

22

social phenomena is not radically altered. then.
although strictly universal social laws are not
inherently impossible. the prospects for establish—
ing such laws in the foreseeable future on the
basis of empirical research do not appear to be
bright.20

Thus. while Nagel provides some justification for the use of
statistical generalizations in social science. he leaves

us somewhat confused as to the relative merits of universal
versus statistical explanation in social science.

In concluding the discussion on explanation. a
position will now be stated which the author shall attempt
to strengthen and reinforce as other concepts are discussed.
for it embodies the underlying principle of this entire

thesis. If we wish to answer questions encompassing both

”truth” and ”meaning., we must proceed from the view that:
Explanation is never ultimate in the mind of a
scientist. What may be considered adequate
explanation today may be relegated to theoretical
purgatory tomorrow. What may seem to some to be a
straightforward empirical relation may be raised by a
theoretician to the level of a postulate and used
(with other postulates) to explain other relation-
ships. Thus. a cross-sectional analysis of the
explanatory concepts of a science provides no more
than a momentary picture. a picture which would be
considerably changed by the flux which is science
were a second analysis made a few years hence.21

 

20Ibid.. p. 508.

lBenton J. Underwood. Psychological Research (New
Ybrk: Appleton-Century-Crofts. Inc.. 1957). pp. 174-75.

23

Perhaps some confusion would be avoided if the first word of
the above quotation were changed to the expression. ”final
explanation” rather than simply. ”explanation.” In any
event. the present position should not be confused with
the position of those who are fundamentally opposed to any
attempts by social scientists to develop logical methods
and explanatory systems similar to those of the natural
sciences. The present desire is simply to caution against
the danger of thinking that methodology pp; pg obviates
thoughtful interpretation. This point will become clearer
as methodological problems related to the subject matter

of social science are discussed.

paws. Causality and Prediction

In the previous section the term law or laws was
used quite frequently. Whether explicitly stated or not.
expressions such as ”scientific laws” denote a predictive
aspect to the average reader. Hempel and Oppenheim. for
example. write: ”For all that a causal law asserts is that
any event of a specified kind. i.e.. any event having certain
specified characteristics. is accompanied by another event

which in turn has certain specified characteristics.”22

 

22Hempel and Oppenheim. op. cit.. p. 326.

24

Similarly. Feigl writes: ”The clarified (purified) concept
of causation is defined in terms of ppedictability according
to a law (or. more adequately. according to a set of laws)."23
It is important to note that these statements assert merely
that one event which we call A is always accompanied by
another event. which we call B. thhing is said about one
event producing the other event. This point is important
in order to avoid metaphysical arguments. In social psychology.
for example. we may be able to change (i.e.. control) the
environment of a person. but we have no way of directly
controlling the personality. Therefore. which variable is
designated as the ”independent variable” and which one is
called the ”dependent variable” is often an arbitrary
decision dictated by our ability to control one set of
conditions rather than the other set of conditions.24

It has already been demonstrated that in the social
sciences we seldom. if ever. satisfy a criterion which
requires that a relationship always hold. or where

characteristics ascribed to a given class must apply to all

members of the class. Thus. the above definitions of laws

 

23Herbert Feigl. ”Notes on Causality.” in Feigl and

Brodbeck. op. cit.. p. 412.

24Ibido 0 p. 417.

25

seem to exclude most or all of the pronouncements of social
science from the status of laws. However. Nagel does not
define laws quite so simply. Instead. he talks about four
types of laws. Briefly stated. these are:25

1. Laws which assert that there are ”natural kinds."
This type of law is not considered to be a causal
law by Nagel. for it makes no assertion about one
object preceding another.

2. Laws which assert an invariable sequential order of
dependence among events or properties. These are
subdivided into causal laws (e.g.. stones thrown
into water produce a series of expanding concentric
ripples). and developmental or historical laws
(e.g.. lungs are never formed prior to the circulatory
system in the hyman embryo).

3. Laws which assert invariable statistical (or probabil—
istic) relations between events or properties.

4. Laws which assert a relation of functional dependence.
in the mathematical sense of function (e.g..
mathematical equations).

If we consider the meaning of scientific laws in this broader

sense as Nagel does. the social scientist is justified in

 

25Nagel. op. cit.. pp. 75-78.

26

using the term. As the previous discussion on explanation
indicated. the statistical or probabilistic laws are common
in the social sciences. It might be argued that social laws
may be of the developmental or historical type. However.

it will be shown later that the combination of time and
space restrictions on social phenomena impose such serious
limitations on generalizations that it is difficult to
imagine ”laws” about social evolution comparable to the
developmental laws in biology.

Therefore. the concept of causation as defined for
present purposes will denote ”predictability according to
a law.” with the understanding that such laws in social
science will. for the most part. be of a probabilistic type.
Before proceeding to the next section. it seems feasible
to relate the concept of prediction to the concept of
explanation.

In terms of the logic involved. there seems to be no
difference between prediction and explanation.26 According
to Hempel and Oppenheim. the difference between the two terms
is simply a pragmatic difference. depending on whether the
explanandum is given and the explanans is provided after

the event has occurred. or whether the explanans are given

 

26Feigl. op. cit.. pp. 417—18.

27

and the explanandum is logically derived before the event
occurs.27 The formal analysis will be the same in both
cases.

It should be pointed out that these terms are not
always viewed as if they are logically interchangeable.
When we consider certain discussions of divergent and con-
vergent phenomena.28 we become entangled with the implicit
or explicit possibility of ”explanation without prediction."29
In essence. this position states that even though we know
that a given set of antecedent conditions will be accompanied
(or followed) by a given phenomena in accordance with a given
law or set of laws. we may be unable to predict the occurrence
of the antecedent conditions. we may know. for example.
that certain kinds of behavior tend to occur in disaster
areas but if we cannot predict the time or place of a
disaster. we cannot predict the time or place of the behavior.
Proponents of this view would argue therefore. that we can
explain the behavior in terms of the disaster but we cannot

predict where or when the behavior will occur.

 

27Hempel and Oppenheim. op. cit.. pp. 322-23.

28Ivan D. London. ”Free-Will as a Function of
Divergence.” Psychologicgl_Review. pp. 1948. pp. 41-46.

29Michael Scriven. ”Explanation and Prediction in
Evolutionary Theory.” Science. vol. 130. pp. 477-82.

28

In the opinion of the present writer. this argument
is not very tenable. If an invariate relationship exists
between two events. failure to predict the occurrence of the
necessary antecedent conditions does not invalidate the
relationship. Even in the natural sciences. no such demands
are made. The fact that the chemist cannot predict the time
and place of every incident of combustion. does not invali-
date the laws of combustion. Newton's laws governing bodies
in motion are not invalidated simply by the inability of
physicists to predict the time and place that each body will
set in motion. When we use examples such as these. the
distinction being made between explanation and prediction
appear somewhat ridiculous.

Therefore. the only distinction between explanation
and prediction which shall be maintained in this thesis.
is the previously cited ”pragmatic" distinction of Hempel
and Oppenheim. However. the discussion of reasonable and
unreasonable requirements of prediction provides a convenient
point of departure for the next section.

EmpiricismL Operationism. verification
gpd Meaning

As others have pointed out. empirical evidence may

answer the question. ”Is it true?” but may not answer the

29

question. ”What does it mean?” However. it is not always
clear just what the differences between empirical evidence
and meaning happen to be. For example. Schlick writes:

”. . . there is no way of understanding any meaning without
ultimate reference to ostensive definitions. and this means.
in an obvious sense. reference to 'experience' or 'possibility
of verification.'”30 Other writers discuss ”experience” or
being ”possible of verification” in relation to empiricism
or operationism. According to Franck. ”. . . the main
principle of empiricism. or even logical empiricism as
Carnap understood it. is the principle of verifiability or
confirmability.”31 In a similar vein. Hempel writes: ”The
fundamental tenet of modern empiricism is the view that all
non—analytic knowledge is based on experience.”32 It
appears. therefore. that the three concepts of empiricism.

verification. and meaning are not distinctly separable

terms in the minds of some writers. The ”purity” of the

 

3OMoritz Schlick. ”Meaning and verification.” in

Herbert Feigl and Wildrid Sellars (eds.). Readings in

Philosophicgl Analysis (New YOrk: Appleton-Century-Crofts.
1949). p. 148

31Philipp Franck. Philosophy of Science (Englewood
Cliffs. New Jersey: Prentice-Hall. Inc.. 1957). p. 335.

32Carl G. Hempel. ”The Empiricist Criterion of
Meaning.” in A. J. Ayer (ed.). Logical Positivism (Glencoe.
Illinois: The Free Press. 1959). p.108.

30

concepts may become even less clear if we add the fourth
concept with which this section deals. This notion.
variously referred to as Operationism. operational definition
or operational meaning was strongly influenced by Bridgman's
bodk. The Logic of Modern Physics. According to this view.
”In general. we mean by any concept nothing more than a set
of operations: the concgpt is synonymops with the correspond-
ipg set of operations.”33

Operationism in the strictest sense. indicates that
any time that we alter the operations used in defining a
concept. we must introduce another concept. ”If we have
more than one set of operations. we have more than one
concept. and strictly there should be a separate name to
correspond to each different set of operations.”34 Yet.
Bridgman recognizes the impractical and frequently unnecessary
limitations which strict adherence to this rule would
impose upon the scientist. He feels that the scientist is
justified in using the same concept. even though two dif-
ferent sets of operations are involved. as long as the numerical
results of the different sets of operations do not differ

beyond the level of experimental error. The last point

 

33 . . .
P. W. Bridgman. ”The Logic of Mbdern PhySics.”

excerpt reprinted from Bridgman's bodk of the same name in
Feigl and Brodbeck. op. cit.. p. 36.

34Ibid.. p. 39.

31

brings us back to a problem already discussed in the section
on explanation.

One wonders how we can be sure we are operating
within the limits of experimental error when the numerical
differences are not based on ratio scale measurement. To
be sure. in the case of replication. where the same instru-
ment is used. we may talk about ”significant differences”
or ”confidence limits.” This is not quite comparable to
the situation to which Bridgman is referring. because the
set of operations has not been changed. For example. if a
specific scale designed to measure ”aspiration level” is
given to two separate populations. the same operations are
used in arriving at a score in both cases. Thus. we are
dealing with the concept ”aspiration level” as measured by
instrument A. If a different instrument is used to measure
the concept ”aspiration level." it will be designated ”aspir-

ation level.f as measured by instrument B. In practice.

the different instruments are usually designated by the name
of the test designer. In any case. it has become a pipg
gp§.ppp of good research in the social sciences that any
report describing the use of such instruments give. in

addition to a description of the instrument. a detailed

description of the operations performed in the administration

32

of the instrument. as well as relevant data on the sample.
reliability and validity. In social measurement. slight
changes in the wording of a question or in the test
situation have been shown to affect the results. Moreover.
if instrument A and instrument B is administered under
”identical” conditions to the same sample. the numerical
results frequently yield at best moderately high correlations.
Two sets of measurements of the physical length of objects
on the earth's surface. involving what Bridgman calls tactual
concepts on the one hand and optical concepts (such as the
angles between beams of light) on the other. are likely to
show very little. if any. numerical difference.

It was noted earlier that to insist upon defining
all scientific terms operationally would be overly restrictive
and would prevent the use of some of the most powerful
theoretical constructs.35 A similar point is emphasized by
Feigl when he writes: ”To demand definition of pygpy
term used in a piece of scientific discourse would not only
be unduly pedantic (beside being incapable of practical

fulfillment and thus utopian) but also quite unnecessary.”36

 

5Hempel. Fundgmentgls of Concept Formation in
Empirical Science (Chicago: The University of Chicago Press.
1952). P. 41.

36Herbert Feigl. ”Operationism and Scientific Method.”
in Feigl and Sellars. op. cit.. p. 499.

33

Feigl goes on to point out that for most purposes. ordinary.
language is sufficiently definite to permit communication.
but questions concerning the exact meaning of concepts may
arise when we shift to higher order constructs.

In the light of the above discussion on operationalism
the following considerations are put forth. First. the
notion of the operational definition of concepts is a
useful principle for social science if the limits of the
utility are recognized. By keeping the underlying principle
of operationism in mind. the social scientist is constantly
alerted to the possible dangers of using ”factually meaning—
less” concepts. The reader is cautioned that high order
theoretical constructs such as the super-ego. the generalized
other. or the reference group. do not fall into the category

of ”factually meaningless.f simply because they cannot be
observed. or measured directly. For these concepts can be
indirectly defined in terms of the observed relationships
between the antecedent and consequent conditions which the
concept (in the proper theoretical context) purports to
explain. The expression ”factually meaningless” refers to
gross violation of the logic of grammar or the formulation

of explanatory schemes which do not. even in principle.

meet the criterion of verifiability. To illustrate this

34

point. let us consider the following example.

There have been many definitions of the concept of
culture in the field of anthropology. These definitions may
differ in terms of how broadly or narrowly they are defined.
but they share certain features concerning the transmission
of knowledge. Suppose that an anthropologist decides to
define culture in such a way that only the ballet. the
opera and the symphony orchestra of western civilization.
qualify as examples of culture. Such a definition is not
simply different from the others but in direct contradiction
of all the other definitions in anthropology. The present
writer would consider such a definition to be ”factually
meaningless” within the field of anthropology because it
violates the syntactical formation-rules of the language of
anthropology.37

The second point mentioned above in relation to
”factually meaningless” terms has to do with the possibility
of verification. It has already been pointed out that
various writers set up this criteria of ”possibility of veri-

fication” in defining both empirical science and meaning.

 

7Herbert Feigl. ”Logical Empiricism.” in Feigl and
Sellars. Ibid.. pp. 3-26. The present writer has taken the
fundamental point from Feigl but the example from anthro-
pology is his own.

35

In the present context. the notion of operationism may be
helpful in reminding the scientist that he is studying
a real world. While certain kinds of models may be useful
in this study. certain other limitations must be kept in
mind. This point was noted by the Columbia Associates in
Philosophy over thirty-five years ago. when they wrote:
But one outstanding difference between the world of
the actually existent in which the physicist works.
and the non-existent or hypothetical world in which
the mathematician is at home. lies in just this
fact. that in the latter one equation is as good

as another. and can be substituted for the other.
whereas in the former. though for purposes of

convenience we must perform this ”reduction._ we

can never take away a single iota from the actual

world with which we commenced.) The geometer starts

with assumptions which he can change at will; the

physical scientist starts from facts which nothing

can ever change.38
Therefore. when we begin to reduce our data to symbols or
to different levels. we must be extremely careful to
recognize the fact that we are no longer dealing with the
same data. and the answers which are discovered may not be
applicable to the original set of data. Within this
context. the nature of explanation and verification may

be confused. It is quite possible to explain certain

facts after they have been translated into a different

 

38Columbia Associates in Philosophy. An Introduction

to Reflective Thinking (New YOrk: Houghton-Mifflin Co..
1923). pp. 139-40.

36

level or form. If this form is not equivalent. the original
facts have not been explained. In short. the social scientist
cannot indiscriminantly substitute one set of operations
for another. More specifically. he cannot build his
explanatory system or model with complete disregard for the
experiential world. and expect such a scheme to fulfill the
canons of science.

It is hoped that this criterion of meaning will not
be interpreted as permission to ”push a very difficult problem
aside and by stigmatizing it as meaningless to discourage
further investigation.”39 It is intended to stress the
importance of recognizing the difference between an.explana-
tory scheme which meets the criteria of confirmability or
rejectability. and a scheme which cannot be tested within
the usual domain of empirical science. because it is
founded upon a logical system such as mathematics. which
is logically consistent independently of the empirical world.
On the other hand. it should be clear that insistence upon
a recognition of the limits and boundaries of the respective
language of different areas of discourse is in no way intended

as an anti-theoretical position.

 

39Feigl. loc. cit.. p. 13.

37

In summary. it is the present contention that empirical
science must utilize the ”possibility of verification”
criteria if we hope to achieve ”factually meaningful”
explanations. In striving for this goal. the underlying
principle of operationism can be very helpful. if kept in
proper perspective and not allowed to become a pip; QEQIEQQ
for all inquiry. It has been pointed out in earlier
discussions that any attempt to define every scientific term
”operationally” is both impractical and unnecessary. In
effect. this is simply a restatement of the central theme
stated in the discussion of explanation. we cannot eliminate
the need for thoughtful interpretation by substituting a

method.

Theory

If we acknowledge the need for convergence between
method and thoughtful interpretation. we take a major step
toward the marriage of theory and reSearch. Explanation
is difficult without a theoretical framework. no matter how
numerous the facts. Facts alone are sometimes contradictory
toour intersubjective knowledge. For we ”know” that when
a human being has not eaten for two days. he will be hungry
and he will eat if food is available. But we also know

that on occasion such human beings will not eat.

38

What then are the facts? The facts are a collection
of data. The ”truth” of the facts can be determined by
controlled observation. The ”meaning” of the facts must
be imposed by experimental laws or theories. As Moore puts
it: ”It is the function of theory to impose order on what
might appear to be chaos. and it does as by being in some
measure abstract.”40 Stating the same idea in terms more
akin to the present discussion: ”The term theory. in science.
denotes that conceptual apparatus which makes explanation
and prediction in the area of experience possible.”41

At the same time. while theories are not tied to
definite observational data. no theory is independent of the
facts. A theory which continually predicts phenomena con-
trary to empirical evidence is not likely to survive. no
matter how abstract or abstruse the theoretical formulation
may be.

Just as a method does not eliminate the need for

interpretation. a theory does not eliminate the need for

 

40Wilbert E. Moore. ”Editorial Introduction.” in

Charles P. Loomis and Zona K. Loomis. Mpdern Social Theories:
Selected American Writers. XXIII (Princeton. New Jersey:
D. Van Nostrand Company. Inc.. 1961).

41Joseph B. Gittler and Ernest Manheim. ”Sociological
.Theory.f in Joseph B. Gittler (ed.). Review of Sociology (New
York: John Wiley and Sons. Inc.. 1957). p. l.

39

fact finding. Each phase is necessary and neither stands
independently. The central theme is again stated. This
time the notion under discussion is theory and the words are
Conant's. ”A scientific theory is not even the first
approximation to a map; it is not a creed; it is a policy--
an economical and fruitful guide to action by scientific

. 42
investigators.”

Scientific Method

The scientific method. like most of the concepts
and methods discussed thus far. is a means. not an end in
itself. It is a way of bridging the gap from experience to
explanation to validation. Exact statements of the scientific
method may vary but in general they include:4
1. Collecting data systematically.
2. Analyzing and organizing these data into generalizations.
3. Arranging these generalizations. which may be drawn
from different collections of facts and generalizations.
into theories.

4. Drawing deductions from the theories in the form of

 

42James B. Conant. Modern Science and Modern Man

(New York: Doubleday Anchor Books. 1954). p. 97.
433. R. Bugelski. A First Course in Experimenpg;
Psychology (New York: Henry Holt and Co.. 1951). P. 46.

40

predictions (such predictions are generally known
as hypotheses or theorems).

5. Testing the hypotheses by experimenting. or other-
wise checking by controlled observation. trying to
see if they are valid or invalid. and thereby.
supporting or discrediting the theories.

A careful scrutiny of the five steps listed above will
reveal the sweeping range of the scientific method. If

we require that any scientist. to deserve the name. muSt be
a master of all phases. we set up a criteria which few
persons can meet in a single lifetime. Thus. Bugelski

writes: ,. . . we can label anyone a scientist if he is

active in any or all phases of the total method.”44
Hewever. regardless of which phase or phases of the

method a scientist works in. he must be cognizant of the

total method. Failure to recognize other phases may result

in excessive labor on ”trivial” data collection at one

end. or the building of ”untestable” theories at the other.
It should be noted that the various steps are stated

in very general language. No boundaries are set up defining

the type of data which should be collected or the specific

 

44Ibid.

41

method which should be used. In short. the method. like
explanation and theory is a guide. In the hands of the care-
ful worker. fruitful results are frequently obtained.45

But in the final analysis. the same language which sets no
limits on specific areas of inquiry. imposes one strong
limitation on the scientist. In any phase of the scientific
method. the investigator must constantly weigh his activity
against the activity of others. Sometimes the ”others”

will be engaged in the same kind of activity and there may
be a relatively high amount of conceptual equivalence.

On other occasions. he will have to translate from the
general to the specific or vice versa. In either case.
there are no ”ivory towers” for those who endorse the

scientific method.

 

5For a similar statement of this position. see
Paul F. Lazarsfeld and Morris Rosenberg (eds.). The Languagp
of Social Research (Glencoe. Illinois: The Free Press. 1955).
p. 4.

CHAPTER III
TOWARD A WORKING DEFINITION OF VALUES

Purpose of this Chapter

This chapter is concerned with the definition of
values and related concepts such as attitude. belief. opinion
and value orientation. After a brief examination of the
literature on the definitions of values. an attempt will
.be made to arrive at a working definition for the present

study.

Problems in the Definition of Vglues

In any reasonably extensive survey of the literature
on values. one cannot help being struck by the variations
in the interplay between such terms as value. attitude.
belief and opinion. The interplay is further complicated
by the sporadic insertion of the concept of ”orientation."
In order to attempt to clarify the situation. it seems
necessary to examine each of these concepts separately.
and then to attempt a synthesis. Therefore. the present
discussion shall examine first the general state of'
confusion. second the respective concepts individually. and

finally the relation of the various terms to a working

42

43

definition of values.

Perhaps the concept which is most intricately woven
with the general notion of value is the concept of the
attitude. In discussing ”The Changing Prominence of Values.”
William L. Kolbl points out the difficulty of distinguishing
between values as objects and as elements of orientation.

In this context. the work of social psychologists such as
Mead and Paris becomes the bulwbrk for viewing values within
a two dimensional framework. The objective aspect of values
is maintained in the cultural super-structure. but it now
becomes possible to speak of a subjective dimension of values
as exemplified in such concepts as”social attitudes” and
”group attitudes.” This formulation involves certain additional
concepts such as personality. socialization. generalized
other. etc. However. the basic question which this intro-
duces is whether an individual perceives values. once they
are internalized. as externally imposed (e.g.. determined

by mutual agreement) or as internally determined standards

or ”givens.”

The importance of this question will be more

 

1William L. Kolb. ”The Changing Prominence of Values
in Modern Sociological Theory.” in Howard Becker and Alvin
Boskoff (eds.). Modern Sociological Theory (New York:
The Dryden Press. 1957). pp. 93-132.

44

obvious when we deal with the general problem of the measure-
ment of values cross culturally. However. in order to focus
the reader on the problem. let us prematurely state the
bearing which this issue has on the measurement of values.
If values are perceived as objects. there are tangible ways
of ascertaining what these values are. In a society with
a high emphasis on religious values. we would expect this
to be reflected in the material culture as well as the
social organization of the society. If there is an extreme
emphasis on the political values. this should be evident
in the social structure. Bellah. for example. has shown
that the dominance of political values in Japan during the
Tokugawa period. was exemplified in the social structure.
the family and the Japanese religion. Shinto.2

On the other hand. when we introduce the subjective
dimension of attitudes as orientations to action. we are
no longer dealing with the observable aspects of culture
but with theoretical constructs. It is not the intention
here to imply that there is a clear and direct link between

the observable aspects of material culture. social organization

 

2R. N. Bellah. Tokugawa Religion: The Values of

Pre-Industrial Japan (Glencoe. Illinois: The Free Press.
1957). Cited in Harry M. Johnson. Sociology: A Systematic
Introduction (New YOrk: Harcourt. Brace and Company. 1960).
p. 88.

45

rituals. etc.. and the scientific determination of the values
of a given group. Nevertheless. it is felt that the

area of ambiguity is increased if we add a third dimension.
unless we are able to spell out some methods of tapping

that dimension. For it then becomes necessary to specify

the connection between cultural values. socialization.
personality and behavior. As we shall see later. these
relationships do not always follow expected paths. In

short. if we observe a particular society in which religion
permeates almost all aspects of behavior. we might say that
religious values are among the most dominant in that
particular system. Obviously. this does not tell us whether
all members of the system unequivically accept this dominance.
However. if the behavior of most of the members of the

system is in accord with this religious dimension. it seems
reasonable to assume that such a value exists in terms of

the system as a whole. When we begin to aSk whether this
value is held by all members of the system. or whether the
value is internalized with equal intensity we become

involved with the problems of individual beliefs. differential
perception and interpretation. and the divergence or con-
vergence of subjective values with the behavior patterns

followed in the face of social sanctions. Add to this the

46

problems inherent in the measurement of theoretical constructs
and the subsequent problems in the validation of such
measurement. and the extent of the problem becomes some-
what clearer.

In the preceding discussion. the terms ”values”
and ”attitudes” have been used as if there was a clear cut
distinction between them. In practice. this is not always
the case. Johnson. for example. states that: ”Values . . .
They are inseparable from attitudes. except perhaps analytical-
1y.”3 Yet. it has already been noted in the consideration
of the Kolb discussion. that Johnson's view is not an
indication of even a reasonable degree of concensus on this
point. Moreover. there have been considerable efforts
devoted to the distinction of attitudes as cognitive elements
and some verbal indicant of these elements. Therefore.

let us consider some of the concepts individually before

attempting any further definition of values.

Attitudes

It does not seem to be an exaggeration to say that
the concept ”attitude” occupies a rather prominant place in

the contemporary literature of social psychology. This

 

3Johnson. op. cit.. p. 87.

47

concept. it must be remembered. is essentially a theoretical
construct. There is no way to directly observe an attitude.
In fact. there is no way to verify the existence of attitudes.
directly. Yet. this concept remains a central one in the
behavioral sciences. One possible explanation of this can
be found in the earlier discussions of Nagel. It will be
recalled that Nagel considers the failure of the social
sciences to use the logical device of stating the ideal
case. and attempting to account for empirical deviations
from the ideal in terms of theoretical postulates and
assumptions. as one of the major reasons for the paucity of
universal propositions. In a later section. this criticism
shall be dealt with directly in terms of the use of typologies.
At the present time. however. there are certain grounds for
viewing the attempts of social scientists to use the con-
cept of attitude. as an attempt. in principle. to use the
basic notion of the ideal case.

As frequently used. the term attitude has a definite
action potential associated with it. Allport. for example.
speaks offa mental and neural state of readiness”4 while

Sherif and Cantril place attitudes ”among those components

 

4Gordon W. Allport. ”Attitudes.” in Carl Murchison

(ed.). A Handbook of Social Psychology (WOrcester: Clark
University Press. 1935). p. 810.

48

of the psychological makeup of the individual which determine
that he shall react not in a passive or neutral way . . ."5
This conceptualization of an attitude as a guide or directive
of behavior persists despite the difficulty of demonstrating
the nature of this relationship empirically. Thurstone. for
example. contends that it is an error to attempt to validate
the measurement of social attitudes by overt behavior. He
writes: ”A man may be entirely consistent in what he says
and in what he does about a controversial issue. and yet
both of the indexes may be dead wrong in reflecting his
attitude. . . . His personal attitude may or may not agree
with what he says and qut he does.”6 The same problem was
pointed out some time ago by Klineberg when he noted that
even though a certain kind of validity of verbal behavior is
often obtained. the kind of validity which deals with the
correlation of verbal and overt behavior. is a rare event.7
The point here is that the use of the concept of attitude

is. in many instances. equivalent to the postulating of an

ideal case (i.e.. persons holding specified attitudes. when

 

5M. Sherif and H. Cantril. ”The Psychology of
Attitudes.” Psychological Revue. 1945. g;, p. 300.

6L. L. Thurstone. The Megsurement of Values (Chicago:
The University of Chicago Press. 1959). p. 187.

7Otto Klineberg. Social Psychology (New York: Henry

Holt. 1940). p. 371.

 

49

exposed to certain antecedent conditions will behave in a
specified. and presumably predictable manner). If we add
to this the probabilistic nature of explanation in the
social science. the resultant form is a combination of the
previously discussed explanations of Hempel and Oppenheim.
and Nagel.8

However. at least one additional factor stands out.
If attitudes are not necessarily correlated very highly with
behavior. and certainly not perfectly correlated. it is
clear that the concept of attitude is not a result of
empirical observation. In short. attitudes are constructs
of an idealistic type and those characteristics which are
associated with or attributed to attitudes are deviations
from the ideal. Thus. it is customary in the social sciences
to speak of persons holding ”strong. weak. or neutral

attitudes” rather than simply ”attitudes.”

Beliefs. Opinions. and Attitudes

Thurstone makes a clear distinction between the
subjective aspect of attitudes and the verbal. expressive
or the measurement dimension. Attitudes for Thurstone are

_. . . the sum total of a man's inclinations and feelings.

 

See chapter two for this discussion.

50

prejudice or bias. preconceived notions. ideas. fears.
threats. and convictions about any specified topic.”9
Thurstone then goes on to define an opinion as the verbal
expression of an attitude. A considerable body of sociologi-
cal literature appears to regard opinions as verbal
expressions of attitudes (or beliefs). It should be noted.
however. that the question of whether the opinion which an
individual expresses is an indication of his true views. is
not solved by this ”fiat” type of definition. Additional
qualifying terms such as public opinion or private opinion
are sometimes used. However. these terms are also used by
some writers to qualify other concepts such as attitudes or
beliefs. so they do not resolve the issue.

Perhaps the simplest way to define a belief is
to quote Loomis when he writes: ”Any proposition about any
aspect of the universe that is accepted as true may be
called a belief.”10 This statement must be amplified for
the purpose of clarification. Johnson. for example. states
that a belief is neither true nor false. in empirical

terms.ll Williams. on the other hand. uses an expression

 

9Thurstone. op. cit.. p. 216.

10Charles P. Loomis. Social Systems: Essays on Their
Persistence and Change (Princeton. New Jersey: D. Van
Nostrand. Inc.. 1960). P. 11.

11Johnson. op. cit.. p. 86.

51

which seems to synthesize Loomis and Johnson. when he talks
about ”non—empirical beliefs.”12 Actually. the difference
between these positions is only superficial. Both Loomis
and Williams acknowledge the existince of non—verifiable
beliefs. This inability to verify beliefs empirically does
not require. as Johnson seems to assume. that beliefs be
defined as neither true nor false. A belief can be complete-
ly false in terms of content. but completely true in terms
of orientation to action for specific individuals or groups.
As Loomis points out: ”The belief that an eclipse of the
sun is an ill omen. that tomatoes are poisonous. that
capitalistic society will eventually decay and become
socialistic. whether true or false. must be taken into
account in explaining action.”13

In the light of the above discussion. it seems that
beliefs. opinions and attitudes are frequently distinguished
analytically. but a relationship between them is usually
maintained. While Thurstone makes a clear distinction between
attitudes and opinions in terms of the overt-covert dimension.

he regards the overt as an indication of the covert

 

12Ro‘binM. Williams. Jr.. American Society (New

YOrk: Alfred A. Knopf. 1960).

l3Loomis. op. cit.. p. 12.

52

dimension.l4 Neither beliefs nor attitudes are directly
demonstrable through empirical techniques. Both are inferences
from verbal or some other form of overt behavior. They
cannot be measured directly. It has already been pointed out
that the same questions hold for the notion of opinion. In
essence. attitudes. opinions and beliefs. insofar as we can
measure them. are subjective components. which we ”know”

only through the verbal and/or non-verbal behavior of
individuals under certain conditions. In this sense. all
three of the concepts are cognitive elements operationalized
by means such as questionnaires. interview schedules. or
direct observation.

Before discussing the concept of values in a more
specific manner. a brief recapitulation seems in order.
First. the relationship between subjective thought processes
and overt behavior is not always clear. direct or consistent.
Second. in tapping the subjective level. it is necessary to
rely on verbal expressions and theoretical constructs.

This does not rule out the possibility of making adequate
predictions of behavior. or accurately tapping the sub-
jective dimension. It merely accentuates the necessity of

constantly keeping in mind. exactly what level we are operating

 

14Thurstone. op. cit.. p. 216.

53

at. With this background. let us now approach the nebulous

domain of the concept of values.

Values and Value Orientations

According to Kolb. the one consistent theme which
permeates all treatments of values in the sociological
literature has to do with the relationship of values to
norms.15 In a general sense. this seems to be true. although
there may be variations on this interpretation when we
consider specific details of any given theorist.

In Parson's book. The Social System. he writes: "An‘
element of a shared symbolic system which serves as a criterion
or standard for selection among the alternatives of orien-
tation which are intrinsically open in a situation may be
called a value.”16 Within this framework. values are tied
to the general notion of orientation of action. in terms
of a choice between norms. More specifically. Parsons
writes: ”Values are modes of normative orientation of

action in a social system Which define the main directions

of action without reference to specific goals or more

 

15Kolb. op. cit.. p. 93.

l6Talcott Parsons. The Social System (Glencoe.

Illinois: The Free Press. 1951). p. 12.

54

detailed situations or structures.”17

At this point. the reader is reminded of the earlier
discussion on the levels of explanation. for Parsons seldom
deals with a singular dimension. Values. he contends. are
grounded in three main directions. One level of values deals
with the individual's existential beliefs about the world.
This level is equivalent to the previously mentioned non—
empirical beliefs of Williams. Actually. Parsons refers to
this level as the ”justification of values.”18 and states
that it is grounded in philosophy and religion. There is a
second level of values. according to Parsons. which deals
with the meaning to the individual as a personality. This
level is referred to as the ”motivation of values.” Finally.
there is the relation of the individual to others in the
society. or what Parsons calls the ”legitimation of social
action.”

In discussing Parsons' earlier definition of values
in terms of ”a shared symbolic system.” Becker reduces this
definition to a choice between norms pointing out that:

”The more profound axiological problem as to the subjective

 

l7Talcott Parsons. Structure and Process in Modern

Society (Glencoe. Illinois: The Free Press. 1960). p. 171.

18Ibid.. p. 174.

55

or objective character of value is not explicitly considered.”19

In this case. the criticism seems to be based on an over-
simplification of Parsons which is a result of taking the
definition of values cited in The Social System. and con-
sidering this definition out of context or without regard
to the over—all theoretical framework. Adler. for example.
follows his quotation of the same definition with the
question: ”How are these covert phenomena to be studied?”20
While Becker's comment does not assume that Parsons'
definition is covert (it Simply fails to specify whether
values are subjective or objective). both descriptions seem
to disregard the numerous aspects of Parsons' theory
whiCh modify a literal interpretation of this definition.
Parsons writes:

On a cultural level we view the organized set of

rules or standards as such. so to speak. from the

actor who is committed to them by his own value—

orientations and in whom they exist as need—

dispositions to observe these rules. Thus a

culture includes a set of standards. An individual's

value-orientation is his commitment to these
standards.21

 

 

19Howard Becker. ”Value.” UNESCO Dictionary of Social
Science. forthcoming.

20Franz Adler. ”The Value Concept in Sociology.”
American Journal of Sociology. Q2 (November. 1956). pp. 274—75.

21
Talcott Parsons and Edward A. Shils (eds.).
Toward a General Theory of Action (Cambridge: Harvard
University Press. 1951). p. 60.

56

In reference to these standards. Parsons writes: ”The
value standards are various recipes or rules (usually
passed from person to person and from generation to
generation) which may be observed by the actor in the course
of this balancing out procedure.”22 In other words. when
dealing with the role of values in Parsons' theory. it
seems unfair to ignore his use of different systems such
as the social. personality and cultural. In interpreting
Parsons. Loomis states:

On the cultural level there are value patterns and

on the personality level there are gratification-

deprivation complexes which are somewhat balanced

by modes of motivation. The actor. motivated to

maximize his gratifications and minimize his

deprivations. does so in accordance with value-

orientations. derived from the cultural value pattern

and internalized by the process of socialization.

to become a part of personality itself.23
In other words. the work of Parsons cannot be simply cate-
gorized as considering values as covert phenomena. This
is what Adler seems to do. but he apparently feels that
the case is self evident by the quotations. In any event.

he does not go beyond the abstracted definition. which

does not deal with the various levels of concern to Parsons

 

ZZIbidol p. 71.

3Charles P. Loomis and Zona K. Loomis. Modern
Social Theories (Princeton. New Jersey: D. Van Nostrand
Company. Inc.. 1961). p. 381.

57

or with the articulation of these levels through such
processes as socialization. When dealing with the pattern
variables. the notion of value orientations is a necessary
component of the actors choice. It must be remembered.
however. that the pattern variables apply at different
analytical levels. Earlier. these were the concrete. the
collectivity. the cultural and the value standard.24 Thus.
we have the value orientations as expressed through the
pattern variables. operating at several levels. ppg of
which is the value standard.

Another indication of the fact that while Parsons
often treats different levels as analytically distinct.
it does not follow that these are empirically distinct.
as indicated by a later work. In this instance. Parsons
distinguishes four levels. These are the primary or
technical. the managerial. the institutional. and the
societal.25 Even though Parsons' makes analytic distinctions
between these levels. Loomis' notion of systemic linkage is

implicit in the statement that: ”The technical or primary

 

24Ibid.. p. 342.

5Talcott Parsons. ”General Theory in Sociology.”
in Robert K. Merton. Leonard Broom. and Leonard S. Cottrell.
Jr. (eds.). Sociology Today (New York: Basic Bodks. 1959).

58

social system cannot. however. subsist alone in a differentiated
society (and all societies are to some degree differentiated).
it must be 'articulated' with the other units in a wider
system.”26 Actually. Parsons' uses the expression ”double
interchange” to convey the idea of interaction between systems.
Thus. it must always be kept in mind that Parsons is dealing
with analytic systems and cautions must be observed to
avoid the reification of such systems as well as the fallacy
that only one system is involved in a concrete event.

To further explicate the problem of defining values
in a singular or ”clear” manner. let us take a brief look at
the writing of Franz Adler on this concept.27 Adler
attempts to delineatefour types of values. In essence.
these are values as absolutes (in the mind of God). values
as objects. values as located in man. and values as actions.
Adler then proceeds to eliminate the first three of these
types as inaccessible to the methods of empirical science.
and argues that values are identical with action. since all
that can be observed is actions. Perhaps the most immediate
temptation here is to equate this type of reductionism to

the early phases of Watsonian Behaviorism. Instead. let

 

26Ibidol p.

27Adler. op. cit.. pp. 272-279.

59

us apply Adler's line of reasoning to a couple of examples.
Consider the food deprived mental patient who refuses to
eat. Should we say that the patient values hunger? What
about the soldier who obeys an order even though he is in

a state of extreme fear? Should we say that he values the
danger more than he values his fear? To attempt to account
for complex behavior entirely on the basis of actions is.
to the present writer. approaching the realm of absurdity.
For example. Adler writes:

Their verbal as well as their non-verbal actions are
their values. It is obviously unjustifiable to give

to either the verbal or the non-verbal action the
greater weight. to claim that either is more expressive
of values or expressive of more real values than the
other. What people say is what they want to pgy:

what people do is what they want to g9. What people
say in Sunday school and what they do during the

week may or may not be consistent; but both sets of
behavior constitute their values.28

But perhaps. it is in Adler's own words that we see the
logical inconsistency of this view most clearly. for on the
same page as the above quotation he writes:

About thirty years ago instinct in the explanation of
social and cultural behavior was dropped because it
was recognized that. since the ”instinct” was but an
inference from some observed behavior. it could not
validly be used to explain it. ”Interests.” ”wishes.”
and ”attitudes.” all of them once inexpendable props
of psychologizing sociology. like the ”instincts” of

 

28Ibid.

60

yore and the ”values” of today. were constructed by
generalization from observed behavior. As such
they could and did serve for the prediction of the
behavior they described. They could be used effective-
ly for explanation in the same sense in which the
law of gravity explains a case of a falling body. that
is. as one case among many similar ones.29
It would perhaps be superfluous to recapitulate the diffi-
culties already discussed in relation to the question of

”laws.' or ”explanation” to say nothing of the assertion
that concepts such as attitudes involve nothing more than a
generalization of observations. Instead. it seems necessary
to point out merely. that most workers in the field of
sociology would probably be gratified (albeit astonished)

to learn that such concepts as attitudes. instincts and
values are on the same level of explanation and prediction
as the law of gravity. This apparent merging of probabilistic
and universalistic forms of explanation not only violates
the previously discussed criteria for explanation in the
social sciences. but introduces logical inconsistency into
Adler's own argument. For the cornerstone of his argument
rests on the identification of sociology with the empirical
natural sciences. It has already been argued by the present

writer that concepts such as attitudes are not only theoreti-

cal constructs. but in addition represent idealized and

 

29Ibid.

61

probabilistic components of an explanatory scheme.

In Nagel's words: ”As the history of science as
well as common experience amply testify. correlations be—
tween empirical data are rarely perfect. and generalizations
based exclusively on such correlations are almost inevitably
bound to be statistical.”30 And in Adler's words: "A
generalization does not contain anything that was not already
present in the cases from which it was drawn. in the principle
or category of generalization. and in the method of generali-
zation.”31 The use of the ideal case is not a generalization
of what is already in the cases and if by ”category of
generalization” or ”method of generalization.” Adler is
referring to what Nagel calls patterns of explanation.”
the concepts under discussion fall into Nagel's probabilistic
type. To equate these concepts or elevate them to the same
level of explanation as the law of gravity is essentially
the same as the equating of statistical generalizations
with invariate laws.

The primary purposesfor including the above discussion

of Adler's conception of values is to illustrate the notion

 

0 .
Ernest Nagel. The Structure of SCience (New York:
Harcourt. Brace and werld. Inc.. 1961). p. 509.

31Adler. op. cit.. p. 278.

62

that the most simplified (at least superficially) definition
of a concept is not necessarily the most meaningful. Any
student of introductory psychology knows that under classical
conditioning. the dog will eventually salivate upon the
presentation of the secondary stimulus. However. few
students at this level are aware of the increasing difficulties
of getting the dog into the harness. the room. etc.. as
such processes as stimulus generalization take place. And
so it seems with the concept of values. The concept. like
the phenomena which it seeks to explain. is not a simple one.
The present writer would contend. therefore. that if one is
to criticize the complex. and seemingly shifting manner in
which Parsons uses values (i.e.. the articulation of different
levels. and external and internal systems). the criticism is
more accurately aimed at his frequent lack of clarity in
writing. rather than at any alleged weakness inherent in
any complex explanatory scheme.

Near the beginning of this section on values and
value orientations. it was pointed out that Parsons formulates
values as rooted in the non-empirical level. the level of
meaning to the individual. and the relation of the individual

to the society. Kolb. in discussing the non-empirical level

63

or ”moral norms.”32 suggests that these norms are based on
”imputed” meanings. which give rise to and legitimizes such
norms. It seems apparent. that this line of thinking. if it
is to be empirically tested. must be done by tapping non-
observable phenomena. Thurstone refers to this problem in
terms of the ”subjective metric.” The issue is raised at
this point to emphasize the previously stated position that
social phenomena are not always or necessarily directly
observable by currently available empirical techniques.
However. Parsons is not the only writer who uses the concept
of values in a variety of ways.

According to Morris33 the term ”values” is usually
defined in one of three ways. The first usage is that of
preference of one object over another. These are referred
to as ”operative values.” The second usage of the term is
in terms of the anticipation of the outcome of the behavior.
In this case. he refers to ”conceived values.” Finally.
there is the level of what ”ought” to be preferred. regardless
of what is in fact preferred. These are referred to as

”object” values. This breakdown does not seem to be very

 

32Kolb. op. cit.

3Charles Morris. Varieties of Human value (Chicago:
University of Chicago Press. 1956). chapter 1.

64

different from Parsons. We might equate operative values
with Parsons' ”motivation of values” and the ”conceived
values” are similar to the legitimation of social action.
The third usage is not so amenable to substitution. While
Parsons' ultimate values may be considered in terms of
”ought” and ”should.” their basis is in the religious and
philosophical realm. while Morris' object values may for
example. be in reference to what one ”ought” to do if he
wishes to maintain his health. In this case. the imperative
aspect is handed down by the physician. not the theologian.
or the philosopher. At any rate. there is a consistent
thread which runs through all three of the definitions used
by Morris. As Catton points out: ”Value may refer to the
preferred. to what is conceived as preferable. or to the
'actually” preferable.”

Actually. most of the sociological uses of the term
”values” utilize some notion of preference. choice. or
standard. There is at least an implicit similarity. based
on this criteria of choice.e¢c.. in most of the definitions.
Thus. it seems that one reasonably consistent use of the
concept of values. deals with some kind of a choice. according
to some standard. In all cases. it seems that the definitions

employed are relativistic. The choice is not absolute. or

65

a function of a single level of operating pressures. Parsons
may postulate five (or six) choices of action which the

actor may make in accordance with his value orientations.
Others tend to place the notion of values in a more discern-
able hierarchical framework. Catton. for example. when
talking about variation of field strength in his ”Theory of
Value.” follows Maslow at one point in describing a hierarchy
of needs.34 The main point here is that values are not
always equally strong. or equally important. This notion is
similar to the concept of ”saliency” as used by Krech and
Crutchfield in relation to attitudes. The concept of
saliency in this sense ”refers to the fact that not all of

a man's beliefs stand out with equal prominence in his
cognitive field.”35 In short. values. like attitudes are
multidimensional rather than unidimensional. Catton makes
this point and argues that there are six dimensions to values.
Included in this formulation are the distance of the object
(spacial. social and temporal). the probability of achieving

the goal. whether the goal is perceived as occuring ”only

 

34William R. Catton. Jr.. ”A Theory of Value.”
American Sociological Review. 24 (June. 1959). p. 311.

35D. Krech and R. S. Crutchfield. Theory_and Problems
of Social Psyghology (New York: McGraw-Hill. 1948). p. 163.

66

once.' or whether there is free selection of the choice
as opposed to coercion. 0n the other hand. the multi-
dimensional aspect of values is not necessarily a function
solely of the cognitive mapping of an individual at a parti-
cular point in time. If values are considered as having a
hierarchical distribution. some values will be dominant

over others at all times. Whether we call these fundamental.
ultimate or something else. the implication is still the
same. However. we are still operating with inferential data
or theoretical constructs asihr as the measurement of these
values is concerned.36 Before dealing with the problem of
the measurement of values. a brief consideration of the
definition of Clyde Kluckhohn. seems in order.

In the words of Clyde Kluckhohn. ”A value is a
conception. explicit or implicit. distinctive of an indi-
vidual or characteristic of a group. of the desirable which
influences the selection from available modes. means. and

ends of action.”37 Kluckhohn contends that any given value

becomes a part of the carrier in the same sense that culture

 

36Clyde Kluckhohn. ”Values and Value-Orientations in

the Theory of Action: An Exploration in Definition and
Classification.” in Parsons and Shills. op. cit.. p. 395.

37Ibid.

67

is ”built into” its carriers. At the same time he writes.
”. . . the social science abstraction 'value' is not abstracted
from neurological properties but from verbal and nonverbal
behavioral events. These internalized symbolic systems do
have a special status as regards methodology. requiring in
part. at least at present. a verstehen rather than an erklaren

type of interpretation.”38

 

Adler bases much of his objection to this usage of
the term value on the apparent subjective nature of such
values and the need for ”verstehen”or ”empathy” in inter-
pretation. rather than natural science observation. The
issue here. seems to hang on the previously discussed question
of the level of analysis. If values are equated with action.
then the concept becomes useless as Adler suggests. How-
ever. this reasoning is based on the assumption that:

”Even a superficial survey of the evolution of some sciences
of culture indicates that their methodological progress is
correlated with their growing emphasis on the study of human

actions rather than on the study of cultural values abstracted

from the actions in which they occurred.”39 The immediate

 

381bid.. p. 396.

39Florian Znaniecki. Cultural Science: Their Origin
apd Development (Urbana: university of Illinois Press.
1952). cited by Adler. op. cit.. p. 279.

68

question which arises is whether or not the abstract analyti-
cal analysis of Parsons or Kluckhohn does. in fact. ignore

the action dimension of values. Kluckhohn writes: ”Acts.

as has been said. are always compromises among motives. means.
situations. and values.”40 While the process of inference

is important in the determining of values. Kluckhohn discusses
such problems as standard or formalized values as well as
non-verbalized or ”real” values. In terms of the analysis

of values. Kluckhohn suggests the use of questionnaires.
projective techniques. simple experiments as well as the
observation of behavior in choice situations and under various
conditions such as crises. It seems. therefore. that this
approach does not ignore the action context in which the

values are deduced.

Toward a working Definition

In the preceeding discussion. certain notions have
continued to be evident. even though the specific context
in which they were used has varied. Whether the discussion
is concerned with beliefs. opinions. attitudes or values.
it is clear that at least two dimensions are involved.

These are variously referred to as the internal and the

 

40Kluckhohn. op. cit.. p. 406.

69

external. the subjective and the objective. or the covert
and the overt dimensions. The point has been made quite
frequently. that all of these concepts are theoretical
constructs which are defined on the basis of the observation
of verbal and non-verbal behavior. Thus we arrive at a

two dimensional conception of the basic concepts. One
dimension has to do with the internal state of the organism.
The second dimension is external to the organism and is
embedded in the super-structure of the social system of

a given society. .pggp dimensions are abstractions based on
the inference of values. attitudes. systems. etc.. from
verbal and non-verbal behavior. All of these concepts

seem to involve some notion of choice or preferential
behavior. At the structural level. thepreferential

aspect may be inferred from the formal standards of value.
as determined by expressed beliefs. attitudes. opinions

and values. At the internal or cognitive level. the preference
of an individual is inferred from his verbal and non-
verbal behavior under specific conditions. The distinction
rests on the emphasis on 295p verbal and non-verbal
behavior at the cognitive level. whereas greater emphasis
is placed on the verbal level when operating at the

structural level.

70

This distinction is more important than it may
appear at first glance. The ”preferred” values of a given
social system may continue to maintain a firm position in
the super-structure even though these values bear little
relationship to behavior. The inability of small businesses
to compete with the growing giant corporations. chain stores.
etc.. has not removed the idea of ”free—enterprise” from
the formalized standards of value in the United States.
Similarly. the idea of ”individualism” continues to be
espoused by both liberal and conservative politicians. even
though numerous factors such as jobs in public office.
insurance rates. bank loans. rental of apartments. etc..
are all contingent upon continual conformity to rules or
predictability. rather than individualism and freedom of
choice. In short. the operating codes of the social system
may be quite divergent from the formal codes without destroy—
ing the values exemplified in the formal codes.

On the other hand. the situation at the internal
level of the individual is not the same. If an individual
continually acts or behaves in a manner which is inconsistent
with his expressed values. he will rapidly be dubbed as a
liar or a hypocrite. Eventually. the same thing may occur

at the social system level. but it is a much slower process.

71

At the individual level. pp; instance of inconsistency of
expressed values and behavior. is often enough to bring
about the rejection. It should be noted that in this case.
the individual's actual behavior is taken as the true
indication of his internal cognitive state and thus. it is
the inconsistent expressed values which are rejected. This
is exactly the opposite of the system level where the formally
stated values remain as the true indicant even though the
operating behavior is inconsistent with the expressed value
system.

In view of the confusion in terminology over con-
cepts such as attitudes. beliefs. opinions and values. no
attempt shall be made to specify precise definitions which
apply to each of these various terms. Instead. a single
definition shall be offered which can be modified to encompass
all of the terms. This is a two part definition which
places one dimension at the cultural level and the other
dimension at the cognitive level. It is felt that different
techniques may be required to measure these dimensions. but
they are not completely unrelated. If we are studying a
society with strong religious values. we should be able to
observe these values in the formal institutions of the

society. and generally in the expressed or public values

72

of the individuals of that society. Where the cultural
values are less intense and greater variation in beliefs is
permitted. it seems reasonable to assume that the degree of
correspondence between ”public and private” values is

» relatively high. As the cultural values become more intense
with greater pressure toward agreement. it is likely that
the public or expressed values of an individual are a less
stable indication of what he really believes.

From the measurement point of view. the present
study deals only with the cognitive level. For the most
part. the items of the instrument concern areas of belief
or opinion. although a few of the items may be viewed as
indicants of past behavior.41 The data of the present
study does not permit us to validate the instrument to a
behavioral criterion. The study is limited. therefore. to
the comparison of the expressed responses of subjects in
three different samples. On the other hand. while the
cultural dimension is not subjected to measurement directly.
it must be taken into account to justify the placing of the
different samples at different points along a hypothetical
continuum of industrialization. Let us now consider the

problem of the cross-cultural measurement of values.

 

41The instrument will be discussed in greater detail

in the following chapter.

CHAPTER IV
METHOD

Purpose of this Chapter

This chapter examines the instrument. samples and
methodology of the present study. The examination of the
instrument goes beyond a simple description of items and
grapples with the underlying assumption that samples
approaching a high degree of industrialization will score
at one end of the sub-scales. while samples with a low
degree of industrialization will score at the opposite
.end of the scales. A series of predictions are set forth
in this chapter. However. it should be kept in mind that
these predictions are not intended as substantive hypotheses.
Instead. the predictions are used as a hypothetical frame—
work for examining the assumptions of the instrument and

the procedure.

The Instrument

The MSU Work Beliefs Check-List (WBCL) is a 44 item

instrument developed by A. O. Haller.l According to the

 

lThe instrument. instructions for administration. and
the scoring key are included in Appendix C.

73

74

instructions: ”This check-list is made up of statements
people often say they believe. You will probably find that
you agree with some and disagree with others. If you agree
with a statement. circle Agpgg: if you disagree with a
statement. circle Disagree. Do not omit any.” It seems
apparent that the instrument is designed to tap what has
been defined in this thesis as the internal cognitive
dimension of values.

The 44 items of the check-list are analytically
clustered into six sub-scales. Since a certain amount of
familiarity with the instrument is necessary to comprehend
the results of the data analysis. each of the sub-scales

shall be briefly examined at this time.

Sub-Scale l: Belief that work has intrinsic versus
instrumental value

1.1 The only purpose of working is to
make money. Agree Disagree

1.2 I believe a man needs to work in
order to feel that he has a real
place in the world. Agree Disagree

1.3 I feel sorry for people whose
jobs require that they take
orders from others. Agree Disagree

1.4 Every man should have a job that
gives him a steady income. Agree Disagree

 

2It will be recalled that beliefs are subsumed under
the definition of values developed in the preceding chapter.

75

1.5 The happiest men are those who work
only when they need money. Agree Disagree

1.6 Doing a good job day in and day out
is one of the most satisfying

experiences a man can have. Agree Disagree
1.7 A regular job is good for one. Agree Disagree

1.8 I feel sorry for rich people who
never learn how good it is to
have a steady job. Agree Disagree

In this scale. as in those that follow. the underlined
responses are scored one point each. The responses which
are not underlined are scored zero points each. Consequently.
the mean score for any single item falls between zero and
one. Moreover. since the only possible scores are zero or
one. the mean score for any given item is equivalent to the
proportion of boys receiving a score of one. The sub-scale
total score is determined by adding the individual item
scores. Therefore. this scale has a possible range of

zero to eight. The higher scores are considered as indica-
tive of the intrinsic value of work. From a close scrutiny
of the items. it appears obvious that the basic assumption
underlying this scale is that persons with a belief that
work has intrinsic value will agree with items espousing the
value of work and disagree with items which imply that work

is simply a means to a financial end.

76

In the opinion of the present writer. all of the
items do not have equal face validity. Thus. the immediate
question which comes to mind is whether we should expect a
unidimensional scale. This question can be answered better
after the results have been presented so it shall be post-
poned until chapter six. It should be noted. also. that
assumptions about the relationship of responses to the level
of industrialization of the sample is reserved for a later

section of the present chapter.

Sub-Scale 2: Positive versus negative evalgation of structured
time

2.1 I don't like people who are always
right on time for every appoint-
ment they have. Agree Disagree

2.2 I feel sorry for people who
have to do the same thing every
day at the same time. Agree Disagree

2.3 I don't like to have to make
appointments. Agree Disagree

2.4 I believe that promptness is a
virtue. Agree Disagree

2.5 I usually schedule my activities. Agree Disagree

2.6 I'd rather let things happen in
their own way rather than
scheduling them by a clock. Agree Disagree

2.7 It makes me feel bad to be
late for an appointment. Agree Disagree

77

2.8 I expect people who have appoint-
ments with me to be right on time. Agree Disagree

In this scale. the items deal with promptness.
appointments. and the scheduling of time. The high scores
indicate a positive evaluation of structured time. In the
case of sub-scale one. there is an assumption that there is
a linkage between the intrinsic evaluation of work and a
high degree of industrialization. A similar assumption
exists for this sub-scale in relation to industrialization

and the positive evaluation of structured time.

Sub—Scale 3: Positive vepsus negative evaluation of
physical mobility

3.1 I would be unhappy living away
from my relatives. Agree Disagree

3.2 I hope to move away from here
within the next few years. Agree Disagree

3.3 People who can't leave their
hometowns are hard for me to
understand. Agree Disagree

3.4 A man's first loyalty should be
to his home community. Agree Disagree

3.5 When a boy becomes a man. he
should leave home. Agree Disagree

3.6 I like to see new things and
meet new people. Agree Disagree

On this scale. the high end of the range indicates

a positive evaluation of physical mobility. 'An underlying

78

assumption here is that more industrialized samples will

have a more positive evaluation of physical mobility.

A close look at the items of this scale will show that

five of the six items deal with family and community ties

rather than any pure notion of physical mobility.

This

point will be discussed in greater detail in the next two

chapters so the reader should keep it in mind for future

reference.

Sub-Scale 4: Positive versus negative evaluation of change

4.1 I like to try new things.

4.2 On the whole. the old ways of
doing things are the best.

4.3 Life would be boring without
new experiences.

4.4 I like people who are willing
to change.

4.5 On the whole. most changes
make things worse.

4.6 The happiest people are those
who do things the way their
parents did.

4.7 New things are usually better
than old things.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

This scale has a positive range of zero to

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

seven .

Once again. the higher scores indicate a more positive

evaluation of change. and it is assumed that the more

79

industrialized samples will have a positive evaluation of

change.

Sub-Scale 5:

determination of events

501

5.6

5.8

I believe that a person can get
anything he wants if he's
willing to work for it.

Man should not work too hard.
for his fortune is in the
hands of God.

A man shouldn't work too hard
because it won't do him any good
unless luck is with him.

With a little luck I believe I
can do almost anything I really
want to do.

A person shouldn't hope for
much in this life.

If a man can't better himself
it's his own fault.

Practically everything I try to
do turns out well for me.

I usually fail when I try
something important.

in the internal determination of events.

Belief in internal versus external

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

High scores on this scale are indicative of belief

The underlying

assumption of this scale is that the more industrialized

samples will tend to believe in the internal determination

of events.

80

Sub-Scale 6: Positive versus negative evaluation of
deferred gratification

6.1 I would rather work than go to
school. Agree Disagree

6.2 Money is made to spend. not to
save. Agree Disagree

6.3 I think there's something wrong
with people who go to school for
years when they could be out
earning a living. Agree Disagree

6.4 One gains more in the long run
if he studies than if he gets a
job. Agree Disagree

6.5 The more school a person gets
the better off he is. Agree Disagree

6.6 Generally speaking. things one
works hard for are the best. Agree Disagree

6.7 When I get a little extra money
I usually spend it. Agree Disagree

Positive evaluation of deferred gratification is
indicated by high scores on this scale. More industrialized
samples are assumed to place a high evaluation on deferred
gratification. It should be carefully noted that the items
of this scale deal with questions of studying. going to
school or spending versus saving money. There are no items
which deal directly with how the money is spent or what
kind of job is involved. The scale is defined therefore.

in terms of broad values generally associated with the

81

middle class in the United States. Essentially. these
values are extended education. thrift and hard work. We
shall return to this point in chapter six.

Each of the six sub-scales is defined in polar terms
such as positive versus negative. internal versus external.
or intrinsic versus instrumental. As pointed out above.
there is an underlying assumption that highly industrialized
samples will fall at one end of the polar continuum. and
non-industrialized samples will fall at the opposite end.

It is to this assumption that we now direct our attention.

The Typological Basis of the WOrk
Beliefs Check List

In an earlier chapter. it was noted that in Nagel's
opinion. the only social science which attempts to use the
logical device of stating the ideal case is economics.

The present writer is in complete disagreement with this
view. As a matter of fact. the literature of sociology and
anthropology suggests a long history of the use of the

ideal type. It is not the purpose of this thesis to provide
a detailed coverage of the history of the ideal type.

It will suffice to point out that the basic concept of the
ideal type or some modification of it. has been prominant

in the writings of Toennies. Durkheim. weber. Sordkin.

82

Becker and Parsons. to mention a few.

While the various typologies are not identical.
there does appear to be certain common features which emerge
from some of them. In general. one end of the continuum
is viewed as rural. non-industrial with a strong emphasis on
kinship. traditional and sacred values. At the other end
of the continuum. there is the urban. industrial type with
weakened kinship ties beyond the nuclear family. and an
emphasis on changing and secular values.

It is this kind of milieu which forms the basis of
the work Beliefs Check List. Out of the literature on
success. achievement. stratification and even mental ilIness.
has grown a certain image of the ”necessary” value orientations
for success in complex industrial societies. This image
generally requires a positive evaluation of structured time.
physical mobility. change and deferred gratification. In
addition. there is usually an emphasis on individualism
and the internal determination of events. It should be
noted that individualism and the internal determination of
events is not necessarily viewed as part of the same value
dimension. Florence Klukhohn. for example. has both a man-
nature orientation and a relational orientation (which

deals with man's relationship to other men).3

 

3Florence R. Kluckhohn. Fred L. Strodtbeck. Variations

in Value Orientations (Evanston. Illinois: Row Peterson Co..
1961).

83

In any case. there is usually an expectation that
the more industrial societies will be close to the above
image in value orientations. Consequently. it is not too
difficult to understand the assumptions concerning the
relationship between level of industrialization and value
orientations for sub-scales two through six. Whether the
specific items of any given scale are unidimensional or in
accord with the orientations designated by a given scale is
another question. This question shall be examined in the
next chapter where the results of the data analysis are
presented.

The assumption that the more industrialized societies
will have an intrinsic rather than an instrumental orientation
to work seems less firmly grounded. Many writers have
contended that With industrialization there is increased
division of labor and specialization. These writers would
argue that an intrinsic valuation of work is more likely
among the older skilled craftsmen who could derive job
satisfaction in a complex article. than among the button
pushers of the large factory. Even if we shift to the
middle-class. the same question arises. Some contemporary

writers such as Riesman4 or Whyte5 paint a picture of the

 

4David Riesman. Nathan Glazer. and Reuel Denney. The
Lonely Crowd (Garden City. N.Y.: Doubleday & Co.. Inc.. 1956).

5William H. Whyte. Jr.. The Organization Man (New York:
Simon and Schuster. 1956).

84

middle class as being more concerned with security and the
opinions of others than with individual innovation. If
the intent of this value orientation is that ali_work.
regardless of what it is. has intrinsic versus instrumental
value. it seems difficult to understand why garbage collectors
or shoe shine boys are consistently given low occupational
evaluation scores in inudstrialized societies. If the intent
is to imply that ”doing the kind of work one enjoys” has
intrinsic value.then the individual's personal criteria are
important. The items of this sub-scale are not geared to
provide this kind of information. It is not the intention
to argue at this point whether one group of writers is
correct and another incorrect. but simply to alert the
reader to the present contention that the assumption of a
stronger belief in industrial societies. that work has
intrinsic value. does not appear to be as firmly anchored
to the literature as the other scales.

The typological tradition in social science. however
imperfect. forms the basis of the underlying assumptions
of the work Beliefs Check List. Perfection. however. is
never an actual feature of social research. As stated
earlier. neither theory nor method provides us with a map

but only guides. In the case of the polar-type used in the

85

check list. a general statement by Loomis and McKinney best
summarizes the purpose.

The polar-type formulations. implicitly at first. but
in recent years with increasing explicitness. have
firmly established the point that the continuum is a
vital notion in the comparative analysis of social
phenomena. The types establish the ”outer limits"
or standards by means of which the process of
change or intermediate structural forms can be compre—
hended from the perspective of the continuum.6

Within this context. let us consider the samples of the

present study. and the hypothetical predictions based on the

assumptions of the instrument.

The Samples

The subjects in all three samples of the present
study were male high school students. The ages included in
the respective samples were not exactly comparable. One
sample included all seventeen-year-old boys in school in
Lenawee County. Michigan. during the spring of 1957. The
original size of this sample was 441 but the final size of

the sample used in the analysis was 439.7 The Turrialba

 

6Charles P. Loomis and John C. McKinney. ”The Appli-
cation of Gemeinshaft and Gesellschaft As Related to Other
Typologies.” in F. Tonnies. Community and Society (translated
and edited by Charles P. Loomis. East Lansing: Michigan
State University Press. 1957). p. 12.

7A few cases were dropped in both the Lenawee and the
Turrialba samples. However. since this was due to the method
for handling ”no response” answers. it shall be explained in
greater detail under the section on method of analysis.

86

sample consisted of all boys in high school in Turrialba in
1959. The age range was thereby expanded to include those
between 14 and 18. The Lansing sample consisted of all
Mexican-American boys between the ages of 15 and 18 living
in the metropolitan area of Lansing. Michigan. in 1958 and
who were attending high school. Thus. it is immediately
obvious that the three samples were not matched samples on
the variable of age. This was in part a practical necessity.
since a strict chronological matching would have reduced the
size of the Lansing and Turrialba samples to a point where
any attempted analyses or comparisons would have been futile.
In addition. it was felt that any attempt to achieve strict
matching of samples on a chronological basis would be of
doubtful utility when the norms governing education. and the
type of education available is not directly comparable in the
respective samples.

The predictions of the present study require the
positioning of the samples at different points along a
hypothetical continuum of industrialization. The Lenawee
sample is viewed as the most highly industrialized and the
Turrialba sample is viewed as the least industrialized.

The Lansing sample consists of Mexican-Americans.~ It is

assumed that this sample includes persons from both ends

87

of the continuum and. therefore. the sample as a whole is
placed in a transitional position between the Lenawee and the
Turrialba samples. If we think in terms ofthe urban rural
breakdown. it is apparent that the samples do not represent
”pure” types and. therefore. the assignment of a particular
sample to a particular position can be questioned. However.
the samples are assigned on the basis of the degree of
industrialization which the system as a whole has achieved.
Therefore. it is felt that even though the Lenawee sample
contains both urban and rural youth. the proximity to such
urban-industrial areas as Detroit. Michigan. and Toledo.
Ohio. together with a paved highway system and the existence
of a flourishing light industry in 1957. tends to push this
sample toward the industrial end of the continuum. In
addition. the high development of the mass media in the
United States. tends to reduce isolation and increase the
probability of the inculcation of industrial value orientations.

While the Turrialba sample does not constitute a
”pure” example of the non-industrialized end of the continuum.
it does approach this end. since it is primarily an agri-
cultural area.

Among the Mexican-Americans. we find both acculturated

and unacculturated persons. Some of the boys were born in

88

Michigan while others were born in Texas. Only three boys
were born in Mexico; however. language and cultural contact
with Mexico is assured through parents.8 Interactional ties
with relatives and friends in Mexico are frequently main-
tained and in some cases return visits to Mexico are made.
The Lansing sample is. therefore. expected to contain
individuals with a range of exposure to the values of
industrial society. from high to low exposure. It is for
these reasons that the Lansing sample was placed between the
Lenawee and the Thrrialba samples along the hypothetical
continuum of industrialization.

The size of the samples used in the analysis was
439. 112. and 87 for Lenawee. TUrrialba. and Lansing respectively.
The subjects were given the choice of agreeing or disagreeing
with a statement and were instructed not to omit any answers.
Nevertheless. some of the items were left blank. On the
basis of the distribution of the blank or ”no answer”
responses. it was decided that only those persons with a
total of 4 or less than 4 such responses out of a total of

44 items would be retained in the samples. This resulted

 

8See Arturo De Hoyos. Occupational and Educationai

Levels of Aspiration of Mexican-American Youth (unpublished
Ph.D. dissertation. Michigan State University. 1961).
esp. pp. 64 and 87.

89

in changing the sample size of Lenawee from 441 to 439 and
changing the Turrialba sample size from 118 to 112. The
Lansing sample was not affected by this criteria.9

In addition. to the three samples discussed above.
the data were analyzed in terms of a fourth sample. This
consisted of the Lenawee. Turrialba and the Lansing samples
combined. The size of this pooled sample was 638. Although
this sample is not discussed in the text of the thesis. the

results are included in Appendices A and B.

Predictions and Method oprpalysis

The predictions of this study are hypothetical rather

than substantive. In other words. while substantive data

are used. it is in the context of testing a series of hypo—
thetical conditions and predictions which should occur if

the sub-scales of the instrument are unidimensional and if
the assumptions about the industrial continuum are valid.

The predictions are thus. intended for methodological
purposes rather than any substantive findings. Most of the

data are analyzed by primary axes factor analysis with

 

91t should be noted. however. that the 87 cases in
the Lansing sample represent a reduction of four cases from
the original 91. This was done because the four cases were
no longer in school. Unfortunately. five such cases remained
in the sample.

90

varimax rotations. Before considering the method of
analysis in any greater detail. the hypothetical predictions
shall be presented.

Hypothetical Model and Predictions

Industrial - Gesellschaft - Lenawee

The Belief Value Areas scores for this sample are
expected to fall at the high end of the score range. Thus.
the predictions are for high means. low variances and low
correlation coefficients. when the Belief Value Areas scores

are compared with each other.
Intermediate - Turrialba

The predictions here are for middle range mean scores.
low variance scores and low correlation coefficients.
Gemeinschaft to Gesellschaft - Lansing

The predictions are for middle range mean scores.
high variance and high correlation coefficients.
Gemeinschagp

The predictions here are forlow means. low variance

and low correlation coefficients.

91

Specific Conditions. Predictions and Rationale

If the items within each sub-scale are tapping a
unidimensional continuum. there should be positive
correlation between items within each sub-scale and a one
factor solution accounting for much of the variance. How-
ever. the correlations and the factor loadings should not
be the same for all three samples. but should differ accord-
ing to where the sample falls along the hypothetical continuum
of industrialization.

Condition A: Intapeitem relationships within each
sub-scale

Prediction A1: There will be a one factor solution for each
sub-scale within each sample.
Prediction A2: The Lenawee sample will have high means.
low variances and low correlation coefficients.
Prediction A3: The Turrialba sample will have middle range
means. low variance and low correlati on coefficients.
Prediction A4: The Lansing sample will have middle range

mean scores. high variances and high correlation

coefficients.

Condition B: All_items ipgespective of sub—scales within
sagples

Prediction B: There will be a seven factor solution with one

92

large factor accounting for much of the variance and

six small factors corresponding to the six sub-scales.

Condition C: Sub-Scale total scores within samples

Prediction C: There will be a one factor solution for all
three samples on the sub-scale total scores within

samples.

General Propositions and Alternative Conclusions

The above predictions are expected to occur if:
the following three hypothetical statements are true:

1. There exists a set of value orientations which
differentiate among those in the industrial versus
those in the non-industrial market.

2. The work Beliefs Check List is an adequate measure
of these orientations.

3. The samples under investigation are distributed
according to these value orientations.

In addition to the gross predictions of one factor
or seven factor solutions. the factor analysis affords us
an opportunity to examine the internal structure of the
instrument. Thus. an effort will be made to interpret the
factors as well as to ascertain whether they are unidimensional.

This will be done in the light of the earlier discussions of

93

concept formation and values. This data together with the

means and variances will enable us to decide whether or

not the scales are measuring what they are supposed to measure

and whether the respective samples actually fall at the

predicted positions along the hypothetical continuum of

industrialization. If the expected relationships are not

obtained. one or more of the three ”ii” propositions may be

rejected within the following framework.

1.

If the instrument fulfills the criteria of uni-
dimensionality of scales but the expected relation—
ships are not obtained. then either the assumption
of the existence of a continuum of value orientations
congruent with industrial society will be rejected.
pg; it will be concluded that the samples under
investigation are not distributed along such a
continuum.

If the instrument does not fulfill the criteria

of unidimensionality of sub-scales. it will be
concluded that some or all of the sub-scales are
invalid.

If the instrument fulfills the criteria and the
results are in accordance with expectation. it will
be concluded that the initial ”if” propositions

are valid.

94
Technigpes for Data Analysis

It has already been pointed out that the data was
analyzed in terms of means. variances. correlations and factor
analyses. All of the analyses were performed by the computer
MISTIC. All correlations were product moment correlations.
and the factor analyses were principle axes with varimax
rotations. The communality estimation was based on Burt's
scheme for modified highest. For this purpose. MISTIC
library routine M23-M (which either places the highest entry
of each column of a correlation matrix in the diagonal.
or modifies this entry and places it in the diagonal) was
used.

Most of the results are discussed in terms of the
principle axes factor matrices. This immediately opens the
controversial door of whether factors should or should not
be rotated. The major justification for the use of unrotated
factors lies in the nature of the predictions. It will be
recalled from the prior section that most of the predictions
involving the factor matrices are stated in terms of one
factor solutions. With the principle axes method. ”each
factor extracts the maximum amount of variance (i.e.. the
sum of squares of factor loadings is maximized on each

factor) and gives the smallest possible residuals.”10

 

10Benjamine Fruchter. Introduction to Factor Analysis
(New York: D. Van Nostrand Co.. Inc.. 1954). p. 99.

95

Consequently. it seemed more likely that a large common
factor. if it existed. would show up in the unrotated factors.
This logic seems particularly valid when varimax rather than
quartimax rotations are used. As Kaiser points out. the
quartimax criterion aims to simplify the row or test of the
factor matrix rather than the column. Consequently. a
general factor will often result from the quartimax rotation.
particularly when there is a large general factor in the
unrotated matrix. Since the varimax rotation maximizes
inequalities in the column. the probability of a large
common factor is diminished.11 This suggests that it may
have been more fruitful to use quartimax rotations. How-
ever. it must be kept in mind that we were interested not
only in whether large common factors exist. but also in
which items were defining the factor in each of the six
sub-scales and in the complete instrument when all of the
scales are included. According to Wrigley. Saunders. and

Neuhaus: ”To summarize these comparisons. present evidence

 

11This discussion is based mainly on Henry F. Kaiser.

”The Varimax Criterion for Analytic Rotation in Factor
Analysis.” Psychomepgika. 23 September. 1958. pp. 187-200.

See also Charles wrigley. David R. Saunders and Jack O.
Neuhaus. ”Application of the Quartimax Method of Rotation

to Thurstone's Primary Mental Abilities Study.” Psychometrika.
23. June. 1958. pp. 151-70. And Harry H. Harmon. Modern
Eagtor Analysis (Chicago: The University of Chicago Press.
1960). pp. 301-08.

96

suggests the quartimax results are the more parsimonious
and may be the more stable when the number of factors for
rotation is increased. and that the varimax results are
more stable when the number of tests in the battery is

”12 Since there are two schools of thought

increased. . . .
as to whether factors should or should not be rotated. it
seemed reasonable to use the principle axes loadings for the
predictions about the number of factors which appeared in a
solution. and to supplement this with the varimax rotations
when trying to ascertain the factorial structure of the
instrument. in terms of specific items. Some unanticipated
problems occurred in relation to the method of ranking the
unrotated factors prior to their rotation by MISTIC. These
problems shall be discussed further in the final chapter
under limitations of the study. For the present. it is
sufficient to say that the end result was a heavier
rehaibce on the principle—axes loadings than was originally
anticipated.

Finally. a few words should be said about the
criterion for how many factors to rotate. Guilford suggests

that: ”When any factor has one or more loadings of plus or

minus .20 or higher. there is enough variance present to aid

 

12W'rigley. Saunders and Neuhaus. op. cit.. p. 168.

97

in the rotations.”13 Where possible. an effort was made

to adhere to this criterion. However. as Wrigley points out:
”When factors are rotated. increasing their number causes

the larger factors to split into smaller . . .”14 Consequently.
not all factors which meet Guilford's suggested criteria.

were rotated. The reason for this will be clearer when

the limitations of the study are discussed in chapter six.

 

13J. P. Guilford. Psychometric Methods (New York:

McGraw—Hill Bodk Co.. Inc.. 1954). p. 500.

14Charles Wrigley. ”The Distinction Between Common
and Specific Variance in Factor Theory.” The British Journal
of Statistical Psychology. lg. November. 1957. p. 86.

CHAPTER V
RESULTS

Purpose of this Chapter

 

In this chapter the results of the data analysis
are presented and discussed in relation to the theoretical
predictions. While there is some effort to interpret the
results in this chapter. a more general attempt to deal with
the findings within the context of the present thesis is
reserved for chapter six.

For purposes of this analysis. all items with loadings
of .20 or greater were retained as significant contributions
to the factor. As previously pointed out. the predictions
will be discussed in terms of the principal axes loadings.
but the varimax rotations will also be considered where such
consideration will aid in the clarification of the factorial

structure.

Condition A: Intep Item relationships within each sub-scale.
Prediction A1: There will be a one factor solution for each
Sub-scale within each sample.

Prediction A2: The Lenawee sample will have high means.

low variances and low correlation coefficients.

98

99

Prediction A3: The Turrialba sample will have middle range
means. low variances and low correlation coefficients.
Prediction A4: The Lansing sample will have middle range
mean scores. high variances and high correlation
coefficients.
[gapeScale l: Belief that work has intense versus instrumental
yells
From Table 1. it can be seen that the first factor
accounts for 50 percent of the total variance in the Lenawee
sample. This factor is defined by items two through seven.
Items one and eight do not have significant loadings on
this factor. The highest loadings are on items two. four.
five and seven. In the varimax rotation. significant loadings
for the first factor are retained on items two. four and
seven. While the rotation providesxuswith a three factor
solution. it should be noted that the only change in the
items with significant loadings is the addition of item
one to the second factor. Item eight still does not have
a significant loading on any of the factors.
The Turrialba sample. as indicated in Table 2. has
a two factor solution. The first factor is defined by items

one. five and seven; while the second factor is defined by

items three. four. five and eight. However. the second factor

is a bi-polar factor.

100

Table 1. Factor matrix for WBCL Sub-Scale l. Lenawee
sample (N-439).

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale l)
I II III I II III
1.1 12 -34 -08 -19 32 00
1.2 54 -10 07 22 41 30
1.3 25 -l6 17 —02 21 27
1.4 37 27 -16 47 16 -01
1-5 45 -22 -20 13 53 03
1.6 27 01 25 11 08 34
1.7 38 32 —01 47 08 13
1.8 12 06 13 09 00 17
% Total
Variance 50 20 10 30 33 17

 

Table 2. Factor matrix for WBCL Sub-Scale l Turrialba
sample (N-112L*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale l)
I II III I II III
1.1 62 08 -13 64 -05 -06
1.2 15 02 -42 20 -01 -40
1.3 -16 43 -02 -06 46 -03
1.4 —08 -21 09 -13 -19 07
1.5 28 32 17 32 25 21
1.6 00 00 00 00 00 00
1.7 39 -03 25 34 ll 29
1.8 16 -50 -00 05 -51 -00
% Total
Variance 31 27 31 31 27 13

 

*Decimal points omitted.

101

In accordance with predictions. the Lansing sample.
as indicated in Table 3. provides a fairly clear cut one
factor solution. Also. in accordance with prediction A2 and

A the factor loadings for the Lansing sample are higher

3.
than for the Lenawee or the Turrialba samples. The first
factor for the Lansing sample accounts for 55 percent of

the total variance and is defined by items two. four. five
six. seven and eight. The same factorial structure is
retained in the rotations and in both the principal axes and
the varimax rotation. items four and seven have very high
loadings.

Table 3. Factor matrix for WBCL Sub-Scale l. Lansing
sample (N-87).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale l)
I II III IV I II III IV
1.1 -05 34 30 16 -05 47 07 07
1.2 49 —31 -16 20 32 -35 34 25
1.3 04 -22 15 22 -04 -04 34 -03
1.4 87 -14 -01 -14 83 -21 18 17
1.5 52 39 -08 28 36 22 07 57
1.6 43 -20 33 19 35 06 49 00
1.7 78 17 16 -31 86 12 -03 07
1.8 33 13 -38 12 20 -17 -10 45
% Total
Variance 55 13 11 9 47 12 13 16

 

*Decimal points omitted.

102

Thus. is appears that the prediction of a one factor
solution is confiremd in the Lenawee and the Lansing samples
but not in the Turrialba sample where a two factor solution
is obtained. Only items five and seven contributed significant-
ly to the first factor in all three samples. However. for
the Lenawee and the Lansing samples. the first factor was
determined. in part. by the same five items (two. four. five.
six and seven).

For sub-scale 1. it seems that items one. three and
eight have the least transcultural relevance. A close lock
at the items of this scale gives us a possible clue to the
reasons. This scale attempts to differentiate between
intrinsic and instrumental evaluations of work. waever.
both items three and eight appear to deal with two dimensions.
Item three states: ”I feel sorry for people whose jobs
require that they take orders from others.” Disagreement
with this item is scored as intrinsic evaluation of work.
However. the item deals with orientation to authority. rather
than orientation to work. To illustrate this distinction.
let us consider the following example.

In setting up classifications for promotions. the
United States Air Force categorizes non-commissioned officers

as either supervisors or technicians. The distinction is

103

written up in terms which acknowledge the fact that some
individuals may be highly competent workers. but do not wish
to supervise others. In short. it seems difficult to equate
the willingness to take orders with the intrinsic evaluation
of work. Examples from art. literature and science seem to
indicate that the opposite is often true.

In a similar manner. item eight seems to be a two
dimensional item. It seems reasonable to assume that a
person may place a positive evaluation on having a steady
job. without feeling sorry for rich people. Since the item
states: ”I feel sorry for rich people who never learn how
good it is to have a steady job.”. the implication seems to
be in line with the stereotype of the ”idle rich.” One
is reminded here of Weber's treatment of the Protestant
ethic. From this point of view. the accumulation of wealth
and the maintenance of a steady job do not appear as dia—
metrically opposed ends.

Item one. which fails to contribute to a one factor
solution in all three samples. states: ”The only purpose
of working is to make money.” The first temptation here is
to attribute the failure of this item to the ”all or none”
wording of the item. Most of the significant items are

worded in terms of one among several alternatives. However.

104

item five is also worded in a way similar to item one. The
meaning of the two items may still be quite different. for
the respective samples. The first item makes a clear
distinction between intrinsic and instrumental evaluation

of work. implying that work is highly undesirable for itself.
The fifth item implies a similar distinction. but the
negative implication of work seems much less clearly defined.
The statement that the only purpose of working is to make
money could be interpreted as the accumulation of excessive
wealth. On the other hand. the statement that the happiest
men are those who work only when they need money could be
interpreted in terms of being satisfied with the fulfillment
of the minimum needs. rather than the accumulation of money
as a value in its own right.

It must be noted that we have been dealing with the
correlation of items to a common factor and not with how the
respective samples responded in terms of the hypothetical
continuum of industrialization. The means and variances
for all three samples on sub-scale l are presented in
Table 4. It appears that the high mean scores and the low
variance scores for the Lenawee sample is consistent with the
predictions for that sample. The mean and variance scores

for the Turrialba and Lansing samplesare less clear cut.

105

Table 4. Means and variances for all samples for sub-

 

 

 

scale 1.

Lenawee Turrialba Lansing

M Var. M Var. M. Var.
1.1 74 19 59 24 56 25
1.2 90 09 97 03 84 14
1.3 80 16 49 25 72 20
1.4 94 06 96 04 97 03
1.5 03 06 88 10 83 14
1.6 86 12 100 00 79 16
1.7 93 07 83 14 93 06
1.8 49 25 50 25 55 25

 

Middle range mean scores were predicted for both of these
samples but high variance scores were predicted for Lansing
and low variance for Turrialba. The mean scores for both
the Turrialba and the Lansing sample appear to have more of
a tendency than the Lenawee sample to approach the midpoint
of .50. but their actual range extends upward to 1.00. While
this extended range is in accord with expectations for the
Lansing sample. it is not consistent with predictions about
the homogeneous nature of the Turrialba sample. Moreover.
variance scores are not consistently high for the Lansing
sample. or consistently low for the Turrialba sample.

We might account for this in the Lansing sample by the
expected cultural variation in this group.

It seems. therefore. that only the Lenawee sample

appears to have scores consistent with the predictions of

106

falling at the high end of the continuum. The scores
indicate that for some of the items. the Turrialba sample
tends to fall as close or closer to the industrial end of the
continuum than the Lenawee sample. On other items this is
not the case. The implication of this is that all of the
items do not represent a single factor for the Turrialba
boys. This conclusion seems consistent with the failure

to obtain a one factor solution in the Turrialba sample.

Sub-Scale 2: Positive versus Negative Evaluation of

Structuped Time

The principal axes loadings in Table 5 indicate that
one factor accounts for 52 percent of the matrix variance
in the Lenawee sample. This factor is defined by all of the
items of sub-scale 2. except items two and six. Actually.
item two has a loading of .19 or just below the .20 level
used in this analysis. In other words. only item six approaches
zero and may be considered as not contributing to this factor.
The highest loadings are on items one. three. five. seven and
eight.

In the Turrialba sample. the first factor accounts
for 38 percent of the matrix variance. As indicated in
Table 6. this factor is defined by items one. two. three. six

and seven. Six. however. is a negative loading. The highest

107

 

 

 

 

 

 

 

 

 

Table 5. Factor matrix for WBCL Sub-Scale 2. Lenawee
sample (N—439).*
Principal Axes varimax
Items Loadings Loadings
(Sub-Scale 2)
I II III I II III
2.1 34 -24 -10 09 31 28
2.2 19 10 02 21 02 05
2.3 38 -09 -34 20 47 08
2.4 24 -22 26 02 40 13
2.5 50 37 —00 62 06 02
2.6 06 33 06 66 07 13
2.7 36 -22 24 14 02 47
2.8 42 -35 23 10 11 58
% Total
Variance 52 22 13 39 20 28
*Decimal points omitted.
Table 6. Factor matrix for WBCL Sub-Scale 2.Turrialba
sample (N-112).*
Principal Axes Varimax
Item Loadings Loadings
(Sub-Scale 2)
I II III IV I II III IV
2.1 59 24 -12 O4 02 19 59 21
2.2 48 01 31 -20 -03 18 15 56
2.3 23 30 -36 -06 —03 —16 49 -06
2.4 -15 -12 -13 00 -12 -05 -10 -17
2.5 -11 38 13 28 49 -00 06 -11
2.6 -23 40 ' 24 -10 41 -31 -09 12
2.7 41 -24 14 23 -08 51 09 14
2.8 00 00 00 00 00 00 00 00
% Total
Variance 38 22 15 8 l8 19 27 18

 

*Decimal points omitted.

108

loadings are on items one. two and seven. A second factor
accounts for 22 percentof the variance while the third
accounts for 15 percent. In addition to not obtaining a
one factor solution.for the Turrialba sample. each of the
factors is complicated by having both positive and negative
loadings. The varimax rotation of the first four factors
did not simplify this situation. Instead. the matrix
variance accounted for by a single factor was more evenly
spread over all four factors. The rotations failed to change
all of the loadings of a given factor t) the same sign. It
is true that after rotation. most of the negative loadings
are below the .20 significance level. However. the
remaining positive loadings of .20 or higher are limited

to one or two items for each of the four factors.

In the light of the failure to obtain a one factor
solution on the principal axes loadings and the additional
confusion of the rotated factors. the unidimensionality of
sub-scale 2 for the Turrialba sample seems questionable.

The Lansing sample again failed to produce a one factor
solution but there is less of a problem with bi-polar factors.
The first factor accounts for 38 percent of the matrix
variance and the second factor accounts for 26 percent so

that the first two factors account for 64 percent of the total

variance. As indicated in Table 7.

109

factor is defined by items one.

two.

five.

six.

the first unrotated

seven.

eight. Only items three and four do not contribute to this

factor.

Table 7. Factor matrix for WBCL Sub-Scale 2. Lansing
sample (N—87).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 2) '
I II III IV I II III IV
2.1 35 21 32 12 41 ~07 17 29
2.2 42 29 ~26 ~17 ~07 34 37 33
2.3 12 ll 22 ~23 27 12 19 ~08
2.4 04 63 04 ~09 04 ~12 62 07
2.5 46 ~31 06 00 24 32 ~26 28
2.6 44 ~28 ~17 ~18 05 50 ~17 22
2.7 57 14 11 29 08 08 07 65
2.8 28 ~19 27 ~09 38 l9 14 06
% Total
Variance 38 26 ll 8 15 18 24 26

 

*Decimal points omitted

It appears therefore. that the prediction of a one
factor solution is supported only for the Lenawee sample.
While certain items contribute significantly to the first
factor in all three samples. the prediction of a one
factor solution is not upheld in either the Turrialba or the
Lenawee samples.

Only items one and seven have significant loadings

for all three samples. Item one states: "I don't like

110

people who are always right on time for every appointment
.they have.” Persons who disagree with this item are scored
as having a positive evaluation of structured time.
Similarly. to score at the same end on item seven. one must
agree with the statement: ”It makes me feel bad to be late
for an appointment.” It is interesting to note what

happens when we attempt to achieve meaningful interpretations
from these data.

Both the Turrialba and the Lansing samples have
app-significant loadings on item four which states: ”I
believe that promptness is a virtue.” Yet both of these
samples have significant loadings on items one and seven which
imply a positive evaluation of promptness. How do we determine
the meaning in such a case? There seems to be no simple
answer to this problem but at least two alternatives seem
plausible.

First. the Lenawee sample is the only one where a
large common factor was obtained. In addition. seven out
of the eight items made significant contributions to this
factor in the Lenawee sample. The failure to obtain a large
common factor in both the Turrialba and the Lansing samples
suggests that either the orientations toward structured time

are less solidified in these samples. or that some of the

items have differential meanings for the three groups of boys.

111

From Table 8. we see that the Lansing sample does

not have generally higher variances than the Lenawee and
Turrialba samples on all of the items. However. it does
appear that more of the items have higher variances than in
the Lenawee and Turrialba samples. In short. there appears
to be more consistently high variance items in the Lansing
sample than in the other two. The prediction that the
Lenawee sample would have high mean scores while the mean
scores for both the Turrialba and the Lansing samples would
be in the middle range is again not clearly supported.
The mean scores for the Lenawee sample are generally toward
the high end of the scale. However. they are not as high
as they were for the same sample on sub-scale 1. In
addition. two of the items on sub-scale 2 fall below the
midpoint of the possible range. Again. the Turrialba
scores cover a wide range from .42 to 100. Thus. while
some of the scores do cluster around the middle of the
possible range. this sub—scale also contains three items on
which 88. 92. and 100 percent of the boys scored at the
industrial end of the scale. There is more consistency on
the means of the Lansing sample since the means fall between
.53 and .77 on all of the items except one.

It appears. therefore. that the three samples do

not show a consistent and clear cut tendency to have the

112

Table 8. Means and variances for all samples for sub—

 

 

 

scale 2.

Lenawee Turrialba Lansing

M Var. M Var. M. Var.
2.1 92 08 64 23 93 06
2.2 45 25 47 25 63 23
2.3 60 24 56 25 71 20
2.4 86 12 88 ll 77 18
2.5 66 22 87 12 53 25
2.6 49 25 42 24 53 25
2.7 83 14 92 07 76 18
2.8 77 17 100 00 77 18

 

predicted kinds of mean scores. although certain items appear
to be consistent with predictions. Thus. only the Lenawee
sample supports the prediction of a one factor solution and
provides some support for the prediction of high mean scores

for the more industrialized sample.

Sub-Scale 3: Positive versus Negative Evaluation of Physical

 

Mobility

The prediction of a one factor solution on sub-scale
three is supported by all three samples. In Table 9. we
see that the first unrotated factor accounts for 56 percent
of the matrix variance. Items one through four contribute
significantly to the factor and item five with a loading of
.19 falls just below our cutting point. Items one and two
have the highest loadings and these are increased by the

varimax rotation.

113

Table 9. Factor matrix for WBCL Sub-Scale 3. Lenawee
sample (N—439).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 3)
I II III I II III
3.1 42 ~28 .12 48 ~16 14
3.2 59 ~14 ~15 60 15 05
3.3 29 31 ~13 13 40 15
3.4 38 21 23 19 13 43
3.5 l9 19 ~05 09 23 12
3.6 01 ~08 ~18 08 05 17
% Total
Variance 56 20 10 47 20 20

 

*Decimal points omitted.

From Table 10 we see that in the Turrialba sample.
the first factor accounts for 65 percent of the matrix variance.
Once again. there are significant loadings on the first five
items.

Table 10. Factor matrix for WBCL Sub—Scale 3. Turrialba
sample (N-112).*

 

 

 

 

 

Principal Axes Varimax

Items Loadings Loadings

(Sub-Scale 3) I II III I II III
3.1 63 02 03 44 37 26
3.2 25 07 11 11 17 20
3.3 47 ~31 ~16 59 07 02
3.4 39 33 ~14 12 51 02
3.5 28 ~04 28 15 97 36
3. ~02 07 ~08 ~03 06 08

% Total

Variance 65 15 10 41 31 17

 

*Decimal points omitted.

114

The same general pattern is illustrated in Table 11
for the Lansing sample. The first factor accounts for 61
percent of the matrix variance. However. the loading for
item three is not significant in the Lansing sample. In
fact. this loading approaches zero. In addition. item six
has a high negative loading in the Lansing sample. As seen
in Tables 9 and 10. the loading for item six approaches zero
in both the Lenawee and the Turrialba samples. Tables
A—7. A~8. and A~9. of Appendix A show that item six is
completely uncorrelated with the other five items in the
Lenawee and Turrialba samples. In addition. these tables
show that item six is inversely correlated with items one.
two. four and five in the Lansing sample. Item six is un-
correlated with item three in the Lansing sample.

Table 11. Factor matrix for WBCL Sub-Scale 3.Lansing
sample (N-87).*

 

 

 

 

Principal Axes varimax
Items Loadings Loadings
(sub-scale 3' I II III I II III
3.1 21 34 07 39 ~12 ~01
3.2 47 ~37 14 15 58 14
3.3 ~06 ~18 05 ~15 13 ~05
3.4 48 ~09 ~33 20 19 52
3.5 42 20 00 43 06 17
3.6 ~67 ~07 ~11 ~57 ~33 ~20
% Total
Variance 61 18 8 40 28 19

 

*Decimal points omitted.

115

The prediction of a one factor solution is thus
supported in all three samples. It appears. therefore. that
this scale is unidimensional for all three samples. The
previously noted pattern of sub-scale l and 2 does not occur
on this scale. That is the tendency for the respective
samples to have mean scores between the midpoint of .50
and the maximum score of 1.00 on all of the items on the
scale. A close lock at Table 12 will verify the fact that
the trend here is for low mean scores on all items except
item six. For the moment. let us consider only the first
five items. The mean scores for the Lenawee sample range
from a high of .52 to a low of .37. The corresponding range
for the Turrialba sample is .59 to .10 and the Lansing

sample range is .31 to .22.

Table 12. Means and variances for all samples for Sub-Scale 3.

 

 

 

Lenawee Turrialba Lansing

M Var. M. Var. M. Var.
3.1 52 25 36 23 31 21
3.2 40 24 59 24 28 20
3.3 37 23 48 25 29 20
3.4 51 25 10 09 23 18
3.5 44 25 36 23 22 17
3.6 95 05 98 02 93 06

 

116

The prediction for high mean scores in the Lenawee
sample is not supported. Actually. the means for the Lenawee
sample tend to fall at the middle of the range or at the point
where the predictions placed the Lansing and Turrialba samples.
The mean scores for both the Turrialba and the Lansing samples
falls below the predicted areas. It appears. therefore.
that except for item six. none of the samples indicate a
strong positive evaluation of physical mobility. On the
contrary. both the Turrialba and the Lansing samples tend
to indicate a negative evaluation of physical mobility. Let
us now consider the items of the scale in order to account
for the lack of correlation between item six and the rest
of the scale.

The first five items deal with moving away from
relatives. homes. hometowns or with loyalty to the home
community. In general. we might expect persons from less
industrialized or gemeinshaft types of cultures to have
strong family and community ties on the basis of the typo—
logical literature. We would expect such groups to have
a negative evaluation of physical mobility i; such mobility
involved separation from the family. This might not be the
case in nomad groups where the entire tribe is physically

mobile. rather than a single individual.

117

Item six states: ”I like to see new things and meet
new people.” Agreement with this item is scored as positive
evaluation of physical mobility. However. a person may see
new things and meet new people without leaving home. This
is of a different order than the other five items of sub-
scale 3. It deals with the evaluation of variety or new
experience. independent from the breaking down of family
and community ties. In short. this item clearly does not
belong in this scale. but probably should have been included
in sub—scale 4 which deals with positive and negative
evaluation of change. It is to sub-scale 4 that we now

turn our attention.

Sub-Scale 4: Positive versus Negative Eyaluation of Change

The factor loadings for the Lenawee sample are
contained in Table 13. Fifty—five percent of the matrix
variance is accounted for by the first unrotated factor.
Significant loadings were obtained on the first six items
but not on item 7. The next largest factor accounts for
only 15 percent of the total matrix variance. and it is a
bi-polar factor. It appears. therefore. that the prediction
of a one factor solution is supported by the Lenawee data.
Turrialba comes closer to a two factor solution. Although

the first factor accounts for 46 percent of the variance.

118

Table 13. Factor matrix for WBCL sub-Scale 4. Lenawee
sample (N—439).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 4)
I II III I II III
4.1 37 ~03 ~18 31 27 04
4.2 31 .11 19 22 ~02 31
4.3 40 ~32 ~11 13 48 15
4.4 49 20 ~18 53 15 08
4.5 33 26 10 36 ~08 22
4.6 41 ~19 26 08 24 46
4.7 12 ~04 01 06 09 08
% Total
Variance 55 15 ll 34 24 23

 

*Decimal points omitted.

we see from Table 14 that the second factor accounts for 25
percent of the total matrix variance. All loadings on the
first factor are significant except items three and seven.
On the second factor. we see that the loadings for items
one. four and five are positive and significant while there
are negative and significant loadings for items two and six.
Items one. four and five are also significant and positive
on the first factor. Since.there is very little difference
between the unrotated and the rotated loadings in terms of
the percent of total variance accounted for by the first
two factors. let us shift to the varimax loadings where the

negative signs have been eliminated. Here. we see from

119

Table 14. Factor Matrix for WBCL Sub-Scale 4. Turrialba
sample (N—112).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub—Scale 4)
I II III I II III
4.1 27 48 05 06 55 03
4.2 68 ~22 ~21 71 05 ~19
4.3 07 03 ~33 07 05 ~33
4.4 20 42 ~06 03 46 ~08
4.5 31 25 05 19 35 04
4.6 63 ~21 25 66 05 27
4.7 08 ~14 ~08 13 ~10 ~07
% Total ~
Variance 46' 25 10 43 28 10

 

*Decimal points omitted.

Table 14 that items two and six with loadings of .71 and .66
respectively. account for most of the variance in the first
factor. The second factor is determined by items one.

four and five. Item two states: "On the whole. the old
ways of doing things are the best.” Similarly item six
states: ”The happiest people are those who do things the
way their parents did.” In both items disagreement is
.scored as positive evaluation of change. Items one. four
and five are stated in terms such as: ”I like to try new
things” or ”I likepeople who are willing to change” or

”On the whole. most changes make things worse.” These items

all deal with change or willingness to change at the present

120

time. Items two and six deal with change from the past to
the present. Thus. one dimension may involve unwillingness
to go back to older ways of doing things while the other
dimension may involve a willingness to change in terms of the
future.

In a highly industrialized society with strong value
orientations toward the future. both of these dimensions may
become fused into a single dimension involving change. The
one factor solution of the Lenawee sample supports this
view. On the other hand. less industrialized societies with
more of a time orientation to the present might yield a two
factor solution on the same items. In the opinion of the
present writer. this is what has occurred in the Turrialba
sample. In other words. it is only in a culture where
change. newness and youth are very strongly emphasized that
the dimensions are not distinguished.

From Table 15. we see that the largest of the un-
rotated factors account for only 37 percent of the total
matrix variance in the Lansing sample. and the first three
factors account for a total of 75 percent of the variance.
It appears. therefore. that there is a three factor solution
for the Lansing sample. In both the unrotated and the rotated

factors. items two and six determine the third factor.

121

Thus. it appears that for the sample of Mexicans in Lansing.
the two items which deal with the past time dimension.
distinguish a different factor just as they did for the
Turrialba sample. The first factor is determined by items
one. three and four. while the second factor after rotation
is determined by items four and five.

Table 15. Factor matrix for‘WBCL Sub-Scale 4. Lansing
sample (N~87).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub—Scale 4)
I II III I II III
4.1 68 ~13 ~19 72 ~03 ~02
4.2 06 ~27 32 03 ~05 42
4.3 63 -11 05 61 09 17
4.4 42 50 02 33 50 ~25
4.5 04 47 39 ~13 60 07
4.6 17 ~32 35 14 ~04 48
4.7 12 05 10 09 12 07
% Total
Variance 37 23 15 36 22 17

 

*Decimal points omitted.

It appears. therefore. that the prediction of a one
factor solution is supported in the Lenawee sample but not
in the Thrrialba and the Lansing samples. The data for means
and variances are presented in Table 16. In accordance with

predictions. the means for the Lenawee boys are high and the

122

variances are low. However. the means for both the Turrialba
and the Lansing samples are also high. although not as

high as the Lenawee means. Consequently. neither the Turrialba
nor the Lansing sample falls at the middle range of mean

scores as predicted. While the variances for Turrialba

are somewhat lower. in general. than the Lansing variances.

they are not as consistently low as the Lenawee sample.

Table 16. Means and variances for all samples for Sub—Scale 4.

 

 

 

Lenawee Turrialba Lansing

M Var. M Var. M Var.
4.1 99 01 97 03 90 09
4.2 85 13 88 10 68 22
4.3 95 04 94 06 80 16
4.4 96 04~ 87 12 77 18
4.5 91 08 63 23. 82 15
4.6 90 09 75 19 72 20
4.7 77 18 88 10 77 18

 

In terms of the predictions of mean scores. only the
Lenawee sample provides strong support for the predictions.
For both the Turrialba and the Lansing samples. we may speak
of their mean scores as being relatively lower than the
Lenawee scores. However. in terms of the absolute numbers.
neither Turrialba nor Lansing provide mean scores which fall

around the middle of the possible range of scores.

123

fipb-Scale 5: Belief in Internal versus External Determination
Events

The unrotated factors. as shown in Table 17. indicate
a one factor solution for the Lenawee sample. since the first
factor accounts for 53 percent of the matrix variance.
Items one. four and six do not have significant loadings.
This factor is. therefore. defined by five of the eight
items. It should be noted that the rotated factors yield
a three factor solution rather than a one factor solution.
with each factor being defined by two or three items.

Table 1?. Factor matrix for WBCL Sub-Scale 5. Lenawee
sample (N—439).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(sub'scale 5' I II III IV I II III Iv
5.1 00 17 ~11 18 01 06 ~10 25
5.2 48 ~22 14 12 06 16 46 00
5.3 48 ~20 12 08 12 23 47 ~02
5.4 10 17 ~01 08 14 06 ~02 16
5.5 55 ~09 ~26 ~08 15 56 20 ~08
5.6 16 09 ~16A 21 01 21 02 24
5.7 24 32 19 07 40 ~01 08 20
5.8 50 25 02 20 50 3O 10 ~05
% Total
Variance 53 17 9 8 24 27 26 10

 

*Decimal points omitted.

As indicated in Table 18. the first unrotated factor

accounts for 48 percent or almost half of the total matrix

124

variance in the Turrialba sample. In this case. the factor
is determined by items two. three. four and five. All of
these items except item four had significant loadings

in the Lenawee sample also. However. it is in the Lansing
sample that the greatest similarity with the Lenawee sample
appears to occur. In Table 19. we see that the first
factor. accounting for 46 percent of the matrix variance.
has significant loadings on items one. two. three. five.
seven and eight. Only on item one is there a difference
between the two samples in relation to which items define
the first factor. Similarly. rotation of the factors for
the Lansing sample produces a three factor solution just as
it did for the Lenawee sample.

Table 18. Factor matrix for WBCL Sub-Scale 5.Turrialba
sample (N-112).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 5)
I II III I II III
5.1 ~36 ~01 20 ~34 ~14 19
5.2 37 31 ~18 28 42 ~10
5.3 67 ~02 15 67 10 14
5.4. 56 22 13 ~59 05 17
5.5 30 01 35 31 00 35
5.6 ~04 07 06 ~05 04 08
5.7 ~07 31 ~02 ~14 29 05
5.8 09 37 09 01 36 18
% Total Variance 48 17 10 47 18 ll

 

*Decimal points omitted.

125

Table 19. Factor matrix for WBCL Sub—Scale 5. Lansing
sample (N-87).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 5)
I II III I II III
5.1 41 50 ~12 ~06 ~34 56
5.2 33 ~52 ~19 62 07 ~15
5.3 48 ~17 ~24 48 22 18
5.4 13 20 ~24 05 00 33
5.5 62 15 25 09 67 14
5.6 05 30 ~26 ~04 ~06 40
5.7 38 09 23 02 45 03
5.8 72 ~17 04 47 57 07
%.Total variance 46 22 10 25 34 19

 

*Decimal points omitted.

From the above discussion. it appears that there
is some evidence to support the prediction of a one factor
solution for sub-scale five. but the evidence is not clear
cut. In all three samples there is one large factor which
accounts for approximately half of the total matrix variance.
However. the specific items which define the factor are not
the same. Certain items such as two. three and five have
significant loadings in all three samples but the rest of
the items which make up the scale vary from one sample to
the next. Moreover. the second factor which accounts for
either 17 or 22 percent of the variance in each of the samples.

contains significant loadings on several items in each case.

126

In short. the strongest support for a one factor solution is
found in the Lenawee sample but there are strong trends in
the same direction for both Lenawee and Turrialba.

The means and variances for sub-scale 5 are
presentedin Table 20. It is apparent that the item mean
scores for the Lenawee sample are spread over a relatively
wider range than are the comparable meanscores for any of
the previously discussed scales. To be sure. the apparent
range of sub-scale 3 is greater since these means range
from a low of .37 to a high of .95. However. a quick glance
at Table 12 will help the reader to recall that once item
six is deleted from sub-scale 3. the range is from .37 to
.52 or a spread of fifteen class intervals. Removal of any
single item mean score will not appreciably reduce the spread
of scores on sub-scale 5 for the Lenawee sample. Thus. while
it is apparent that most of the boys of the Lenawee sample
have scores which indicate a belief in the external deter-
mination of events. the scores are not consistently at the
high end of the scale.

The means of both the Turrialba and the Lansing
samples tend to cluster closer to the middle range giving
some support to the prediction of mid—range mean scores for

the Turrialba and Lansing samples. However. the considerable

127

Table 20. Means and variances for all samples for Sub-

 

 

 

Scale 5.

Lenawee Turrialba Lansing

M Var. M Var. M Var.
5.1 90 10 69 21 87 11
5.2 59 24 52 25 44 25
5.3 86 12 53 25 82 15
5.4 68 22 73 20 51 25
5.5 78 17 41 24 49 25
5.6 79 16 79 16 75 19
5.7 46 25 64 23 34 23
5.8 82 15 71 21 57 24

 

fluctuation of mean and variance scores within the same scale
prevent us from interpreting the data as conclusive support
for the original predictions concerning placement of the
three samples along the hypothetical continuum of industrial-

ization.

Sub-Scale 6: Positive versus Negative Evaluation of Deferred

Gratification

As indicated in Table 21. one large factor accounts
for 59 percent of the total matrix variance in the Lenawee
sample. All of the items except item two have significant
loadings on this factor. Item two states that: ”Money is
made to spend. not to save.” The wording of this item is

such that agreement with the item implies a negative

128

Table 21. Factor matrix for WBCL Sub-Scale 6.Lenawee
sample (N—439).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 6)
I II III I II III
6.1 21 17 ~04 ll 24 07
6.2 12 32 ~09 ~01 36 01
6.3 44 ~20 ~21 53 04 ' ~01
6.4 60 ~20 ~07 61 05 20
6.5 46 ~06 26 31 03 44
6.6 38 17 15 18 24 33
6.7 28 35 ~08 10’ 44 10
% Total Variance 59 20 9 44 24 19

 

*Decimal points omitted.

orientation to saving money. under any conditions. There-
fore. we might expect this item to be correlated with the
scale which measures deferred gratification among the most
industrialized sample. Actually. the item does have a
significant loading but it is on the second factor rather
than the first. Shifting to the rotated loadings. we see
that the second factor is defined by items one. two. six and
seven. It should be noted that the varimax rotations indicate
a three factor solution with the largest factor accounting
for 44 percent of the variance.

Failure to obtain a one factor solution is indicated
by the data for Turrialba. as presented in Table 22. It is

true that the first factor accounts for 48 percent or almost

129

half of the total variance but the second factor accounts for
26 percent of the variance and. therefore. cannot be ignored.
Moreover. the rotated factors indicate a three factor solution
with each of the three factors accounting for slightly over
one-fourth of the total matrix variance.

Table 22. Factor matrix for WBCL Sub-Scale 6. Turrialba
sample (N-112).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 6)
I II III I II III
6.1 32 ~17 ~26 07 ~01 44
6.2 44 50 02 13 64 11
6.3 56 ~15 ~25 23 11 58
6.4 47 ~14 34 58 08 10
6.5 58 ~23 ll 52 05 36
6.6 22 09 02 13 17 10
6.7 18 55 ~04 ~08 57 03
% Total Variance 48 26 10 27 31 27

 

*Decimal points omitted.

In the Lansing sample. the first unrotated factor
accounts for 39 percent of the total matrix variance. The
second factor accounts for 22 percent of the variance and
the third accounts for 14 percent. It seems clear that the
prediction of a one factor solution is not supported in the
Lansing sample. However. there appears to be considerable

overlap of items which contribute to the first two factors.

130

Table 23. Factor matrix for WBCL Sub-Scale 6. Lansing
sample (N-87).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 6)
I II III IV I II III IV
6 1 79 ~09 ~07 09 75 21 20 04
6.2 20 41 ~14 ~29 02 14 20 51
6.3 36 ~22 ~30 ~29 44 ~30 13 23
6.4 51 ~43 26 ~05 65 02 ~31 ~10
6.5 07 32 28 ~31 ~07 24 ~22 41
6.6 47 42 28 20 22 67 08 13
6.7 24 28 ~38 19 09 14 53 ~06
% Total Variance 39 22 14 10 36 2O 15 15

 

*Decimal points omitted.

Thus. items two. six and seven have significant and positive
loadings on both factors one and two.

Even though the Lenawee sample is the only one which
yields one factor which accounts for more than half of the
matrix variance. all three of the samples have significant
loadings on all but one of the items on the first factor.
The specific item which does not have a significant loading
is different in each of the samples. but the fact that only
one item fails to contribute indicates a trend toward a uni—
dimensional scale. However. since the first factor accounts
for only 48 and 22 percent respectively in the Turrialba and
Lansing samples. the case for a unidimensional scale is

weakened.

131

The three samples do not appear to differ in relation
to mean scores for the items of sub-scale 6. The general
tendency for all three samples is toward the high or

deferred gratification end of the continuum. Neither the

Table 24. Means and variances for all samples for Sub—

 

 

 

Scale 6.

Lenawee Turrialba Lansing

M Var. M Var. M Var.
6.1 68 22 93 07 85 13
6.2 63 23 52 25 57 24
6.3 90 09 84 13 89 10
6.4 87 11 87 12 89 10
6.5 90 09 98 02 89 10
6.6 93 07 73 20 90 09
6.7 56 25 46 25 51 25

 

Turrialba nor the Lansing samples have mean scores clustered
around the middle range. In both samples the general tendency
is for relatively high mean scores. Only items two and seven
with mean scores close to the midpoint of .50 are in the
predicted range for the Turrialba and Lansing samples. As
previously pointed out. the wording of item two implies a
negative orientation of saving money regardless of the
condition. In all three samples. the mean is slightly above
the midpoint and the variance is relatively high. Item

seven. which is the other consistently high variance and

middle range mean item. is similar to item two in that both

132

deal specifically with money and the question of spending.
The other items of the scale deal with jobs. school. or working
hard for something. but they do not specifically deal with
money.

It appears. therefore. that there is some
support for a one factor solution in the Lenawee sample and
a similar trend for the Trurialba and Lansing samples. The
means for all three samples indicate a tendency toward
deferred gratification in expressed value orientations since
the means are generally above the midpoint of the possible

range of scores.

Summary of the firstpgour predictions

The first prediction (Al) was for a one factor
solution for each sub-scale within each sample. Only on
sub-scale three was this prediction supported by all three
samples. On sub-scale five. there was some support in all
three samples but not as strong as the support for the third
scale. The prediction of a one factor solution for sub-
scale one was supported by both the Lenawee and the Lansing
samples. On sub-scale six. there was support for a one
factor solution in the Lenawee sample with similar trends
in both Turrialba and Lansing. Support for the one factor

prediction was found among the Lenawee boys only. for

133

sub-scale two and four. In short. the Lenawee sample was
the only one where there was relatively consistent support
for the prediction of a one factor solution. As the prior
discussion of this chapter indicate. even this consistent
trend toward a one factor solution among the Lenawee boys
should not be construed as an indication that all of the
items of any specific scale contributed significantly to
the large common factor.

The second. third and fourth predictions are all
based on assumptions about where the three samples fall
along a hypothetical continuum of industrialization. The
second prediction (A2) states that the Lenawee sample is at
the industrial or gesellshaft end of the continuum and will.
therefore. yield high means. low variance and low correlation
coefficients. In general. there was support for this pre-
diction on all scales except sub—scale three. where the mean
scores were closer to the middle range.

The third prediction (A3) states that the Turrialba
sample is at an intermediate position along the hypothetical
continuum of industrialization and will. therefore. yield
middle range mean scores. low variance and low crorelation
coefficients. In general this prediction was not supported.

With the exception of sub-scale three the Turrialba sample

134

tended to score above the middle range and toward the high
or industrial end of the range. It should be noted that
there was some support for this prediction among all three
samples on sub-scale four which deals with the evaluation of
change. While sub-scale three which deals with the
evaluation of physical mobility is actually below the middle
range of scores in general. this does give support to the
view that the less industrialized samples will not score at
the high end of the continuum.

The fourth prediction (A4) is slightly more complex
than the others. The Mexican-American boys are viewed as
being influenced by two cultures and occupying a transitional
position (i.e.. gemeinshaft to gesellshaft) along the
hypothetical continuum. The predictions were for middle
range mean scores (just as for Turrialba) but for high
variance and high correlation coefficients due to the increased
range of scores. The pattern here was similar to the
pattern for Turrialba. Only on the third sub-scale was
there a consistent (except for item six) tendency for mean
scores to fall below the midpoint of the possible range.
There was some support for middle range mean scores on sub-
scale four but this was not strong or consistent. The

general pattern for the rest of the sub-scales was for

135

scores to approach the high or industrial end of the scale.

Prediction B

When all 44 items of the werk Beliefs Check List
are factor analyzed irrespective of sub—scales. it was
predicted that there would be a total of seven factors. One
large factor was expected to account for much of the variance
and six small factors were expected to correspond to the six
sub—scales.

Table 25 indicates that most of the 44 items have
significant loadings on the first factor for the Lenawee
sample. The one obvious exception to this is on sub-scale
three where item six is the only item with a significant
loading. It has already been pointed out that item six is
the only item of this scale where there was a high mean
score. Since relatively high mean scores tended to be a
consistent pattern for most of the other scales. the
factorial structure shown in Table 25 supports the previous
indications that most of the items yield high mean scores
for the Lenawee boys. The first five items of sub-scale
three represent an exception to this pattern. but the
exception is consistent with the failure to obtain significant

loadings on these items.

136

 

 

 

 

Table 25. Factor Matrix for WBCL - All Items
Lenawee Sample (N 439)”

Item Principal Axes Loadings
l 2 3 4 5 6 7 8 9 10 11
1.1 27 ~24 05 09 -10 ~09 16 ~18 03 ~20 20
1.2 28 17 -05 ~07 ~28 ~20 03 03 07 ~13 05
1.3 27 -03 06 09 ~22 -08 -17 ~08 11 -O4 -18
1.4 21 37 ~05 ~16 ~19 11 ~02 ~02 ~14 04 ~10
1.5 44 14 ~06 02 ~16 00 23 01 04 ~20 ~08
1.6 19 16 ~01 -02 -03 -12 ~12 -10 21 04 00
1.7 15 35 04 02 -08 -15 11 06 -08 12 ~07
1.8 ~00 16 04 ~20 ~08 -13 ~04 ~04 06 01 07
2.1 50 12 ~03 16 ~07 10 ~16 06 ~22 ~05 05
2.2 10 10 36 00 ~06 01 ~04 ~15 12 09 ~10
2.3 35 ~10 20 11 01 ~15 ~15 02 ~16 03 ~01
2.4 33 ~03 02 03 ~01 ~07 ~00 07 ~09 14 19
2.5 22 13 39 -21 28 -02Q -19 04 07 -15 -04
2.6 33 ~05 43 ~20 15 ~09 ~01 '22 ~02 ~15 ~07
2.7 27 15 23 18 ~05 15 19 22 04 03 01
2.8 25 13 13 09 03 25 07 30 ~05 00 06
3.1 02 ~34 02 ~16 ~07 ~21 00 ~01 ~22 12 ~05
3.2 ~11 ~23 ~16 ~30 01 ~20 ~01 04 ~13 ~08 ~14
3.3 ~10 ~10 ~10 ~19 15 11 02 16 ~03 -13 ~05
3.4 05 -41 -16 04 ~01 03 ~01 21 05 01 -10
3.5 ~10 ~07 ~15 ~25 11 16 07 10 09 06 05
3.6 24 14 ~14 ~12 11 ~16 ~05 ~07 01 ~07 ~02
4.1 27 21 -24 -04 00 -07 ~09 03 -10 -01 -03
4.2 25 ~22 ~13 13 15 02 ~11 ~03 27 12 -05
4.3 21 15 ~32 ~04 01 ~19 05 12 11 03 05
4.4 25 15 ~22 ~14 ~11 09 ~08 13 23 ~08 05
4.5 41 -15 -13 -10 -09 23 -26 -09 07 -01 -04
4.6 36 ~07 ~20 04 09 ~15 ~10 21 07 10 03
4.7 ~04 10 ~07 ~21 05 02 ~03 04 03 23 02
5.1 03 24 ~07 ~20 ~11 21 ~09 ~07 ~18 07 ~05
5.2 20 ~32 ~02 03 ~18 ~04 01 14 ~07 08 ~02
5.3 39 ~25 ~06 04 ~19 08 26 ~06 02 11 ~21
5.4 -04 -03 ~00 -31 -05 06 06 06 01 -06 09
5.5 48 ~29 ~10 ~13 ~02 06 ~02 03 02 ~03 ~02
5.6 10 02 ~12 ~25 04 06 07 ~13 -12 ~00 -04
5.7 09 -04 10 -23 -17 05 11 -14 04 10 19
5.8 38 -22 11 -14 -07 05 -01 ~05 ~02 08 25
6.1 21 ~24 06 ~07 13 15 20 ~16 06 ~10 ~08
6.2 21 00 20 ~02 ~03 ~01 02 ~01 ~05 14 ~13
6.3 46 -10 ~11 18 21 10 -19 ~17 ~14 ~03 09
6.4 33 13 -22 11 30 ~03 19 ~16 ~15 ~09 01
6.5 30 25 ~16 ~05 23 07 10 ~12 09 09 ~11
6.6 30 08 ~02 00 25 ~18 20 05 01 14 02
6.7 31 ~06 32 ~16 11 ~04 07 ~09 07 12 02

1 Total

Variance 18 9 7 5 5 4 4 4 3 3 2

 

*Decimal points omitted.

Table 25 (Continued)

137

 

 

 

Item Vagimaxyppadinqs
I 2 3 4 5 6 7
I.) 09 -04 05 -02 04 ~16 37
1.2 04 41 02 -02 05 -I7 15
1.3 23 16 06 04 -]0 -24 10
1.4 06 95 03 )6 03 )2 -02
1.5 10 30 03 )7 22 -05 32
1.6 11 21 09 00 08 -12 '06
I.7 -12 3] 06 14 17 -13 -02
1.8 -08 23 10 -09 -02 00 '08
2.] 42 2] 07 24 12 -I3 08
2.2 '04 07 29 18 -13 -09 01
2.3 24 00 27 02 0] -28 11
2.9 19 09 12 02 12 -I) 16
2.5 09 0] 57 05 04 03 -17
2.6 06 00 59 00 02 '04 )3
2.7 00 06 )5 40 07 -05 20
2.8 12 04 12 33 07 08 09
3.1 01 -07 07 -34 -10 -07 20
3.2 -03 -04 '0] -47 -03 08 06
3.3 00 '10 -01 -i# 02 25 -03
3.4 19 -22 -13 -i9 -04 300 22
3.5 -01 '09 -05 -I# 03 33 -00
3.6 13 19 09 -I] '26 -03 -04
A.) 20 28 -04 -01 23 -02 -05
4.2 39 -17 -00 -05 14 -07 .09
4.3 09 25 -ll -12 32 -05 02
4.4 23 3] -05 00 08 12' 02
4.5 53 ll 04 -00 ~06 I I 13
4.6 32 06 0] -I3 25 -14 08
4.7 ~01 l I 03 -O9 03 16 ~10
5.) IO 28 00 09 -05 23 -09
5.2 )8 -04 ~01 -11 -10 -11 31
5.3 )7 03 -02 09 06 -01 54
5.4 -06 10 07 -15 -07 24 06
5.5 4) 03 ,12 -12 09 05 35
5.6 04 12 04 ~11 09 22 06
5.7 -04 15 )2 -03 -12 12 2]
5.8 26 05 23 -05 -OA 02 3)
6.1 10 -16 14 0] 10 i7 30
6.2 05 05 22 08 -0] -06 12
6.3 51 '09 08 10 23 -08 05
6.4 13 01 -00 08 52 0] 07
6.5 )3 14 06 ii 40 13 -01
6.6 00 03 18 '00 #2 -07 11
6.7 05 0] 45 .9; 05 00 l8

% Total ‘

Variance IO 8 8 6 2 15 8

 

”Decimals omitted.

138

The same general pattern emerges from Table 26.
except that some of the items of sub-scale three are signi—
ficant but negative in the Lansing sample. It should be
noted that the first factor accounts for only 19 percent
of the total matrix variance in the Lansing sample. Although
many of the items have significant loadings. less than one-
fifth of the variance is accounted for by the factor as a
whole. On the other hand. the second factor is only about
one-half the size of the first factor in both cases.

The pattern is less clearly demonstrated for Turrialba.
From Table 27. we see that the first factor accounts for
only 13 percent of the matrix variance. Once again. many
of the items have significant loadings but there is less
consistency within any specific scale. Sub-scale three is
either uncorrelated or negatively correlated with the factor.
This pattern is consistent with the previously discussed

data on Turrialba.

Summary of Prediction B

The factor analysis of all items of the Work Beliefs
Check List did not provide strong support for the prediction
of a seven factor solution. It is true that there was a

consistent trend for most of the items to have significant

L139

Table.26. Factor Matrix for WBCL ~ All Items
Turrialba Sample (N z 112)“

 

 

 

 

Item Principal Axes Loadings
1 2 3 4 5 6 7 8 9 10 11
1.1 40 09 ~05 ~14 ~25 29 02 ~16 ~11 ~06 03
1.2 ~06 00 ~08 ~06 ~41 08 13 ~07 02 02 03
1.3 12 24 02 ~05 24 ~15 11 17 ~08 ~25 17
1.4 ~20 ~01 07 03 ~23 07 ~12 20 ~07 20 06
1.5 40 24 ~09 ~17 ~12 ~06 ~16 ~20 ~17 ~09 ~07
1.6 ~00 ~00 00 ~00 ~00 00 00 ~00 00 ~00 ~00
1.7 13 ~07 01 ~26 19 30 ~16 ~21 07 01 03
1.8 03 ~20 06 22 ~05 38 O3 05 12 07 ~11
2.1 30 ~16 ~24 23 26 02 02 10 ~22 ~09 ~01
2.2 38 ~04 ~35 07 15 03 ~13 ~14 ~02 ~08 27
2.3 ~02 11 ~15 19 32 ~06 19 07 ~03 ~06 ~10
2.4 ~22 ~01 ~02 ~03 ~29 ~14 23 ~17 ~12 17 ~01
2.5 18 ~23 35 23 01 ~24 ~11 03 12 09 ~04
2.6 08 ~03 21 ~12 19 ~29 ~14 ~19 10 03 ~15
2.7 16 ~32 ~24 09 ~04 16 22 ~09 ~08 08 14
2.8 00 00 00 00 ~00 ~00 00 00 ~00 ~00 ~00
3.1 ~13 52 31 ~04 ~08 ~02 16 08 00 ~20 01
3.2 14 10 13 ~19 ~10 25 19 18 ~14 ~00 ~02
3.3 ~31 18 24 ~03 ~00 09 ~20 28 10 ~07 04
3.4 ~09 31 13 :03 ~06 ~01 15 16 ~05 13 01
3.5 ~18 08 28 ~07 08 30 ~14 ~09 ~12 03 01
3.6 ~03 ~11 8 ~40 13 ~06 03 08 09 00 02
4.1 27 26 00 ~05 10 25 12 ~03 14 13 12
4.2 50 ~15 ~04 ~34 03 ~03 ~14 11 ~27 10 ~25
4.3 13 ~34 42 ~14 06 13 02 ~04 11 ~12 ~10
4.4 17 07 09 ~09 30 14 05 09 26 20 17
4.5 35 34 04 ~10 04 ' 06 07 ~03 11 ~13 03
4.6 50 ~02 ~17 ~01 ~07 04 ~10 38 ~08 03 ~12
4.7 ~06 ~15 ~09 ~29 ~00 ~16 ~08 21 05 ~17 15
5.1 ~20 ~26 ~12 ~00 29 ~07 08 ll 04 ~10 01
5.2 42 16 06 ~13 ~18 ~11 ~18 13 ~02 02 02
5.3 39 29 02 44 ~08 ~01 ~12 17 17 02 02
5.4 ~33 ~18 ~04 ~26 ~04 ~06 38 16 ~14 ~09 06
5.5 23 3 ~13 22 ~06 ~10 29 ~01 19 09 ~25
5.6 04 ~35 ~01 12 ~15 24 ~04 36 ~01 ~03 10
5.7 12 ~13 ~05 ~21 ~01 ~19 05 09 01 27 08
5.8 24 O7 02 ~12 19 ~02 14 13 ~08 25 ~09
6.1 ~01 ~34 31 16 ~03 11 ~05 ~05 ' ~13 ~23 05
6.2 51 ~14 ~04 ~01 ~18 ~11 21 ~02 17 ~24 10
6.3 37 ~03 40 16 04 01 24 ~05 ~22 '~18 ~11
6.4 28 ~01 27 06 00 ~18 07 ~11 ~11 21 40
6.5 30 ~21 43 14 ~06 ~21 00 02 ~15 12 10
6.6 19 ~09 20 ~11 22 17 28 ~06 18 09 ~05
6.7 25 ~25 ~03 ~20 ~32 ~13 02 03 41 ~13 ~04

% Total

Variance l3 7 7 6 5 5 4 4 4 3 3

 

* . . .
Dec1ma1 pOInts omitted.

Table 26 (Continued)

.140

 

 

 

 

Item Varimax ppadings
I .2. 3 1+ 5 6 7
I.I 46 05 ~04 II 08 BI 03
1.2 08 15 ~I0 ~08 ~22 26 I7
1.3 05 II -0I 06 26 -26 08
I.4 -08 II -02 ~00 -25 I3 -II
1.5 49 05 ~06 20 ~00 -II 02
I.6 00 00 00 00 00 00 00
I.7 22 ~14 ~05 ~09 2I 08 ~32
1.8 -II ~14 IO 09 06 4I ~09
2.I OI ~42 02 20 22 ~01 I4
2.2 25 -38 -I3 23 I3 ~05 09
2.3 -23 ~08 -I0 09 29 -I5 15
2.4 -II I9 -04 -I8 ~21 07 25
2.5 ~04 ~08 53 I3 ~I0 -II 00
2.6 07 -00 24 ~06 OI ~35 ~IO
2.7 02 ~32 ~00 ~07 08 30 24
2.8 00 00 00 00 00 00 00
3.] ~05 62 ~02 08 I0 -07- ~01
3.2 2] 2] 02 ~08 I9 24 -0I
333 -I6 27 ~03 OI -I0 -03 ~35
3.4 -04 36 ~05 02 07 ~02 04
3.5 ~09 I7 03 ~02 08 I4 -4I
3.6 I4 00 02 ~37 08 -I3 ”-II
4.] 20 I2 ~08 IS 35 II ~00
4.2 55 ~24 13 -07 12 -I0 ~03
4.3 09 -05 45 ~22 IO I6 -20
4.4 08 ~02 04 00 37 ~03 ~12
4.5 32 20 ~03 18 27 ~04 05
4.6 42 -2I 02 2] 05 05 II
4.7 I2 ~10 ~05 ~27 -I3 ~15 ~03
5.] ~25 ~26 ~02 ~22 09 -I2 00
5.2 48 08 II I9 ~07 -09 02
5.3 II 04 IO 64 04 OI 12
5.4 -I7 08 ~08 -52 -00 05 I7
5.5 02 ~02 ~01 I7 II 04 43
5.6 -0I ~24 I2 ~0I -II 36 -04
5.7 18 -08 06 -20 -03 -12 12
5.8 I7 -00 05 ~02 29 -I0 07
6.I -I4 -IO 39 -O3 -06 2] -I4
6.2 37 -II 22 04 05 09 38
6.3 08 II 49 I3 27 I0 I3
6.4 I2 07 37 09 07 ~10 II
6.5 08 00 60 08 ~04 -04 06
6.6 06 00 2] -I5 40 II OI
6.7 35 -09 I6 -I6 -23 08 I8

% TotaI

Variance 9 7 7 7 6 5 5

 

xDecimaI points omitted.

141

Table 27. Factor Matrix for WBCL ; All Items
Lansing Sample (N z 87)

 

 

 

 

Item Principal Axes Loadings
1 2 3 4 5 6 7 8 9 10 11
1.1 11 55 -32 -06 -01 -04 18 32 -05 -23 09
1.2 30 -39 ~24 -12 09 05 06 -19 -17 03 18
1.3 08- 26 ~05 O8 34 13 -18 -18 -16 10 10
1.4 79 -38 -Ol 15 10 -02 -00 ~07 -10 12 -03
1.5 51 07 -34 ~01 -32 ~18 03 ~03 -26 -02 -05
1.6 45 -O6 -15 21 13 17 ~02 -17 05 -21 24
1.7 82 -17 14 -16 08 -14 27 16 03 10 -01
1.8 29 -20 -02 ~17 -37 ~09 -25 -08 -09 -03 12
2.1 32 20 -16 -05 09 -18 -35 06 ~01 15 09
2.2 29 24 -10 01 21 04 16 -29 07 -17 -11
2.3 05 47 10 -07 -OO -10 -19 -03 -25 12 -28
2.4 32 18 17 28 -00 -24 ~07 -30 -10 -05 10
2.5 28 -O6 -28 -18 -07 17 -13 19 O9 -15 -24
2.6 -02 31 -41 -23 08 26 11 01 12 -04 -10
2.7 46 -03 -36 -18 06 -22 -02 -18 19 -18 05
2.8 37 -05 -15 -30 08 23 -11 05 -20 15 -09
3.1 -08 21 -06 33 30 ~13 03 -06 ~04 03 -03
3.2 -26 19 ~13 12 ~25 28 -36 03 04 -13 15
3.3 -09 -35 29 07 -16 09 ~04 21 14 11 10
3.4 -26 45 07 12 -24 -06 03 08 ~03 18 13
3.5 -42 16 15 12 25 O8 -15 20 -10 -11 -26
3.6 75 -29 15 ~18 11 O7 18 13 08 05 -07
4.1 67 ~00 -01 26 -O7 -09 ~04 24 -05 -05 -19
4.2 05 45 16 11 10 01 37 11 -13 03 16
4.3 48 04 -06 08 13 26 -15 32 O3 O4 05
4.4 24 -19 -12 25 -12 20 -25 06 ~02 34 ~06
4.5 -03 20 -34 02 -28 22 06 -23 14 33 11
4.6 13 55 -17 10 32 -04 -07 14 02 -04 18
4.7 20 23 -01 06 -05 -11 -10 -28 27 -07 -19
5.1 29 02 43 -31 ~04 11 -10 07 08 -22 23
5.2 16 24 -00 19 -41 00 42 02 -18 05 -03
5.3 r 17 48 03 -03 -16 42 O7 -13 12 11 -06
5.4 02 13 11 -24 O9 12 01 04 ~04 -12 -02
5.5 37 20 23 15 -09 24 -05 04 28 -13 12
5.6 02 07 30 -14 22 06 ~06 -11 -08 26 21
5.7 25 05 30 12 -19 -12 -26 ~13 07 -29 -13
5.8 47 28 22 21 -21 09 14 -02 08 -07 02
6.1 36 51 19 -30 ~02 -14 -06 -02 11 19 06
6.2 32 06 12 03 -05 12 ~05 08 -35 -10 10
6.3 09 38 ~07 00 -90 -30 ~23 22 ~07 01 14
6.4 33 25 03 -08 02 -34 -07 10 38 25 -05
6.5 51 -14 -05 46 O7 15 ~07 02 -06 -04 -05
6.6 30 30 34 -11 13 24 -02 -24 -08 06 -24
6.7 07 20 11 -37 -19 02 -08 -12 -30 -15 ~01

% Total

Variance 19 11 6 5 S 4 4 4 3 3 3

 

* O
Decimal points omitted.

Table 27'(Continued)

142

 

 

 

 

Item VarimaxfiLQadinqs
I 2 3 4 5 6 7
1.1 -05 36 -45 -04 -29 01 14
1.2 42 -16 -I6 04 I8 -20 -I3
l.3 -04 20 -I6 21 20 I6 25
1.4 81 -00 08 35 05 06 06
1.5 37 29 21 12 -3I -18 -31
1.6 36 00 -15 39 00 00 06
1.7 90 10 02 -05 -ll 08 01
1.8 23 09 09 03 -00 00 '53
2.1 14 48 -09 17 15 00 -12
2.2 22 I7 -24 06 -07 08 24
2.3 -14 43 -01 ~01 -05 26 03
2.4 19 33 24 22 -I8 04 11
2.5 20 03 -32 12 09 02 -27
2.6 -09 05 -61 -05 -01 05 04
2.7 44 30 -26 00 06 -22 -14
2.8 33 02 -32 04 20 18 -18
3.1 ~13 I7 02 I8 -00 -14 41
3.2 -47 00 -12 28 -00 I3 -25
3.3 00 -33 34 02 03 06 -15
3.4 -41 19 02 -05 -34 12 05
3.5 -43 -01 II 02 21 ll 27
3.6 84 -07 -OO 02 03 I7 -03
4.1 52 23 07 38 -20 00 -05
4.2 -00 13 -08 -08 -35 21 43
4.3 33 08 -18 37 08 21 -00
4.4 12 -08 00 46 04 -Ol -22
4.5 -17 -02 -38 11 -26 -01 -17
4.6 -07 47 -27 15 00 10 35
4.7 05 29 -00 11 -11 06 -01
5.1 27 04 l4 -I4 08 49 ~14
5;2 06 -Ol -06 00 -67 02 00
5.3 -06 05 -34 I3 -30 48 01
5.4 02 02 -10 -16 08 25 04
5.5 18 06 02 28 -I9 40 01
5.6 04 03 10 -10 I8 29 15
5.7 11 21 34 I7 -08 20 -18
5.8 28 I4 04 22 -45 32 04
6.1 18 53 -O8 -18 -12 4O -00
6.2 22 05 00 16 -07 23 -05
6.3 -09 53 00 03 -O6 00 -04
6.4 22 48 04 -05 -08 03 00
6.5 39 -03 04 59 -06 -03 08
6.6 19 12 -02 03 -01 55 15
6. 00 16 ~08 -24 -04 29 -24

% Total

Variance l6 8 6 6 6 7 5

 

*Decimal points omitted.

............................................

143

loadings (except for sub-scale three) on the first factor.

It is also true that the first factor was twice as large as
the second largest factor in both the Lenawee and Lansing
samples. However. the amount of variance accounted for by
any single factor did not exceed twenty percent of the

total matrix variance. In addition. there is a rapid decline
in the amount of variance accounted for by each factor so
that by the time we reach the fourth factor. we are dealing
with less than six percent of the total matrix variance.
Consequently. as shown in Tables 25. 26. and 27. the matrix
variance accounted for by the fourth factor may be no more
than one or two percent higher than that accounted for by

the tenth or eleventh factor. Therefore. while the trends
appear to be in the predicted direction. the total matrix
variance is spread over so many factors that it is difficult
to extract any definitive support for the prediction of a

seven factor solution.

Prediction C

This prediction states that the sub-scale total
scores within samples will produce a one factor solution
for all three samples.’ Since the earlier predictions have
been based on the assumption that each sample would score

in a consistent manner on all six sub—scales. there is an

144

implied basis for assuming that the six scalesare not
orthogonal. It is this rationale which forms the basis of
the sixth prediction.

Tables 28. 29, and 30 show that the factor analysis
of the sub-scale total scores produced one large general
factor in all three samples. As indicated in Table 28.
the first unrotated factor accounts for 61 percent of the
total matrix variance in the Lenawee sample. All of the
sub-scales except sub—scale three have significant loadings
on this factor.

Table 28. Factor matrix for WBCL Sub-Scale total scores.
Lenawee sample (N-439).*

 

 

 

 

Principal Axes Varimax
Sub—Scale Loadings Loadings
I II I II
1 43 -10 13 42
2 60 -29 . O6 67
3 -05 36 29 -23
4 34 23 38 18
5 43 33 50 21
6 6O 02 32 50
% Total Variance 61 20 30 50

 

*Decimal points omitted.

The largest single factor for the sub-scale total
scores was obtained in the Lansing sample. From Table 29

we See that the first factor accounts for‘65 percent of the

145

matrix variance among the Lansing boys. All of the loadings

are relatively high except for sub-scale three.

Table 29. Factor matrix for WBCL Sub-Scale total scores.
Lansing sample (N-87).*

 

 

 

 

Principal Axes Varimax
Loadings Loadings
Sub-Scale
I II III I II III
1 71 -28 -13 16 76 05
2 69 -19 -O4 26 68 04
3 -19 42 -20 Ol -33 39
4 59 25 -26 31 41 47
5 45 36 20 59 l3 l6
6 65 15 22 59 39 05
% Total
Variance 65 17 7 28 48 13

 

*Decimal points omitted.

The same pattern with lower (but still significant)
loadings is evident fnom Table 30 for the Turrialba sample.
Again. the only non-significant loadings are on sub-scale
three. It will be recalled that a one factor Solution was
obtained on sub-scale three in all three samples but tho
mean scores were much lower than they were for any of the

other scales.

146

Table 30. Factor matrix for WBCL Sub-Scale total scores.
Turrialba sample (N-112).*

 

 

 

 

Principal Axes Varimax
Loadings Loadings
Sub-Scale
I II III I II III
1 30 30 -15 43 09 -09
2 35 -38 —20 16 -54 01
3 -22 50 -02 O6 51 -19
4 54 27 09 57 04 21
5 41 -ll 33 22 -15 47
6 59 10 -13 58 -20 07
% Total
Variance 51 28 10 44 30 15

 

*Decimal points omitted.

Summary of Prediction C

The data for the sub-scale total scores appear.
therefore. to lend further support to the prior indications
that sub-scale three approaches a unidimensional scale which
is orthogonal to the other five scales. In addition. the
Lansing sample provides the strongest support for a one
factor solution and the highest loadings for the sub-scale
contributions to a large common factor. The weakest support
is in the Turrialba sample where the second factor accounts

for 28 percent of the matrix variance.

CHAPTER VI
SUMMARY AND CONCLUSIONS

Even when limited to the cognitive level. as we have
done in this thesis. the measurement of values is not a
clear-cut operation. In the earlier discussion of concept
formation and scientific method. it was pointed out that
explanation requires the use of precise terms and measurement.
Values were defined in such a way as to include only the
cognitive dimension. An instrument was used with assumptions
based on the use of the ideal type. and predictions were
made according to these assumptions. The data was analyzed
by means of factor analysis. a technique which has sufficient
mathematical underpinnings to maintain a respectable position
in the world of contemporary social science. With the aid
of a high speed digital computer. numerous factor matrices
were produced.— And yet. no Vautomatic? meaning was produced.
Despite all the numbers. in the final analysis meaning had
to be imposed on the data by the author. The method for
achieving this varied from the use of existing literature
to outright speculation.

Yet. we are often inclined to place high level

measurement on a sacred alter. and to equate such measurement

147

148

with precision. However. one of the curious quirks of

science appears to be that the very concept of ”measurement.”
despite all of the overtones of precision. does not denote

the same thing to all workers in the field. Thus. a sharp
line is drawn between measurement and meaning. As pointed

out earlier. whenever the data is transformed into symbols

of a different level. the underlying principle of operationism
must be kept in mind. lest we equate the symbol with the
original source of data.

In this study. written responses were transformed
into numbers referred to as factor loadings. It is these
loadings which we interpret. But in order to make sense
out of the loadings. it is necessary to refer back to the
non-quantitative domain of theory. As previously stated.
the same argument can be advanced when it comes to the
interpretation of theory. In short. the dialogue between
theory and method is no less complete. whether it involves
high level theory or high level measurement or both. Within
this context. let us consider the relationship of the
results of this study to the hypothetical predictions.

The most immediate implication of the results of
the data analysis is the lack of uniformity of results. The
prediction of a one factor solution for all three samples

was supported only on sub-scale three. It was suggested in

149.

the last chapter that the five items which contributed to
this factor were more indicative of family and community

ties than physical mobility. The present writer suggests
that a more appropriate title of this dimension would be
Vpositive versus negative evaluation of family and community
ties.? While there was some support for a one factor solution
among all three samples on sub-scale five. only the Lenawee
sample showed a relatively consistent trend in this direction
on all six sub-scales. Consequently. the first inclination
is to conclude that the WOrk Beliefs Check List. insofar

as it taps a common factor. reflects the value system most
prevalent in the Lenawee sample.

In addition. only the Lenawee sample had mean scores
which placed them at a position along the hypothetical
continuum consistent with predictions. In both the Lansing
and the Turrialba samples. there was a consistent tendency
for the boys to score toward the industrial end of the
continuum. on all scales except sub—scale three. In
addition. there was a tendency toward greater variation and
fluctuation among the Lansing and Turrialba samples.

If we relied on the mean scores alone. the
implication would be that the Lansing and Turrialba boys
held strong industrial values. HOwever. the lack of uni-

dimensional factors among these samples indicates a lack

150

of concensus on the specific items of the various scales.

It seems more plausible. therefore. to assume that the items
of the Work Beliefs Check List are a less ambiguous indication
of the value orientations of the United States than of other
cultures. This seems particularly true when we consider
the instrument as a whole. since only the Lenawee sample
showed a relatively consistent pattern of approaching uni-
dimensionality and mean scores toward the industrial end
of the continuum.

The first prediction of a one factor solution was
supported in a relatively consistent manner for the
Lenawee boys on all six scales. HOwever. as shown in the
discussion of the results. this does not mean that the
evidence for a one factor solution was equally strong on
all six scales. or that all of the items make significant
contributions to the factor. On the other hand. the
failure to achieve a large common factor on all of the
scales in the Lansing and Turrialba samples does not mean
that there were no single factor solutions for any of the
scales. It simply means that the predictions were not
uniformly or consistently upheld.

The second prediction about the Lenawee sample having

mean scores toward the high or industrial end of the continuum

151

was generally supported. although the third sub-scale on
physical mobility indicated more of a tendency for this sample
to fall around the middle of the range. The third and fourth
predictions which placed the mean scores for the Turrialba

and Lansing samples in the middle range were generally not
supported. While it is true that the mean scores for some

. of the items were lower than for the Lenawee sample. the
results were erratic and non consistent for any given scale

with the exception of sub-scale three.

Discussion of the Alternative Conclusions

In chapter four. a set of conditions was stated and
it was argued that if the expected relations were not obtained.
one or more of the propositions would be rejected. However.
the results are not simple and easy to fit into the
appropriate slot. Hence. while the instrument does not
appear to fulfill the criterion of unidimensionality in all
three of the samples on all of the scales. it cannot be
argued that none of the predictions were supported. In
short. unidimensionality was found on some of the scales in
all of the samples.

It seems. therefore. that the most reasonable

conclusion is that the instrument as a whole appears to

provide more clear cut and consistent data among the most

152

highly industrialized samples. Perhaps. as a function of

the partial but incomplete internalization of these industrial
values by the intermediate and transitional samples. the
results are difficult to describe in terms of unidimensionality.
As a result of this. the present conclusion is that while
there is Strong evidence that the Turrialba and Lansing
samples express industrial values on many of the items

of the scale. and while there does appear to be unidimension-
ality on some of the scales. this is not true for all of the
scales or for the instrument as a whole. It should also be
noted that the sub-scale dealing with the evaluation of
physical mobility having strong evidence for a unidimensional
scale among all three samples. is the only one in which the
non-industrial samples (more precisely. the intermediate

range samples) scored at the non-industrial end of the

scale. The possible reasons for this were discussed in the

results chapter and will not be repeated at this time.

Limitations of the Study

One of the most obvious weaknesses of the study lies
in the particular samples used. Since there was no sample
which could be placed at the extreme non-industrial end
of the continuum. there is no way of knowing whether such a

sample would have shown a fairly consistent tendency to

153

support the predictions. Since the Lenawee sample did show
such a tendency to support the predictions for the industrial
end of the continuum. data from such a fourth sample would
enable us to make better judgments about whether the tendency
for the Lansing and Turrialba samples to score toward the
industrial end of the continuum is actually a function of
the level of industrialization.

Another weakness lies in the method of analysis.
Since the study was concerned with the question of uni-
dimensionality of scales. it is possible to question the
use of factor analysis. rather than the scalogram analysis.
Guttman. for example. contends that the factorial structure
is derivable from a scale analysis but knowledge of the factor
analysis will not enable us to ascertain the scale patterns.l
However. the proponents of factor analysis argue just as
strongly for the merits of this technique. Ideally. both
techniques should have been used but this was not possible
within the operating limits of time and money.

It has already been pointed out that within the camp
of those who endorse factor analysis. there is considerable

dissention over whether or not factors should be rotated.

 

lLouis Guttman. fRelation of Scalogram Analysis to
Other Techniques.? in Samuel A. Sbouffer et. a1.. Measurement
and Prediction (Princeton. New Jersey: Princeton University

Press. 1950). p. 192.

 

154

Moreover. among those who advocate rotation. there are
numerous unresolved issues over the specific methods of
rotation (i.e.. oblique versus orthogonal). Similar
problems exist even after a decision has been made to use
orthogonal rotations. The reader is reminded of the
discussion in chapter four concerning quartimax and varimax
rotations.l Since the rationale for the choice made in this
study were presented in this earlier discussion. it will
not be repeated at this time. However. it was noted that
certain unanticipated problems occurred in the rotation.
The factors were rotated in the order of size. that
is the two largest. then the three largest. etc. However.
the ranking procedure which the computer lab used (at least
on this data) todk the sign of the eigenvalues into account.
This resulted in some factors being ranked below other
factors even though the "low? factor accounted for a
larger percentage of the total matrix variance. Since it
always turned out that the three largest factors were
positive. there was no problem in the rotation of these
factors. However. the fourth largest factor was frequently
ranked last and. therefore. was rotated only when all of
the factors were rotated. As pointed out in the section
on data analysis. increasing the number of factors being

rotated results in the larger factor splitting. In the

155

present study. this splitting was so extreme that there
appeared to be almost as many factors as there were items.
Consequently. most of the rotations involve only three
factors. even though there were fourth factors which had
enough significant loadings so they should have been
rotated. The net effect of this was an increased reliance
on the unrotated factors for interpretation.

A further limitation of the study involves the
ordinary problems of translation. and conceptual meaning
which arise when an instrument has to be translated into a
different language. Since the present writer is not qualified
to address himself to this problem. he can only acknowledge

the fact that this is an additional limitation of the study.

Implications for Future Research

In using or designing an instrument like the Work
Beliefs Check List. one of the first modifications which
seems to be needed has to do with the assumption that all
six of the sub-scales are equal indicants of the internali-
zation of industrial values. It has already been suggested
that the third scale appears to be tapping orientations toward
family and community ties rather than orientations toward
physical mobility. Even if orientations toward structured

time. work. change. deferred gratification. etc.. are

156

related to the level of industrialization. this does not
mean that all of the scales are equally weighted or equally
salient to the members of the sample. Strong negative
evaluation of the severance of kinship and community ties
by more gemeinshaft groups is consistent with theory and
the underlying assumptions of the instrument. However.
theory also leads us to expect weaker reactions to dimensions
which have no kinship basis. even though the reactions may
be in the same direction. Consequently. some modification
of the theoretical expectations of mean scores should be
made. In other words. rather than expecting the middle
range samples to score around the midpoint. it might be more
consistent with the present data if the expectations were
for scores between the midpoint and the high end of the
continuum. but with considerable inter—item variance.
Another modification of the instrument involves the
specific items of the respective scales. Some of the items
showed little or no discriminating power and should be
eliminated from the scale or included with a different
scale. These items have been discussed in some detail in
the results chapter where the outstanding example was
item six of the third sub-scale. This item did not appear
to belong with the other items on the level of face validity

and the results strongly supported this for all three samples.

157

However. the most important implication for future
research lies in the relationship of concept formation.
theory construction. and measurement to each other. Con-
sidering the amount of post—analysis projection of meaning
necessary to attempt to interpret the data. it seems legitimate
to ask whether the high level techniques used for data
analysis have increased the precision in the measurement of
values. In the opinion of the present writer. the justifi-
cation of the technique cannot be based on the present
data per gg but lies in the potential for developing and
improving the measurement of transcultural value variables
in future research. If the Work Beliefs Check List is
modified in accordance with the present findings. it should
yield increasingly less erratic results. and with fewer items
and less labor. If subsequent studies support this view.
the technique may aid in the development of a more precise
and a more simplified instrument for the measurement of
values. At this stage. presumably. the boundary conditions
for interpreting the data will be more accurately stated.
This writer strongly believes that such a stage will be
reached gnly by recognizing the dialogue between quantitative
and qualitative variables in social research. and the in-
escapable linkage of concept formation. theory construction

and measurement.

BIBLIOGRAPHY

Adler. Franz. PThe Value Concept in Sociology.” American
Journal of Sociology. 62 (November. 1956).

Allport. Gordon W. 9Attitudes.? in Carl Murchison (ed.),
A Handbook of Social Psychology. ‘Worcester:
Clark University Press. 1935.

Becker. Howard. FValue.9 UNESCO Dictionary of Social
Science. forthcoming.

Bridgman. P. W. FThe Logic of Modern Physics.9 excerpt
reprinted from Bridgman's bodk of the same name in
Herbert Feigl and May Brodbeck (eds.). Readings
in the Philosophy of Science. New Ybrk: Appleton-
Century—Crofts. Inc.. 1953.

Bugelski. B. R. A First Course in Experimental Psychology.
New York: Henry Helt and Co.. 1951.

Catton. William R.. Jr. PA Theory of Value.9 American
Sociological Review. 24 (June. 1959).

Columbia Associates in Philosophy. An Introduction to
Reflective Thinking. New York: Houghton—Mifflin
Co.. 1923.

Conant. James B. Modern Science and Modern Man. New Yerk:-
Doubleday Anchor Books. 1954.

De Hoyos. Arturo. Occupational and Educational Levels of
Aspiration of Mexican-American Ybuth. Unpublished
Ph.D. dissertation. Michigan State University. 1961.

Feigl. Herbert. VNotes on Causality.9 in Herbert Feigl
and May Brodbeck (eds.). Readings in the Philosophy
of Science. New York: Appleton-Century-Crofts.
Inc.. 1953.

. PLogical Empiricism.” in Herbert Feigl

and Wilfrid Sellars (eds.). Readings in Philosophical

Analysis. New York: Appleton-Century—Crofts. 1949.

 

158

159

"Operationism and Scientific Method.9 in
Herbert Feigl and Wilfrid Sellars (eds.). Readings
in Philosophical Analysis. New York: Appleton-
Century-Crofts. 1949.

 

Frank. Philipp. Philosophy of Science. Englewood Cliffs.
New Jersey: Prentice-Hall. Inc.. 1957.

Fruchter. Benjamine. Introduction to Factor Analysis. New
York: D. van Nestrand Co.. Inc.. 1954.

Giltter. Joseph B. and Manheim. Ernest. FSociological
Theory.? in Joseph B. Gittler (ed.). Review of
Sociology. New York: John Wiley and Sons. Inc..
1957.

Guilford. J. P. Psychometric Methods. New Yerk: McGraw-
Hill Bodk Co.. Inc.. 1954.

Guttman. Louis. fiRelation of Scalogram Analysis to Other
Techniques.9 in Samuel A. Stouffer et.al. .
Measurement gnd Prediction. Princeton. New Jersey:

Princeton University Press. 1950.

Harmon. Harry H. Modern Factor Analysis. Chicago. Illinois:
The University of Chicago Press. 1960.

Hempel. Carl. Fundamentals of Concept Formation in Empirical
Science. Chicago: University of Chicago Press. 1952.

. FThe Empiricist Criterion of Meaning.9
in A. J. Ayer. (ed.). Loqical Positivism.. Glencoe.
Illinois: The Free Press. 1959. '

 

. VThe Logic of Functional Analysis.7 in
Llewellyn Gross (ed.). Symposium on Sociological
Theory. Evanston. Illinois: Row. Peterson and
Company. 1959.

 

Hemple. Carl G. and Oppenheim.Paul. PThe Logic of Explanation.”
in Herbert Feigl and May Brodbeck (eds.). Readings
in the Philosophy of Science. New York: Appleton-
Century-Crofts. Inc.. 1953.

Johnson. Harry M. Sociology: A Systemstic Introduction.
New Ybrk: Harcourt. Brace and Company. 1960.

160

Kaiser. Henry F. VThe Varimax Criterion for Analytic
Rotation in Factor Analysis.” Psychometrika. 23
September. 1958.

Klineberg. Otto. Social Psychology. New York: Henry Helt.
1940.

Kluckhohn. Clyde. FValues and value-Orientations in the
Theory of Action: An Exploration in Definition and

Classification.? in Talcott Parsons and Edwards A.

Shils (eds.). Toward a_General Theory of Action.

Cambridge: Harvard University Press. 1951.

Kluckhohn. Florence R.. Strodtbeck. Fred L. Variations in
Value Orientations. Evanston. Illinois: Row
Peterson Co.. 1961.

Kolb. William L. VThe Changing Prominence of Values in
Modern Sociological Theory.9 in Howard Becker and
Alvin Boskoff (eds.). Modern Sociological Theory.

New Ybrk: The Dryden Press. 1957.

Krech. D. and Crutchfield. R. S. Theory and Problems of
Social Psychology. New Ybrk: McGraw—Hill. 1948.

London. Ivan D. VFree-Will as a Function of Divergence.?
Psyghological Review. 55. 1948.

Loomis. Charles P. Social Systems: Essays on their
Persistence and Change. Princeton. New Jersey:
D. Van Nostrand. Inc.. 1960.

Loomis. Charles P. and Loomis. Zona K. Modern Social Theories.
Princeton. New Jersey: D. Van Nostrand Company.
Inc.. 1961.

Loomis. Charles P. and McKinney. John C. 4The Application
of Gemeinshaft and Gesellschaft as Related to
Other Typologies.9 in F. Tonnies. Community and
Society. translated and edited by Charles P. Loomis.
East Lansing: Michigan State University Press.
1957.

 

Moore. Wilbert E. ”Editorial Introduction.9 in Charles P.
Loomis and Zona K. Loomis. Modern Social Theories:
Selected American Writers. Princeton. New Jersey:
D. Van Nbstrand Company. Inc.. 1961. XXIII.

161

Morris. Charles. Varieties of Human value. Chicago:
University of Chicago Press. 1956.

Nagel. Ernest. The Structure of Science. New York:
Harcourt. Brace and WOrld. Inc.. 1961.

Parsons. Talcott. ”General Theory in Sociology.9 in Robert
K. Merton. Leonard Broom and Leonard S. Cottrell.
Jr. (eds.). Sociology Today. New York: Basic Bocks.
1959.

. Structure agd Process in Modern Society.
Glencoe. Illinois: The Free Press. 1960.

 

. The Social System. Glencoe. Illinois: The
Free Press. 1951.

 

Parsons. Talcott and Shils. Edward A. (eds.). Toward a
General Theory of Action. Cambridge: Harvard
University Press. 1951.

Riesman. David. Glazer. Nathan. and Denney. Reuel. The
Lonely Crowd. Garden City. New Ybrk: Doubleday
and Co.. Inc.. 1956.

Rosenberg. Morris and Lazarsfeld. Paul F. (eds.). The Language
of Social Research. Glencoe. Illinois: The Free
Press. 1955.

 

Schlidk. Moritz. VMeaning and Verification.? in Herbert
Feigl and Wilfrid Sellars (eds.). Readings in
Philosophical Analysis. New York: Appleton-
Century-Crofts. 1949.

Scriven. Michael. FExplanation and Prediction in
Evolutionary Theory.” Science. vo1. 130.

Sears. Robert. fiTranscultural Variables and Conceptual
Equivalence.9 in Bert Kaplan (ed.). Studying
Personality Cross-Culturally. Evanston. Illinois:
Row. Peterson and Co.. 1961.

Sherif. M. and Cantril. H. VThe Psychology of Attitudes.?
Psychological Revue. 1945. 52,

Siegel. Sidney. Nonparametric Statistics. New York:
McGraw-Hill Book Co.. Inc.. 1956.

162

Thurstone. L L. The Measurement of Values. Chicago:
The University of Chicago Press. 1959.

Underwood. Benton J. Psychological Research. New Ybrk:
Appleton-Century—Crofts. Inc.. 1957.

Whyte. William H.. Jr. The Organization Man. New York:
Simon and Schuster. 1956.

Williams. Robin M.. Jr. American Society. New York:
Alfred A. Knopf. 1960.

Wrigley. Charles. VThe Distinction Between Common and
Specific Variance in Factor Theory.” The British
Journal of Statistical Psychology. a9. November. 1957.

Wrigley. Charles. Saunders. David R.. and Neuhaus. Jack O.
FApplication of the Quartimax Method of Rotation
to Thurstone's Primary Mental Abilities Study.?
Psychometrika. 23. June. 1958.

APPENDIX A

CORRELATION MATRICES FOR ALL SAMPLES

164-

Table A-1. Item Intercorrelations fgr Sub-Scale l.
Lenawee Sample (N = 439)"

 

 

 

 

 

Item 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1.1 13 08 -O6 16 ~04 -06 -01
1.2 15 17 23 14 18 12
1.3 01 12 14 O6 01
1.4 16 06 22 04
1.5 11 10 -Ol
l.6 09 07
1.7 05
1.8
*Decimal points omitted.
Table A-2. Item Intercorrelations forLSub-Scale l.
Turrialba Sample (N - 112)k

Item 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1.1 20 -01 ~09 21 OO 25 11
1.2 -06 -O4 -06 OO -07 -06
1.3 -13 02 OO -08 -23
1.4 -08 00 ~02 04
1.5 00 O6 -14
1.6 00 00
1.7 02
1.8

 

xDecimal points omitted.

165

Table A-3. Item Intercorrelations for Sub-Scale 1.
Lansing Sample (N : 87)x

 

 

Item 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1.1 -13 -03 -17 21 07 03 -14
1.2 8 43 21 24 25 17
1.3 02 -Ol 19 -O7 -04
1.4 41 37 69 21
1.5 14 35 32
1.6 31 05
1.7 21
1.8

 

* . .
Decimal p01nts omitted.

Table A-4. Item Intercorrelations fgr Sub-Scale 2.
Lenawee Sample (N a 439)"

 

 

Item 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
2.1 03 19 17 08 12 17 19
2.2 09 -01 14 14 05 07
2.3 22 17 15 11 08
2.4 02 08 02 17
2.5 40 10 09
2.6 15 14
2.7 29
2.8

 

* . .
Decimal poxnts omitted.

.166

Table A-5. Item Intercorrelations for Sub-Scale 2.
Turrialba Sample (N a 112)*

 

 

Item 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
2.1 26 28 -11 04 -08 19 00
,2.2 01 -13 10 03 21 00
2.3 -01 02 -02 -06 00
2.4 -07 ~01 -01 00
2.5 23 -02 00
2.6 -21 00
2.7 00
2.8

 

‘Decimal points omitted.

Table A-6. Item Intercorrelations £or Sub-Scale 2.
Lansing Sample (N a 87)

 

H.
n
(D.
B

 

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
2.1 07 13 17 11 02 27 17
2.2 10 26 O9 23 29 -02
2.3 14 06 01 ~06 14
2.4 -l9 -19 O7 -10
2.5 26 22 20
2.6 17 20
2.7 07
2.8

 

*
Decimal points omitted.

167

Table A-7. Item Intercorrelations far Sub-Scale 3.
Lenawee Sample (N : 439)‘

 

 

Item 3.1 3.2 3.3 3.4 3.5 3.6
3.1 28 02 14 01 01
3.2 14 15 12 06
3.3 16 12 04
3.4 08 ~08
3.5 -05
3.6

 

‘Decimal points omitted.

Table A-8. Item Intercorrelations for Sub-Scale 3.
Turrialba Sample (N = 112)*

 

 

Item 3.1 3.2 3.3 3.4 3.5 3.6
3.1 17 29 25 18 -04
3.2 08 09 O9 02
3.3 10 10 00
3.4 07 04
3.5 -04
3.6

 

*Decimal points omitted.

.168

Table A-9. Item Intercorrelations ior Sub-Scale 3.
Lansing Sample (N = 87)

 

 

Item 3.1 3.2 3.3 3.4 3.5 3.6
3.1 -02 -10 05 13 -21
3.2 06 21 11 -34
3.3 -05 -03 07
3.4 17 -28
3.5 -30
3.6

 

*Decimal points omitted.

Table A-lO. Item Intercorrelations for Sub-Scale 4.
Lenawee Sample (N = 439)

 

 

Item 4.1 4.2 4.3 4.4 4.5 4.6 4.7
4.1 07 18 19 12 10 04
4.2 06 13 17 15 11
4.3 17 Ol 21 09
4.4 21 11 04
4.5 13 00
4.6 02
4.7

 

”Decimal points omitted.

.169

Table A-ll. Item Intercorrelations for Sub-Scale 4.
Turrialba Sample (N - 112)"

 

 

Item 4.1 4.2 4.3 4.4 4.5 4.6 4.7
4.1 11 -04 26 21 03 -06
4.2 14 02 12 43 12
4.3 12 03 ~06 01
4.4 13 08 ~07
4.5 15 04
4.6 03
4.7

 

*Decimal points omitted.

Table A-12. Item Intercorrelations for Sub-Scale 4.
Lansing Sample (N a 87)"

 

 

Item 4.1 4.2 4.3 4.4 4.5 4.6 4.7
4.1 09 40 26 -16 O4 17
4.2 03 -14 05 24 -08
4.3 21 -01 21 01
4.4 30 -09 -04
4.5 -03 16
4.6 09
4.7

 

7'6 . .
Dec1mal poxnts omitted.

.......

.......

170

Table A-l3. Item Intercorrelations for Sub-Scale 5.
Lenawee Sample (N a 439)k

 

 

Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8
5.1 -05 -03 O6 -00 11 06 ~01
5.2 25 04 18 05 04 13
5.3 -02 25 O6 10 16
5.4 05 04 09 07
5.5 11 O4 27
5.6 05 06
5.7 20
5.8

 

*
Decimal points omitted.

Table A-14. Item Intercorrelations for?Sub-Scale 5.
Turrialba Sample (N s 112)

 

 

Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8
5.1 -19 -21 16 -06 09 06 -01
5.2 23 -14 04 04 06 12
5.3 -37 25 05 -11 09
5.4 -11 09 01 05
5.5 -02 -02 06
5.6 04 -04
5.7 17
5.8

 

*Decimal points omitted.

171

Table A-15. Item Intercorrelations for Sub-Scale 5.
Lansing Sample (N = 87)*

 

 

Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8
5.1 ~03 09 25 31 26 20 23
5.2 30 -01 01 -13 04 8
5.3 12 29 07 09 31
5.4 06 01 ~01 -01
5.5 -06 25 43
5.6 ~02 03
5.7 3
5.3

 

sh
'Decimal points omitted.

 

 

Table A-16. Item Intercorrelations for Sub-Scale 6.
Lenawee Sample (N : 439)*

Item 6.1 6.2 6.3 6.4 6.5 6.6 6.7
6.1 09 06 10 O9 O7 13
6.2 00 02 00 O9 19
6.3 32 17 O9 08
6.4 25 21 09
6.5 17 11
6.6 14
0.7

 

J.
Decimal points omitted.

172

Table A-17. Item Intercorrelations for¢Sub-Scale 6.
Turrialba Sample (N : 112)"

 

 

Items 6.1 6.2 6.3 6.4 6.5 6.6 6.7
6.1 01 26 09 22 ~01 04
6.2 21 15 14 10 38
6.3 18 31 17 ~04
6.4 34 12 -01
6.5 07 -01
6.6 11
6.7

 

* . . .
Dec1mal p01nts omitted.

Table A-l8. Item Intercorrelation far Sub-Scale 6.
Lansing Sample (N u 87)“

 

Items 6.1 6.2 6.3 6.4 6.5 6.6 6.7
6.1 10 35 46 -05 39 23
6.2 13 -10 27 17 22
6.3 21 ~02 -12 8
6.4 -02 11 -14
6.5 23 -14
6.6 19
6.7

 

* .
Decimal points omitted.

173

Table A-l9. Sub-Scale Total Score Coirelations
Lenawee Sample (N : 439)

 

 

Sub-Scale l 2 3 4 5 6
l 26 -14 16 21 18
2 -13 12 17 36
3 08 16 -01
4 22 20
5 24
6

 

*
Decimal points omitted.

Table A-ZO. Sub-Scale Total Score Correlations
Turrialba Sample (N z 112)*

 

 

Sub-Scale 1 2 3 4 5 6
l 01 09 26 02 23
2 -31 10 O9 20
3 O4 -18 ~07
4 25 30
5 21
6

 

* o O 0
Dec1ma1 p01nts omitted.

.174

Table A-21. Sub-Scale Total Score Cgrrelations
Lansing Sample (N : 87)

 

 

Sub-Scale 1 2 3 4 5 6
1 50 -29 44 21 38
2 -21 39 20 45
3 11 01 -10
4 33 36
5 43
6

 

*Decimal points omitted.

175

Table A-22. WBCL Intercorrelations for all Items
Lenawee Sample (N = 439) *

 

 

Item 1.1 l.gyl.3 1.4 1.5 1.6 1.] 1.8 2.1 2.2 2.3 2.4 2.5 2.6 2.] 2.8
1.1 -- I3 08 -06 16 -04 -06 -Ol 10 -01 10 ll -00 08 07 -06
1.2 -- 15 17 23 I4 18 12 10 01 IO 07 -00 10 06 07
1.3 -- 01 12 14 06 01 16 20 19 07 04 07 O4 04
1.4 -- 16 06 22 04 I6 06 -01 04 02 04 09 09
1.5 -- ll 10 -01 25 07 08 ll 04 13 18 12
1.6 -- 09 07 09 IO 02 07 13 -02 03 00
1.7 -- 05 07 06 04 07 04 06 13 08
1.8 -- -03 -02 -02 -02 10 01 00 -08
2.1 -- 03 l9 17 O8 12 l7 19
2.2 -- O9 -01 l4 I4 05 07
2.3 -- 22 I7 15 ll 08
2.4 -- 02 08 02 17
2.5 -- 40 10 09
2.6 -- 15 14
2.7 -- 29
2.8 '-
3.1
3.2
3.3
3.4
3.5
3.6
4.1
4.2
4.3
4.4
4.5
4.6
4.7
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
6.1
6.2
6.3
6.4
6.5
6.6
6.7

176

 

 

Table A-22. WBCL Intercorrelations for all Items
Lenawee Sample (N = 4391*

Item _3.1 3.2 313 3.4;y3.5 3.6 4.1 4.2 4.3 4.4 4.5 4.6 4.7
1.1 11 -O4 -10 01 -O6 06 -Ol 09 02 O3 08 O9 -13
1.2 -05 O3 -07 Ol -12 07 11 ~03 15 20 O9 O9 -02
1.3 -01 -OO -10 08 -IO 09 05 O9 05 O6 13 O3 -06
1.4 -05 -O8 -O4 -16 -OO 16 15 -05 O4 13 O6 O4 06
1.5 -07 -O7 -O4 -O4 -O3 ll 14 04 I6 12 15 12 -08
1.6 ~05 -Ol -15 -08 -Ol 12 02 O7 07 11 O9 12 -00
1.7 -04 -08 -IO -14 -O7 02 14 -07 11 O3 -O6 -01 05
1.8 -02 -Ol 01 -O6 -02 O4 02 -ll 11 05 01 -O3 01
2.1 -01 -12 -08 -OO -13 12 20 O3 O9 14 20 16 -03
2.2 -00 -14 -06 -13 -O8 02 -12 -OO -04 -O6 02 -O8 02
2.3 12 -O3 -O6 O4 -O9 05 O9 O6 -02 -Ol 11 15 -07
2.4 02 -O7 -O3 -02 -O4 O9 O8 O7 10 05 O9 12 OS
2.5 -05 -05 02 -15 OO IO 01 02 -O4 -01 O7 04 00
2.6 04 O3 O3 -02 -IO -01 L02 05 -O3 O3 04 OS -05
2.7 -07 -20 -O7 -02 -O3 -02 O7 02 -Ol 06 Ol 09 -OS
2.8 -08 -12 O4 02 02 ' OO -01 -O3 -01 IO 07 06 -00
3.1 -- 28 02 I4 01 01 -02 Ol -03 -11 O3 O6 02
3.2 -- I4 15 12 06 -OO -05 -02 -06 00 OO 03
3.3 -- 16 12 O4 - Ol -08 O3 05 -02 -02 01
3.4 -- O8 -08 ~06 15 OO 01 12 15 -08
3.5 -- ~05 ":04 193 O4 O7 -02 OO 12
3.6 -- I7 05 IO 16 O4 16 -00
4.1 -- O7 18 19 12 IO 04
4.2 -- 06 I3 17 15 II
4.3 -- 17 01 21 09
4.4 -- 21 11 04
4,5 -- 13 00
4.6 -- 02
11.7 --
5.1

5.2

5.3

5.4

5.5

5.6

5.7 .

5.8

6.1

6.2

6.3

6.4

6.5

6.6

6.7

177

 

 

Table A-22. WBCL Intercorrelations for all Items
Lenawee Sample (N = 4391*
Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 6.1 6.2 6.3 6.4 6.5 6.6 6.]
1.1 -09 11 16 OO 17 -Ol 11 15 22 01 I6 09 -O4 04 10
1.2 04 Ol 05 -Ol 11 02 IO 07 -Ol 00 O3 O8 02 O6 04
1.3 05 11 17 -O3 13 -05 O3 O4 01 01 11 -OO 06 ~05 06
1.4 25 -08 06 Ol 09 10 05 01 -O7 O6 05 04 IO -04 06
1.5 03 02 27 -01 17 ll 01 14 O7 O6 O7 17 12 14 04
1.6 -01 -02 -01 -O4 O3 02 O7 05 -Ol 00 O7 O4 12 O6 04
1.7 07 -02 02 -O4 -O4 01 -02 ~02 -O8 04 -OO 08 II 19 -00
1.8 08 -O6 -O6 01 -O6 O9 08 02 -12 02 -O7 -O9 03 03 13
2.1 13 10 12 -O7 20 03 -Ol 16 02 O7 32 13 II 14 06
2.2 02 -05 01 -04 OO -06 12 05 O4 12 -OO -07 OO 04 18
2.3 -01 IO O7 -11 13 -Ol 01 17 O7 16 15 12 02 04 16
2.4 03 O8 13 01 I9 -02 05 13 O7 O8 14 O7 02 I4 13
2.5 05 -O4 -O7 O7 O3 02 -02 O7 10 08 10 -Ol 09 08 18
2.6 -08 O9 O7 08 17 O4 02 I7 O7 13 12 OO 01 I3 25
2.7 -04 05 14 -05 OO -08 05 08 05 14 O4 05 O9 O6 13
2.8 04 05 02 -Ol 08 -05 05 O3 03 -01 IO 10 O9 08 06
3.1 -05 17 ll 05 O4 05 O9 O7 02 -OO 02 -05 -IO -02 II
3.2 -07 O3 02 IO 08 05 03 -Ol 03 -O3 -05 -O3 -O6 -O4 -05
73.3 03 -O4 -O4 02 O4 05 -OO -05 08 -07 -Ol -01 -05 -02 -05
3.4 -08 17 14 -O4 14 -08 -05 O8 10 -O4 O4 -05 -05 02 02
3.5 05 -03 -05 15 04 IO 02 02 .03 -12 -O9 -O4 02 -OO 00
3.6 02 -00 -03 -OO 11 06 OO 03 O7 02 O7 16 13 18 07
4.1 10 -OO 02 -03 05 05 -O3 O6 -O3 05 ll 15 I8 05 -01
4.2 -13 IO l3 -14 21 01 -O3 O7 15 02 23 07 OO 10 05
4.3 03 O4 O7 06 10 -00 ~04 01 -O6 -O3 O8 15 18 15 -06
4.4 07 O3 05 09 IO -03 O6 02 -OO O4 01 O9 11 -02 06
4.5 10 12 18 O3 27 IO 03 22 IO 06 29 O3 14 Ol 08
4.6 00 09 IO -05 23 -01 -O9 I3 -01 O6 17 O7 13 16 05
4.7 07 -02 -05 O7 -O6 l3 I4 -05 -O7 02 -05 -02 06 Ol -05
551 -- -05 -O3 06 -OO 11 06 -Ol 05 IO 01 02 05 -Ol -01
5.2 -- 25 04 I8 05 O4 13 O4 O6 06 -OO -13 01 -01
5.3 -- -02 25 06 IO 16 19 I3 16 O7 15 09 10
5.4 -- 05 O4 O9 O7 04 -00 -O9 -05 -Ol -04 -01
5.5 -- II 04 27 18 IO 22 11 O6 11 15
5.6 -- 05 O6 O9 -02 02 12 O7 05 -00
5.7 -- 20 O6 -04 -Ol -02 01 OO 13
5.8 -- 10 IO 21 OO 04 IO 17
6.1 -- O9 06 IO 09 O7 13
6.2 -- OO 02 OO O9 19
6.3 --. 32 l7 09 08
6.4 -- 25 21 09
6.5 -- 17 H
6.6 -- I4
607 --

Table A-23.

.178

Turrialba Sample (N = 112)

WBCL Intercorrelations for all Items

 

 

Item 1.1 1.2 1.3 1.4 1.5 1.6 1.] 1.8 2.1 2.2 2.3 2.4 215 2.6 2.7 2.8
1.1 -- 20 -01 -09 21 oo 25 11 06 21 -08 01 -12 -03 09 00
1.2 -- -06 -04 -06 oo -07 -06 -12 -06 -15 27 -07 -08 15 00
1.3 -- -13 02 oo -08 -23 10 07 18 -11 -09 07 -04 00
1.4 me&~-08 oo 02 04 -16 -14 -02 18 04 -08 -06 00
1.5 -- oo 06 -14 08 18 -04 -05 02 -03 -11 00
1.6 -- oo oo oo oo oo oo oo oo oo 00
1.7 -- 02 01 09 -06 -17 -04 14 04 00
1.8 -- 07 -13 -05 -11 03 -16 03 00
2.1 -- 26 28 -11 04 -08 19 00
2.2 01 -13 -10 03 21 00
2.3 -- -01 02 -02 -06 00
2.4 -- -07 -01 -01 00
2.5 -- 23 -02 00
2.6 -- -21 00
2.7 -- 00
2.8 '-
3.1
3.2
3.3
3.4
3.5
3.6
4.1
4.2
4.3
4.4
4.5
4.6
4.7

oos16~611ruoho-—

O‘O‘O‘O‘O‘O‘O‘ mmulmmmmm
o o o o o o o o
Nam-PWNd

L179

Table A-23. WBCL Intercorrelations for all Items.
Turrialba Sample (N = 112)

 

O ('1'

00.

3.1 3.2 3.3 3.4 .3.5 3.6 4.1 4.2 4.3 4.4 4.5
05 l9 -14 -O3 02 02 20 15 01 -06 10
01 09 OS 05 01 -02 303 ~06 -O4 -O7 -01
20 13 05 04 -O6 13 05 02 -O4 07 I7

-02 I7 12 07 07 -O3 -04 -08 -O6 -09 -17

-02 15 -21 -07 -08 -05 ll 30 -09 -06 30
00 00 00 00 00 00 00 00 00 00 00

-ll 11 -09 -09 14 12 22 13 08 10 04

-I9 00 -04 03 04 -13 06 -03 ll -03 07

2.1 -22 02 -I4 -07 -O7 -10 -12 20 -12 O9 -04
2.2 -22 -12 -16 -19 -11 -I4 16 12 -12 II 14
2.3 06 -O4 -12 -O7 -02 02 O8 -O9 -O8 02 -05
2.4 00 -04 -Ol 03 OO -05 -O6 -05 -IO -15 -13
2.5 -09 -O6 01 -05 -O4 -05 -07 02 12 08 -09
2.6 05 -10 Ol -10 01 -02 -08 20 14 07 -01
2.7 -26 O9 -24 -12 -05 -O4 -05 10 O6 08 -16
2.8 00 OO 00 00 OO 00 OO 00 OO 00 00
3.1 -- I7 29 25 18 -O4 12 -20 -12 -O4 12
3.2 -- 08 O9 O9 02 O9 15 O8 10 10
3.3 -- IO 10 -OO -06 -15 -05 12 -06
3.4 -- 07 O4 05 -O7 -16 O4 13
3.5 -- -04 -II -08 19 O7 00
3.6 -- -02 I6 24 14 -10
4.1 -- II -04 26 21
4.2 -- 14 02 12
4.3 -- 12 03
4.4 -- 13
4.5 "
4.6
4.7
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
6.1
6.2
6.3
6.4
6.5
6.6
6.7

.180

WBCL Intercorrelations for all Items.

Turrialba Sample (N = 112)

 

 

u@_5J525354555£5J58 a1a2aia4a5a6
1.1 -25 25 12 -10 -oo -02 06 02 05 25 13 04 02 07
1.2 -11 -05 -05 02 03 05 -01 -11 -05 06 -07 -07 -02 -10
1.3 05 05 14 -05 -06 -03 oz 05 -07 13 14 07 -oo -01
1.4 -05 05 05 06 -08 21 02 -14 -06 -12 -09 04 -03 -13
1.5 -24 32 1o -16 -04 -12 -04 07 -10 10 15 10 -05 -03
1.6 00 oo oo oo oo oo oo oo oo oo oo oo oo 00
1.7 11 04 -09 -11 -11 -05 -04 02 06 04 -07 -04 -O6 16
1.8 oz -18 09 -08 oo 24 -04 -06 14 oo 05 -03 oo 20
2.1 06 03 11 -11 05 08 -01 09 08 14 13 04 04 05
2.2 -02 oz 11 -23 08 04 03 02 -01 16 03 11 -01 01
2.3 10 -06 10 08 15 -05 -06 oz - 1o -06 01 -08 -12 12
2.4 -08 -09 -2o 14 15 -06 06 -07 oo -04 -02 09 -05 02
2.5 02 09 15 -12 06 06 -02 -02 20 09 18 23 34 12
2.6 -01 -08 -06 -14 06 -24 14 03 03 02 08 07 11 06
2.7 01 -09 -oz 04 11 09 05 03 05 17 05 08 -04 04
2.8 00 oo oo oo oo oo oo oo oo oo oo oo oo 00
3.1 -14 05 11 11 -02 -22 -11 -09 -01 -03 12 02 -04 03
3.2 -13 -01 -03 07 07 11 06 10 -oz 07 18 -01 02 11
3.3 -00 -07 -02 -1o -22 05 -06 -oo 13 -25 -06 -15 -14 -02
3.4 -04 -04 01 -oo 03 -13 -oo 15 -14 -04 06 04 04 -07
3.5 -10 -03 -15 -1o -24 01 -14 -05 06 -25 -03 02 -04 11
3.6 05 oo -13 07 -16 -07 18 06 -04 oo -06 -05 -oz 07
4.1 -11 06 18 -1o 14 05 -01 14 -05 06 08 10 -02 15
4.2 -06 21 -01 -09 02 oz 14 25 01 15 15 02 16 03
4.3 -Ol 05 -02 01 -08 14 -04 -Ol 21 12 29 01 24 18
4.4 -04 04 10 -06 01 -01 04 15 -11 -01 -03 15 -05 24
4.5 -08 14 15 -09 12 -12 -08 07 -14 18 11 13 03 11
4.6 -03 27 20 -12 15 17 13 12 -08 23 08 02 08 02
4.7 09 07 -2o 26 -01 14 oo -13 oo 01 -17 01 -os -05
5.1 -- -19 -21 16 -06 09 06 -01 04 oo -03 -10 ~09 03
5.2 -- 23 -14 04 04 06 12 -13 18 11 15 14 06
5.3 -- -37 25 05 -11 09 -05 19 07 10 14 -13
5.4 -- -11 09 01 05 07 -10 -10 -06 ~08 04
5.5 -- -02 -oz 06 -12 12 12 -10 -02 14
5.6 ‘ -- 04 -04 20 04 02 -07 10 -01
5.7 -- 17 - 13 06 -12 14 18 05
5.8 -- -10 oo 14 09 06 14
6.1 6+ 01 26 09 22 -01
5,2 -- 21 15 14 10
6.3 -- 18 31 17
6.4 -- 34 12
6 5 -- 07
6 6

6 7

181

Table A-24. WBCL Intercorrelations for all Items.
Lansing Sample (N = 87)*

 

o 0001'?
OOVmU'I-I-‘WN-‘g

did—l—d—d—

NNNNNNNN
o o o o o o o
mVO‘W-tWN—i

coco. no. )0.
VmW#WNd O‘W-PWN-i

#P-P-P-P-P'F WWWWWW

no... .0000.

o
Nam-FWN-d

O‘O‘O‘O‘O‘O‘O‘ WWWWWU'IU'IU'I

l.l lil.31.41.51.61.7 1.18 2.12.2 243-2g4 2.5 2,6 2;] 2.8
-- -13 ~03 -I7 21 07 03 -l4 13 15 I6 O7 14 38 15 01
-- 08 43 21 24 25 I7 00 06 -28 -02 09 -10 34 28

-- 02 -Ol 19 -07 -04 14 I7 29 O9 04 09 -05 03

-- 41 37 69 21 20 25 -12 35 20 -18 34 35

-- 14 36 32 24 16 ll 18 24 -00 31 18

-- 31 05 08 20 -l4 19 O9 O3 31 13

-- 21 10 17 03 17 20 -08 38 28

-- 12 -11 -Ol 11 17 -ll 19 17

 

-- 07 13 17 ll 02 27 17
-- 10 26 09 23 29 -02

-- 14 O6 01 -O6 14

-- -19 -19 07 -10

-- 26 22 20

-- 07

182

Table A-24. WBCL Intercorrelations 1tor all Items.

Lansing Sample (N = 87)

A

 

 

o o o o o o
mNO\U1.p'WN—‘

o o o o o o
VO‘U‘I-P'WN‘

oxoxo‘oxmo‘ox U'IU'IU‘IU'IU'IWWUW

Item _3. 1 3.2 3.3 3.4 g3.5t 3.6 4.1 4.2 4:3 4.4 4:5 4.6 4.7
1.1 04 13 -21 15 02 -06 16 29 15 -21 06 49
1.2 -11 -15 -07 -28 -22 25 16 -10 10 13 03 -13
1.3 25 04 -23 03 08 ~07 04 13 08 03 11 31
1.4 -01 -31 12 -35 -36 69 56 -13 38 35 -09 -12
1.5 -09 -06 —25 -04 -35 24 44 -05 16 11 18 13
1.6 04 06 -18 -26 -21 31 20 07 25 19 05 07
1.7 -11 -44 07 -28 -41 82 50 10 32 17 -13 03
1.8 -19 04 06 -11 -31 21 15 -23 08 17 05 -09
2.1 08 -03 -13 04 -19 10 21 -09 21 17 ~01 24
2.2 15 -12 -15 -04 -06 17 13 14 11 -13 13 17
2.3 15 05 -27 23 21 -07 03 11 ~06 08 03 18
2.4 13 03 -08 -03 -11 17 17 15 01 03 02 15
2.5 -16 12 -11 -14 -oo 29 28 -16 29 09 03 -07
2.6 04 07 -21 -03 05 -08 -09 09 -oo -02 32 19
2.7 -03 -19 -29 -2o -35 27 25 -16 13 01 08 07
2.8 -11 -09 -14 -35 -24 28 17 -03 28 16 -05 03
3.1 -- ~02 -1o 05 13 -21 06 20 -05 -05 -07 25
3.2 -- 06 21 11 -34 -21 -13 04 09 23 15
3.3 -- -05 -03 07 05 -21 -07 17 -03 -29
3.4 -- 17 -28 -08 26 -01 -09 26 15
3.5 -- -3o -19 13 -02 . 02 -18 08
3.6 -- 35 01 44 I7 -13 -07
4.1 -- 09 4o 26 -16 04
4,2 03 -14 05 24
4,3 -- 21 -01 21
4.4 -- 30 -09
4.5 -- -03
4.6

4 7

Table A-24.

183

WBCL Intercorrelations for all items.
Lansing Sample (N = 87)*

 

 

Hm 5J53535A555£5J5£ 6J62636A65666J
1.1 01 26 24 O6 O4 -I4 -14 O9 22 13 34 12 -03 01 10
1.2 02 -06 -13 -06 -O7 03 -O8 -O6 -10 O7 -16 -O6 14 -05 07
1.3 07 -18 II O6 -01 23 -15 15 IO 04 IO -06 10 13 01
1.4 12 ~04 -O9 -O6 19 O3 14 22 IO 22 -07 I3 52 14 -06
1.5 01 34 02 -O3 O3 -20 08 22 15 16 12 12 22 O4 16
1.6 15 05 05 -05 22 -O4 I9 19 02 19 -01 -Ol 35 Ol -11
1.7 31 15 -Ol 09 I8 05 IO 32 27 22 O4 33 33 21 00
1.8 28 00 -13 -15 O6 -05 22 16 08 II -03 O4 O4 -O8 26
2.1 03 -13 10 O9 09 05 01 O4 27 I3 33 33 19 O6 09
2.2 -00 -00 13 IO 18 -II 05 16 22 02 -05 02 IO 29 10
2.3 -01 IO 16 03 -03 04 14 17 30 12 25 17 -15 28 24
2.4 04 15 oz -05 1o 12 28 19 15 19 15 15 23 17 06
2.5 13 -05 03 -Ol 06 -18 01 07 -Ol 03 -05 O9 O9 -02 03
2.6 -15 -OO 27 I3 01 -O7 -19 -02 I9 -02 02 02 -05 -02 13
2.7 11 -IO 01 O3 02 -08 13 ll 14 O6 13 22 I3 -02 -02
2.8 -04 -O7 23 O6 10 06 -Ol 03 15 I9 06 O6 O6 17 17
3.1 -27 01 -OO -03 03 -Ol -02 02 -O7 -08 09 Ol 01 -02 -18
3.2 00 -O3 16 O4 06 ~11 O9 -O9 -10 -04 14 -18 -IO -04 -01
3.3 17 -15 -09 -O3 08 -O4 -O3 -07 ~09 08 -O9 -O9 -01 -12 -13
3.4 -12 23 I9 05 -05 O7 -05 O8 15 -14 20 03 -15 01 16
3.5 -05 -19 -ll 02 -13 -Ol 03 -22 -O9 -05 02 -I6 -16 -OO -03
3.6 31 06 -Ol O9 18 05 IO 22 I4 13 -IO 19 33 21 00
4.1 10 15 O3 -O3 26 -II 25 39 18 24 II 23 47 13 -03
4.2 03 31 18 01 24 11 -O7 25 26 IO 21 O6 -O9 09 II
4.3 16 -03 I4 -02 20 II 05 28 12 I6 10 IO 28 12 -08
4.4 -13 O4 02 -05 IO -00 05 O3 -08 O8 -O3 O6 32 -01 -16
4.5 -18 18 31 -II 05 -OO -28 01 13 -II 01 01 -08 -O6 01
4.6 -00 -O3 17 O6 15 O6 -O9 15 25 O4 18 26 IO 13 -04
4.7 -04 -Ol 23 -05 10 -O7 28 19 23 -08 O6 23 06 17 06
5.1 -- -08 09 25 31 26 20 23 33 16 -O3 O8 -O3 21 25
5.2 -- 3O -01 Ol -13 O4 38 11 IS 03 03 IO 07 04
5.3 -- 12 29 O7 O9 31 22 13 Ol II 01 42 06
5.4 -- O6 01 -01 -Ol 04 O3 08 -07 -I4 19 17
5.5 -- -06 25 43 22 11 07 21 28 18 06
5.6 -- 02 03 20 O9 -O4 04 -I3 24 01
5.7 -- 28 IO 09 19 II II 17 14
5.8 -- 23 20 -02 20 20 I7 -01
6.1 -- IO 35 46 -05 39 23
6.2 -- 13 -O9 27 17 22
6.3 -- 21 -02 -12 08
6.4 -- -02 11 -14
6.5 -' 23 '1“
6 6 -- I9
6 7

.184

Table A-25. Item Intercorrelations for Sub-Scale l.

Pooled Sample, Lenawee, Turrialba, and

Lansing (N a 638)*

 

 

 

 

 

Item 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
1.1 08 08 -09 20 -02 O4 -01
1.2 08 18 18 18 13 10
1.3 -02 O9 O9 04 -05
1.4 14 10 21 06
1.5 10 13 02
1.6 08 05
1.7 06
1.8
*Decimal points omitted.
Table A-26. Item Intercorrelations for Sub-Scale 2.
Pooled Sample, Lenawee, Turrialba, and
Lansing (N: 638)

Item 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
2.1 09 21 08 OO 06 13 07
2.2 08 -00 08 13 10 05
2.3 16 11 11 05 07
2.4 -01 02 O3 11
2.5 33 13 14
2.6 10 12
2.7 25
2 8

 

*Decimal points omitted.

.185

Table A-27. Item Intercorrelation for Sub-Scale 3.

Pooled Sample - Lenawee, Turrialba, and
Lansing (N : 638)*

 

 

Item 3.1 3.2 3.3 3.4 3.5 3.6
3.1 21 05 18 07 -03
3.2 14 11 11 01
3.3 10 ll 05
3.4 12 -11
3.5 -08
3.6

 

*
Decimal points omitted.

Table A-28. Item Intercorrelation for Sub-Scale 4.

Pooled Sample - Lenawee, Turrialba, and
Lansing (N : 638)*

 

 

Item 4.1 4.2 4.3 4.4 4.5 4.6 4.7
4.1 11 26 26 07 09 06
4.2 09 06 12 23 08
4.3 20 02 18 05
4.4 23 09 ~01
4.5 15 00
4.6 ‘ 02
4.7

 

*Decimal points omitted.

186

Table A-29. Item Intercorrelations for Sub-Scale 5.
Pooled Sample - Lenawee, Turrialba, and
Lansing (N : 638)*

 

 

Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8
5.1 -08 00 09 09 12 04 05
5.2 25 01 15 O3 05 19
5.3 -09 32 05 01 19
5.4 03 05 08 07
5.5 06 03 30
5.6 04 04
5.7 21
.1_5.8

 

* . . .
Dec1mal p01nts omitted.

Table A-30. Item Intercorrelations for Sub-Scale 6.
Pooled Sample - Lanawee, Turrialba, and
Lansing (N : 638)

 

 

Item 6.1 6.2 6.3 6.4 6.5 6.6 6.7
6.1 06 09 13 10 O3 11
6.2 07 03 05 12 23
6.3 ‘ 27 14 10 06
6.4 21 16 04
6.5 12 05
6.6 15
6.7

 

xDecimal points omitted.

187

Table A-3I. WBCL Intercorrelations for all Items.
Pooled Sample: Lgnawee, Turrialba, and
Lansing (N . 638)

 

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 2.1 2.2 2.3 2.4 2.5y2.6 2.7
-- 08 08 -09 20 -02 O4 -01 ll 04 O6 10 00 11 O8
-- 08 18 18 18 13 10 Ol 00 -OO 08 04 04 I4

-- -02 O9 09 O4 -05 20 15 20 03 -02 08 -01

-- 14 10 21 06 07 05 -02 09 04 -00 10

-- IO 13 02 19 09 05 10 08 07 I6

-- 08 05 01 09 -02 10 15 -02 ll

-- 06 09 08 02 02 02 O7 13

-- 02 -05 -02 -02 09 -03 O3

ddddd—l—l—l—
o cort-
mumm-I-‘WN-‘g

-- O9 21 08 00 06 13
-- 08 -00 08 13 10

-- 16 11 ll 05

-- -Ol 02 03

-- 33 I3

-- 10

b.‘ o

NNNNNNNN
O
(”\IO‘U'I-PWNH

o o o o o
mer-wN—

ppf-g-FF: wwwwww
I 0 O O O O 0 O O
CDNO‘U'I-PWNH \ImU'I-P'WN—

Q 0...
\1 oxU'I-I-‘wN-H

OO‘O‘GGChCh U'IU'IU1U‘IU‘IU'IU'IU'I
O

188

 

 

Table A-3I. WBCL Intercorrelations for all Items.
Pooled Sample: Lenawee, Turrialba, and
Lansing (N = 638)*
Item 3.1 3.2 y333 3.4 3.5 43.6 4.1 4.2 4.3 4.4 4.5 4.6 4.7
1.1 11 02 -I3 O8 -01 03 IO 14 07 -Ol 12 23 -10
1.2 -05 03 -O4 -05 -II II 12 -O3 13 I4 03 O3 -02
1.3 IO -01 -II 14 -05 05 O6 O7 05 O9 20 12; -04
1.4 -05 -06 OO -16 -O3 21 17 -O7 07 IO -02 -Ol 06
1.5 -04 -02 -11 -OO :06 II 26 08 13 IO 21 13 -06
1.6 -04 04 -II -13 -03 I6 09 O9 12 11 01 O8 04
1.7 -05 -O9 -09 ~09 -05 I4 22 -Ol 13 O8 01 OO 04
1.8 -07 -01 01 -O6 -05 05 05 -12 O9 05 O3 -O3 -01
2.1 -02 -12 ~12 08 ~10 05 O8 O3 O4 13 I6 21 -04
2.2 -O4 -14 -IO -14 -IO 02 -02 02 -05 -O6 05 01 06
2.3 10 -03 -IO 04 -O6 02 05 03 -05 -OO 06 IO -05
2.4 04 -O4 -02 -Ol -O3 10 10 O9 O6 02 02 O7 08
2.5 -O7 01 O3 -16 Ol 13 IO 02 O8 O3 -02 02 03
2.6 04 -OO -01 -01 -O7 -02 -O4 07 OO 02 O8 O6 04
2.7 -09 -13 -II -07 -O7 O4 II 01 05 05 -O3 07 -00
2.8 ~09 -O6 03 -O9 -02 O6 05 -Ol 06 O7 -O3 02 01
3.1 -- 21 05 I8 07 -O3 O6 02 -02 -03 O8 10 -01
3.2 -- 14 II 11 Ol -02 -Ol 03 02 02 O4 04
3.3 -- 10 II 05 01 -O9 01 O9 -05 -07 00
3.4 -- 12 -11 -OO 14 02 O6 22 18 -08
3.5 -- -O8 -05 05 O8 09 -Ol 02 07
3.6 -- 19 O6 20 15 -O3 O7 O3
4.] -- ll 26 26 O7 O9 06
4.2 -- O9 O6 12 23 08
4,3 -- 21 02 I8 05
4.4 -- 23 09 '01
h,5 -- 15 00
4.6 " 02
4.7 "
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
6.1
6.2
6.3
6.4
6.5
6.6
6.7

Table 31.

.189

WBCL Intercorrelations for all Items.
Pooled Sample: Lenawee, Turrialba, and
Lansing (N = 638)*

 

 

 

Item 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 6.1 6.2 6.3 6.4 6.5 6.6 6.7
1.1 -08 I7 19 OO 16 -02 06 I4 15 O8 19 O8 -O4 O7 10
1.2 -01 -OO -03 00 05 O3 O8 03 -Ol 02 -03 04 05 -Ol 05
1.3 10 07 22 -03 14 -OO -03 O8 -05 06 I3 01 02 05 03
1.4 15 -05 03 00 05 11 05 -OO -05 O4 00 05 I4 -04 02
1.5 -04 15 18 -O3 12 Ol 01 17 O3 09 IO 14 ll 08 09
1.6 -02 -00 -OS -02 03 01 ll 08 02 02 O4 O3 18 00 00
1.7 14 02 02 -05 01 -OO -03 05 -O4 07 -OO 08 09 20 00
1.8 08 -08 -03 -03 -O4 10 O7 02 -O6 O3 -04 -06 03 06 13
2.1 16 O6 20 -O8 20 O4 -O6 13 -Ol 11 27 12 O6 16 06
2.2 00 -05 O4 -O7 02 -05 08 04 O7 11 -01 -02 01 06 13
2.3 02 07 O9 -O7 10 -02 00 12 O7 11 13 09 -O3 O9 11
2.4 00 07 03 03 17 -OO 10 12 06 08 ll 08 06 11 09
2.5 01 -02 -O7 06 01 -OO 02 06 ll 06 07 O3 13 02 14
2.6 -06 05 O8 04 13 -03 OO 11 O6 O9 10 02 OO 10 20
2.7 -02 01 05 -01 01 -05 O8 08 07 12 05 O8 10 02 09
2.8 -03 02 -02 02 Ol -02 07 Ol 08 OO 06 08 IO 02 05
3.1 -07 15 I3 06 08 00 O4 07 ~03 -OO 06 -O3 -08 02 05
3.2 -10 02 -02 11 O4 O4 08 00 03 -02 01 -O4 -O4 -O4 -04
3.3 02 ~06 -O7 01 -02 O4 00 -O4 07 -09 -04 -O4 -O4 -05 -06
3.4 00 17 20 -02 20 -05 -07 14 OO -02 08 -03 -08 O7 05
3.5 01 -03 -06 10 OO 07 ~00 01 -OO -13 -06 -04 -Ol 04 -03
3.6 06 01 -O6 03 O7 05 05 O7 O7 03 02 14 16 13 06
4.1 04 O7 O6 -02 16 01 O6 21 -Ol 10 IO 14 22 IO -02
4.2 -09 I6 09 -O8 18 O3 01 I6 13 05 20 06 01 O7 06
4.3 05 04 06 OS 12 05 00 II -02 O4 12 IO 20 I4 -02
4.4 02 06 09 05 14 -Ol 05 10 -07 05 00 O9 12 O9 02
4.5 06 I4 27 -03 26 O3 -08 16 -Ol 09 20 05 04 IO 08
4.6 02 I3 18 -O3 25 O4 -04 17 -O3 10 I6 09 IO 14 06
4.7 03 -Ol -06 O8 -05 IO 15 -03 -01 -Ol -06 02 06 -OO -01
5.1 -- -O8 00 O9 O9 12 O4 05 02 10 01 00 -00 O8 02
5.2 -- 25 02 15 O3 05 19 01 IO 07 03 -O7 04 03
5.3 -- -O9 32 05 01 I9 07 l6 14 08 O7 13 09
5.4 -- O3 05 08 07 03 -01 -O7 -05 ~03 OO 02
5.5 -- 06 03 30 05 13 I9 08 05 18 15
5.6 -- O4 04 ll 01 Ol 08 O4 O6 02
5.7 -- 21 06 -01 -02 O3 06 -OO 12
5.8 -- 05 10 I6 05 06 I4 12
6.1 -- O6 O9 13 10 03 11
6.2 -- 07 03 05 12 23
6,3 -- 27 I4 10 06
5.1. -- 21 I6 04
6.5 -- 12 05
6 5 -- 15
6 7

19o. ‘. -.

Table A-32. Sub-Scale Total Score Correlations
Pooled Sample - Lenawge, Turrialba,
and Lansing (N‘: 638)

 

 

Sub-Scale 1 2 3 4 5 6
1 26 -O9 26 21 22
2 -16 15 15 35
3 14 14 -03
4 30 24
5 26
6

 

3': . . .
Dec1ma1 pOints omitted.

APPENDIX B

FACTOR MATRICES FOR POOLED SAMPLES

Table B-1. Factor matrix for WBCL Sub-Scale 1.

pooled sample:
(N-638).*

193

Lenawee. Turrialba. Lansing

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 1)
I II III I II III
1.1 14 -4O 08 -10 42 01
1.2 49 -01 -19 21 18 45
1.3 13 -16 -09 —05 16 14
1.4 _41 3O 12 50 -02 16
1.5 447 -26 11 18 46 17
1.6 30 06 -25 10 01 38
1.7 38 12 21 42 15 07
1. 13 10 -04 12 -03 12
%»Total Variance 50 21 11 3O 26 25

 

*Decimal points omitted.

Table B-2. Factor matrix for WBCL Sub—Scale 2 pooled

sample:

Lenawee. Turrialba. Lansing (Ne638).*

 

 

 

 

Principal Axes varimax
Items Loadings Loadings
(Sub-Scale 2)
I II III I II III
2.1 23 29 -20 02 4O 13
2.2 22 02 -O7 16 14 08
2.3 29 21 -31 14 45 05
2.4 12 21 -1O -02 24 10
2.5 52 -31 03 59 -02 12
2.6 54 -30 -ll 62 10 05
2.7 37 22 26 12 08 48
2. 38 21 29 13 05 51
% Total Variance 48 21 14 37 21 25

 

*Decimal points omitted.

194

Table B—3. Factor matrix for WBCL Sub-Scale 3 pooled

sample: Lenawee. Turrialba. Lansing (N—638).*

 

 

 

 

 

 

 

 

 

Principal Axes varimax
Items Loadings Loadings
(sub'scale 3) I II III I II III
3.1 43 04 -23 39 25 -15
3.2 45 28 -OO 51 O9 11
3.3 23 09 20 19 08 25
3.4 45 -29 01 14 51 04
3.5 27 -05 18 13 22 21
3.6 -09 20 01 05 -21 03
% Total Variance 58 17 10 39 35 11
*Decimal points omitted.
Table B-4. Factor matrix for WBCL Sub-Scale 4 pooled
sample. Lenawee Turialba. Lansing (-638).*
Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 4)
I II III I II III
4.1 49 -21 -15 50 05 24
4.2 30 34 -05 10 44 05
4.3 46 -O9 -26 51 14 10
4.4 52 -22 26 25 O3 57
4.5 29 12 31 -O6 23 38
4.6 36 34 -O4 13 47 09
4.7 07 06 -1O 09 09 —04
% Total variance 54 18 14 32 26 29

 

*Decimal points omitted.

195

Table B—5. Factor matrix for WBCL Sub-Scale 5 pooled
sample. Lenawee. Turrialba. Lansing (N—638).*

 

 

 

 

Principal Axes Varimax
Items Loadings Loadings
(Sub-Scale 5)
I II III I II III
5.1 07 23 27 35 06 08
5.2 39 -16 -24 -32 31 20
5.3 52 -30 O7 -15 59 04
5.4 04 24 01 14 -07 19
5.5 57 —03 23 13 57 19
5.6 10 ll 13 16 09 07
5.7 17 32 -17 02 -05 40
5.8 55 26 —14 -01 30 55
'% Total Variance 52 19 12 15 41 28

 

*Decimal points omitted.

Table B-6. Factor matrix for WBCL Sub-Scale 6 pooled sample.
Lenawee. Turrialba. Lansing (N-638).*

 

 

 

 

Principal Axes varimax
Items Loadings Loadings
(Sub-Scale 6)
I II III I II III
6.1 23 03 -01 16 13 11
6.2 23 36 -O6 O7 43 00
6.3 47 -17 -28 57 O7 08
6.4 54 —25 O7 44 -01 42
6.5 36 -09 19 18 06 38
6 6 34 16 15 O9 28 29
6.7 27 40 -O3 O6 47 04
% Total variance 54 24 9 34 29 24

 

*Decimal points omitted.

196

Table 1B7. Factor Matrix for WBCL Sub-Scale Total Scores
Pooled Sagple, Lenawee, Turrialba, Lansing

 

 

 

 

 

(N = 638)
Sub-Scale Principal Axes Loadings
I II III IV V VI
1 45 -O6 19 -20 10 01
2 51 -36 05 -05 -I3 02
3 -OO 37 28 O9 -O6 00
4 51 32 -12 -12 -O4 -05
5 47 27 -13 O4 01 O7
6 60 -14 q; 26 O6 -03
% Total
Variance - 62 22 7 6 A; O

 

J

“Decimal points omitted.

APPENDIX C

THE MSU WORK BELIEFS CHECK LIST AND SCORING KEY

198

YOUR NAME

 

THE MSU WORK BELIEFS CHECK—LIST

Ipstructions:

This check—list is made up of statements people often say
they believe. YOu will probably find that you agree with
some and disagree with others. If you agree with a statement.
circle Agraa; if you disagree with a statement. circle
Disagree. Do not omit any.

Be sure your name is on the top of this sheet.

199

The only purpose of working is to
make money.

I believe a man needs to work in
order to feel that he has a real
place in the world.

I feel sorry for people whose jobs
require that they take orders from
others.

Every man should have a job that
gives him a steady income.

The happiest men are those who
work only when they need money.

Doing a good job day in and day
out is one of the most satisfying
experiences a man can have.

A regular job is good for one.

I feel sorry for rich people who
never learn how good it is to have.
a steady job.

I don't like people who are always
right on time for every appointment
they have.

I feel sorry for people who have
to do the same thing every day at
the same time. ‘

I don't like to have to make
appointments.

I believe that promptness is a
virtue.

I usually schedule my activities.
I'd rather let things happen in

their own way rather than scheduling
them by a clock.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

200

It makes me feel bad to be late
for an appointment.

I expect people who have appoint-
ments with me to be right on time.

I would be unhappy living away from
my relatives.

I hope to move away from here
within the next few years.

People who can't leave their home-
towns are hard for me to under-
stand.

A man's first loyalty should be to
his home community.

When a boy becomes a man. he
should leave home.

I like to see new things and meet
new people.

I like to try new things.

On the whole. the old ways of
doing things are the best.

Life would be boring without
new experiences.

I like people who are willing to
change.

On the whole. most changes make
things worse.

The happiest people are those who
do things the way their parents
did.

New things are usually better than
old things.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

201

I believe that a person can get
anything he wants if he's willing
to work for it.

Man should not work too hard. for
his fortune is in the hands of God.

A man shouldn't work too hard
because it won't do him any good
unless luck is with him.

With a little luck I believe I can
do almost anything I really want to
do.

A person shouldn“t hope for much
in this life.

If a man can't better himself it's
his own fault.

.Practically everything I try to do

turns out well for me.

I usually fail when I try some-
thing important.

I would rather work than go to
school.

Money is made to spend. not to
save.

I think there's something wrong
with people who go to school for
years when they could be out
earning a living.

One gains more in the long run
if he studies than if he gets a
job.

The more school a person gets the
better off he is.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

202

Generally speaking. things one
works hard for are the best.

When I get a little extra money
I usually spend it.

Agree Disagree

Agree Disagree

2.1

203

SCORING KEY (Tentative) 1957-1960

MSU Work Beliefs Check-List

Underlined responses are scored one point; all
are scored zero points.
There is a score for each sub-area. six scores

The only purpose of working is to
make money. Agree

I believe a man needs to work in
order to feel that he has a real
place in the world. Agree

I feel sorry for people whose jobs
require that they take orders from
others. Agree

Every man should have a job that
gives him a steady income. Agree

The happiest men are those who work
only when they need money. Agree

Doing a good job day in and day out
is one of the most satisfying

experiences a man can have. Agree
A regular job is good for one. Agree

I feel sorry for rich people who
never learn how good it is to
have a steady job. Agree

I don't like people who are always
right on time for every appoint-
ment they have. Agree

I feel sorry for people who have to
do the same thing every day at the
same time. Agree

others

in all.

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

204

I don't like to have to make
appointments.

I believe that promptness is a
virtue.

I usually schedule my activities.
I'd rather let things happen in
their own way rather than

scheduling them by a clock.

It makes me feel bad to be late
for an appointment.

I expect people who have appoint-
ments with me to be right on time.

I would be unhappy living away
from my relatives.

I hope to move away from here
within the next few years.

People who can't leave their home-
towns are hard for me to under-

stand.

A man's first loyalty should be to
his home community.

When a boy becomes a man. he should
leave home.

I like to see new things and
meet new people.

I like to try new things.

On the whole. the old ways of doing
things are the best.

Life would be boring without new
experiences.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

5.1

5.2

5.4

5.5

5.7

5.8

205

I like people who are willing to
change.

On the whole.
things worse.

most changes make

The happiest people are those who
do things the way their parents
did.

New things are usually better than
old things.

I believe that a person can get
anything he wants if he's willing
to work for it.

Man should not work too hard. for
his fortune is in the hands of
God.

A man Shouldn't work too hard because
it won't do him any good unless
luck is with him.

With a little luck I believe I
can do almost anything I really
want to do.

A person shouldn't hope for much in
this life.

If a man can't better himself it's
his own fault.

Practically everything I try to do
turns out well for me.

I usually fail when I try some-
thing important.

I would rather work than go to
school.

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

206

Money is made to spend. not to save.

I think there's something wrong with
people who go to school for years
when they could be out earning a
living.

One gains more in the long run if
he studies than if he gets a job.

The more school a person gets the
better off he is.

Generally speaking. things one works
hard for are the best.

When I get a little extra money
I usually spend it.

Agree

Agree

Agree

Agree

Agree

Agree

Disagree

Disagree

Disagree

Disagree

Disagree

Disagree

Fifth-dis .. {vat-LG

R8011 USE UNLI.

 

 

 

'4-

"11111111111111111111“