University of Southern California Dissertations and Theses

Group Factors And Individual Internalization Of A Value

(USC Thesis Other)

Group Factors And Individual Internalization Of A Value

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation h as b een
m icrofilm ed exactly as receiv ed 6 7 — 1 3 ,7 6 0
ROBERTS, M artha A nne, 1 9 2 8 -
GROUP FACTORS AND INDIVIDUAL INTERNALIZATION
O F A V A LU E .
U n iv e r sity o f Southern C a lifo rn ia , P h .D ., 1967
S o cio lo g y , g e n e ra l
University Microfilms, Inc., Ann Arbor, Michigan
Copyright (c) by
MARTHA ANNE ROBERTS
1967
GROUP FACTORS AND INDIVIDUAL
INTERNALIZATION OF A VALUE
by
Martha Anne Roberts
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Sociology)
June 1967
UNIVERSITY O F S O U T H E R N C A LIFO R N IA
THE G RADUATE S C H O O L
UNIVERSITY PARK
LOS AN G ELE S. CA L IF OR N IA 9 0 0 0 7
This dissertation, written by
...Martha. Anjae. . .R a h e r J . s...............
tinder the direction of h ^ X . . ..Dissertation C o m
mittee, and approved by all its members, has
been presented to and accepted by the Graduate
School, in partial fulfillment of requirements
for the degree of
D O C T O R O F P H I L O S O P H Y
Dean
D ate June JL 1 . 9 . 6 . 7
I TAT I ON COM.VIITTKK
hrrt CeLre~l.ut.jDl --
Chairman
TABLE OF CONTENTS
Page
LIST OF TABLES iii
Chapter
I. INTRODUCTION .................................. 1
II. THEORETICAL BACKGROUND ....................... 20
The Importance of Consensus to
the Group
Interpersonal Aspects of
Internalization of Values
The Group and Internalizetion
III. THEORETICAL APPLICATION OF THE HYPOTHESES . . 67
IV. THE SAMPLE AND METHODS..........................79
V. RESEARCH RESULTS .............................. 121
Hypotheses I and II
Internalization Measurement
Hypotheses III through VI
Internalization and Other Variables
71. INTERPRETATION AND EVALUATION................. 154
Methodological Considerations
Confirmation of Isolated Hypotheses
General Theoretical Considerations
VII. SUMMARY AND CONCLUSIONS........................ 199
APPENDIXES...............................................209
BIBLIOGRAPHY 235
LIST OF TABLES
Table Page
1. The Four Samples by A g e ..................... 31
2. The Four Samples by Education.............. 83
3. The Four Samples by Sex ................. 84
4. The Four Samples by Occupation Level . . . 35
5. Sample Xr^ for Internalization by
Treatment Demands for Group Action . . . 124
6. Generalized Other Scores by Sample .... 136
7. Awareness of Group's View by Sample .... 138
3. Awareness of Roles by Sample............. 136
9. Internalization and Conformity Tau
Correlations ............................. 141
10. Status of the experimental Hypotheses for
nach Sample and the Final Conclusions . . 145
11. Socioeconomic Class for All Four Samples . . 140
12. Internalization by Age Groupings
All Four Samples............................150
13. Internalization by Sex All Four Samples . . 150
14. Internalization by education Levels
All Four Samples............................152
15. Consistency-Guilt College Sample ..... 222
16. Consistency-Guilt High School Sample . . . 223
17. Consistency-Guilt Adult Sample ...... 224
18. Consistency-Guilt Senior Citizen Sample . . 225
19. Consistency-Certainty College Sample .... 226
iii
Table Page
20. Consistency-Certainty High School Sample . . 227
21. Consistency-Certainty Adult Sample ..... 228
22. Consistency-Certainty Senior Citizen
Sample.......................................229
23. Certainty-Guilt College Sample . . . . . . . 230
24. Certainty-Guilt High School Sample .......... 231
25. Certainty-Guilt Adult Sample ................. 232
26. Certainty-Guilt Senior Citizen Sample . . . 233
27. Consistency by Samples ................234
28. Identification with Section by Samples . . . 234
iv
CHAPTER I
INTRODUCTION
One principle that is considered fundamental to the
explanation of group behavior is that coordinated group
action depends on some sort of consensus in the group.
Quite properly people have questioned how much consensus
and what kind of consensus is necessary for particular
types of coordinated activity, but generally it is agreed
that in all forms of activity there is a requirement that
there be some degree of consensus of some sort (coercive,
voluntary, disinterested, or whatever). Thus it becomes
necessary for the group which is to acheive action to
develop and sustain consensus within the group. Obviously
the concept of social control arises to explain the
influence of the group over the individuals in maintaining
consensus, and a great many studies have been undertaken
to explore and comprehend the conformity of individuals to
group values, frames of reference, and the like. Sherif
and Asch are among the pioneers in this field.
Most studies of the control or influence of the
group over the individual have emphasized the compliance or
conformity aspects which tend to disregard the concepts of
1
2
socialization and internalization. One of the greatest
potential controls that a group can exert over an indi
vidual is by socializing him into internal acceptance of
its norms, values, and frames of reference. Not only has
the element of internalization often been overlooked in
explaining social control, but when it is studied it is
usually considered in terms of individual rather than
group concepts. It is explained by imitation, identifi
cation, and personal reward and punishment concepts.
Seldom is internalization considered in terms of being
part of a group process.
At the same time, it is commonly assumed that
people acquire their values from the groups with which
they associate. They take on the viewpoints represented
by their membership groups--or at least their reference
groups. The child takes on the values of his peer group;
the adult takes on his work group's frames of reference,
his religious group's values, his leisure group's atti
tudes, etc. It is believed that the individual picks up
these factors to the point where he accepts them as his
own ideas and standards; that is, he has Internalized
them.
Thus explanations of social control and conformity
often overlook the processes of self control on the basis
of internalizing the group views; the explanations of
3
internalization usually omit any reference to groups as
playing a part in the internalization process, and still
it is believed that the Individual gains most of his
values and views from his membership and/or reference
groups, there appear to be blind spots in several theo
retical considerations which might be filled in and even
tied together if one could develop and prove a systematic
explanation of how the group is related to individual
internalization processes.
It is to fill this gap that the present study is
being proposed. More specifically, it is the intent of
this study to relate internalization to the need for con
sensus for group action and to explain internalization in
its aspects of being a group, as well as an individual,
process. Before stating the specific hypotheses which
identify the relationship between these concepts it is
first necessary to define the concepts.
Internalization is a term which, by definition,
relates the group and the individual although it does
not specify the conditions of this relationship. Stated
most simply, internalization means the acceptance of group
values, norms, or goals as one's own. When these elements
which characterize the group are accepted as one's own,
it follows that they will be deeply integrated into the
orientations and behavior of the individual. He will act
on the basis of these internalized group values and norms
even when he is not with the group and there is little
prospect of the group even being aware of his behavior.
Internalization, however, is not the result of his
acceptance of group qualities, but rather it is the
process of this acceptance. A comprehensive definition
of the term would include some explanation of this process,
but since that is the underlying content of the whole
dissertation simply "accepting values as one’s own" will
provide the basic definition used here. The remainder of
the dissertation will provide further embellishments.
Upon first thought the term "consensus" appears to
be on obvious concept. It means agreement. However, upon
further consideration it becomes somewhat elusive. What
does agreement mean? Certainly it needn't mean total
unanimity. Even if there is unanimity, is there consensus
when most of the group is holding mental reservations on
what they are agreeing to verbally? The lock step of the
chain gang implies total consensus of one sort but typi
fies a lack of consensus of another sort.
The purposes of this study do not permit the
behavioristic definition of consensus which would say the
chain gang has consensus because they are all doing the
same thing. One of the basic premises is that people are
able to unite their action because there is consensus.
5
Therefore, defining consensus in terms of coordinated
behavior would be proving the premise by definition.
On the other hand, it is not considered that a
strictly limited definition of consensus in which there is
agreement on private, unexpressed views will prove very
fruitful. A compromise position is considered desirable.
What one states as his views is usually the basis on which
he is prepared to act. Therefore, the level of consensus
of concern in this study is what one expresses rather than
what one internally holds to be true, and consensus is
defined as "the greatest sharing of perspectives within a
group."
The use of "perspectives" implies that inherent
qualities are being agreed upon. Thus situations would
be defined similarly, and presumably coordination of action
would be possible and probable.
By using the term "greatest sharing" in the defi
nition there is no numerical limitation to a condition of
consensus in the group. It might mean that everyone
agreed, or it might mean only two out of a whole grcup
agreed while all the others had differing views. In the
latter case one would assume consensus was so low for the
total group that coordinated effort on the basis of shared
perspectives would be difficult if not entirely unlikely.
Although the term "value" appears in the title of
this work it is not considered to be one of the basic
experimental variables. It is used more to specify and
locate the internalization process than to be "manipulated
or "held constant" or "measured" in the experiment itself.
Still, it is of importance in terms of the internalization
process, and therefore it should be defined. Unlike in
ternalization, "value" is not necessarily tied to groups
or society. It relates directly to the individual. For
purposes of this dissertation it will be defined as "quali
ties imputed to objects which would tend to direct the
individual’s goal-oriented behavior."
"Objects," as it is used here, implies anything
the person can refer to. Thus if one "values" his nation
highly he will direct his behavior in a positive way
toward his nation. If he "values" the freedom of speech
of his nation but not the foreign policy of his nation,
this too will direct his behavior. Values, then, can be
applied to all classes of objects, single units or whole
categories. In each case the value held by the individual
will be related to the goal orientations behind the be
havior of the individual.
Other definitions critical to the study will
follow in the discussion of the experiment itself. How
ever, there is one other basic term that needs to be
clarified before the hypotheses are stated and the experi
ment is explained. This is the concept of the group.
7
Rather than defining the group in terms of specific
structure or membership, it is considered more desirable
to apply the term to "any number of individuals who are
capable of carrying out coordinated action." This implies
not only dynamic overtones but the added element that the
group is more than the sum of its parts.
Individuals are not automatically capable of
joining with any other individual in coordinated action.
A relationship between them is necessary; some sharing of
views through interaction is usually necessary; rules of
behavior are often required. The condition of being
capable of carrying out coordinated behavior presupposes
the presence of factors which are typical of groups, and
it also implies that interaction will be (or has been)
present.
For present purposes the group is not defined
broadly in terms of interaction only, nor is it limited
to an awareness of membership, but rather it is related
to conditions among certain individuals where sufficient
"group factors" are present so that the individuals are
able to carry out coordinated action. Some of these
"factors" that occur with the development and presence of
groups will be defined and discussed in the explanation of
the experimental procedures of this study. First the
basic hypotheses should be identified.
8
To specify the relationship between internalization
and demands for action and consensus in the group, the
following hypotheses have been devised:
I. As situations call for increasing demands
for group action (and consequently increasing
need8 for consensus) concerning a value,
there will be increasing consensus on the
values adhered to by the individuals in the
group.
II, As situations call for increasing demands
for group action (and consequently increasing
group pressures toward consensus) concerning
a value, there will be greater strengthening
of the individual's internalization of a
value.
These hypotheses are expected to relate demands
for group action to need for and evidence of increasing
consensus within the group. Consensus in turn is expected
to bring into play group factors which will further the
internalization of the value by the individuals in the
group. If the above hypotheses are found to be true, then
it would seem possible and probable that the group per se
will play some role in the process of internalization.
It then becomes necessary to explain the nature of the
factors within the group or the group situation which
would play a part in the internalization process. To
9
explore the kinds of group factors which might be Involved,
the following hypotheses have been devised:
III. Individuals who are identified as having a
group role will display (a) more internali
zation, and (b) greater conformity with the
consensus value.
IV. As viewing the group as a generalized other
increases, there will be (a) more internali
zation, and (b) greater conformity with the
consensus value.
V. As identification with the group increases,
there will be (a) more internalization, aryl
(b) greater conformity with the consensus
value.
VI. Identification with the group and viewing
the group as a generalized other will
increase as there is an increase in demand
for group action.
Although there are probably other group factors
which also might be important to the internalization
process, the scope of this study will be limited to con
sidering: having a group role, viewing the group as a
generalized other, and identification with the group.
Hypotheses III through V relate each of these group
factors to internalization and conformity for the indi
vidual. Hypothesis VI then rechecks the assumption that
10
these group factors will be present more often when there
is a demand for group action. This last hypothesis rounds
out relationships among the group factors and provides
another way to test the relationship between demands for
group action and internalization within the individual.
Since these hypotheses involve group processes,
it was felt necessary to use groups which were already in
existence and in which the members were accustomed to
interaction and to being part of the group. Further, it
was determined to try to obtain groups which represented
a rather broad sampling of groups since groups vary
grtsatly on the basis of membership characteristics, envi
ronmental setting and the like. On the other hand
obtaining groups that were widely divergent might prohibit
the assumption that the experimental variables were the
basis for any differences which might be observed.
By way of a compromise four different types of
groups were sampled. High school social clubs comprised
one sample composed of 5 clubs totaling 95 participants.
College sociology classes were not ideal examples of on
going groups that had built up interaction patterns, but
their easy accessibility and apparent meeting of the
minimum requirements for "groups" provided sufficient
cause for using them as a second kind of sample. Eight
classes totaling 210 students were used. For a sample
11
of adult groups 3 fellowship and service groups from
group-work agencies were used. These groups included 59
participants. Two leisure time activities clubs of senior
citizens provided a sample of 41 subjects from an older
age category. Four hundred five participants were included
in all four samples.
In order to determine if and how group factors
would influence internalization it became necessary to
devise some technique which would provide for a kind of
continuum of "groupness. • ’ That is, if groups were thought
to be important there would have to be some way of meas
uring internalization with and without the conditions of
group Influences. Therefore an experimental situation was
demanded, one which could be varied In such a way that
some participants could be said to be not exposed to group
influences in any appreciable amount, others who would be
exposed to them, and others who would experience a still
greater potential for influence from group factors.
Each on-going group was divided into three types
of smaller sections: the first required no interaction
among the participants, the second a certain amount of
discussion, and the third some discussion and also the
creation of a group product. These experimental differ
ences thus provided differing degrees of demand for group
action.
12
To be able to make comparisons among the sec **ions
each section was asked to consider the same question which
involved a value choice. All participants were asked
what they considered to be most important in getting ahead
today, and there were six possible qualities from which to
choose. After choosing the quality a letter to an imagi
nary newspaper about getting ahead was written to provide
for some reinforcement of the choice and the reasoning
behind it. All of this was done individually for the
first type of section. In the second section the parti
cipants discussed the possible qualities for getting
ahead, then the final choice and the letter were on an
individual basis. For the third type of section the par
ticipants were asked to discuss the qualities, write down
theI>r individual choices of a value and then as a xroup
produce a letter to a newspaper.
The final aspect of the experiment was the com
pletion of a questionnaire that provided the researcher
with some measurement of the variables considered impor
tant for the testing of the hypotheses. The basic, and
the most difficult, measurement was that of internali
zation. A detailed explanation of the theoretical bases
and the particular techniques will follow later in the
paper.
Internalization was generally defined as the
process of accepting values as one's own. Since it was
13
felt that if one had accepted the value of a quality as
one's own the individual would consistently regard that
quality as higher than others, would be fairly certain of
its importance, and would feel guilty if he acted in ways
which would contradict his belief in the importance of
the value. Consistency, certainty, and guilt were meas
ured in the questionnaire as indicators of internalization
of a value.
In addition to the above theoretical and opera
tional approach, two other theoretical views on interna
lization were attempted in operational measurement.
Techniques were devised which were designed to measure
when a person stopped taking the role of others and also
when a person stopped being influenced by others views
and began to control his behavior on the basis of his own
value judgments. Each of these was considered as a
possible measurement of internalization.
Nominally consensus was considered as a sharing
of perspectives. Operationally it was defined as the
number choosing the most shared (modal) choice of a
quality within a section plus an added dimension of
purity of this choice. That is, the consensus score was
based on how many chose the quality which was the most
chosen (whether it be chosen by only two participants or
by all six participants in the section) and whether this
14
choice stood out in value for the participants or merely
held a slight lead over another or several other choices.
The conformity of each participant was considered
not just in terms of whether the person went along with
the modal choice but also whether this modal choice for
his section had received a high or a low degree of con
sensus. Since it could not be assumed that any of the
sections would necessarily achieve a high degree of con
sensus (based on the experimental hypotheses) it was
considered that a person who went along with a choice that
held low consensus in the group was not displaying as much
conformity as one who went along with a high consensus
choice. Also, an individual who did not go along with a
choice which held low consensus was not displaying as much
lack of conformity as one who had not chosen a quality
which had received high consensus in his Beetion.
Thus conformity was not defined merely in terms of
the individual but also in terms of the consensus condi
tions of the section or group in which he was participating.
Conformity was not taken as an indication of possible
internalization but merely the amount of alignment of the
individual with the most shared group view in terms of the
amount and purity of sharing this view in the group.
The hypotheses of this study assume that the
demand for group action will relate to the amount of con
sensus in the group and the internalization within
15
individuals in the group. They further assume that
specific factors of the group will relate to the confor
mity and internalization of the individual participants.
Again, the specific group factors taken into account are:
having an identified group role, viewing the group as a
generalized other, and identification with the group.
Since ascribed roles can be either developed by
the particular group or defined by factors outside the
group such as those set up by the society as a whole, it
was decided that roles need not be assigned by the selec
tion of the group of the person who was to fill the role.
Arbitrarily assigning a person to a role would be the
best way to discover if playing an identifiable group role
would affect internalization without having to consider
possible influences of qualities of individuals who are
group-chosen or self-chosen to fill group roles.
Since "role" was considered to be a set of nor
mative expectations attached to a specific social position,
the operationalizing of the assignment of individuals to
an identified group role was merely to announce that a
person in each section who had drawn a particular number
would act as group coordinator. The group coordinator
would help the group carry out the experiment and transfer
materials to and from the experimenter as was necessary.
Viewing the group as a generalized other was
defined as consisting of at least two components:
16
(1) perceiving a "group" viewpoint rather than only being
aware of the views of particular others, and (2) perceiv
ing the group as a unity composed of interrelated parts
among which the individual would identify himself. Test
ing to see if an individual is aware of a group viewpoint
is a rather common practice in social psychological experi
ments. However, there have been few, if any, attempts to
test whether an individual sees the group as an inter
related structural unity. For this experiment the concept
of group-participation-roles was used to determine if the
subject was aware of these roles within the section in
which he participated, if he could identify people who
were playing such roles, and if he could fit himself into
one of the participation roles.
The final group factor specified in the hypotheses
was that of identification. For purposes of this study
identification was not considered in all its deep psycho
logical implications but merely as the pleasurable associ
ation by the individual with the group and the members in
it. Since the experiment divided each natural group into
three or more sections it was necessary to specify both
identification with the smaller section of the experiment
and also with the larger on-going group.
Operationally "identification with the section"
was scored on the basis of pride in the section,
17
enjoyment in being one of its members, and desire to
associate with the same section members if the experiment
were re-done. This was the basic type of identification
called for by the hypotheses, but for some purposes it
was useful to test identification with the larger group
in addition. This was done using two separate variables:
how well the participant liked the larger group and how
long he had been in the group.
The basic purpose of this study is to attempt an
empirical evaluation of the theoretical position that
group action demands consensus which will be associated
with group factors influencing the internalization process
in the individual. Chapter two will develop the theore
tical considerations leading up to the formation of the
hypotheses, and chapter three will place the hypotheses
into their potential for broader theoretical development
and for practical application. Then in chapter four the
specific techniques, methods, and sample characteristics
are explained.
At the point of explanation of the techniques a
second purpose of this study becomes apparent. There is
a lack of adequate techniques to measure the kinds of
concepts and processes involved in this type of theore
tical consideration. The virtual nonexistence of a
measurement for internalization seriously hampered this
18
study and probably is responsible for a good part of the
general lack of clear-cut, tested theoretical development
in regard to the internalization processes. Techniques
for measuring other concepts basic to the experimental
hypotheses were also lacking. Therefore, necessity
forced the addition of a second purpose of this study:
to develop and test techniques to measure concepts such
as "internalization*1 and "the generalized other."
In chapter five which explains the results of
the study and even more in chapter six which discusses
the evaluation and implications of the results it will
become clear that the second purpose of developing
measurements is closely entwined with the primary purpose
of testing the hypotheses. In fact the interrelationship
is so close that it inevitably results in the recognition
that only in so far as the second (measurement) purpose
has succeeded can the first (hypothesis testing) be
properly evaluated.
Difficulties arising from lack of previous de
velopment of methods and/or theory is not an unusual
problem in the social science fields and particularly in
the area of social psychology. It becomes easy to blame
the lack of theoretical development on the lack of ade
quate techniques to test theories, and likewise it is easy
to blame the lack of development of techniques on the
19
fuzzy condition of the theoretical concepts and hypotheses.
However, placing the blame Is a tangent to the problem and
will not lead to progress In either theory or methodology.
What is needed are aggressive attempts to struggle with
and conquer both the theoretical issues and the methodo
logical barriers. This study tries to come to grips with
both. It is not expected that this research attempt will
in any measure "conquer" these areas which need develop
ment, but it may add additional force to the beginning
struggle.
CHAPTER IX
THEORETICAL BACKGROUND
The purpose of this research is to clarify and
emphasize the relationship between consensus in the group
and the internalization of a value by the individual. The
theoretical background for the experiment will be dis
cussed in three sections: (1) the importance of consensus
to the group, (2) interpersonal aspects of internalization,
and (3) the group and internalization. The coverage of
these three topics should lead to the position taken in
this dissertation that the need for consensus in the group
increases the strength of interpersonal pressures which
lead toward internalization and also adds group pressure
which is not present in situations where group consensus
is not present or necessary.
The Importance of Consensus
to the Group
Although there is widespread agreement that some
sort of consensus is a prerequisite for group existence
and action, probably one of the most clearcut statements
of this relationship is made by Shibutani. He asserts,
"The extent to which independently motivated men are able
20
21
to coordinate their respective activities depends upon the
degree of consensus that exists among them. Consensus
refers to some kind of mutual understanding, a sharing of
perspectives.The emphatic first statement that coordi
nation of activies depends on consensus is somewhat
softened by the need to define consensus as "some kind"
of mutual understanding which leaves the concept in rather
hazy surroundings. In addition, the limiting of consensus
to "mutual understandings" or "shared perspectives" might
preclude the necessity of consensus on values or norms
even though it does cover some kind of common frame of
reference.
Generally this statement by Shibutani leads back
to Mead *s emphasis of the joining of the lines of men’s
activities through development of common meanings. Under
this approach it is considered that meaningful interaction
occurs when objects become common to the participants and
can be translated into significant symbols. Thus human
"meaningful" interaction would depend on some sort of
consensus about meanings or on some common frame of refer
ence about objects.
When considering group action and consensus it
1Tamotsu Shibutani, Personality and Society
(Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1961),
p. 40.
22
soon becomes obvious that within acting groups consensus
is not always present. There are many areas of disagree
ment within groups, and certainly when there is even some
sort of agreement by some members of the group it seldom
approaches unanimity on any subject. Newcomb recognizes
this problem although he agrees in general with the
necessary relationship between consensus and group
activity. He points out,
People who are members of the same group are espe
cially likely to acquire similar frames of reference,
because it is particularly important to group members
to be able to communicate about things of common
interest. . . . And yet, as everyone knows, members of
the same group often disagree2about things about which
they are able to communicate.
In his explanation of this lack of agreement he speaks of
membership in other groups which also have the same item
as common object but have different attitudes, and also
he speaks of personal interests opposing the group
interest.
It may be deduced that there is a difference in
consensus needs depending on the importance of what the
group is agreeing on, such as: a common definition of the
object, the "proper'1 attitude of the individual toward the
object, or the value of the object to the individual.
However, this still does not nullify the concept that
2
Theodore M. Newcomb, Social Psychology (New
York: The Dryden Press, 1950), p. 221.
23
there must be some consensus before group action is
possible since apparently there must be consensus on de
finitions of objects before communication can take place
(even though the communication indicates disagreement on
attitudes or values). It also would logically follow that
the group might not be able to take action beyond the
discussion stage without some consensus of attitudes and
values.
Perhaps Etzioni can add another dimension to this
discussion of the importance of consensus to the group
when he considers different types of groups and various
types of , , consensus-8pheres,,. These consensus-spheres he
divides into: general values, means, participation,
performance obligations, and cognitive perspectives. In
regard to cognitive perspectives, which seem to correlate
best with the areas of consensus implied by Shibutani and
Newcomb, Etzioni states: ". . . since this sphere has
more empirical anchorage than the others, it may be assumed
that consensus will be relatively higher here than in the
3
other five spheres."
In terms of consensus in different types of groups
Etzioni points out that "normative organizations require
3
Amitai Etzioni, A Comparative Analysis of
Complex Organizations (New York: The Free Press of
Glencoe, Inc., 1961), p. 130.
24
both a high degree and a wide range of consensus. . . .
Utilitarian organizations require high consensus mainly
with respect to performance obligations, participation,
and cognitive perspectives. . . . Coercive organizations
are so structured that they can operate in the face of
4
widespread or even total dissensus. ..."
In the light of this organization-type and
consensus-sphere perspective it might seem necessary to
qualify the broad general statement that consensus is
necessary for group action. As Newcomb pointed out it is
possible for cooperative groups to lack consensus in some
spheres and still maintain some coordinated action--even
if it is only discussion. As Etzioni says it is possible
in a coercive organization or situation for action to be
carried out even if there is no real underlying consensus.
Even in the coercive situation, however, it would seem
necessary to have overt consensus although there might be
no covert consensus among the participants. In fact, in
so far as the coercive situation can actually function
adequately there would be a greater necessity for almost
unanimous overt consensus.
The original statement that consensus is necessary
for group action would often need to be made explicit in
terms of what kind and how much consensus and in what
4Ibid.. p. 136
25
areas of group perspectives and interaction according to
the particulars of the situation. However, it is still
maintained that in any situation some degree of some kind
of consensus in regard to some group perspectives or inter
action would be required for the group to interact and
take coordinated action effectively.
In most cases of group action there is such a
prevailing consensus that one seldom has occasion even to
become consciously aware of the general agreement in these
areas. As Sherif has pointed out: "In social life we
find not only values (positive and negative) ready-made
for us, but also more or less well-defined classifications
of objects and persons which permit us to apply the values
without hesitation.
The great number and great strength of these
ready-made frames of reference and values were certainly
emphasized by Durkheim in his discussions on collective
re pre s entat ion s•
If men did not agree upon these essential ideas at
every moment, if they did not have the same concep
tion of time, space, cause, number, etc., all con
tact between their minds would be impossible, and
with that, all life together.6
5Muzafer Sherif, The Psychology of Social Norms
(New York: Harper and Brothers Publishers, 1936), p . 139.
^Emile Durkheim, The Elementary Forms of Religious
Life (New York: The MacMillan Company,1915), p . 17.
26
The importance of existing culture is often noted
by those who study the socialization process. Rommetveit
has commented:
We shall approach the socialization process from the
assumption that any human being is born into an
already existing social structure, tied to the
carriers of cultural traditions with ties and pre
destined to social pressure from those carriers of
cultural traditions.7
The concept of society's ready-made frames of
reference also forms the backbone of many recent linguistic
studies such as those by Hoijer and Whorf which point up
the relationship between what and how the language ex
presses things and the frames of reference and perspective
used by individuals in the language group. Even though
the individual is seldom aware of how much these common
perspectives influence his thinking and behavior, they are
certainly no small influence on the individual and also on
the perspectives used in any particular group situation.
While recognizing the pervasiveness and importance
of these societally-common "representations," it should
also be noted that in each specific interaction situation
there normally is also consensus on attitudes and values
which are unique to the group or situation. That is, the
process of interaction in the particular situation
^Ragnar Rommetveit, Social Norms and Roles
(Minneapolis, Minn.: University of Minnesota Press,
1954), p. 4.
27
develops consensus on certain values and frames of refer
ence for areas which have not been previously defined by
the society as a whole.
We sometimes find this necessity for establishing
situational consensus even in areas where one might sup
pose there would be general societal consensus. For
example, Spaulding (as well as others) has pointed to the
lack of consensus in regard to social classes in America.
"That no clear and sharp consensus exists concerning the
class system to be applied to the American people has been
O
recently demonstrated by a number of studies." Thus,
while it is necessary to recognize the societal norms and
values as very important factors influencing interaction
it is also necessary to consider the situationally-
specific consensus areas and their influence in inter
action.
If one can postulate that consensus is necessary
for group action no matter whether the consensus is de
veloped as part of the current interaction process or
delivered to the individual as part of the cultural
heritage, then it would seem likely that this need for
consensus would affect the forces operating to influence
the individual to adopt a particular frame of reference,
^Charles B. Spaulding, "Social Class and Social
Perception," Sociology and Social Research. XL (September-
October, 1956), 20.
28
value, or course of action. In most discussions of how
an individual acquires a value the process is explained in
o
"interpersonal" rather than "group*' terms. In the next
section the interpersonal aspects of internalization will
be considered which will then pave the way to a final con
sideration of the relationship between the consensus needs
of the group and the individual internalization of a par
ticular value.
Interpersonal Aspects of
Internalization of Values
Before attempting to discuss the internalization
process the meaning of a few terms should be explained.
Although it is empirically somewhat difficult to separate
"frame of reference" from "value" and "norm" the concep
tual framework used here is intended to make such a dis
tinction. From this point on "frame of reference" will
be taken to mean the conceptual or organizational theme
from which one attaches meaning and value to objects. The
frame of reference is merely the organizational pattern
which is used and into which the values and norms are
fitted or attached. Perhaps Sherif has stated it more
o
In this paper "interpersonal" is used to distin
guish person to person relationships while a "group" has
added qualities including some degree of consensus, poten
tial for coordinated action, collective sanctions (such
as prizes or expulsion), and potential for being concep
tualized by the individual as some sort of unity.
29
clearly:
. . . in the course of the life history of the indi
vidual and as a consequence of his contact with the
social world around him, the social norms, customs,
values, etc., become interiorized in him. These
interiorized social norms enter as frames of refer
ence among other factors in situations to which they
are related, and thus dominate or modify the person's
experience and subsequent behavior in concrete
situations.
Thus the frame of reference is derived from the social
world and acts as the gateway for the interiorization of
social norms and values.
Social norms have been defined in various ways
according to the theoretical approach of the definer.
Rommetveit presents an interesting comparison of the
''norms" of Newcomb and those of Festinger. Newcomb differs
from Festinger's "behavior standards" by describing norms
as "shared frames of reference." Rommetveit points to
theoretical and practical problems with the use of either
of these definitions and concludes by making distinctions
between "norm senders" and "norm receivers" which fit more
practically into his theoretical and research approach.**
Since "norm" is not one of the primary concepts of the
present research, perhaps it will suffice to define a norm
as a standard of behavior which is perceived as shared by
others. This "shared" aspect of norms once again leads
Sherif, p. 43. **Rommetveit, pp. 19-22.
30
to an understanding of consensus and its importance in
explaining society and social interaction.
The concept of "value" seems to fall somewhere
between the individual*s frame of reference and what is
perceived as a social norm. It was noted earlier that
frames of reference provide the groundwork for norms and
values. Social values, like social norms, tend to be
acquired by the individual as he is led to using parti
cular frames of reference again and again. Since this
implies a social process it may be assumed that social
norms and values developed this way become consensus
objects in the society.
This does not mean that individual behavior stan
dards and values are non-existent, but it does assume that
in so far as individuals perceive behavior standards and
values as being shared or receiving some form of consensus
these behavior standards and values will become norms and
social values. The theme of this paper is aimed at ex
plaining social factors as well as the individual and
thus the values considered here will be viewed with the
"social" overtones although it is recognized that "indi-
vitual values" do exist.
More specifically, values will be regarded here
as qualities imputed to objects ("objects" including
intangibles as well as the more concrete objects) which
31
would tend to direct the individual's goal-oriented be
havior, Sargent and Williamson emphasize some aspects of
values and their relationship to other concepts which are
appropriate here:
Values are a type of norm and are closely related
to attitudes. At the same time they are kinds of
motives, since they represent orientation or striv
ing toward a goal. In the broadest sense, we may
think of values as attitude-related attributes that
are projected upon people, objects, and situations,
• • • Values have been defined as "desiderata, i.e.
anything desired or chosen by someone sometimes."
. . . Generally, however, we tend to think of values
as the more enduring clusters of wants that a given
individual or group works toward fairly consistently.
• . , values tend to be in the nature of general
life themes or goals, whereas attitudes are somewhat
more content-oriented.
In keeping with this "life theme" explanation the
values used in the empirical research reported on in sub
sequent chapters relate to the question: "What is the
most important quality for getting ahead today?" Presum
ably getting ahead could be classified as a general life
theme or personal goal. Various objects (e.g. knowing the
right people, having brains, etc.) would be related to the
goal and would have differing values depending upon how
each was seen in relationship to the goal.
The term interiorization is equivalent to the
term preferred in this paper: internalization. Interna
lization is often described as the product of a process
S. Stanfeld Sargent and Robert C. Williamson,
Social Psychology (3d ed.: New York: The Ronald Press
Company, 1966), p. 261.
rather than the process itself. For instance, it might be
described as: when the individual controls himself rather
than having others control him, or when role-taking stops.
However, it seems more profitable for use here to define
it in terms of the process rather than the product. Using
the concepts in the definitions above it would be the
process of increasing self control as opposed to external
control or the process of reducing role-taking in favor of
action on the basis of one's own values as one makes the
group's value8 his own. These are the basic definitions
of internalization as presented by Parsons and Turner.
Parsons gives one clue to locating the process of
internalization, or the product of when it has occurred,
when he says that internalization is "to act in conformity
with it [a behavior standard 3 becomes a need-disposition
in the actor's own personality structure, relatively in
dependently of any instrumentally significant consequenses
i 3
of that conformity." The association of internalization
with the occurrence of self control rather than external
control is a common assumption in various writings where
14
internalization is considered.
Turner's approach presents a different avenue to
13
Talcott Parsons, The Social System (Glencoe,
111.: The Free Press, 1951), p. 37.
l^For example: Shibutani, p. 198, and Daniel R.
Miller and Guy E. Swanson, The Changing American Parent
(New York: John Wiley and Sons, 1958), p.92.
33
the location of the internalization process or its results.
Turner says, "One form of this autonomy is indicated when
a person is said to have interiorized a social norm, mean
ing that an earlier process of role-taking has become
truncated.Although Turner gives a different way to
find or identify internalization the definition is not very
different from Parsons'. Both assume that as internaliza
tion occurs the individual will rely more and more on him
self and less and less on the influence of other indivi
duals or conditions of the social environment.
Because the purpose of this paper is to explore
what the internalization process is and how it takes place,
the definition for purposes of this research will be a
broad one, simply stating that internalization is the
process by which the societal or group frames of refer
ence, values, and norms are transferred to the individual
and become his own. The explanation of how this happens
will be the major task of this paper.
Before taking up this explanation of the process
it would be well to point out the great importance of
internalization. In fact it is difficult to overemphasize
its importance. Even though one must recognize that
physiology and individually unique experiences do play
^Ralph H. Turner, "Role-taking, Role Standpoint,
and Reference-group Behavior." American Journal of
Sociology. LI (January, 1956;, 326^
34
some part in the explanation of the behavior of the indi
vidual, an enormous amount of his behavior can and should
be explained in terms of the societal and group frames of
reference, norms and values which he has picked up from
the world about him. In many cases the surrounding values,
frames of reference, and attitudes are of greater conse
quence than the actual experiences through which the indi
vidual goes. "As Horowitz's material shows, the child's
attitude toward the Negro is the result not of contact with
Negroes but of contact with the prevailing attitude toward
,.16
Negroes."
Even the language we use subtly directs our think
ing and conceptual organization. "The central idea of the
Shapir-Whorf hypothesis is that language functions, not
simply as a device for reporting experience, but also, and
more significantly, as a way of defining experience for
its speakers."^ It Is difficult to estimate how great an
influence our social world has upon what we like to think
of as our own individual thought and behavior.
Perhaps the most obvious explanation of how inter
nalization occurs is in terms of actual pleasure-pain or
punishment-reward experiences. Cats may be viewed as
16Sherif, p. 140.
^Harry Hoijer, "The Sapir-Whorf Hypothesis," The
American Anthropological Assn.. LVI, No. 6 (December,
19545, 93.
35
having one type of value if one's experiences have been
with the soft fur and purring qualities of cats. They may
be viewed in a quite different value relationship if ex
periences have been of the scratching and spitting sort.
This kind of experience explanation may seem
obvious, but it is not completely accurate. First, one's
personal experiences are very limited and do not include
a vast array of objects which may be viewed with rather
well defined values. For example, very few people can
claim to have had personal experience wit' ghosts, but
almost everyone thinks of ghosts as having characteristics
to which positive or negative values are attached.
Second, a previously established value may so
define the experience that the result is more indicative
of the previous value than of the individual's experience.
For example, if a person is really convinced that the
golden rule has value, many rebuffs from the people he
tries to help will not persuade him to alter his values.
A somewhat more sophisticated explanation of the
internalization process is in terms of associations with
other meaningful objects. (Mead used the term object to
Include anything to which a meaning could be given and
this is the frame of reference for "meaningful objets"
used in this paper.) In explaining the internalization
of a social value Sherif elaborates on the "meaningful"
approach:
36
. . . a social value is first external to the indi
vidual, Being external to the individual, it has
for each individual who first confronts it, an ob
jective reality. The individual comes to perceive
this reality through a meaningful sentence, such as
"Stealing is a sin," or through a concrete example .g
of the behavior of parents, teachers, playmates, etc.
This kind of "meaningful" explanation of interna
lization seems to have a good deal of merit, but it still
leaves the problems of where one gets the meanings and
what it is about certain meanings and meaningful experi
ences that leads them to be internalized while others are
not so thoroughly adopted by the individual.
The next phase of the explanation of internaliza
tion seems to require the introduction of interaction
relationships between individuals. At several points,
Parsons explains internalization of the symbolic act in
terms of an interaction relationship between ego and alter.
Perhaps a condensed and simplified description of this
interaction relationship factor can be seen by using the
term "identification."
Parson himself seems to use "identification" to
summarize the ego and alter involvements in the interna
lization process. "Identification . . . means taking over,
i.e. internalizing, the values of the model. It implies
that ego and alter have established a reciprocal role
19
relationship in which value-patterns are shared." As
*®Sherif, p. 125
19
Parsons, p. 211.
37
will be noted again later in a discussion on conformity
and internalization, one of the basic ingredients for
going that extra step from conformity into internalization
is identification.
Although identification can certainly play a part
in getting a person to internalize a value it doesn't tell
the whole story either. Unless the person makes some kind
of transition into his own value system the values of
others will still be somewhat external to him. If one's
hero seems to place value on a certain breakfast food, as
long as (or whenever) one identifies with the hero one may
value the breakfast food, but he may not value it if some*
one else becomes the hero or if one doesn't recognize a
particular breakfast serving as the food chosen by the hero.
Freud seems to call attention to this limitation
of the identification process when he says:
It is easy to state in a formula the distinction
between an identification with the father and the
choice of the father as an object. In the first
case one's father is what one would like to be.
and in the second he is what one would like to
have. The distinction, that is, depends upon
whether the tie attaches to the subject or to the
object of the ego.^°
Carrying this a step further: identification with another
person doesn't necessarily mean wanting the ego-objects
or values of the person identified with. Wanting to be
20
Sigmund Freud, Group Psychology and Analysis
of the Ego (New York: Liveright Publishing Corporation,
1951), p. 62.
38
like someone may not imply that one must want to have
what they have or must value what they value. Thus, al
though identification can be seen as a potential contri
butor to the internalization process, it must be recog
nized as not always leading to internalization of the
values of the person identified with.
The term identification is generally conceived in
connection with the identification of one individual with
another individual, and this is the meaning appropriate
here. However, the additional concept of identification
with a group will be discussed later as an essential ele
ment in the theoretical base for the present empirical
research.
To complete the cycle in the explanations of the
internalization process one should again consider the
experience element. Now, however, experience is not
considered from the pleasure-pain emphasis but from the
factor of use of a frame of reference or value and how
this may increase internalization processes. Newcomb
contends that even though experience or use may not be a
sufficient cause for internalization it is a necessary one.
As applied to the child who is just acquiring frames
of reference, this means that he comes to use--not
merely to absorb--the frames of reference which he
finds the people around him using. He finds that
these frames of reference work: they provide common
understandings upon the basis of which he can com
municate with others. . . . And only when they have
been found useful in leading to motive satisfaction
39
21
are they absorbed, or 1 1 interiorized" ....
This explanation fits rather well into a society
or group based explanation of internalization since it is
usually the society or group that sets up the situation in
which an individual finds himself and also provides the
major determiner of what will prove useful to the indivi
dual. However, it should be pointed out that many author
ities do not necessarily agree that an attitude or value
has to be used to be internalized.
00 23
Maccoby and Whiting both insist that children
can learn roles as part of their early socialization
process which they will not have opportunity to use for
years (and perhaps never use) but still remain part of
their potential social repertoire of behavior. Thus
presumably the accompanying attitudes, norms, and values
can be internalized to a considerable degree even without
use. For the most part, however, use or practice or
experience would be favorable to internalization even if
it is not a necessary part of the process.
^Newcomb, p. 224.
22
Eleanor E. Maccoby, "Role-taking in Childhood
and Its Consequences for Social Learning," Child
Development. XXX (1959), 251.
^John W. M. Whiting, "Resource Mediation and
Learning by Identification," Personality Development in
Children, ed. Ira Iscoe and Harold W.Stevenson (Austin:
University of Texas Press, 1960), p. 122.
40
All of the explanations of the internalization
process mentioned above are considered to have some valid
ity. It is conceivable that each explanation could prop
erly and adequately describe some particular internaliza
tion of a specific thing in a specific situation. On the
other hand, it is also considered true that no one of these
explanations could be sufficient to describe all interna
lizations. Therefore, this study will accept a pluralistic
explanation theory of the internalization process, i.e. all
of the above explanations either singly or in combination
are needed to explain internalization and there may be
additional factors beyond these individual and interper
sonal relationship factors.
Before discussing some of these possible additional
factors it is first necessary to consider the problem of
conformity and internalization. Extending from the range
of deviance-conformity there is a narrower range between
conformity and internalization. When socially acceptable
standards and values are evident in any situation the
individual has a range of possible behaviors from complete
deviation from the acceptable standards and values to
complete conformity with them. However, the nature of
social control is such that even complete conformity with
what is considered socially acceptable may not indicate
that the individual has internalized these acceptable
standards and values. Thus conformity also implies a
41
possible range from compliance to external pressures to
internalization.
Parsons points to this kind of continuum in one
explanation of internalization:
There is a range of possible modes of orientation
in the motivational sense to a value*standard.
Perhaps the most important distinction is between
the attitude of "expediency" at one pole, where
conformity or non-conformity is a function of the
instrumental interests of the actor, and at the
other pole the "introjection" or "internalization"
of the standard . . . .
With this explanation of internalization as one
of the poles on the compliance continuum comes the neces
sity of considering internalization as something of which
one can have a greater or lesser degree rather than some
thing that one has or does not have. Rommetveit points up
this necessity of considering degrees of internalization
when he discusses the range of possibilities between anti
cipation of sanctions at one extreme and there being "no
2 S
norm-receiver at all any longer" at the other.
A great many studies have considered the problem
of conformity versus deviance or non-conformity. Since
Sherif's and Asch's experiments there have been all sorts
of variations on the measurement of conformity using many
different variables and several theoretical perspectives.
Although it should not be considered that these studies
necessarily explain the basic aspects of internalization,
^Parsons, p. 37. ^Rommetveit, p. 28
42
it still should be noted that moving in the direction of
conformity also implies the possibility of moving toward
internalization. As external control factors direct one
toward compliance they may also direct one toward personal
acceptance of the values or standards being complied with.
In line with this possible relationship between
conformity and internalization it might add to an under
standing of internalization to consider what factors have
been found to be important variables in explaining confor
mity. Glake and Mouton have provided a good summary of the
findings of these studies of such factors. Among the
psychological and physiological properties of the person
are listed: (1) prior experience with the task, (2) per
sonality differences and prior experience, especially in
volving relationships with dominating authority figures
and having a dominant life style, (3) physiological char
acteristics such as age, sex, sleep deprivation, food dep
rivation, and degree of anxiety, and (4) psychological
factors such as submissiveness, lack of self-confidence,
nervous tension, authoritarianism, degree of intelligence
and originality, amount of need of achievement and need for
2 6
social approval, etc.
2 6
Robert R. Glake and Jane Sygley Mouton, "Confor
mity, Resistance, and Conversion," Conformity and Deviation,
ed. Irwin A. Berg and Bernard M. Bass (New York: Harper
and Brothers, 1961), pp. 8-20.
43
Although these factors have proven important for
situations involving conformity, it was not considered that
they would be the prime variables in the internalization
process. Therefore, in the present research these factors,
with the exception of age, have hopefully been randomized
and would not then greatly influence the experimental re
sults.
Since the individual must conform before he could
be classified as having internalized the group's views, it
might seem that the individual who is most susceptible to
conformity pressures would be most apt to internalize the
things he conformed with. However, this is not necessarily
the case. As was implied earlier in the discussion on
internalization the person who internalizes a value goes
beyond mere use of the value and makes it a real part of
himself. Perhaps the differences between conformity and
internalization are best illustrated in terms of Reiman's
description of the three processes of social influence.
He differentiates between compliance, identification, and
internalization.
Compliance can be said to occur when an individual
accepts influence from another person or from a group
because he hopes to achieve a favorable reaction from
the other.
Identification can be said to occur when an individual
adopts behavior derived from another person or a group
because this behavior is associated with a satisfying
self-defining relationship to this person or group.
Finally, internalization can be said to occur when an
44
individual accepts influence because the induced be
havior is congruent with his value system,27
On the basis of this differentiation it may be concluded
that conformity may be on various levels and achieved by
various means. Internalization is just one aspect of
conformity.
In spite of being just one aspect of conformity,
internalization itself can not be thought of as a single
cohesive unit. Internalization also occurs in degrees
along a continuum, and it occurs under a variety of condi
tions.
For purposes of this paper and experiment then,
internalization will be thought to consist of the process
of personal acquisition of values under conditions of per
sonal experience, interpersonal relationships, or group
qualities which lead the individual to accept the values
as his own. It is now time to consider some of the poten
tial group factors in the process of internalization and
consider the relationship of the group's need for consensus
to internalization within the individual.
The Group and Internalization
Part one of this chapter discussed the importance
of consensus for group cooperation and action. Part two
was devoted to briefly exploring the process of
^Herbert C. Kelman, "Processes of Opinion Change,"
Public Opinion Quarterly. XXV (Spring, 1961), 62, 63, 65.
45
internalization in terms of the individual and interper
sonal relationships. Now the two areas need to be combined
into the basic theoretical framework for the present empir
ical study.
If group action depends upon consensus it would
seem likely that it would be of utmost importance to the
group to induce general agreement among its members. One
of the most efficient and long lasting methods of getting
this agreement would be for the members of the group to
internalize the same values, norms, and frames of reference.
By achieving internalization there would be little need for
continually directing efforts of control to maintain con
sensus, and the goal oriented activity of the group would
be given more complete and intensive attention by the group.
Thus it would appear that the group would have a rather
great stake in inducing internalized agreement among its
members.
In spite of the fact that internalization could be
a great asset to building consensus which is so vital to
the group, the process of internalization is very seldom
explained in terms of a group situation. The customary
explanations have been reviewed above and include consider
ation of factors of the individual and factors of person-
to-person relationships. With internalization being of
such potential importance to the group it would seem
strange, if not almost illogical, for groups to leave this
46
process entirely up to individual and interpersonal factors.
There are many who insist that a group is nothing
more than individuals and person-to-person relationships.
Perhaps this is true. However, it is the contention of
this author that there is more to a group than the sum of
the individuals and that the individual conceptualizes the
group as more than its individual members. Specifically,
some of these group aspects include: the use of the gener
alized other, identification with the group, the role re
lationship of the member within the group, and collective
sanctions of the group.
Piaget's study of moral judgments in children points
up two different stages of group-to-individual relation
ships which are important for internalization. At one
stage the moral judgments are based on heteronomous stand
ards where norms are seen in terms of absolutes which are
handed down from parents or superiors and are taken as in
violate. Later on the autonomous stage appears where
standards are seen as more flexibile and depending for
their authority on the mutual agreement of those affected
by them.28
In the heteronomous stage the play group as a whole
is of little concern and is far outshadowed by the superior
authorities. In the autonomous period it is the present
28Sherif, pp. 180-182
47
group which is of utmost importance. In the first stage
the group influence would occur when society was viewed as
a supreme authority to be followed without question or
change. In the second the group influence would be more
in terms of the ties of role relationships and the need for
develoing consensus within the group with which the indi
vidual could identify and which would further develop ap
propriate sanctions.
Margaret Mead has pointed out that entire cultures
may be similarly separated in their approaches to moral
standards. She emphasizes the greater potential for mass
movements in the "age-grade standard" cultures as opposed
2Q
to those using "parental standard" approaches. However,
in either of the two stages or approaches, group factors
would be important in the determination of the moral judg
ments of the individual--or in the determination of his
values.
As has been pointed out earlier, however, momen
tarily using a group's value or conforming to a group value
is not necessarily the same as internalizing the value.
In fact the culture or subculture may even emphasize con
formity without internalization.
jq
Margaret Mead, "Social Change and Cultural
Surrogates," Personality in Nature. Society, and Culture,
ed. Clyde Kluckholn and Henry A. Murray (New York:
Alfred A. Knopf, 1956), pp. 660-662.
48
Miller and Swanson found that there are differences
in child-rearing practices between entrepreneurial
and bureaucratic parents* The entrepreneurial
parent, being concerned with "getting ahead," used
techniques which are effective in establishing in
the child*8 internalized standards of behavior, and
in emphasizing an active and manipulative approach
to life. . . . This same drive toward individual
accomplishment is perceived as detrimental, however,
to the individual whose first concern is to get along
with his fellow workers. Accordingly, the bureau
cratic parent does not emphasize "internalization"
techniques in child-rearing behavior.3^
If this is true it would apparently be possible for whole
sections of a culture, or even an entire culture, to con
tinue without anyone internalizing anything except the
value of doing what is expedient to getting along.
This possibility of great conformity even when
there is no coercion in group behavior with little or no
internalization is probably more common than many people
realize. It poses a real problem for an experiment which
sets out to measure internalization, and it also restricts
the application of the findings of such an experiment.
Findings could not be generalized to situations where
there was conformity without internalization. To be sure,
as was mentioned earlier, there are elements in common
between conformity in general and specific internalization
processes, but there are dangers in assuming that what is
true in the internalization process is true for all
3®Robert J. Havighurst and Bernice L. Neugarten,
Society and Education (3d ed.: Boston: Allyn and Bacon,
Inc., 1962), p. 112.
49
conforming behavior or vice versa.
Even though conformity and internalization are not
identical an examination of conformity studies may reveal
the extent and type of group influences over the indivi
dual. Probably the cornerstone of conformity experiments
is Sherif's autokinetic study. In regard to group effects
on judgments about an apparently moving light he concludes
that individuals with previously established personal
norms tend to merge their norms when put in a group situa
tion.
But the convergence is not so close as when they first
work in the group situation, having less opportunity
to set up stable individual norms. . . . If, for the
group there is a rise or fall in the norms established
in successive sessions, it is a group effect; the
norms of the individual members rise and fall toward
a common norm in each session.31
Since a major interest of this work is to uncover
the nature of these group influences a further quotation
from Sherif might give greater understanding in this area
also. Here he is giving his findings from questioning the
subjects in regard to their adjusting to the group norms:
This subject wrote, ". . .My judgment in each case
was already made, and 1 did not change to whatever
the other person said. But on subsequent observa
tions my judgments were adjusted to their judgments
. . . ." Despite the above case, every individual
need not be aware of the fact that he is being in
fluenced in the group situation, or that he and the_
other members are converging toward a common norm. z
It would seem necessary then, in measuring and
3* Sherif, p. 104, 32Ibid.. p. 108.
50
uncovering group influences, that subconscious as well as
conscious factors would have to be considered. Examination
of subconscious factors is exceedingly difficult and
results are often highly suspect. The author recognizes
these problems but feels that it is necessary to /aake some
attempt to uncover unconscious influences if group influ
ences are to be properly accounted for.
In Sherif1s experiments consensus was a dependent,
rather than an independent, variable. It appeared that
consensus occurred very quickly and very effectively in
situations where the group as a whole was confronted with
the autokinetic effect first. Even when individuals had
established their own norms first, however, it appeared
that consensus tended to develop in the group situation.
These results might lead one to suspect that the tendency
toward consensus in a group was very strong.
Asch's experiments used the amount of consensus as
one of the independent variables and came up with some
different kinds of findings. Where consensus (artificially
produced) was unanimous in contrast to the subject’s
judgment about the discrepancy between lines, there appeared
to be strong pressures for the individual to conform to
the group judgment even though it was obviously in error.
However, when the consensus was even slightly reduced by
various means, giving the subject's views some support,
51
the group pressures seemed to be much less effective.^3
The results of these experiments then would lead one to
discount the Importance of consensus and group pressures
unless there was complete unanimity within the group.
This author Is Inclined to explain these seeming
differences In the Sherif and Asch experiments, not so
much In absolute terms regarding the Importance of consen
sus, but In terms of the consensus factors of the different
situations. On the surface both kinds of experiments
seem to be reality-testing situations, but in Sherif's
experiments the judgment has to be made about a highly
illusive object and in Asch's experiments the judgment
was made In comparison to a group judgment that in some
cases was obviously wrong. For Sherif's subjects the group
judgment must have appeared as a more certain reality;
while for Asch's subjects the group judgment stood in
contrast to apparent reality. Therefore, it would appear
that the group factors are of importance in reality testing
situations, but the unstructured nature of the object
being tested is of great consequence in what the strength
and results of group pressures will be.
Other studies have indicated that group pressures
* 3 0
S. E. Asch, "Effects of Group Pressure upon the
Modification and Distortion of Judgments," Readings in
Social Psychology, ed, E. E. Maccoby, T. M. Newcomb, and
eT Li Hartley (ed ed.: New York: Holt, Rinehart and
Winston, Inc., 1958), pp. 179-182.
52
exist and are very influential whether or not the experi
ment involves an ambiguous situation* In testing various
kinds of group conditions in terms of their ability to
induce behavioral results Pelz concludes that "the process
of making a decision and the degree to which group consen
sus is obtained and perceived" are two factors which alone
could produce as much influence over the individual as was
found in earlier experiments by Lewin*s co-workers* The
earlier experiments had compared group discussions to
hearing lectures by authorities to determine which had more
long-lasting influence over the individual. Pelz' research
indicates decision making and group consensus were two
factors greatly responsible for the greater influence of
the group over the individual.^
Katz and Lazarsfeld emphasize a variety of factors
and processes in explaining the importance of group influ
ence over individuals. In a summary of their arguments
they point to such things as: the satisfactions of accept
ance and achievement of status that come with conformity,
the dependency of the individual on the group for defini
tions of reality and standards of behavior, the importance
of interaction in developing shared standards and behavior,
the selective factors which may lead to similarity of
34
Edith Bennet Pelz, "Some Factors in 'Group
Decision'," Readings In Social Psychology* ed. E. E.
Maccoby, T. M. Newcomb, and E. L. Hartley (New York: Holt
Rinehart and Winston, 1958), p. 218.
53
values prior to interaction, and reasons why groups insist
3 * 5
on uniformity. ^
With all the attention directed by these and
other authors toward explaining group influences on the
individual it is most remarkable that there has been no
effort to explain group influences on the internalization
process when it apparently is such a potential in the
essential consensus making process.
Although Parsons recognizes the social importance
of individual internalization of values, the closest he
comes to relating the internalization process to the
social system is to compare it to institutionalization:
"Internalization of values in the personality is thus the
direct counterpart of their institutionalization in the
social system.It is the contention of this author
that not only are the processes similar, but internaliza
tion is a part of institutionalization and institution
alization is a part of internalization. That is, the
group processes and factors which lead to institution
alization of social norms, values, and behavior patterns
also lead to internalization of norms and values by indi
viduals.
^^Elihu Katz and Paul F. Lazarsfeld, Personal
Influence (Glencoe, 111.: The Free Press, 1955), p. 63.
•^Talcott Parsons and Edward A. Shils (ed.),
Toward a General Thaorv of Action (New York: Harper and
Row, Publishers, 1951), p . 240.
54
Based on the assumptions that there are group
pressures on the individual, that group pressures are
greater when the need for consensus for group action is
greater, that group action demands consensus, and that
group pressures may at least reinforce, if not entirely
account for, internalization processes the two basic
hypotheses of the present study are:
I. As situations call for increasing
demands for group action (and conse
quently increasing needs for consensus)
concerning a value there will be in
creasing consensus on the values ad
hered to by the individuals in the group.
II. As situations call for increasing
demands for group action (and conse
quently increasing group pressures
toward consensus) concerning a value
there will be greater strengthening
of the individual's internalization of
a value.
Perhaps one of the most obvious increases in
pressures that comes when the group achieves a large meas
ure of consensus is merely the addition of the number of
congruent interpersonal relationships. Backman, Secord,
and Pierce have stated and tested this aspect of consensus:
55
Interpersonal congruence theory • • • . suggests
that the greater the number of significant other
persons who are perceived to define an aspect of
self congruently, the greater the resistance to
change. The hypothesis was tested by choosing,
for each individual, a self-ascribed trait that he
believed five significant other persons generally
attributed to him and a self-ascribed trait that
he believed they did not. Strong pressure exerted
toward changing these traits by means of a false
personality assessment resulted in a greater change
in the low consensus trait.3'
Obviously this can not be considered as a group effect
per se. but its occurrence is much more likely where a
number of people agree in an effort to coordinate group
action.
Another pressure, that does come directly from
the group situation, is involved with the group roles for
each member. Homans has pointed out that the leader is
O Q
the member who most lives up to the norms of the group.JO
Later Homans modifies this statement somewhat:
In The Human Group I said that a man of high status
would conform to a high degree to all the norms of
his group, but this was certainly an overstatement.
. . . To keep his high status a man must provide rare
and valuable services to others, but so long as he
does that, the other members may allow him some leeway
in lesser things. He may even take the l e e w a y .39
3?Carl Backman, Paul F. Secord and Jerry R. Pierce,
"Resistance to Change in the Self-Concept as a Function of
Consensus Among Significant Others," Sociometrv. XXVI
(March, 1963), 102.
3®George C. Homans, The Human Group (New York:
Harcourt, Brace & c World, Inc” 19 50), p. 147.
3^George C. Homans, Social Behavior: Its
Elementary Forms (New York: Harcourt, Brace & World, Inc.,
1961), p. 339.
56
However, in spite of this qualification of his earlier
statement Homans still follows the view that the leader
will conform quite closely to the most valued norms of the
group.
Harvey and Consalvi have conducted an experiment
which tested the group pressures on the leader, the second
ranking member, and the lowest status member of informal
groups. When the critical subjects were unknowingly shown
a different set of light flashes than the rest of the
group it was found that the second ranking member con
formed much more than the lowest member or the leader,
and the leader conformed slightly less than the lowest
member.^
Although the leader might be allowed to deviate
more than others in minor matters it would appear logical
that the leader and any member playing an important group
role would need to conform rather highly to important
group values if the group were to function efficiently.
Assuming the leader does conform to most group
values it is possible to question the direction of in
fluence and ask if it is not the leader who is influencing
the group rather than vice versa. Harvey and Consalvi
also considered this problem and found the higher the
^ 0. Harvey and Conrad Consalvi, "Status and
Conformity to Pressures in Informal Groups," Journal of
Abnormal and Social Psychology. LX (March, 1966),185,
rank of the source the more influence on the group judg
ment, but the difference was significant only between
leader and lowest source subjects.^
Sherif also considered this possibility and found
. • • in our experiments the leaders were constantly
observed to be influenced by their followers--if not
at the moment, then later in the series and in
subsequent series, • • . Even if the group norm
gravitates toward a dominating person, the leader
represents a polarization in the situation, having
a definite relationship toward others which he
cannot change at will. If the leader changes his
norm after the group norm is settled he may cease
thereupon to be followed, as occurred several times
strikingly in our experiments.^2
It would seem then that there is some foundation for the
theory that the group influences the leader more than the
leader influences the group. At least it is clear that
the leader would be very susceptible to group pressures.
The previously cited references consider con
formity of the leader rather than internalization, but
it seems unlikely that a member would remain in a group
long enough to be accepted as a leader and to achieve the
role of leader without wishing to be identified with the
group and make its definitions and standards his own.
Conformity without some internalization by the leader
seems highly unlikely.
Under the assumption that those who play leader
ship roles will conform to group standards and accept
A1ibid., p. 186. Sherif, p. 105
58
group values and that they will be strongly influenced by
the group merely because they are playing a leadership
role, it can be stated that playing a group role would be
one of the group factors which influence internalization.
This, then, becomes one of the minor supporting hypotheses
which help explain why group factors would be expected to
influence internalization. For purposes of this study the
role hypothesis may be stated:
III. Individuals who are identified as
having a group role will display (a)
more internalization and (b) greater
conformity with the consensus value.
Consideration of the individual's view of group
pressure leads to another major group factor explored In
the present experiment: the awareness of a group as a
unity. Perhaps an awareness could be called an individual
rather than a group aspect of internalization, but perhaps
also the group only exists in the eye of the beholder.
That is, certain aspects of group pressures may operate
only when the individuals define the collectivity as a
"group" rather than a collection of individuals.
The family is considered a group by sociologists,
but it may not always be so considered by the members
within it. In an experiment by Mussen and Distler child
ren were presented doll play situations and expected to
complete the situation involving a mother, a father, and
59
a child. The scores were recorded as to whether the mother,
the father, or "they" were perceived as nurturant, punitive,
or powerful. Although the purpose of the study precluded
statistical analysis of the differences in the "they"
scores and the scores for each parent, the results indicate
a very obvious drop in the "they" scores over either
parent score,^
These results might give a clue that parents are
perceived primarily as individuals and not in combination
as a group. Further study might give valuable insights in
this regard. As far as this author knows no studies of
this nature have been made. The point to be made here,
however, is that certain group factors may not operate to
influence individual behavior unless the individual per
ceives that they are a group or a "generalized other."
The use of and pressure of the "generalized other"
in a group situation depends on the individual's perception
of the others in more general terms than as other indivi
duals. In this instance once more the conception of
consensus is vital. Here the individual perceives others
(even society) as having an established value with which
he aligns his own value.
Mead's term "the generalized other" includes this
^Paul Mussen and Luther Distler, "Masculinity,
Identification, and Father-Son Relationships," Journal
of Abnormal and Social Psychology. LIX (November^ 19$9),
_n
60
potential for group influence over the individual. Other
authors have included this idea under the term "reference
group." Shibutani's definition of a reference group de
scribes the group factor that is being sought here: "The
concept of 'reference group' . . . signifies that group
whose presumed perspective is used by an actor as the
frame of reference in the organization of his perceptual
field."44
Because there is some apparent duality in the term
reference group, Shibutani distinguishes between groups
which have comparative functions and others which have
normative functions. "It is important to recognize . . .
that a group which serves as a point of comparison is
quite different from a group whose culture constitutes
45
one's point of view." It is this normative function of
groups which is vital to the present theory and experiment.
It is the contention here that when the individual
perceives the group as a unity (perhaps not with total
consensus but presenting a united view) this group picture
will exert a group pressure on the individual to adopt and
internalize the group value. This reference group function
in the internalization process is dependent upon at least
44Tamotsu Shibutani, "Reference Groups and Social
Control," Human Behavior and Social Processes; An
Interactlonist Approach, ed. Arnold M. Rose (Boston:
Houghton Mifflin Company, 1962), p. 132.
45Ibid.. p. 133.
61
two factors: (1) the perception of the group as an inte
grated whole rather than as individuals, and (2) the iden
tification of the individual with the group.
The first component of a reference group, i.e.
the perception of the group as an integrated whole, cor
responds to Mead's use of the term "generalized other" and
provides another group factor which the present study
assumes would help support the major hypotheses. This
supporting hypothesis is stated:
IV. As viewing the group as a generalized
other increases there will be (a) more
internalization and (b) greater conformity
with the consensus value.
The second component of reference groups, i.e.
the identification of the individual with the group, is
considered as another group factor which may add to inter
nalization pressures. Specifically the hypothesis is
stated:
V. As identification with the group
increases there will be (a) more
internalization and (b) greater
conformity with the consensus value.
The identification factor apparently is the major
differential in whether the group provides a "comparative
function" or the "normative function" of a reference group.
It would be possible to progress from usine- a group for
62
comparison of one's standards into usinf* the group's stand
ards as one's own, but as this progress occurs there
would almost necessarily be occurring the process of in
creased identification by the individual with the group.
While it might be possible to identify with the group and
still not internalize a particular value, internalization
would have to be accompanied by identification. All of
this is necessarily based on the premise that the indivi
dual first perceives the group as a group so that he could
go through the process of internalizing a group value.
Note that this discussion of internalizing a group
value does not preclude possibilities of other internali
zation processes where groups are not involved, but it does
imply that where the individual perceives a group as an
integrated unity and identifies with the group this will
allow for group pressures toward internalization which can
not be accounted for by individual and interpersonal re
lationship factors.
The group perception and identification theory
presented here was explored somewhat in a study described
by liewcomb. Catholic students were divided into three
group. In the experimental group the subjects were told
that their attitudes as Catholics were desired. The other
two groups were not identified as Catholics. The "Catholic"
point of view was much more widespread in the group iden
tified as Catholics. "Average differences between the
63
experimental and the control groups were so great as to be
almost certainly not attributable to chance."^
In Newcomb's experiment the reference group was
not so much the immediate group of individuals as it was
the larger group involved in the generalized other concept
of "Catholic," Although the tested group influence in the
present experiment is primarily the experimental set of
individuals, this paper does not wish to imply that the
group pressures of a reference group need to come from
some organized or tangible group, A reference group may
be any group which is perceived by the individual as a
unity presenting some frame of reference.
For purposes of studying internalization processes
in terms of group pressure one would have to have a ref
erence group. Thus the groups used in this experiment are
considered not only as "groups capable of consistent, co
ordinated action," but it is hoped that they are also
capable of being "groups whose perspective is used by
members in their own organization of their perceptual
f ield,"
Relating the special factors of groups which might
influence internalization to the earlier hypotheses demands
one more theoretical link: joining the incrasing charac
teristic of "groupness" with an increase in the group
^Newcomb, p, 228,
64
factors which might influence the internalization process.
Thus the final hypothesis states:
VI. Identification with the group and
viewing the group as a generalized
other will increase as there is an
increase in demand for group action.
It is conceivable that Identification with a per
ceived group and viewing it as a generalized other could
occur without being directly related to the group need for
consensus. However, the above hypothesis implies that the
group need for consensus for group action would strengthen
these factors and might call them out when they would not
otherwise be present. Since the third experimental group
variable, group roles, would probably not occur at all
without the need for group consensus for action it was
omitted in this hypothesis.
Reviewing all the experimental hypotheses of this
dissertation it will be noted that the basic assumption is
that when group action is demanded there will be group
pressures, in addition to interpersonal influences, to
achieve consensus with the group and to bring individuals
to internalize the value under consideration. In an effort
to specify factors which would apparently be involved in
these group pressures it has further been hypothesized
that playing a group role, use of the eroup as a gener
alized other, and identifying with the group will lead to
65
greater conformity with the group and to increased inter
nalization of the value involved. To tie the need for
consensus to carry out group action to the group factors
it was finally hypothesized that there would be a direct
relationship between increased need for group action and
the occurrence of the group factors which might further
the internalization process.
One further problem that should be covered in this
discussion on the theoretical basis for the present exper
iment is that of conflicting reference groups. In most
modern complex societies there are bound to be many ref
erence groups having different values from which the indi
vidual may choose. In a discussion of the basis on which
a choice is made Shibutani lists two hypotheses: ’ ’ One
widely entertained hypothesis comes from the experimental
study of small groups: a person tends to comply with the
norms of the group that he finds most attractive.
"Another promising hypothesis, not necessarily
inconsistent with the first, is derived from psychoanalysis
and from the work of Cooley and Mead: the choice of defi
nitions depends upon one's sentiments toward the signifi
cant others who serve as representatives of reference
43
groups." The approach taken in this paper conceded
47
Shibutani, "Reference Groups . . ., p. 140.
48lbid., p. 141.
66
interpersonal relationships as significant factors in
internalization. It has further considered that the group
as a whole may be taken as a significant or "generalized"
other and that the individual may identify with the whole
group rather than just specific significant others.
The present study can not eliminate the possibility
that the experimental subjects will use some reference
group other than the group present in the experiment. This
might negate the occurrence of a group influence which is
the primary focus of the study. However, in so far as the
hypotheses include consideration of the occurrence of
specific group factors such as use of the generalized other
and identification with the group, it is assumed that if
these factors occur in the experiment there is a strong
liklihood that the subjects are using the experimental
group as a reference group. If the experimental group is
being used as a reference group then the study's basic
hypotheses would be expected to apply.
The basic task of the experimental aspects of the
study, then, remains that of determining, if group factors
are related to the internalization process.
CHAPTER III
THEORETICAL APPLICATION OF
THE HYPOTHESES
One of the basic problems that has troubled socio
logists from the beginning is how to explain the origin of
social networks and cultural patterns. Why and how do
individuals display such persistent and uniform behavior
patterns that sociologists are able to study, and to come
to conclusions about, society and social behavior? Are
there social instincts in man that drive him into patterns
of social behavior? Is society an external force that
molds and controls man by some exterior design? Today, we
usually agree that there are factors about the interaction
of men in a social setting that produce these uniformi
ties, but what are these factors and how do they operate?
The explanation of the uniformities in man is an
underlying element in psychology also. The lack of com
plete uniqueness of individuals can only partically be ex
plained on the basis of common biological backgrounds.
The cultural determinism approach under which one would
logically expect to find "national character" seems to be
inadequate in the face of such variation within the
67
68
cultural unit. How can one explain the common elements
and still account for uniqueness In man?
One task of the social psychologist Is to offer
postulates and evidence that will at least partially clear
up this problem--to give some explanation that will at
once explain for the sociologist why and how individuals
display persistent and uniform behavior patterns and ex
plain for the psychologist why and how the social environ
ment produces such commonness (or lack of uniqueness) for
particular categories of people.
Quite obviously this is an Olympic task that is
much beyond the capabilities of this dissertation. To
bring this study down to somewhat manageable experimental
proportions many areas of theory were necessarily omitted
and many variables were (hopefully) either excluded from
the experimental situation or were assumed to be randomized
in such a fashion that they did not influence the results
with any consistent directionality. As a consequence, not
only are the hypotheses limited in inclusiveness, but also
the results will be confined in their application to cover
only the circumstances and variables which were handled by
the experimental situation.
While, on the one hand, the limitations of the
theory and research of this study must be constantly kept
in mind; on the other hand it is necessary to point to
the vast potential contribution that can be made if even
this limited research area can be given some concrete sta
bility. Dependable findings in this area could have impor
tant implications for sociology to understanding social
organization, individual and group deviancy, the operation
and stability of small groups, group conflict and its re
solution, socialization, and social change. For psychology
such findings might aid explanations of cognitive disso
nance, motivation, perception, learning, and individual
change and disorganization. The following paragraphs will
attempt to point out how proving the experimental hypoth
eses would be important to each of the abtoftre mentioned
areas of sociology and psychology.
A relationship between group factors and interna
lization might help explain aspects of social organization.
If group factors do increase internalization, where these
factors were most evident and effective the groups would
be more long-lasting and influential. More of the members
would have reached a deeper level of attachment to group
values. If the participants were in greater and deeper
agreement on values this might in turn lead to further
development of group factors. For instance, if society
needed effective organization and cohesion to defeat an
enemy in war this need or requirement on the part of the
group would produce organizational patterns and pressures
which would aid the bringing of consensus among the
citizens in regard to values (e.g. the value of their
70
cause, their military efforts, etc.). This agreement on
values within the group would in turn result in tighter
organization and cohesiveness within the society.
Accepting the basic premise that certain group
factors produce greater internalization of uniform values
in the group members might make it possible to help explain
deviancy. People who were farther removed from these group
factors would be more likely to become deviant (or at least
conform only on the basis of external pressures without
internal motivation based on common values). For instance,
if a new employee were given a handbook to read rather than
a class in company policies and procedures where group fac
tors would occur, one might expect that the individual
would violate the procedures and policies more often than
he would if he had been trained under a group influence.
The same experimental hypothesis might be useful
in explaining deviant groups as well as ir viduals. If
one were sure that certain group factors tended to produce
greater pressures on individuals to internalize certain
common values, then one might predict that a group which
possessed more of these group factors would resist outer
pressures to conform to a general societal value more than
would other groups which did not have as many of these
group factors. For instance, if a delinquent gang had
group titles and observable roles which were lacking in
another otherwise similar group, it might be predicted
71
that the first group would have factors leading its members
to internalize the value of deviant behavior to a greater
degree. In this case the first group would not give in as
easily as the second to pressures of police, welfare wor
kers, or society at large in their efforts to get the group
to give up its deviant beliefs and practices.
Once again, the proof that group factors increased
consensus and internalization would be of great aid in ex
plaining the differences in group behavior and stability
of various groups which differed in the amount of and
operation of these factors. If in one political party (or
prison work group, or Girl Scout Troop, or bowling club)
it was found that the members were very aware of the roles
they played and the roles played by others and how all
fitted into a pattern which made the group operate, this
might well help explain why this group would be expected
to have greater internal agreement and motivation to
achieve what the group values than would be found in a
political party (prison gang, etc.) where these factors
were not present. The former group could be expected to
be more productive and enduring than the latter group.
The development of conflict between employer-
employee groups (between line and staff groups, between
delinquent gangs, between churches, etc.) might be anti
cipated by applying the hypotheses. If one saw an in
creasing development of group factors and feelings in
72
several groups, one would expect greater agreement within
the group and greater internalization of what the group
valued. If the values of the groups did not agree, conflict
might be expected whenever the groups came in contact with
each other. From a premise relating group factors and in
ternalization one might also be led to try to resolve con
flict by getting groups to agree on values. One could
produce agreement by manipulating the group factors which
influence the values accepted by the individuals in each
group until members in both groups thought alike.
In this discussion of the pertinence and applica
tion of the basic premise it should be noted that if speci
fic group factors influence internalization the socializa
tion process also must be explained in terms of these group
factors. This explanation of socialization would go beyond
the fact that a person picks up values from other indivi
duals in certain groups (e.g., family) or that a person
picks up values by participation in society in general.
This addition to socialization theory would point out that
the individual also picks up values on the basis of speci
fic situational factors within the group and in the indi
vidual. For instance, if the group has clearly identified
roles (to which the person is ascribed or which he has
achieved), or if the group activity or goals provide a
need for consensus within the group, or if the individual
receives a great satisfaction from this particular group,
73
or if he sees the group as a unity of which he is a part,
then it can be expected that within the group there are
factors which will produce greater internalization of the
values of this group.
Certainly most sociologists would not assume
broadly that an individual picks up values equally from
all groups in which he participates; however, a certainty
that particular group factors influence the acquisition of
values from the group would lead to pinpointing where and
when and how groups socialize individuals.
If one were able to assess that certain groups were
increasing in their influence over larger and larger num
bers within a society, on the basis of our basic premise
one might expect that the values of these groups would be
accepted in greater numbers and with greater intensity,
and as a consequence the organization and operation of the
society would change in the direction of these groups and
their values. Quite obviously there are other factors of
importance here (such as the amount of agreement between
the various groups, the nature of the values being inter
nalized, etc.), but even this one premise, if it were
true, would greatly aid in the development of theory and
application of knowledge about social change.
Of course, social change is not always in the
direction of increasing consensus and internalization, and
the basic premise would also be useful in locating
74
Instances where there was a reduction of group factors
which would indicate a lessening of group influences over
internalization of values. In these cases one might expect
either a growing decline of internalization or a greater
diversity in what was internalized by the individuals in
the society; either of which would be expected to further
disorganization within the society and further social
change that results from disorganization.
The preceding paragraphs have indicated a few of
the potential sociological applications of a premise that
group factors do operate to influence the amount of agree
ment within a group or society and also influence the
strength of internalization of the values of the group or
society. If it were possible to prove the influence of
these group factors on the internalization process, several
areas of psychology would benefit as well.
Some cases of cognitive dissonance might be ex
plained if one were able to show that although two people
were members of the same two groups who had differing
values but that they were differentially affected by the
group factors operating with the two groups. That is, if
one person identified equally with the two groups, had
equivalent roles in both groups, and received group sanc
tions equally from both groups, while the other person was
unequally affected by factors in the two groups, then one
might be able to explain why the first person experienced
75
cognitive dissonance while the other person did not. Cer
tainly this is not to say that this situation is the only
condition that can lead to cognitive dissonance, but it is
one possible explanation.
In the area of motivation an awareness that an
individual had been influenced by group factors to inter
nalize particular values and/or goals might help explain
why the individual used so much of his energy and resources
in efforts to maintain the values or achieve the goals.
In addition it would also help explain why his motivational
pattern changed as his internalized values changed when he
left the influences of one group and came under the influ
ences of other groups. This does not necessarily mean
that the original internalized value would be no longer
internalized, but this internalized value might be pushed
into the background with the acquisition of other inter
nalized values.
Because perception is so easily selective or
distorted in favor of one's own value system the explana
tion of perception would be furthered if one could point
with some confidence to the fact that groups influence the
strength of internalization of specific values by the
individual and consequently influence his perceptions.
Also, an awareness of the influence of group factors in
producing greater consensus among individuals' internalized
values would help explain why there is such uniformity of
76
perceptual patterns within particular groups of individuals.
Theories of learning may easily overlook the pos
sibilities of group factors. In considering the rewards,
punishments, reinforcements, etc. on an individual basis it
is easy to overlook the fact that groups as a whole may
provide rewards, etc. beyond the interpersonal rewards
offered within the group. If approval by an individual is
reinforcing or leads to greater internalization, then it
might also be possible that if the individual perceived
the group as a whole approves of him, his behavior, or his
values there would be even greater internalization result
ing from these group approval factors. If these group
factors did increase internalization then one might expect
a greater potential for learning where these factors were
present.
The prediction of change and disorganization for an
individual could be made more accurate if one were able to
explain how strongly he had internalized certain values
from his previous contact with particular groups. If the
group factors were weak and internalization was slight one
might expect a much greater potential for change within the
individual. In addition, a knowledge of present and/or
future group pressures would help explain in what direction
the individual might be expected to move on the basis of
what he would be apt to internalize from these groups.
As was mentioned previously, dissonance within the
77
individual can perhaps be better explained if one were able
to demonstrate the influence of group factors on internali
zation, Even at a subconscious level, i.e., below the
cognitive level, one might expect problems where there are
conflicting group pressures to internalize or use differing
values. With increasing pressures and increasing conflict
at increasing levels of internalization a situation of dis
organization might seriously affect his capability for
effective absorption of social influence and effective
social interaction.
The preceding psychological areas are just a few
which might be influenced by the establishing of a sound
knowledge that there are group factors which influence
internalization. If one could prove that such group fac
tors do exist and operate in particular ways and to a
specified degree, most of the primary areas of psychology
would need to take this additional finding into account.
Thus the real question for the basic premise of
this study is not whether such a premise is of importance
to the fields of sociology, psychology, and social psy
chology, but can such a premise be proved? On the surface
it may not sound like too difficult a task to show that
there are group factors that influence internalization of
values by individuals; however, most of the remainder of
this dissertation will struggle with the problems that such
an undertaking presents. The final outcome will not be a
78
condition of any great degree of certainty in regard to
the findings and implications of the empirical study that
was undertaken. The only outcome that can reasonable be
expected at this stage of theoretical and methodological
development is that the research will give clues as to
some possible ways to tackle the problem and some indi
cations of if and how groups influence internalization.
CHAPTER IV
THE SA M PL E AND M ETH OD S
For methodological and statistical reasons (which
will become apparent later) it was necessary to obtain
four separate samples. Basically the samples differ in
age and will be referred to by this characteristic. The
youngest sample consisted of five clubs of high school
youth, three boys clubs from the YMCA and two girls clubs
from the YWCA. The college-age sample was composed of
eight sociology classes at California State College at
Long Beach (three introductory classes, three social psy
chology classes, an ecology class, and a social institu
tions class). The adult sample included two clubs of
single adults (one club composed of those age 2l to 31 and
the other from age 32 to 50) from the Jewish Community
Center, and a Y's Men's Club from the YMCA. In the senior
age sample there are a current events discussion group
from the Jewish Community Center and a Women's Golf Club
from Leisure World.
There was no real effort to select the groups at
random. All of the above mentioned clubs and groups are
active in the Long Beach, California area. Since it was
79
80
felt that in an experiment which was trying to identify
and measure group effects the likelihood of occurrence of
these factors would be much greater in on-going groups.
In fact there was some question about using the college
classes which might not have had the kind and quantity of
interaction to build up group feelings and characteristics.
However, the experiments were conducted late in the spring
semester, and the groups were fairly small (for modern
college classes in a state college), so they were used
under the assumption that enough interaction had taken
place to build up some group conditions.
Trying to get a list of such group, randomly se
lecting the ones for the experiment, and then getting the
acceptance of the group to participate in such an experi
ment would be an extremely difficult task. For the stage
of experimentation that this study represents it was not
felt that the benefits of randomized cluster sampling
would be sufficient to offset the difficulty of the task.
Therefore, the groups were chosen partly on the basis of
availability and partly because it was felt that the groups
within each sample had enough characteristics In common to
be pooled into statistical analysis and still would be
broad enough to somewhat represent a larger category of
people than just the specific members of the specific
groups involved. Still the differences between the samples
seemed to be great enough to require their separate
consideration as four distinct samples.
The major reason for handling the four samples
separately is that it was felt the age differences made
them quite different from each other, and this variable
might well be quite important to internalization and sus
ceptibility to influence from a peer group. Therefore,
theoretically and statistically they should be considered
separately. Table 1 indicates that there are wide differ
ences in the mode and range of age in the samples.
TABLE 1
THE FOUR SAMPLES BY AGE
Age *
Sample
High
School
College Adult
Senior
Citizen
1 - 1 7 71 1
18 - 29 24 183 20
30 - 44 21 23
45 - 59 5 16 8
60 - over 1 31
*The age categories in Table 1 are roughly in 15
year intervals. Since there is a rather sharp distinction
in education and status at 18 it was decided that expanding
the first category to 17 would fit better into culturally
defined age groupings. This consequently narrowed the
second interval, but once again there seemed to be a cul
tural basis for distinguishing a young adult level between
18 and 29.
82
To be sure, there is some overlapping of age for
the samples, but the central weighting of age in each
sample is quite different. In addition, the groups used
are identified by themselves and others as basically dif
ferent age groups: The high school clubs clearly had
graduation points where members nearly always left the
club upon reaching that level. Even older adults who are
nowadays often present in the college classroom identify
the college class according to an 18-23 age category.
Two of the adult groups had specific age limits
for membership, and even where specified limits do not
exist such groups usually have some clustering of age,
although this is admittedly not as clear-cut as was true
for the younger groups. The senior citizens' groups are
labeled as such by outsiders and the members as well. In
every phone contact with such a group the experimenter was
told, "Now you know that these are older men (or women),
don't you?" The case for rather clear-cut differences in
ages between the samples seems to be rather obvious.
As one would expect there were educational differ
ences accompanying the different age samples. Table 2
indicates that the high school sample was rather clearly
limited to that educational level. The college group by
definition was somewhat beyond the high school level, al
though some were only in their first year of college. For
the adult and senior citizen samples there was an expected
83
range of educational levels.
T A B L E 2
TH E F O U R S A M P L E S BY E D U C A T IO N
Highest
Grade
Completed
Sample
High
School
College Adult
Senior
Cit izen
8th Grade 75 2 7
High School 20 14 11 10
1 Yr. College 36 7 7
2 Yr. College 95 16 4
4 Yr. + 65 21 9
It appeared that access to the higher age groups
was more likely where the level of education was generally
higher than would be expected for the population as a
whole, but this slight weighting in favor of advanced edu
cational level for most of the groups (even the high school
participants were usually expecting to attend college) kept
this factor more nearly balanced for the groups and tended
to accent the differences in age between the samples. Thus
the differences between the samples could be considered as
largely based on age differences.
Two other variables, occupational level and sex,
were also roughly equal for the four samples. Table 3
shows a fairly equal distribution of males and females in
84
the various samples.
T A B L E 3
THE FO U R SA M P L E S BY SE X
Sample
Sex
High
School
College Adult
Senior
Citizen
Male 55 88 35 3
Female 40 121 23 37
The occupational level of the head of the house
hold shown in Table 4 indicates some approximation amonR
the samples with the adult groups having a more equal dis
tribution among the levels, which one might expect. Per
haps the seemingly high figures for the level eight cate
gory need some explanation. This category includes the
retired and the unemployed (participants who listed occu
pation as "student" were considered in this category).
The question asking for the occupation of the head of the
household several times received the "student" response
from the high school and college groups. Probably some
didn't understand it was the head of the household's occu
pation that was being called for, and in the college group
some of the students legitimately consider themselves as
on their own and thus the head of the household, and their
major occupation at the time of the experiment would be
"student."
TABLE 4
THE FO U R S A M P L E S BY
O C C U P A T IO N A L L E V E L
Occupat ional
Level *
Sample
High
School
College Adult
Senior
Citizen
1 (Highest) 9 20 4 5
2 20 44 13 7
3 20 45 12 8
4 13 29 16 3
5 17 29 2 3
6 2 15 3 1
7 3 3 0 2
6 (Lowe st) 7 25 9 9
^August B. Hollingshead and Fredrick C. Redlich,
Social Class and Mental Illness (New York: John Wiley
and Sons, 1958), Chapter 4 and Appendix 2.
86
The experiment was conducted in Lone; Beach which
contains a rather high number of retired people which also
influenced the number of participants classified in the
level eight column. When senior citizens groups were con
ducting the experiment they were instructed to list the
former occupation of the head of the household even if he
were retired--or deceased. In spite of the instructions,
however, a number put "retired" when occupation was called
for.
In spite of this rather unusual weighting in the
level eight category for all the groups, the occupational
level of the participants' head of the household is rather
clearly indicated in each sample, and there is some reason
to consider the samples as covering a rather wide occupa
tional range and being roughly equivalent to each other in
this variable.
Although age, education, sex, and occupation were
considered in relation to the samples and also in connec
tion with their influence on the hypotheses, the major
variables of the study included other factors. Internali
zation and consensus were the most vital variables of the
two major hypotheses. In the supplementary hypotheses that
spelled out specific group factors there were concepts of
the generalized other, identification, identifiable group
roles, and conformity. All of these variables required
operational definitions. To explain the operational
87
definitions it is first necessary to understand the details
of the experiment as it was conducted with each group in
the four samples.
The actual operation of the experiment was con
ducted in the following manner: First the group or club
was divided into sections of six members each. To insure
a random distribution of the members into each section the
members drew balls from a bowl. Each ball was marked with
an S (section) number and an M (member) number which then
became the code number for that person and was marked on
all papers he received and returned.
After the balls were handed out it was explained
that the M-3 (member number three) in each section would
act as coordinator: getting materials from the experi
menter and seeing that the group carried out its assign
ment. The assigned role was thus appointed on a random
basis within each section, and as would be expected it was
carried out with varying degrees of efficiency and earnest
ness.
It was further explained that for the basic prog
ress of the experiment each section was to consider which
one of five qualities (assigned by the experimenter) would
be the most important for getting ahead today. Then letters
would be written to imaginary newspapers telling what the
participant felt would be important for getting ahead, and
finally there was a questionnaire in which the participants
88
would describe themselves, the experiment in their section,
and further opinions.
The differences between the sections would be in
the demands for group action. Section one would not dis
cuss the quality or anything else during the experiment.
Section two would discuss the quality for a while, then
each member would mark on the available sheet which quality
he personally felt was important, and finally the partici
pants would write their own letters to the newspapers, etc.
Section three would not only discuss the quality but would
also write one newspaper letter from the whole group.
Questionnaires were completed individually by the subjects
in all sections after the letters had been written.
If the group had more than enough members for the
three sections, additional sections were added if possible.
Section four was to do the same as section three. If there
were enough for a section five it would do the same as
section two, and section six would be like section one.
Any odd number which would be left over from the section
assignments were given "X" (extra) numbers and were to do
the experiment individually just as section one.
When the situation seemed to warrant it the re
searcher explained briefly that this was an experiment to
find out about groups and their responses were not impor
tant individua11v and would not be used to "psychoanalyze"
them or try to uncover anything very personal about them.
89
The researcher was introduced as a professor working for
her doctor's degree in sociology, and basically nothing
was said about the real underlying purpose of the experi
ment itself. This does not mean that the participants did
not realize some of the underlying themes and possible
purpose within the experiment. For instance, the measure
ment of consistency of opinion was rather thinly veiled in
the questionnaire. Still there is no reason to suppose
that any but a very few participants gave other than what
they felt was the honest response even if they were some
what aware of what was being measured.
Materials provided for the participants included
a check list of the qualities which they might choose as
being the most important one for getting ahead today.*
These qualities were: hard work, pleasant personality,
brains, knowing right people, good luck. A paper with in
structions and blank space for the letter to the newspaper
was also provided. The most important material received,
filled in, and returned to the experimenter was the six
page questionnaire. Samples of these three materials may
be seen in the appendix.
Before explaining the contents of the question
naire some of the rationale behind the basic procedure of
*The basic question used in the experiment and the
choice of possible qualities were taken from a previous
study: Morris Rosenberg, Occupations and Values (Glencoe,
111.: The Free Press, 1957), pi 140.
90
the experiment should be pointed out. The group was
divided into sub-sections to provide for treatment differ
ences. The smaller number also permitted more primary
interaction among the participants--where interaction was
called for by the treatment variable. The careful randomi
zation of assignment to the sections and equal numbers
within each section were called for on the basis of method
ological and statistical considerations.
The treatment variable of requiring a different
type of interaction situation in each group was designed
to provide a brief and rough continuum of a "groupness"
factor. Section one where no interaction was to occur
provided a "non-group" extreme. The request for brief
discussion but no group task (other than discussion) which
occurred in section two provided a minimum of the require
ments for a group or a demand for a minimum of "coordinated
activities" which Shibutani considers as the basis of a
group.^ 3y adding the responsibility of composing one
letter to the newspaper from section three it was felt
that this would provide additional demands for specific
"coordinated activity" which would then represent the
greatest potential for the operation of "group" factors.
It is not supposed that these three treatment dif
ferences provide an interval measurement of all possible
^Tamotsu Shibutani, Society and Personality (Lngle-
wood Cliffs, N. J.: Prentice-Hall, Inc., 1961), pT 33.
91
points on the group factor continuum. On the other hand,
it does provide a rough ordering of perhaps the more impor
tant areas along that continuum.
The choice of the discussion topic, i.e. what is
the most important quality for getting ahead today, was
determined primarily because it was thought to be a ques
tion: (1) in which most people were interested, (2) about
which most people would have some ideas, and (3) of which
most people and groups had not previously thought in quite
these terms. In line with Kelman's observations on inter
nalization mentioned earlier it was felt that the interna
lization process would be much more likely to occur if the
experiment was based on some topic which was considered
3
important to the individuals involved. It would tend to
engage the person more deeply and directly in the consi
deration of the topic and also would heighten the impor
tance of and the possibility of reference group influences
over the individual.
The original study from which the question was
taken gives some indication of the importance placed on the
idea of getting ahead today--at least its importance to a
nationwide sample of college students: 584 felt it was
“very important" to get ahead in life, 445 felt it was
“fairly important," and 106 felt it was "not important."^
3Suppa, pp. 43-44.
4
Rosenberg, p. 9 5
92
A second consideration in the selection of the
question was that of whether the participants would have
previously thought about the topic. If someone had pre
viously begun to internalize some value related to getting
ahead today this might influence the results of an experi
ment which was attempting to produce and measure the inter
nalization process.
At first an attempt was made to devise some new
experience which would not have had a chance of being in
ternalized. Although Pollis seemed to get results with
this approach the task was considered almost impossible
by the present experimenter.^ with any new experience or
area of thought the individual will tend to try to use
previous frames of reference (which might have been inter
nalized to some unknown degree). Additionally most new
experiences that might be used in this type of an experi
ment would be so insignificant to the individual that he
would not become very deeply involved in the task and thus
there would be much less chance of internalization occur
ring. After a pilot test with a hopefully "new" experi
ence, it was decided that this approach had too many prob
lems to be handled adequately at this early stage of
experimentation with internalization.
^Carolyn Sherif, Muzafer Sherif, and Roger E.
Nebergall, Attitude and Attitude Change (Philadelphia:
W, B. Saunders Company, 1965), pp. 5(37-211,
93
Changing to some topic that had probably been con
sidered previously by the participant meant that there was
a likelihood that internalization had already begun for
some of the participants. It might then be difficult to
determine if the experimental treatment differences had
been the cause of differences in internalization rather
than the differences occurring from previous internaliza
tion.
For the most part two factors seemed to overcome
this problem: First, the random assignment of individuals
to the various treatment sections would also tend to ran
domize previous internalization differences in the members
of the various sections. Thus any previous internalization
would not tend to influence the results of one section any
more than another, and differences between the sections
could be more directly attributable to the treatment dif-
f erences.
Second, the topic itself seemed to have some ad
vantage in overcoming this previous internalization problem.
Although the topic is one in which most people have an
interest and feel it would have value to them personally,
it is also a topic which most people have probably not
considered in quite the way it was presented in the experi
ment. Even though "getting ahead" is considered to be of
value, a specific quality for getting ahead is not usually
considered. Ordinarily getting ahead is thought of in
94
terms of either a group of qualities for getting ahead
generally or in terms of some specific situations where one
quality might be important at one time and other qualities
at other times.
During the experiments many subjects expressed a
desire to choose several of the qualities or asked the
researcher to explain specifically what "getting ahead"
meant. In order to reduce the effect of previous interna
lization (and also for ease of analysis of results) the
subjects were told they must choose one quality (on the
first check sheet, but not on the newspaper letters) and
they would have to decide for themselves what "getting
ahead" meant, i.e., it meant whatever they thought it meant.
Hopefully then, the topic of getting ahead was one that
was important to most of the participants, was an area
about which they had some notions and would be open to
further ideas, but would not be one where some particular
quality for getting ahead would have been internalized
much, if at all.
Even if a person had begun to internalize a par
ticular value the experimental treatment might still be
important. That is, he might be partially accepting a
certain value or viewpoint as his own but still be some
what susceptible to outside influence. There are degrees
of internalization, and this whole experiment is designed
to try to catch a difference in degree of internalization
95
as it relates to other variables rather than to observe
when internalization has fully occurred (i.e., when the
process is completed and the individual can no longer be
Influenced by factors other than his total acceptance of
his internalized value).
Writing the letter to the imaginary newspaper was
not so critical in providing measurements for the experi
ment, but it did serve four experimental purposes. First
it gave the researcher some time between the returning
of the first quality choice sheet before the questionnaire
was handed out. This time was needed to rig the question
naires to go along with the indicated choice of quality,
which will be explained in more detail later.
The second advantage of the letter was that it
provided an open ended question which allowed the parti
cipant to explain himself, give further indication of which
quality (or qualities) were important, and reveal the
"purity" of his thinking about the qualities, i.e., whether
one quality was obviously dominant, whether there were
other conditional--or perhaps equally important--qualities,
or whether no quality was considered of any particular
dominance.
The third advantage, which was part of the original
motivation to add the letter to the experiment, was that it
provided a reinforcement possibility which might help
strengthen and clarify the person's thinking and
96
internalization of the quality for getting ahead. It was
recognized that internalization is usually thought to occur
over some period of time, and if even a small degree of
internalization were to be induced some effort would have
to be made to intensify and focus the attention of the
individual on the problem and provide some sort of commit
ment to his opinion. The letter was designed to help per
form these functions.
Actually the letter originated as a possible tech
nique which would help provide for the treatment differen
ces between the sections; that is, it presented a possible
group task for the section which needed some activity
beyond the mere discussion stage to stimulate group factors,
and at the same time provided an equivalent, non-group
task for the other sections. In this way it would be more
likely that any measured differences between the sections
would be due to group factors and not to the writing of the
letter. In other words, it would avoid the kind of "treat
ment effect" which was found in the now famous Western
Electric Studies where it was discovered that the subjects
were responding to the treatment itself rather than to the
theoretically important variables.
Writing a group letter may not be an exactly
equivalent task to writing a letter individually, but the
two are so slightly different in terms of the task involved
that it is here assumed that differences in the sections
97
would be due to social environmental differences and not
to basic task difference.
For purposes of measurement of the important var
iables in this study the questionnaire carried by far the
greatest burden. Naturally it included usual questions
regarding age, sex, education, and occupation and education
of the head of the household. These were placed in the
middle of the questionnaire to provide a break from the
obviously experimental questions. Since consistency of
opinion was one of the attempted measurements these back
ground questions briefly interrupted the frame of reference
being used for previous questions. Although the "halo"
effect probably still occurred to some extent, there was
an attempt to avoid it by inserting other questions between
consistency questions.
The major variables that were operationally defined
and measured in the questionnaire were: internalization,
consensus, the generalized other, and identification with
the sections and with the larger group. The rationale
behind, and the measurement of, each of these will be dis
cussed in turn.
Internalization is probably the most critical and
at the same time most difficult measurement of this study.
As was mentioned in chapter two it has been defined in
different ways which imply different operational techni
ques for its measurement. Some of these were experimented
98
with in different parts of the questionnaire.
Turner*s idea that internalization occured when
role-taking ceased^ stimulated the question which asked:
"In giving your answers to these last two stories (which
attempted to illustrate cases of getting ahead with accom
panying choices of possible causes) do you feel you would
be surer of your answers if you knew what other people
felt the most likely answer was?" The possible responses
to be checked included replies from "would feel much surer
if I knew others* views" to "would not feel any surer if I
knew others* views."
Here the individual was answering the questionnaire
alone, and this question presumably would point to whether
he felt quite sure of himself and thus had theoretically
internalized the value involved, or whether he was still
uncertain and would welcome the opinions and suggestions
of others. This is not precisely equivalent to a cessation
of role-taking, but it does indirectly ask if the person
would like to be able to take-the-role of others.
Parsons has developed the idea that internalization
occurs when the individual now acts under his own direc
tions and sanctions rather than under the pressure of ex
ternal sanctions or threat of sanction by others.^ This
is a rather widely held view of internalization, and it,
6Supra, p. 33. ?Supra, p. 32.
99
too, was incorporated into a part of the questionnaire.
An interesting study by Asch shows that individuals
will often vary their responses after hearing the opinions
O
of supposed experts. In fact this is presumably a part
of the development of norms in Sherif's famous autokinetic
effect experiment and the conformity to group views in
Asch*s line comparison experiment, where the views of
others or even one’s own previous perception (in Sherif’s
study) become a kind of "expert" view which directs the
individual.
To test Parson’s approach in the present experi
ment it was decided to rig this experts' choice effect.
The final page of the questionnaire told the participants
that this same question about the most important quality
for getting ahead had been put to a large number of voca
tional counselors and even to quite a few successful people
who were to be found in r ho's "ho. The outcomes from
both these groups indicated that "___________" was believed
to be the important quality. These studies and their
results were invented by the researcher, i.e., they are a
falsehood--which was always revealed to the participants
when the experiment was completely over. The blank line
in the sentence above is used to indicate that there were
®S. E. Asch, "Studies in the Principles of Judg
ments and Attitudes: II Determination of Judgments by
Group and by Ego Standards," Journal of Social Psychology.
XII (1940), 433-465.
100
two possible last pages; each had a different quality as
the one which both the counselors and the Who's Who indi
viduals had chosen. In this way it was possible to line
up a questionnaire which had a quality different from the
one which the experimental participant had indicated on
Q
his original check sheet.
After the disclosure of the choice by the experts--
which was rigged to be different from the participant's
choice--there were two questions. One asked that if the
experts' answer didn't agree with the participant's answer
would the participant change his answer? The possible
responses included: probably would change, might change,
might change--but doubt it, probably would not change. The
second question asked the participant to pretend he had
changed after hearing other groups had disagreed with him,
then how would he feel. The alternative responses were:
would not feel anything
would feel I had improved my answer by changing
would feel I was justified in changing my answer
would feel a little uneasy in changing my answer
would feel guilty if I changed my answer because I
would be going against what 1 had already marked
down
would feel guilty because I would be changing to what
I feel is an incorrect answer
Since these questions are direct and obvious it is
to be expected that the answers to them might differ from
Q
As was previously noted, this "arranging" was
possible during the interval when the participants were
writing their newspaper letters.
101
responses if indirect techniques, such as are usual in
this type of measurement, would have been used. Thus one
might expect to find less inclination to change the answer
in the direction of the "experts" than might be true where
a more subtle approach was used.
In spite of this influence of the direct approach,
however, it was felt that if a person had not internalized
the value of his own particular quality choice he might
say he would change to the expert view and not feel the
least guilty about doing so. On the other hand, if the
person had rather deeply internalized the value he would
not change his answer in the face of differing "expert"
opinion and would feel guilty if he did change. This
technique was therefore considered to be one possible way
to measure internalization.
On a rather common sense level the present re
searcher decided that if internalization meant deeply
accepting a value as one's own the expression of this
deep acceptance might be revealed in three ways: First
the person would be likely to hold to this value in the
face of other possible alternatives and through a variety
of situations. That is, the particular frame of reference
he used where this value was dominant would be likely to
be applied on a consistent basis even though certain situ
ational factors might differ slightly. Thus his choice of
one quality would be "consistent" throughout the
102
questionnaire.
Second, if the person felt rather deeply about
this, his own, value he would probably be rather certain
that this was the proper value. The concept of ethnocen-
trism implies that as people come to hold their ways as
having value and being the proper ones they become very
sure of their own views and their "rightness" while the
views and ways of others are discredited and even despised.
Thus "certainty" was thought to be a second indication
that internalization might be present.
Finally, the concept of guilt occurring if one
violated or went against his own choice seemed to have
merit. If one had deeply accepted a value as part of him,
he might well be expected to feel guilty if he went
against this value. Certainly the whole psychoanalytic
approach to mental health and mental illness seems to
support this view. Even the response that one would feel
guilty because he was changing his answer (not because his
answer was correct) shows some internalization. Deutsch
and Gerard say:
Normative social influence from oneself to conform
to one's own judgment may be thought of as an
internalized social process in which the individual
holds expectations with regard to his own behavior;
conforming to positive self-expectations leads to
feelings of self-esteem or self-approval while non
conformity leads to feelings of anxiety or guilt. u
Norton Deutsch and Harold B. Gerard, "A Study
of Normative and Informational Social Influences upon
103
Because the measurement of internalization is such
a difficult thing to ascertain and because there are prac
tically no other reported attempts to devise techniques to
measure internalization,!* it was decided to test this
three factor theory of internalization by constructing
scales of the sub-variables and then pool them into a
single internalization score.
The total internalization score would thus be com
posed of measurements in three dimensions: consistency,
certainty, and guilt* On a common sense or intuitive basis
it was believed that each of these factors would occur if
internalization had taken place. However, it was also con
sidered possible that any one of these might occur even
though internalization had not taken place. Therefore, it
was felt that a pooling of the scores would be a better
indicator of internalization than any one separately.
A person might choose one quality just because he
liked the sound of it or had heard it most recently and
Individual Judgment," Journal of Abnormal and Social
Psychology. LI (1955), 656.
**0nly two attempts are known to the present
author. Herbert Kelman uses one technique as described in
"Compliance, Identification, and Internalization: Three
Basic Processes of Attitude Change." Journal of Conflict
Resolution. II (March, 1958), 51-60, but it was not easily
adaptable to the present type of study. Daniel Miller and
Guy Swanson developed a very unusual internalization index
based on child rearing practices which is reported in Tha
£^ag^lng American Parent (New York: John Wiley and Sons,
104
might stick with this in opposition to the other possible
alternatives, but if he had not internalized the value of
this quality it would not be very likely that he would be
especially certain of his answer and might not feel guilty
in changing it.
Likewise a person might choose one quality because
sane person he identified with felt this was important and
would be very sure that this was the right quality. Then
when the qualities were presented in disguised form he
might change his answer because he did not perceive this
to be the choice of his identification object since he had
not accepted that quality on the basis of his own interna
lized value. It is assumed here that if he had interna
lized the value he would tend to apply the same frame of
reference even if the situation were slightly different.
Thus a person might indicate he was very certain of the
quality--through identification influences--but not con
sistently choose the quality from lack of internalization.
Again, a person might feel guilty about changing
his answer in the face of conflicting ' ’ expert” opinion
just because he had internalized the value of sticking to
his own opinions, even if he weren't very certain of his
views and might also give a different quality choice if
it were somewhat disguised and he did not recognize it in
a different setting because he had not inernalized one
quality to the point where set the frame of reference.
105
On Che basis of these possibilities of obtaining
a high score on any of the separate dimensions without real
internalization having occurred it was felt that by pooling
the scores a better index of internalization would be
created. A high total score would be the function of high
scores on all three sub-variables which would mean the
person was consistent, certain, and would feel guilt if he
changed his stated position. It was felt that the occur
rence of all three conditions would Indicate the presence
of internalization.
The two previously mentioned approaches (the Turner
"stopping of role-taking" and the Parsons "responding to
inner direction rather than outer direction") seemed to be
logically incorporated into the present author's "trio
technique" approach. A response that indicated one did
not feel it would be helpful to know the views of others
indicates perhaps both the stopping of role-taking and a
rather high degree of certainty of one's own choice. Thus
this question was incorporated in the "certainty" score.
Not changing one's answer in the face of contrary
"expert" opinion would also seem to indicate certainty of
one's own cho&Ae as well as a likelihood of inner direction.
Obviously the second question in the "expert" part of the
questionnaire is directly intended to get at feelings of
guilt as well as to indicate inner direction. It consti
tuted the third component in the trio approach.
106
To construct a scale of internalization consis
tency (i.e., being consistent in one's choice of a quality
as a possible indication of internalization) four slightly
different parts of the questionnaire were fitted together.
Near the beginning of the questionnaire the participant was
asked to state again his choice of the important quality
for getting ahead. The possible choices were identical
with the original check sheet. This was to see if the per
son had changed his mind in the interim period after he
originally checked the quality, while he composed his letter
to the newspaper and was waiting for additional materials
to be handed out. It was expected that few pMple would
change their minds, but there was a possibility that they
might--or even that they might have forgotten what their
first choice was. If the choice indicated on the question
naire matched the choice on the original check sheet the
person was given four points toward his internalization
consistency score.
Later in the questionnaire all possible pairs of
the qualities were listed and the participant was asked to
check his preference in each pair. Since there were five
qualities there was a possibility of his preferring his
original choice over another quality in each of four dif
ferent pairs. Each preference for his original choice gave
him one point, making four points possible toward his inter
nalization consistency score from this part of the
107
questionnaire.
Earlier mention was made of two stories in the
questionnaire. These stories were an attempt to use a more
indirect technique to measure consistency. They consisted
of sketchy accounts of a worker's promotion to foreman and
a young girl's launching an apparently successful singing
career. After each story there were five different pos
sible explanations for the success of the story's subject,
each explanation being intended to match one of the quality
choices: hard work, pleasant personality, brains, knowing
right people, an d g o o d I n c ’'.
Of course these qualities were not openly labeled
the original choices of qualities in an effort to disguise
the fact that the person was again choosing between the
five qualities. This disguise was no doubt precariously
thin, however, since it was difficult to write an explana
tion which would be taken for an equivalent of one of the
original qualities without making it apparent that this
was being done, especially when the participants had been
considering these qualities for a rather extended period.
Still the stories did present a slightly different
setting for the qualities and might induce a person to
alter his answer if he had not strongly internalized his
value choice and now saw a greater likelihood that one of
the other qualities would be more important in this parti
cular (story) situation. If the participant's choice of
108
explanation for the story matched hia original choice of
the quality for getting ahead he was given an additional
four points (four points possible for each story) toward
his internalization consistency score.
Putting these four parts, with each worth a pos
sible four points, together resulted in the internalization
consistency score. Obviously there would be a possible
range from 0 to 16. This, then, was one component of what
might lead to a total internalization score. Certainty
and guilt would be added to complete the total internaliza
tion score.
In addition to the two components of the certainty
scale mentioned earlier (i.e., not being surer if one knew
others' views and not changing to the "experts'" view)
there was also a direct question just after the person was
asked to once more indicate his choice of the qualities:
"How sure are you that this is really the most important
quality?'1
The possible responses ranged on a six point scale
from "very sure" to "very uncertain." Thus the points on
this question ranged from six (for very sure) to one (for
very uncertain). Four points were possible for "very
certain" responses on the other two questions making up
this scale, and thus there was a possible fourteen points
for a "very certain" indication that internalization was
present.
109
There seemed to be only one possible measurement
of a feeling of guilt, since this involved manipulation
which would seem very improbable to the participant if
repeated in another form. The responses on the guilt
question were scored from 1 (for "feel nothing in changing
the answer") to 6 (for "feel guilty because the answer
would be incorrect"). Then to give a rough equivalent to
the consistency and certainty scores the guilt score was
multiplied by three (making a possible 18 points for the
greatest guilt internalization) before it was added to the
other scores for the total internalization score.
Next to the measurement of internalization, prob
ably the most critical measurement for testing this
paper's hypotheses is that of determining group consensus.
Finding the complete picture of depth of agreement and
specific areas and points of agreement within a group
would be most useful but also would be quite difficult and
perhaps not absolutely necessary for the present study.
Thus a more superficial operational definition of consensus
was developed.
The one area in which consensus was to be measured
was agreement on the quality which would be most important
for getting ahead today. Obviously it would be rather easy
to find which, if any, quality was most chosen in any
particular section, and additionally to see how many agreed
with the consensus quality. However, it was decided that
110
an additional factor (her* called "purity of emphasis")
might be added to the original number of those who agree
with each other, giving a little greater consideration to
the possibility of there being secondary choices within
the section and in the minds of the section members.
This additional factor might be measured by deter
mining the purity of emphasis as revealed in the letters
to the newspaper. A rather crude purity of emphasis scale
was set up where four points were scored if only one qual
ity was discussed in the newspaper letters, three points
if there was some general qualification(s) to the main
quality, two points where two or three qualities were
given nearly equal weight, and one point if the letter
seemed to be a generalized discussion involving several
points without giving much weight to any particular one(s).
To obtain a purity of emphasis score for the whole
section was a simple matter in the cases where the whole
section together wrote the letter to the newspaper. For
the other sections a purity of emphasis score was estab
lished for each person's letter and then the average score
was used for the section. The total consensus score then
consisted of the number of people who had chosen the modal
quality (i.e., the one which had received the most choices)
plus the purity of emphasis score for each section.
One of the major problems in setting up the experi
ment to consider this matter of consensus was the
Ill
possibility of demanding consensus by the design of the
experiment. No doubt this problem was not totally elimi
nated. The instructions that one section was to discuss
the qualities and another not only discuss them but pro
duce a group letter on what was most important for getting
ahead probably suggested to at least some of the partici
pants that agreement was being called for. However, no
statement was made that they ought to agree. When asked
if they needed to agree the researcher told them no, and
the instructions on the quality check list specifically
said this was to be their Individual choice of a quality.
There still might be pressures felt within the group for a
need to agree, but it was hoped that there was no pressure
to agree on the part of the experimenter or the experiment
per se.
Permitting the members to express their opinions
rather freely in the newspaper letters and then using the
purity of emphasis factor in the final determination of
the consensus score presumably helped avoid experimental
design pressures to agree and roughly helped to account
for variations in the depth of agreement within each sec
tion. There is no claim here that the design of the experi
ment was completely free from producing pressures for con
sensus that would not be present in everyday discussions
and group actions; nor is there a claim that the Insertion
of the purity of emphasis measurement comes close to
112
handling the problem of the depth of agreement within a
group in a complete fashion* The only contention here is
that these factors were considered and were handled as well
as it was seen possible under the other requirements of the
experimental design and under the lack of readily available
techniques to control, measure, or account for these
factors.
The nominal definition, apart from the operational
definition, of the "generalized other" is far from being
obvious and explicit. At one point it was suggested to the
researcher that the term be omitted and other concepts
such as group identification, etc. be used instead. This
omission was considered, but it was finally determined
that there was something in the "generalized other" concept
that would not be adequately handled by other term^ and
perhaps an operational definition and attempted measurement
would help clarify the concept's nominal meaning.
Two of the elements which were felt to be important
to the larger term of generalized other were singled out
to be inserted into the measurement techniques of the
questionnaire. First there was the element of perceiving
the group as a whole. There have been experiments which
attempted to measure an individual's idea of how the group
1 7
as a whole felt or acted. For example, Pelz,
i o
1 Edith Bennett Pelz, "Some Factors in 'Group
Decision'," Readings in Social Psychology, ed. E. E.
113
Newcomb,Katz,*** and Wallen*^ all asked the subject to
indicate what he felt the group believed.
Generally these experiments are designed to test
(or end up showing) that the individual perceives the group
as having feelings usually more closely in line with his
own than the actual facts warrant. There seems to be
little doubt that this is the case, but this fact does not
alter the argument here that individuals do see the group
as a whole. Presumably if one were being influenced by the
"group" view he would also be more sensitive to how his
view fitted into the group pattern.
Considering that individuals do think of the group
as a whole as having a view, it is here argued that this
is a component of the concept of the generalized other.
The person is thus using the entire group much as he would
a "significant other." Certainly the whole idea of the
occurrence of reference groups, as discussed in chapter
two, appears to be based on the assumption that this view
Maccoby, T. M. Newcomb, and E. L. Hartley (New York: Holt,
Rinehart and Winston, 1958), p. 217.
13Theodore M. Newcomb, Personality and Social
Change (New York: Dryden Press, 1943), p. 45.
*Slartin R. Katz, "A Hypothesis on Anti-Negro
Prejudice," American Journal of Sociology. LI11 (September,
1947), 100-104.
*^Richard Wallen, "Individuals' Estimates of Group
Opinion," The Journal of Social Psychology. XVII (1943),
269-274.
114
of the group as a whole occurs and is important to the
individual.
Two questions from the questionnaire were used to
build a score for this awareness of the group's view or
this part of the "generalized other" phenomenon. The par
ticipant was asked to indicate how sure he thought the
others in his section were of their choices. No points
were given if he replied "haven't any idea." One point
was counted if he checked: "They were so different I can't
tell how sure most of them were of their choices." If he
indicated any idea of how certain the others were (from
"very sure" to "very uncertain") he was given two points
under the assumption that he was perceiving the group as a
whole as being sure or uncertain.
Not only would the Individual using the generalized
other see the group as a whole, but he would see how his
views fitted into the consensus of the group. After the
participant was asked to repeat his original quality choice
he was further asked to "check the number of people in
your section that you think agree that the quality you
have chosen is the most important one." If his choice
was accurate, i.e., if he correctly indicated the number
of people who agreed with him, he was given two points.
If he missed by one person (over or under his estimate)
he was given one point, and an error beyond one in either
direction of his estimate was given no points.
115
The points resulting from the responses to these
two questions (how many agreed with you and how sure were
the others in the group) were pooled to provide a total
"awareness of others* view" score,
A second element in the generalized other concept
was thought to be the ability to take part in a group
under some level of awareness that the group was an inte
grated unit, each person fitting into the total design,
with the person himself fitting into the total picture.
In describing the generalized other George H, Mead is
reported to have said:
The organized community or social group which gives
to the individual his unity of self may be called
"the generalized other," The attitude of the
generalized other is the attitude of the whole
community.
This getting of the broad activities of any given
social whole or organized society as such within
the experiential field of any one of the individuals
involved or included in that whole is, , , , the
essential basis and prerequisite of the fullest
development of that individual's self , . , .1®
The points in these two quotations that should be empha
sized here are: the organized community and experiencing
the social whole or organized society. In an earlier
discussion Mead compares "play" and "the game," In that
discussion he says, "the child who plays in a game must
be ready to take the attitude of everyone else involved in
*^George Herbert Mead, Mind. Self and Society
(Chicago: The University of Chicago Press, 1934;, pp, 154,
155.
116
that game, and that these different roles must have a defi
nite relationship to each other.*7 It would seem that
this perception of the roles and their relationship to
each other are an important component of one's awareness
of the organization within the group, which in turn was a
component of the concept of the generalized other.
Although the period of interaction in the experi
ment was admittedly too short to expect much development
of this picture of an integrated unit to develop for the
members, it was felt that since these people had been used
to interacting with each other they might fall into an
integrated pattern that was based on previous interaction.
Thus they might be able to be aware of this total design
of the section discussion. To measure this possibility it
was decided to use the concept of an integrated interper
sonal role pattern. This was explained to the participants
in the questionnaire in the following way:
Sometimes in group discussions different people take
particular roles. These roles might include such
things as: "suggestion giver," "peace maker,"
"questioner," "disagreer," "expert," etc. Not count
ing the role of section coordinator which was assigned,
can you identify people in your section who played
roles in your discussion?
A variety of possible responses followed this
question, and if the person indicated that there were roles
and he could identify some he was asked to do so. If he
17Ibid.. p. 151.
117
said there were roles played in his section he was further
asked what role he played, was he aware of these roles
during the discussion, and if so how much did they influ
ence him.
A "role awareness" score was devised so the parti
cipant would receive no points if he chose one of the
answers that indicated a lack of awareness of any role in
the discussion, two points if he said there were roles but
couldn*t identify any, and three points if he identified
roles. Then points were added or subtracted in the fol
lowing manner: +1 if he identified his own role, -1 if
he said there were roles but couldn't identify his own,
+1 if he was aware of roles during the discussion, +1 if
the roles influenced him quite a bit, and +2 if they influ
enced him very much. The total points provided a "role
awareness" score for each participant.
Since it was felt that the concept of the gener
alized other included at least these two elements of aware
ness of others' views and a role pattern awareness these
two scores were pooled into a total "generalized other"
score. As with the other measurement techniques in this
study the tools are very crude and may be questioned on
the basis of either theoretical or empirical validity, but
they are intended as an attempt for a beginning of measure
ment into some of these relatively unexplored fields--
however faltering these early steps may be.
118
Perhaps it would be wise to point out here that
the playing of a group role is one of the variables impor
tant to testing the basic hypotheses* The roles for this
purpose were considered to be the randomly assigned group
coordinator role. This role was identified publicly, and
the person carrying out this role was easily identified
by the section members--including the coordinator himself.
The identification of interpersonal roles from the group
discussion is considered to be important to an awareness
of an integrated group pattern but is not considered to be
well developed enough in this experiment to count playing
these roles as having an important effect on internali
zation. Thus in attempting to measure the influence of
playing a publicly identified group role on internalization
the M-3 (who was assigned as coordinator) was compared to
the rest of his section members, none of whom were assigned
roles publicly.
Still another concept which needed an operational
procedure for the testing of this paper's hypotheses was
that of identification. Because of the design of the
experiment there are actually two possible identifications
involved. First there is an identification by the members
with the larger group (i.e., not the sub-sections into
which they were randomly placed), and then there is the
identification with the section in which he took part in
the experiment.
119
For a measurement of his identification with the
larger group the questionnaire included a request for the
participant to mark on a line the place he fitted on the
continuum from "don't like" the larger group to "like it."
Further he was asked to indicate how long he had been a
member which might also be counted as a rough indication
of whether he probably felt he really belonged to the
larger group.
The important identification, however, was felt to
be that of the participant's identification with the sec
tion which was supposed to have influenced (or not influ
enced) the degree to which he internalized his quality
choice. Once again it was felt that a pooling of several
components could achieve a more valid score than a score
based on a response to one question. At the beginning of
the questionnaire (immediately after the discussion situ
ations) the participant was asked: (1) how well he felt
his section did, (2) how much he enjoyed being a member,
and (3) if he were to do it over would he want to do it
with the same people or not.
It was presumed that if he felt his section did
quite well, he really enjoyed being a member, and he would
re-do it with the same people he could be considered to
rate highly in identification with his section. Thus the
total identification-with-section score was an addition of
the points attached to each of these three questions (0-6
120
points for the pride question, 0-6 for the enjoyment, and
1-5 for the re-do question).
Nearly all the methodological techniques described
in this chapter were devised by the researcher. This is
not the most satisfactory approach for getting hard and
fast results. Since these techniques have not been stand
ardized and solidly validated the results which will be
described in subsequent chapters are somewhat suspect
just on the basis of having been obtained by new tech
niques. It would appear that the concepts and hypotheses
which form the basis of this study are not new, but the
tools by which one could test and measure them are so
lacking that the researcher was forced to create some.
Thus, although the new techniques were not desired, they
were almost a necessity.
In the opinion of the author the concepts and
variables of this dissertation are so fundamental to under
standing the relationship between groups and individuals
that they must soon be tested properly if sound knowledge
of this area is to be developed. The author found they
were not easy to pin down in tangible measurable form, but
it was felt that making some progress toward measurement
(even though it would be to find out what doesn't work)
was worth the effort.
CHAPTER V
RESEARCH RESULTS
In general results tending to confirm the experi
mental hypotheses were lacking in all four of the samples.
There were occasional exceptions, and in one or two places
results indicated the reverse of the original hypotheses.
In general, however, the experiments undertaken to locate
and specify a relationship between group factors and inter
nalization revealed that there may be some reason to
question the existence and operation of group factors in
the internalization process.
Hypotheses I and II
Considering the hypotheses in order, the reader
will recall that number I stated: "As situations call for
increasing demands for group action (and consequently in
creasing needs for consensus) concerning a value there
will be an increasing consensus on the values adhered to
by the individuals in the group." By design the experi
ment forced group action differences among the sub-sections
of each group by requiring some not to discuss anything,
other to discuss only, and still others to discuss and
write a joint statement of their views. Thus it was
121
122
possible to consider this variable as a "treatment" dif
ference which could be examined by analysis of variance.
On the other hand, the crude measurements of consensus on
qualities chosen by the section members did not seem to
warrant using techniques that ftssumed an interval scale of
measurement. Thus the type of data and the nature of the
question being asked by the hypothesis seemed to call for
the use of Friedman's analysis of variance which is a non-
parametric analysis.1
For the college sample, which was by far the
largest, the results of comparing the consensus in each
section in terms of the three treatment differences indi
cate a Xr^ of -11.17 which is significant between the
.0029 and the .0013 levels. However, the results proved
the reverse of the hypothesis; that is, the greatest con
sensus was found in the sections where there was no
discussion, the next greatest in the sections where there
was discussion only, and the least consensus where the
most group action was demanded.
In the high school and adult samples the Xr^s were
.3 and 1.5 respectively and were definitely not significant.
Because the nature of this analysis called for comparing a
single consensus score among the three treatment sections
* Sidney Siegel, Nonoarametric Statistics for the
Behavioral Sciences (New York: McGraw-Hill Book Company,
Inc., 1956), pp. 166-172.
123
in each on-going group the number of comparisons was quite
limited. The number of comparisons for the high school
sample was four, and the adult N was three. The number of
senior citizens' groups was so small that there was no
justification for using the analysis for this sample.
On the basis of the data one would have to con
clude that the first hypothesis would be rejected, but
there are definite indications that at least under some
conditions the reverse hypothesis (i.e., that there is
greater consensus when there is no demand for group action)
2
may be true.
With lack of support for the first hypothesis the
theoretical basis for the second hypothesis becomes some
what shaky. The second hypothesis states: "As situations
call for increasing demands for group action (and conse
quently increasing group pressures toward consensus) con
cerning a value there will be greater strengthening of the
individual's internalization of a value." If increasing
demands for group action do not produce greater pressures
toward increased group consensus, there would be little
reason to believe that increasing demands for group action
would create group pressures which would lead to greater
internalization of a value. The results of attempts to
^For a summary of the status of the experimental
hypotheses for each sample and the final conclusions the
reader may find it convenient to refer to Table 10 on
page 145.
124
test the second hypothesis indicate this, generally, is the
case.
Once again the Friedman analysis of variance was
used to compare differences in internalization among the
treatment sections which differed in demands for group
action. The Xr^ for each sample is listed in Table five.
TABLE 5
SAMPLE Xr2 FOR INTERNALIZATION
BY TREATMENT DEMANDS FOR
GROUP ACTION
2
Sample
High School 6.60
College 2.37
Adult .82
Senior Citizen 2.58
In the case of internalization it was not the
college sample but the high school sample that had statis
tically significant results. The high school sample was
statistically significant beyond the .05 level. However,
once again this statistical significance does not mean the
hypothesis was confirmed. This time the results did not
indicate the reverse trend but rather a confusing mixture.
The section which had no discussion ranked the lowest In
internalization, but the section which had discussion and
the group task was next, with the discussion-only section
the highest in internalization.
The number of comparisons made in each test ranged
from fifty for the college sample to six for the senior
125
citizen sample. As before each comparison rated the three
treatment sections, but in this case it was possible to
compare members rather than being limited to comparing
whole sections as was necessary in considering consensus.
The internalization score for member one in section one
was compared to the score for member one in section two
and member one in section three. It should be remembered
that the assignment of individuals to sections and to
member numbers was done on a random basis.
Although the experimental data demand the rejection
of the second hypothesis a full understanding of the re
sults must take into account the technique used to measure
internalization. The experiment was designed to include
several measures of internalization, and thus it was
necessary to determine which meas^ure would be used to test
the hypotheses.
Internalization Measurement
It will be remembered from chapter four that a
total internalization score was to be obtained by pooling
the scores of three presumed components of internalization:
consistency, certainty, and guilt. It was assumed that
there might be some cases where an individual would score
high on one of the sub-variables and still not truly have
internalized the value, but the pooling of the scores
would give a high internalization score where all three
126
component* were present.
To check the internal validity of the total inter
nalization score each sub-variable was tested for corre
lation with each of the other sub-variables. Kendall's
tau from grouped data was the correlation technique used.3
The lack of correlation among these sub-variables was
rather startling. There was almost no correlation between
consistency and guilt. The tau correlation between these
variables for the college and senior citizen samples was
.01, and for the high school and the adult samples it was
.09.
Practically, this meant that people who were con
sistent in sticking to the same quality throughout the
questionnaire would not necessarily feel guilty in changing
their choice of quality after learning that the supposed
"experts" had chosen a different one. Equally the results
would indicate that people who had not been consistent in
sticking to their choice of a quality might well indicate
they would feel guilty if they changed their choice after
learning of the "experts" opinions. There just did not
seem to be much relationship between consistently indicating
a choice of a particular quality and feeling guilty at
3Hubert M. Blalock, Jr., Social Statistics
(New York: McGraw-Hill Book Company, Inc.,1960), pp. 321-
324. As far as the present author was able to determine
there is no test for significance of the tau when the
grouped data technique is used.
127
changing that choice when it was learned that "experts"
had chosen another quality.
The correlations between the consistency and the
certainty sub-variables were generally just a bit higher
among all of the samples. For the high school sample there
was absolutely no correlation between these sub-variables,
but for the college sample the tau was .04; for the adult
sample it was .07, and for the senior citizens it was .09.
In spite of a very slight general improvement in correla
tions, it still must be considered that there is very
little corelation between consistency and certainty as it
was measured in the experiment.
Again these results would seem to indicate that an
individual might consistently stick with a particular
value in preference to other values, but he would not
necessarily feel certain that this value was any better
than the others. Likewise he might indicate he was fairly
certain that this value was preferable to other possible
values, but he would not necessarily be consistent in his
choice of this value.
The possibility of there being low correlation
between the sub-variables might be very great if by far
the greatest numbers fell in the middle of the continuum
on both the variables. That is, if most of the people
were both average in certainty and average in consistency
there might be too few people distributed at the extremes
128
to be able to indicate any correlation between the vari
ables. This does not explain the lack of correlation in
the present study, however, since the distribution pattern
of scores is not over-predominantly clustered in the
center.^ There is a central tendency of course, but it
does not completely dominate the distribution so that the
correlation would be low as a consequence.
The final pairing of the sub-variables produces
considerably higher correlations. The tau for certainty
and guilt was .27 for the high school sample, .24 for the
college, .09 for the adult, and .07 for the senior citizen.
It would appear then that especially in the high school
and college samples an individual who was fairly certain
of his choice of a value would be more likely to feel
guilty if he changed his choice after hearing the "experts*"
views. On the other hand, an individual who was rather
uncertain of his choice would be less likely to feel guilty
if he changed. Of course there were exceptions, but the
results indicate that more often than not certainty and
guilt went together.
Two other techniques for measuring internalization
were tested in the experiment: the Parsons and the Turner
approaches. The Parson approach was tested in two
questions: whether the person would change to the experts*
^See Tables 15-22 in the Appendix.
129
view and whether the person would feel guilty if he did
change. The Turner approach was tested by asking whether
the person would feel surer of his choice of a quality if
he knew the opinions of others. These approaches were
also tested to determine if there was any correlation
between them.
The correlation between guilt and needing to know
others' views were rather sporadic among the samples. The
high school and the adult samples indicated a rather strong
correlation: .35 for the high school and .20 for the
adult. However, the college and senior citizen samples
were much lower: .05 for the college and .01 for the
senior citizen. Hxcept for the high school sample there
was greater correlation between the certainty sub-variable
score and guilt than there was between the single "needinc
to know others' views" score and guilt.
When the Parsons "change to the experts" score was
compared to the Turner "need to know others' views" score
there was stronger and more consistent correlations among
the samples. The tau for the high school sample was .42,
for the college .23, for the adult .19, and for the senior
citizen .11. The individuals who needed to know the views
of others in order to be more sure apparently changed to
the experts' views more easily, and those who didn't need
others would change to experts' views less frequently,
Generally there was stronger correlation between the
nc
Parsons and Turner approaches than between the throe com
ponents of the total Internalization score.
Since the sub-variables of the total internalisa
tion score had such generally low correlations, there
seemed little justification for the theoretical asaiaaptior
that a pooling of these sub-units would produce a score
that would take advantage of the high sub-scores on all
three sub-variables for the people who had really interna
lized the value. With such low correlations it appeared
that a high score on the total internalization sieaeure-
ment might be even more likely the contribution of one
high sub-score rather than the pooling of three high •u-'*
scores. Thus the total internalization score was not used
to test the hypotheses where a measurement of strength of
internalization was required.
From all checks on the internalization rMsiureegrti
the certainty score seemed to be the best single measure*
ment of internalization. Among the correlations hetweer
the three components the certainty-guilt taus were the
highest with the certainty-consistency next. Certainty
was the common component in the two highest matching*.
In the Parsons-Turner correlations the gullt-
"need to know others' views" correlations were lower thar
the "change to experts"-"need to know others* views. • *
Since the change to experts and the need to know others'
views were two of the three elements which produced the
130
Parsons and Turner approaches than between the three com
ponents of the total internalization score.
Since the sub-variables of the total internaliza
tion score had such generally low correlations, there
seemed little justification for the theoretical assumption
that a pooling of these sub-units would produce a score
that would take advantage of the high sub-scores on all
three sub-variables for the people who had really interna
lized the value. With such low correlations it appeared
that a high score on the total internalization measure
ment might be even more likely the contribution of one
high sub-score rather than the pooling of three high sub-
scores. Thus the total internalization score was not used
to test the hypotheses where a measurement of strength of
internalization was required.
From all checks on the internalization measurements
the certainty score seemed to be the best single measure
ment of internalization. Among the correlations between
the three components the certainty-guilt taus were the
highest with the certainty-consistency next. Certainty
was the common component in the two highest matchings.
In the Parsons-Turner correlations the guilt-
"need to know others' views" correlations were lower than
the "change to experts"-"need to know others' views."
Since the change to experts and the need to know others'
views were two of the three elements which produced the
131
certainty score there Is some indication of internal vali
dity for this measurement of internalization.
Because the certainty score had the advantages of
showing some internal validity and the highest correlations
with other sub-variables, it was chosen to be the index
of internalization which would be used to test the experi
mental hypotheses. Throughout the rest of this disserta
tion the "internalization score" refers to the combined
scoring of the subject's responses to three questions:
How sure are you? Would you be surer if you knew the
choices of others? Would you change your answer after
learning the "experts'" views?
Hypotheses 111 through VI
Returning to the testing of the experimental hy
potheses, it will be remembered that the auxiliary hypothe
ses were designed to explain particular group factors which
might help produce the results which were expected under
the first two hypotheses. With neither of the major hy
potheses confirmed it is to be expected that the tests of
auxiliary hypotheses would not indicate that particular
group factors would lead to greater conformity to the con
sensus and to greater internalization. Basically this was
found to be true, but an examination of the results of the
tests of the auxiliary hypotheses may give clues as to
the points where the major hypotheses broke down as well
132
as clues to other points relating to the general area of
study.
Hypothesis 111 stated: "Indivduals who are iden-
tified as having a group role will display (a) more inter
nalization and (b) greater conformity with the consensus
value.*' The reader will remember that the identified
group role in the experiment was the role of group coordi
nator for the section and was assigned on a purely random
basis by telling the M-3 (member number three) in each
section he was to act as coordinator.
To test the differences between the M-3 and the
other members of the section who were not given identi
fiable roles the Wilcoxon Matched-Pairs Signed-Ranks Test
was used.'’ The score for M-3 was compared in each case to
the median score for the other section members. The cal
culation of the median omitted considerations of the M-3
score since the M-3 was assigned a role, and the purpose
of the test was to compare the role versus the non-role
scores. The section median was used in the Wilcoxon test
as a single score to which the M-3 score was compared.
The Wilcoxon test made it possible to compare each
M-3 against the others in his section rather than having
to make some general comparison between all M-3s and all
non-role participants. In a pre-test with a slightly
^Siegel, pp. 75-83
133
different experimental technique it was discovered that
there were great internalization differences between
sections. Therefore, to control for factors which might
produce between-group differences each person with an
identified role was compared directly only to other members
of his section.
A limitation of this technique was that it did not
take into account that the treatment differences among the
sections might have made the assigned role tasks of varying
importance to the group in terms of their different group
needs. However, under the hypothesis being tested the
influence of having an identifiable group role should have
produced internalization differences regardless of the
treatment differences among the sections.
In testing the hypothesis in regard to internali
zation it was found that the high school sample scores
showed the M-3 as generally internalizing more than the
other members in his section. The level of significance
was between .025 and .01. In the other samples the dif
ferences were definitely not significant, and in the adult
sample there even seemed to be a slight tendency for the
M-3 to internalize less.
With regard to conformity it was found that in none
of the samples was the score of M-3 significantly different
from the median score of the others in his section. In the
senior citizen sample there were too few sections to make
134
a comparison. With the adult sample there were so many
ties between the M-3 and the median score that it was im
possible to carry out the Wilcoxon test. For the high
school sample there was a slight tendency for the M-3 to
conform less than the others, and for the college sample
there was an even stronger tendency for the M-3 to conform
less, although it must be remembered that these results
are not statistically significant.
Previously it was mentioned that there were rather
important differences between the sections in terms of
internalization. The uniformity within the sections and
the differences between the sections in the conformity
scores were quite apparent upon inspection of the raw data.
The large number of ties between the M-3 and the section
median for each of the sample Wilcoxon tests further gives
evidence of the great similarities within each section.
Analysis of the data would certainly seem to indicate that
there were other factors that were accounting for confor
mity much more than was the factor of playing an identi
fiable role.
Hypothesis IV states: "As viewing the group as a
generalized other increases there will be (a) more inter
nalization and (b) greater conformity with the consensus
value." The results of the tests of this hypothesis indi
cate that there was a correlation between use of the
generalized other and the internalization and conformity
135
variables in the senior citizen sample, Kendall's tau
between the generalized other and internalization was ,37,
Between the generalized other and conformity it was ,19,
For the other samples, however, the correlations
were quite low, and the college and adult samples even had
low negative correlations between the generalized other
and internalization ( -,03 and -.05 respectively). For
using the generalized other and conformity these two
samples had low positive correlations (.01 and .03 respec
tively). For the high school sample there was low positive
correlations in both instances: ,05 for generalized other-
internalization and .04 for generalized other-conformity.
Prior to the experimentation it was felt that there
might be such low scores on the generalized other measure
ment that it might not lend itself well to statistical
analysis. This did not prove to be the case however, as
may be seen in Table six. A heavier weighting at the lower
end of the scale does skew the distribution somewhat, but
there is rather wide dispersion of scores with a somewhat
surprising number of high scores.
The rather wide range and distribution of the
generalized other scores may be due to the pooling of two
scores which were skewed in opposite directions, however.
It was felt that the generalized other consisted of at
least two elements: an awareness of the group's view and
an awareness of an Integrated group pattern which was
136
TABLE 6
GENERALIZED OTHER SCORES BY SAMPLE
(NUMBERS AND PERCENTAGES)
Score
Sample
Total
No. %
High
School
No. %
College
No. %
Adult
No. %
Senior
Citizen
No. %
High 1
1.1
3 1.5 0 0 0 0 4 1.1
22 24.7 26 12.8 7 18.9 5 14.3 60 16.5
7 7.9 37 18.2 5 13.5 3 8.6 52 14.3
29 32.6 43 21.2 12 32.4 18 51.4 102 28.0
27 30.3 60 29.6 13 35.1 4 11.4 104 28.6
Low 3 3.4 34 16.8 0 0 5 14.3 42 11.5
137
tested by an awareness of roles being played during the
experimental activities. An examination of Table 7 will
reveal that there was generally a rather high awareness of
the group’s view, and Table 8 will reveal that there was
generally low awareness of roles in the sections.
Since the figures in Tables 7 and 8 include members
of all three sections the lack of any activity or discus
sion in section one would obviously tend to make it impos
sible to see roles within the section. This would partially
account for the large numbers at the low end of the role
awareness scale. It is interesting to observe, however,
that the lack of discussion in section one did not prevent
many of the partcipants from having some opinion of the
group's view (Table 7).
The fact that all the members of each section knew
each other from previous Interaction as part of an on-going
group may help account for the fact that they felt they
could make some statement about the views of others. Still
it is unlikely that the larger group of which they were all
members had discussed the question which was posed by the
experiment. Therefore, they probably would not have been
able to have direct knowledge of the views of others on
this particular question.
Although the generalized other score needs more
research and theoretical development, it is used here as
the only available technique to test the hypothesis. On
138
TABLE 7
AWARENESS OF GROUP'S VIEW BY SAMPLE
(NUMBERS AND PERCENTAGES)
Score
Sampie
Total
No. %
High
School
No. %
College
No. %
Adult
No. %
Senior
Citizen
No. %
High 47 52.2 56 27.2 19 33.3 24 63.2 146 37.3
18 20. 51 24.8 15 26.3 0 0 84 21.5
21 23.3 61 29.6 16 28.1 5 13.2 103 26.3
0 0 13 6.3 1 1.8 3 7.9 17 4.4
Low 4 4.4 25 12.1 6 10.5 6 15.8 41 10.5
TABLE 8
AWARENESS OF ROLES BY SAMPLE
(NUMBERS AND PERCENTAGES)
Score
Sample
Total
No. %
High
School
No. %
College
No. %
Adult
No. %
Senior
Citizen
No. %
High 2 2.2 5 2.5 1 2.7 1 2.7 9 2.4
25 27.5 49 24.1 9 24.3 6 16.2 89 24.2
13 14.3 28 13.8 4 10.8 2 5.4 47 12.8
Low 51 56. 12! 59.6 23 62.2 28 75.7 223 60.6
139
the basis of its results one would have to reject the
hypothesis.
The final group factor which it was thought might
influence the internalization process was identification
with the group. Hypothesis V stated: "As identification
with the group increases there will be (a) more internali
zation and (b) greater conformity with the consensus
value." The design of the experiment made it a bit diffi
cult to specify this identification factor. If identifi
cation was to play a part in the internalization processes
going on in the section activities it was felt that the
primary Identification measurement should be identification
with the section. Since all participants were members of
an on-going group it was felt that to a lesser degree iden
tification with the larger group might also further inter
nalization and conformity. Therefore, the hypothesis was
tested with both types of identification.
Across the samples the results indicated an in
creasing tendency for negative, rather than positive,
correlation between internalization and identification with
the section. The high school sample tau was a positive
.10; then the college sample jumped to a -.08. The adult
sample went to a -.11, and the senior citizen sample went
on to a -.13. It would thus appear that for people out of
high school the less they identify with the immediate group
with whom they are interacting the more they internalize a
140
value. Of course this Is contrary to the experimental
hypothesis, but it might help explain some of the other
results obtained in the experiment such as the lack of dif
ferences in internalization between the treatment sections.
With regard to the conformity section of the iden
tification hypothesis the results are almost the reverse.
That is, the tau of the high school sample is -.04; the
college sample moved up to a -.01; the adult sample went
up to a +.07, and the senior citizen sample dropped
slightly to a +.04. None of these are very high correla
tions, but the trend does seem to be the reverse of the
trend found with internalization.
If both of these trends are accurate it would mean
that as one moves from high school groups through older
adult groups identification with the group would produce
increasing conformity (from a negative to a positive in
fluence) and decreasing internalization (to the point that
the group is a negative influence toward internalization).
When the results of comparing identification with
the larger group to internalization and to conformity are
examined the seme trends are somewhat apparent, but they
are not as clear-cut as was true for identification with
the section. Table 9 indicates that there is increasing
negative correlation between group identification and
internalization and a trend toward the positive between
group identification and conformity.
141
TABLE 9
INTERNALIZATION AND CONFORMITY
TAU CORRELATIONS WITH
IDENTIFICATION WITH
THE LARGER GROUP
BY LIKING AND
BY TIME
Identif ication
and:
Sample
High
School
College Adult
Senior
Citizen
Internalization
By: Liking .01 -.05 -.26
vO
0
.
1
By: Time .16 .03 -.22 .13
Conformity
By: Liking
o
o
.
i
.
o
.24 —
By: Time
o
.
t
.09 .12 .21
142
Since internalization decreases while identifica
tion increases with age, one is lead to the conclusion that
any correlation between identification and conformity as a
person gets older will be due to external types of controls
rather than to the inner control that comes with interna
lizing the value of the group. This also causes rejection
of the basic assumptions behind the hypothesis.
The final experimental hypothesis stated: "Identi
fication with the group and viewing the group as a gener
alized other will increase as there is increase in demand
for group action." To test this hypothesis the Friedman
Analysis of Variance was once again used, but this time it
was necessary to compare only treatment sections two and
three. It was felt that the section which did not even
have a discussion could not accurately be labeled a "group,"
and thus it would be unwise to compare it with the others
in regard to the occurrence of "group" factors.
The occurrence of the use of the generalized other
was significantly more common in the sections which called
for group task action in addition to discussion in the
high school sample. The Xr^ was 11.56, significant beyond
the .001 level. However, there were definitely no signi
ficant differences in the other samples, and the adult and
senior citizen samples produced results slightly in the
negative direction, i.e., less use of the generalized
other in the group with the task and discussion.
143
Considering the second aspect of the final hypothesis,
i.e., that an increasing demand for group action will cor
respond with an increasing identification with the section,
it was discovered that both the college and the adult
samples produced significant results in the hypothesized
o
direction. The Xr for the college sample was 4.08, and
for the adult sample it was 4.0. Both were significant
beyond the .05 level. The high school result of .53 was
not significant, and the senior citizen result of .11 was
slightly in the negative direction, but the figure was
almost exactly what might be predicted on a chance basis,
i.e., definitely not significant.
This is the only part of all the experimental
hypotheses that might be accepted on the basis of having
the results of at least two of the samples confirm the
hypothesis. Unfortunately the acceptance of this aspect
of the final hypothesis does not confirm much of the
general theoretical basis for the hypotheses. Even though
the results seem to indicate that an increasing demand for
group action will lead to greater identification on the
part of the group members, other results indicate that
this increasing identification will not increase, and may
even reduce, the internalization processes. The results
also indicate that this increased identification will not
even correspond to a greater conformity by the group
members to the consensus choice of the group.
144
Summarizing the results of the experiment, it may
reasonably be said that they provide practically no support
for the hypotheses, and their almost consistent rejection
of the hypotheses provides a rather strong basis for
questioning the whole theoretical base behind the hypothe
ses. Table 10 gives a review of the status of each hypoth
esis for each of the samples and the final conclusion in
regard to the hypothesis.
A further discussion of the implications of the
results will follow in the next chapter, but it should be
noted here: In so far as one can accept the validity of
the experiment, its techniques, and its results there is
every reason to doubt that group factors will lead to
greater consensus among group participants, greater con
formity bv the participants to a group point of view, or
greater internalization of values by group members.
In addition it may be concluded that: Playing a
group role, using the group as a generalized other, and
identifying with the group lead neither to internalization
nor to conformity. Identification probably increases with
increased demands for group action, but use of the group
as a generalized other does not.
TABLE 10
STATUS OF THE EXPERIMENTAL HYPOTHESES FOR
EACH SAMPLE AND THE FINAL CONCLUSIONS
Hypotheses High School College Adult Sr. Citizen Conclusion
1, Increasing demands
for group action in
creasing consensus
reject reject
(reverse
is true)
reject
m m
reject
11. Increasing demands
for group action -*• in
creasing internaliza
tion
reject
(sig. mixed
results)
reject reject reject reject
111, Playing group
role -4
(a) more internaliza
tion
(b) greater conformity
accept
reject
(trend to
reject
reject
the reverse)
reject
m
reject reject for
adults
reject
IV. Increasing use of
generalized other
(a) more internaliza
tion
(b) greater conformity
reject
reject
reject
(trend to
reject
reject
the reverse)
reject
accept
accept
reject
► —
reject
TABLE 10--Continued
Hypotheses High School College Adult Sr. Citizen Conclusion
V. Increasing Identi
fication with group
(a) more internaliza
tion
(b) greater conformity
reject
(samples mov
reject
(samples mov
reject
e from slight
reject
e from slight
reject
positive to
reject
negative to
reject
negative)
reject
positive)
reject
reject
VI. Increasing demands
for group action
(a) more identification
(b) greater use of the
generalised other
reject
accept
accept
reject
accept
reject
(trend to th
reject
reject
e reverse)
accept for
most adults
reject for
adults
147
Internalization and Other Variables
Although there was nothing in the experimental
hypotheses which accounted for factors of socioeconomic
class, age, sex, or education these variables were impli
citly assumed to have no relationship to the process of
internalization. These variables have proven to be of im
portance in great numbers of sociological investigations,
however, and thus it was considered necessary to see if
there was some relationship between internalization and
each of these variables.
The correlation between internalization and socio
economic class, age, and education was tested through the
Kendall Tau. The chi square technique was used to test the
relationship between internalization and sex since the tau
requires both variables to be on an ordinal scale.
The Hollingshead measurement of socioeconomic class
used in this experiment takes occupation and education into
account, and results of the scores can be grouped into five
socioeconomic classes. When all the four samples were
counted together the distribution of members among the
classes closely followed a normal distribution, as may be
seen in Table 11,
It may well be argued that the distribution of most
populations along the socioeconomic continuum does not nec
essarily follow a bell-shaped curve, but the purpose of the
present study was not to test socioeconomic distributions.
148
The only consideration here is whether socioeconomic class
accounts for any differences in internalization. The
samples included sufficient numbers from each of the socio-
economic strata that all levels could be included in test
ing whether or not a relationship existed.
TABLE 11
SOCIOECONOMIC CLASS FOR
ALL FOUR SAMPLES
Class Level Number
I 46
II 68
III 124
IV 82
V 52
Total 372
The subjects used in this experiment displayed no
correlation between class and internalization. The tau
was .02, and a chi square test on the same data produced a
5.79 which would be significant only at the .95 to .90
level. It is quite clear then that for the subjects used
there was no relationship between class and internalization.
There was a correlation between age and internali
zation. Table 12 shows the age-internalization distribu
tion, and even though there was a heavy concentration of
subjects in the 18-29 age category the tau correlation
149
was .16.
This apparent correlation between age and inter
nalization confirmed the value of considering the four
samples separately throughout the testing of the hypothe
ses since the samples differed rather considerably in age
level.
The chi square for internalization by sex was 1.65
which would be significant only between .5 and .7. Even
though the variable of sex was definitely not significant
the results of this comparison present a rather interesting
trend. In Table 13 it will be noticed that the direction
of difference between the sexes is reversed as it moves
along the internalization continuum. For the "rather low"
and "average" scores there are more males than females, but
moving into the "rather high" and "high" levels of inter
nalization the females increasingly outnumber the males.
Since the results are decidedly not very different
statistically from what might occur on a chance basis, not
much weight can be placed on the seeming tendency of the
females to rank higher than males in internalization.
Still, the trend is consistent in the results produced in
this experiment.
Education was one of the factors considered in the
measurement of social class, although it was weighted less
than occupation in the calculation of the social class
score for each individual. Since social class showed no
150
TABLE 12
INTERNALIZATION BY AGE GROUPINGS
ALL FOUR SAMPLES
Age
Grouping
Internalization
Low
Rather
Low
Ave
rage
Rather
High
High Tota'
1-17 0 4 19 31 14 68
18 - 29 0 1 27 94 103 225
30 - 44 0 0 4 19 20 43
45 - 59 0 0 6 5 17 28
60 - over 0 0 3 9 16 28
Total 0 5 59 158 170 392
TABLE 13
INTERNALIZATION BY SEX
ALL FOUR SAMPLES
Sex
Internalization
Low
Rather
Low
Ave-
range
Rather
High
High Total
Male 0 4 31 69 72 176
Female 0 1 29 87 96 213
Total 0 5 60 156 168 389
151
correlation to internalization it might be expected that
education would do likewise. This is not the case. The
tau correlation between education and internalization was
.18 which indicates some correlation is present. When the
chi square test was run on the same data the results were
43.9 which is significant beyond the .001 level.
From Table 14 it may be seen that the relationship
is fairly consistently in the direction of greater inter
nalization as education advances from 8th grade to college
graduates. The differences between the number with an
"average" internalization score and those with a "high"
internalization score range from -3 for the 8th grade
graduates to +41 for those with 4 or more years of college.
That is, there were three more in the average column than
the high column for the 8th grade graduates, but there were
41 more in the high column than in the average column for
the college graduates.
Summarizing the results of the correlation between
these customary sociological variables and Internalization
it may be said that only the social class variable can be
excluded entirely as having no relationship to internali
zation. While there is no unequivocal basis for saying
sex is related to internalization there are some indications
that it might be. Age appears to be related to internali
zation, and education is even more so. Thus for the
subjects of this experiment it can be said that as age
152
TABLE 14
INTERNALIZATION BY EDUCATION LEVELS
ALL FOUR SAMPLES
Highest Internalization
USTAQO
Completed
Low
Rather
Low
Ave
rage
Rather
High
High Tota
8th grade 0 5 20 35 17 77
12th grade 0 0 6 28 17 51
1 yr.
college
0 0 6 15 28 49
2 yrs
college
0 0 16 42 56 114
4 yrs
or more
0 1 8 36 49 94
Total 0 6 56 156 167 385
153
and/or education increased there was a greater likelihood
that there would be greater internalization of a value.
If group factors played an important part in the
internalization process one would expect that they would
be most apt to show up in the samples with the higher age
participants. Even in this regard the basic hypothesis
of this study is rejected. The results of the study give
no reason to believe that the higher internalization scores
for the older age levels were even slightly a function of
group factors.
CHAPTER VI
INTERPRETATION AND EVALUATION
The statistical evidence from the experiment has
shown almost no support for the postulated hypotheses.
Since the basic hypotheses were devised on the basis of
some rather widely held assumptions in sociology and social
psychology the lack of their empirical support deserves
some careful consideration. If it is to be assumed that
the hypotheses are incorrect there are a number of studies
and theoretical statements which will need to be re-exam
ined. However, it may not be entirely necessary to reject
the validity of the hypotheses on the basis of their lack
of support in the present study.
There are two major possibilities which would ex
plain the lack of support for the experimental hypotheses.
Mention has been made throughout the study of methodologi
cal problems which have plagued the experiment. It is
entirely possible that one could question whether or not
the experimental hypotheses have been accurately tested.
However, if one assumes there has been accurate testing
then one is forced to the conclusion that the hypotheses
themselves are invalid. Both of these possibilities will
154
be considered in this chapter*
155
Methodological Considerations
Methodologically there are three major considera
tions in determining the accuracy of the testing of the
hypotheses: (1) Was the sample unbiased and representative
of the population? (2) Was the design of the experiment
such that the important variables were actually present so
that they could be tested, and (3) were the specific tech
niques used for measurement sufficiently accurate to pro
vide a valid, reliable indication of the presence and
amount of the variables? These issues will be discussed
in turn.
Since the selection of the participants in the
experiments was not a random selection (the groups were not
even selected on a random basis), it may be argued that the
findings should not be applied beyond the boundaries of
those who actually took part in the experiment* It is
admitted that the results do not apply directly to a larger
population from which the subjects were randomly chosen,
but it is not necessarily admitted that the findings are
so limited that they give no implications for the society
as a whole.
First, the four samples covered a rather wide
range of groups as well as a diversity of ages, occupa
tions, education, Interests, and so forth. They were not
156
randomly selected from these larger categories, nor were
they chosen to be truly representative of these categories.
Still, they included people from a rather wide variety of
background factors. This would make it unlikely that they
presented such a bias in sampling that they were unrepre
sentative of the larger society to any important degree.
Second, the results of the study appeared to be so
universally not significant for all the samples which were
analysed separately that it would seem very unlikely that
any large segment of society would produce results which
would prove the hypotheses to a very significant degree.
If the results had not been so consistent one might suspect
that other factors such as differences in the characteris
tics of the participants in the samples might be distorting
the results to produce the variations.
Thus it was assumed that while caution must be used
in where and how the application of the findings might be
used, there is no reason to believe that the results from
this study would be particularly atypical of the larger
society.
One of the moat obvious pitfalls of the experiment
was trying to locate the internalization process. Inter
nalization does not continually occur, and thus it would
be possible to study experimental situations at length
and still not have the internalization process caught in
the net. Since the author knows of no previously devised
157
technique for recording the presence or absence of inter
nalization under situations which were required for this
type of experiment, there is no way to know whether any
internalization occurred in the experimental situations,^
There may be some rather important reasons for sus
pecting that internalization did not occur in the experi
ment, It is commonly assumed that the process of Interna
lization is a very gradual process and as such would neces
sarily involve some considerable amount of time to develop.
Little consideration has been given by authorities to the
actual amount of time that is necessary to accomplish in
ternalization, however.
It may be that internalization can occur rather
quickly under favorable conditions. The experimental situ
ation attempted to induce some of these possibly favorable
conditions such as having already established groups with
ready-built personal relationships and having a task (writ
ing the letter) which would reinforce the beginnings of
internalization. Also, the experiment was not intended to
produce complete internalization, but rather tried to
uncover variations in the degree of It, It was assumed
that internalization was a continuum rather than a point
^The reader will recall from Chapter V that the
only other known experiment where a conventional definition
of internalization was measured was made by Kelman where
actual group situations were not required.
158
and that one might have greater or lesser degree* of it
even without complete internalization.
In spite of attempt* to compensate for the time
factor, the experiment was brief. It may very well be
chat there would seldom be any measurable increase or de
crease in internalization in the forty-five minutes to an
hour and a half that the experiment took. A much more
favorable type of experiment would have been a longitudinal
study in which the group influences could have had consi
derable time to be effective.
Adding a considerable amount of time to permit full
effectiveness of group influences and the increased like
lihood of the occurrence of internalization processes
would have limited the study in other ways, however. For
instance, the observation of effects of having a group role
would have been blurred by the group*s developing and
changing its roles and its role participants. Factors such
as personality qualities rather than purely the factor of
playing a group role might have been uncontrolled, but
important, variables affecting internalization in the role
players. Measuring the use of the generalized other would
have been difficult over an extended time period. Main
taining differences in the demands for group action among
the three "treatment" sections would have been very diffi
cult to stabilize over time.
Many considerations made it necessary to attempt
159
to catch and measure internalization in a short interval of
interaction rather than design a longitudinal study. Since
there seemed to be no proof that internalization could not
occur in a short time span it was decided to try testing
the hypotheses in brief experimental situations. This
decision may easily have resulted in throwing the baby out
with the bath, but at least it is possible to say from the
experiment that in a short time span there seemed to be no
relationship between group factors and individual interna
lization. Perhaps future experiments can be designed to
permit time enough to allow group factors and internaliza
tion processes to get into full swing.
Still another explanation that comes to mind to
explain the possible lack of capturing "internalization"
relates to the general prescriptions of reference group
theory. The whole concept of reference groups demands the
assumption that some groups would be important to the in
dividual in providing normative influence over him and
getting him to internalize its values, norms, etc. This
necessarily means that other groups will not perform this
function for the individual, and thus this kind of group-
individual situation might not produce results which would
be expected to show up for reference groups.
It is possible that the groups chosen did not act
as reference groups for the individuals in them, or even
that there was sufficient mixture among the members of its
160
being a reference group for some and not being a reference
group for others that no group influence would come through
in a study such as the present one. There are some points
in the study that tend to deny this circumstance, however.
The study measured identification with the group (with both
the section and the larger, on-going group), and if a group
were being identified with one would strongly suspect that
it would be a reference group. The findings point to the
fact that even when individuals did identify with the group
there was no significantly greater internalization. When
they did not identify there was no significantly less
internalization.
Kelman does distinguish the identification process
and its results as distinct from the internalization pro
cess and its results. In his study of compliance, identi
fication, and internalization it was shown that the influ
ence of identification (with an individual, not a group)
tends to die out as salience (with the person identified
with) is reduced. If these findings can be applied to
identification with groups as well as to individuals then
it still might be possible to have individuals identify
with the group but not accept the group standards as their
own.
^Herbert C. Kelman, "Compliance, Identification
and Internalization: Three Basic Processes of Attitude
Change," Journal of Conflict Resolution. II (March, 1958),
57.
161
As was discussed in Chapter 11, however, it seems
highly plausible that identification is one of the factors
which might lead to internalization, so one might still
expect an increase in the degree (if not the end product)
of internalization where identification occurred, and this
did not prove to be the case. The author is thus inclined
to feel that the results cannot be satisfactorily explained
on the basis of the lack of occurrence of a reference group.
This view is further supported by the comparison of
the samples. As has been mentioned previously the college
sample was not ideal in terms of providing an on-going
group in which the members were quite familiar with each
other and in which group factors were probably already
established on the basis of much previous interaction.
Considering this, one might make a case that the college
classes were rather unlikely prospects as reference groups.
The high school clubs and the adult groups had a much
greater likelihood of serving as reference groups for their
members. Still the results of the study do not show that
the college classes were very different from the other
kinds of groups in the correlations between group factors
and internalization within individuals. The argument that
the lack of confirmation of the hypotheses is due to the
lack of occurrence of reference groups does not appear to
be very probable.
Another problem involved with the basic design of
162
the experiment is that of the subject discussed.by the
participants in the experiments. A change of subject was
made from the pre-test to the final experimental design in
an effort to find something the participants would become
personally involved in. It will be remembered that a
previous study from which the basic question was chosen
had indicated that "getting ahead" was important to most
of the subjects of that study.^
In spite of this prior information that would lead
one to expect rather great interest in considering factors
which would be Important to getting ahead, the participants
of the present study may not have considered it so impor
tant. For instance, in the original study there was an
indication that getting ahead was considered to be impor
tant for most of the men and not very important for most
of the women.^
The present study did not attempt to measure the
importance of getting ahead for each individual but was
primarily concerned with their choosing a quality which
would be of most value in getting ahead. Even though there
is no evidence to support the view that the participants
of this study did (or did not) feel getting ahead was of
importance in itself, there does not seem to be any
3
Supra, p. 91.
Morris Rosenberg, Occupations and Values (Glencoe,
111.: The Free Press, 1957), p. 95.
163
significant reason to believe that most of them would not
feel it was important or that the qualities suggested in
the experiment would not be of value in the process of
getting ahead.
In attempting to induce internalization it would be
beneficial to produce ego-involvement with the experimental
task and value judgments. Sherif et al. point to two tech
niques to achieve this: selecting subjects who have al
ready shown by action that they are interested and involved
in the question or by adding additional motivation at the
time of the experiment by instructions which would enhance
the importance of the task or similar techniques.^
Neither of these techniques was directly used in
the present study. Although the importance of the experi
ment itself as a piece of sociological research was
stressed, the importance of getting ahead was not stressed
by the experimenter or the experimental design. If one can
assume that the younger age groupings were more concerned
with success and getting ahead than the senior citizens
then any differences in the older group must be considered
in this light. However, the only real difference between
the senior citizen groups and the others was in their
significant correlation between the use of the generalized
^Carolyn W. Sherif, Muzafer Sherif, and Roger E.
Nebergall, Attitude and Attitude Change (Philadelphia:
W. B. Saunders Company,1965), pp.68-70.
164
other and (1) internalization and (2) conformity. Thus it
would seem that the addition of this factor of ego-involve-
ment leads to further rejection of the hypotheses rather
than toward their acceptance.
A further point should be made in regard to the
choice of using "getting ahead" as the focus of attention
in the experiments. In an experiment exploring the use of
reference groups Turner discovered: "Many more respondents
can accept the groups as standard for their ethical-moral
behavior than for 'success' or 'richness of life'."^ Since
the present study involved "success" types of value choices
rather than "ethical-moral" choices, there may be some
reason to suspect that the topic itself was not ideally
suited to promotion of reference group influences on the
individual. However, Turner's study showed little accept
ance of group standards for either "success" or "ethical-
moral" behavior.
Still another reason to question the choice of the
topic was that it probably was not of real relevance for
the group itself. That is, the only reason it was of par
ticular importance to the group to achieve consensus or
^Ralph H. Turner, "Reference Groups of Future-
Oriented Men," Social Forces. XXXIV (December, 1955), 133.
The reader might be interested in Turner's further con
clusion which states that "consistently different patterns,
though short of statistical significance, link ethical-
moral relevance with equality of participation and success-
relevance with subordination within the group." p. 135
165
to arrive at some group action on the topic was because the
researcher imposed certain conditions on the group. If a
topic could have been chosen that would really relate to
the actual goals of the group it is much more likely that
a need for consensus and action would be of greater urgency
to the group. With greater group imperatives might have
come a greater influence by the group over the individual.
This importance-to-the-group factor was recognized
in advance of the experiment, but there seemed to be no
topic that would be equally relevant to all of the groups
that were encompassed in the experiments. In addition it
was felt that if groups acted as reference groups for in
dividuals their influence would not necessarily be limited
to topics that would be of vital concern to the achievement
of group goals. In evaluating the experimental evidence,
however, it must be said that it does not appear that groups
arbitrarily influence individuals over a wide scope of
topics and values. Whether they do influence members in
areas that are of vital concern to the particular group can
not be determined from this experiment.
Summarizing the influence of the topic chosen for
the experimental tasks it might be said that: (1) the
topic probably was of personal interest to most, but not
necessarily all, the participants; (2) if ego-involvement
in the topic of getting ahead is related to younger age
group* the results of the study further support rejection
166
of the hypotheses; (3) there is some reason to question
whether group influences are as great in regard to "success”
values as other types of values, and (4) a topic more
closely related to the goals of the specific group might
have produced more demands for group influence over the
individual.
By implication then, the results of the present
study would not only tend to reject the experimental
hypotheses but to add support to the possibilities: (1)
that groups do not necessarily play an important part in
all areas which are of interest to the individual, (2) that
"success” values may be one of these areas where groups are
not of great influence, and (3) that any topic not of
specific concern to a particular group may put it outside
the range of group influence over the individual.
The above implications are highly speculative and
cannot be said to be directly supported in any measure by
the present study. Still they do offer some possible
variables which might be given special consideration in
future experiments on the basis of the present findings.
Beyond the general design of the experiment there
were problems in more specific areas of measurement. Dis
cussion in previous chapters has shown that the measurement
of internalization was in itself a knotty problem. Any
time a concept must be measured from inference, such as is
the case with motivation, attitudes, internalization, etc.,
167
the techniques used will be highly suspect. The scarcity
of previous attempts at measurement of internalization make
it difficult to draw any definite conclusions in regard to
the accuracy of the techniques used in this study.
Certainly there are reasons to question the validi
ty involved in these measurements. For instance, did the
fact that the questionnaire obviously kept repeating ques
tions concerning the participant's choice of the quality
important for getting ahead influence the way he answered
the questionnaire? Other studies indicate that suspecting
the purpose of an experiment may influence the results
obtained, although the direction of the influence appears
to be uncertain. Sherif et aJL. point to conflict between
the results of studies in this regard.
In fact, some studies show greater resistance to
change of judgment when the subject can easily
infer from the procedures that attitude change is
the topic of research. There are other examples
of greatest change on the part of subjects led
to believe that their performance on a rating
task is of great importance and who surely are
not unaware that repeating the task has something
to do with the investigator's research interest.'
The results of the present study do not seem to
indicate much influence from this particular factor (aware
ness of repetition of the task), however. The consistency
scores in all samples had a fairly normal distribution with
7Sherif, p. 70
168
a quite heavy weighting in the middle range of scores.®
The lack of disguise in the questionnaire does not seem to
have influenced the results in the measurement of consis
tency of response greatly.
In another way the lack of disguise may have in
fluenced the results, but evaluation of the influence is
difficult. When the participants were asked to indicate
whether they would change their views after hearing the
"experts" disagreed with them and whether they would feel
guilty if they did change, the results might have been
different if an indirect technique could have been used
instead. A person might sav they would or would not change
or would not feel guilty, but actual behavior might have
been different.
It would have been possible for the questionnaire
to merely ask the individual to repeat once more his choice
of quality for getting ahead after hearing the experts*
views, but it was felt that most of the subjects would
easily see through this disguise and would answer the same
as if they were directly asked whether they would change
or not. Since some participants (especially in the college
sample) even suspected that the experts* views were rigged,
there is reason to believe that seeing through an effort
to disguise the testing of the influence of the experts
Q
See Tables 15-18 in the Appendix.
169
would not be very effective.
Although there is perhaps some reason to question
the results on the basis of the design of the question
naire, it is generally felt that the design was not greatly
at fault per se. However, there is Bomewhat greater reason
to suspect a lack of congruency between the nominal defini
tions and the operational definitions of the concepts.
There was an attempt to measure the validity of
internalization measurements. The validity of the author's
triple factor approach to internalization (consistency,
certainty, and guilt) was not supported by the results in
general. The lack of correlation between the three factors
would seem to indicate that eithar the measurements were
faulty or that measuring consistency, certainty, and guilt
together will not produce a score that is indicative of the
individual's internalization. The author is inclined to
place heavier emphasis on the lack of relationship between
the operational technique and the concept rather than on
the lack of ability to produce reasonably accurate measure
ments of consistency, certainty, and guilt.
On the other hand, the relationships between the
Parsons and Turner approaches to internalization were
rather high and thus give some weight to their validity of
measurement. It is admitted that these two measurements
have great room for improvement, but the results indicate
that there may have been at least a rough correspondence
170
between the operational techniques and the concept of in
ternalization.
The measurement of consensus is another area in
which there may have been a lack of correspondence between
the measurement technique and the nominal aspects of the
concept itself. Once again it was necessary to measure on
the basis of inference rather than through direct tech
niques. If one could have assumed that consensus must be
present when a group was able to take on and carry out a
task, the job of measurement would have been easier. Since
this was a relationship that was hopefully being tested in
the study, however, it was impossible to use this sort of
technique.
Under the limitations of the experimental design
and the hypotheses of the study, the techniques for measur
ing consensus do not seem to be able to live up to Hopkins'
definition of "normative consensus":
"Normative consensus" contains at least four dimen
sions. There is the normative component, referring
to ideas and sentiments that the participants in a
group (a) are expected as members to hold, express,
and act in accordance with, and (b) have internalized
or "integrated" with the need-dispositions of their
personality systems. . . . There is, third, the com
ponent of consensus, referring descriptively to the
extent to which various sentiments and ideas are in
fact common to participants. In addition, there are
the contents of normative consensus, which may be
described in many ways.9
a
Terence K. Hopkins, The Exercise of Influence in
Small Groups. (Totowa, N. J.: The Bedminster Press, 1964),
P. 11.
171
In a way the study was designed to test whether
norative consensus actually occurs. That is, do groups
necessarily have views the members are expected to hold,
and are these views internalized? The study's operational
definition of consensus did include Hopkins' last two com
ponents to some degree. There was an attempt to include
intensity (or purity) of the consensus view as well as the
number who agreed on the choice of which quality was im
portant for getting ahead. In spite of this, there may be
some question as to the accuracy in calling the counting
of people who agree a real measurement of consensus even
when there is additional consideration of ranking how far
this choice stands out from other choices.
In fairness to this technique, however, it must be
said that many studies have operationally defined consen
sus purely on the basis of the number of people who agree
in the group. The addition of the purity of the choice
attempts to add a little more sophistication and control in
trying to align the nominal and operational definitions.
Similarly the operational definitions of other
concepts such as the generalized other, identification with
the group, and having an identifiable group role are not
presented as being anything other than a crude measurement,
but in all cases it is felt that at least an attempt is
made. Hopefully there will be future efforts which can
provide greater correspondence between the nominal and the
172
operational aspects of these concepts, but for the moment
these seem to be as good a technique as is possible under
the circumstances. In this case one must assume that the
techniques are measuring something similar to the desired
variable and then judge the results accordingly.
While recognizing that there may be many flaws in
the methodology employed in the present experiment and
keeping in mind that all results are at least somewhat
under question on the basis of validity and reliability of
measurement, let us pretend that the results are accurate.
Presuming the accuracy of the findings what kind of impli
cations does this lead to in regard to the theoretical con
siderations? How does one explain the necessary rejection
of the hypotheses of the experiment?
Before plunging deeply into general theoretical
considerations it must first be said that there were occa
sional areas where the hypotheses seemed to be confirmed.
Perhaps these were the rare occasions which fall into the
category of the five or less percent of the times when the
results do prove significant merely by chance. However,
there is a greater likelihood that rejecting the results
as false would be a greater error than accepting them as
true. Thus one must recognize and attempt to account for
these instances.
173
Confirmation of Isolated Hypotheses
The one case where the hypothesis appeared to be
confirmed in most of the samples tested was in relating
the increase in demand for group action in the treatment
section to an increase in identification with the section.
The results of both the college and adult samples confirmed
this hypothesis, while the high school and senior citizen
samples* results rejected it.
The difference in the samples can not be explained
on the basis of differences in identification in the dif
ferent samples.I® In identification with the section the
high school sample revealed almost the same distribution
(percentage-wise) as did the adult sample. The senior
citizen sample had a much higher percentage in the low
identification category than did the other samples. Appar
ently there was little identification here, and the demand
for group action did not alter this condition.
«/hen people in the college and adult samples had
the experience of being in situations demanding greater
action on the part of the group there seemed to be a more
pleasurable and closer association with the group. For
these people the feeling of identification is associated
with the activities of the group itself. Apparently for
the high school and the senior citizen samples
^See Table 28 in the Appendix.
174
identification with the group is related to factors other
than the amount of group activity. One might speculate
that perhaps identification with the group is based more
on the interpersonal relationships among the participants
than on group activities for high school and senior citizen
groups.
Whatever the reasons for the differences in the
samples, it must be noted that even where there was in
creased identification with the group in association with
increased demands for group action there was not a corres
ponding increase in either internalization or conformity
on the basis of the increased identification. When working
with groups it might therefore be possible to increase
identification with the group for most adults by increasing
the action demands for the group, but even then one should
not expect to achieve any greater influence of the group
on what or how strongly the individual will internalize
value s.
The senior citizen sample was unique in showing an
association beteeen increasing use of the generalized other
and both internalization and conformity. This does not
seem to be a function of differences in the use of the
generalized other in this sample.** Neither are these
**Table 6 on p. 136 indicates that this sample is
not very different from other samples except for a
slightly heavier weighting in the middle scores.
175
significant results related to differing demands for group
action. The senior citizen sample results forced rejection
of the hypothesis of that relationship. It may be merely
that for senior citizens using the generalized other im
plies a greater awareness and sensitivity to the world
about them which would also be related to being sure of
one's own views (internalizing) and to being able to con
form to the world about them.
In two instances the results from the high school
sample differed from those of the other samples by confirm
ing rather than rejecting the hypotheses. First, the mem
bers of the high school sections who played identifiable
roles internalized values more than others in their sec
tions. They did not necessarily conform to the group view
more, but the value of the quality chosen was internalized
more deeply. In the other samples playing an identifiable
role was not associated with the individual's internaliza
tion.
The explanation for the high school difference is
not immediately apparent, but it might be that for the high
school sample playing the role did not function so much as
a group factor as a relationship to the person conducting
the experiment. The role which was given the title of
group coordinator consisted more of a liason officer rela
tionship between the section and the researcher. It is
especially likely in the high school sample that this
176
relationship might have provided the awareness of a signi
ficant other (in a power position for the experiment)
rather than a greater susceptibility to group foie influ
ences. This might account very logically for a greater
internalization but no greater conformity to the group by
the role players in this sample.
The second instance of the high school sample con
firming the hypothesis when the other samples failed to do
so was in the relationship between increasing demands for
group action and greater use of the generalized other.
This fits in with a trend toward the reverse in the higher
age groups. The high school sample confirmed the relation
ship; the college showed no relationship between the vari
ables, and in the adult and senior citizen samples there
was a tendency toward a reverse relationship. It will be
remembered that for this test only the two sections where
there was discussion were compared to each other. For the
high school sample having to write a group letter tended
to be associated with greater use of the generalized other,
but this was less and less so with samples of increasing
age.
There is less correlation between group demands for
action and use of the generalized other as one moves from
the high school to the college to the adult samples. Based
on these results one might suspect that the reverse of the
original hypotheses in these cases would be more and more
177
the case as one moved along the age continuum up to the
point of senior citizenship status. Thus an interpretation
of the experimental results in this regard would lead to
the tentative supposition that: As one progresses from
high school into adult status greater demands for action
in membership groups will be apt to be associated with less
use of the generalized other and less use of the generali
zed other mav be associated with slightly greater interna
lization of values. Group needs seem in this instance
almost inversely related to some internal processes within
the individual which might be functional for the group.
Since the age variable seems to play so great a
part in the findings as presented above it is very likely
that an explanation for the results would lie more in in
dividual than in group differences. When each of the
samples was expected to do, and apparently did, the same
sort of experimental assignments it is difficult to see how
group differences could account for this sort of difference
in the results. The answer must then lie in the indivi
duals themselves, or perhaps in how the individuals related
to the group factors and situations.
With regard to the relationship between group de
mands for action and identification this age factor does
not have the same influence. As was mentioned previously
there was a significant relationship between group demands
for action and identification for the college and adult
178
samples, but there was no relationship for the high school
sample or the senior citizen sample. It will be remembered
from Chapter V that the relationship between identification
and internalization showed a trend from positive to nega
tive correlation as one moves up the age continuum. The
relationship between identification and conformity was the
reverse; that is, there was a shift from a negative to a
pesitive correlation as one moves up the age continuum.
With regard to identification then, one might tentatively
suppose that: As one progresses from high school into
adult status greater demands for group action in membership
groups will be more apt to lead to Identification bv the
individual with the groua and the greater the identifica
tion the less the Internalization but the greater the con
formity on the part of the individual.
These two suppositions of group-individual rela
tionships based on differences along an age continuum must
be taken only as very tentative hypotheses which might
arise from analysis of the data from this research. Both
would need much further consideration and investigation
before they could be taken too seriously. Still, the re
sults do show that the presupposed relationships between
groups and individuals do not generally hold up, and some
further explanation is needed.
If it should occur that the suppositions which have
been drawn out of the data analysis are true this would
179
mean that the factors of use of the generalized other and
of identification do not occur and operate in similar ways.
Separate investigation of different group factors is called
for in future investigations. The results of this study
indicate that there may be differences in such factors and
their operation. Considering them erj masse might well
blurr results because of counter influences of the various
factors.
General Theoretical Considerations
It may be concluded that the few instances where
the hypotheses were confirmed point to theoretical consi
derations of some importance, but of far greater conse
quence is explaining the lack of support for the hypothe
ses. Even the few cases where the hypotheses were confirm
ed sometimes seem to indicate a still further lack of
support for the broader theoretical base for this Btudy.
In fact in some instances just the reverse of the hypothe
sis appears more likely.
One explanation for the lack of relationshi'p be
tween group factors and individual internalization might
be that what is called an influence by a reference group
is only the coincidental group membership of several sig
nificant others. Shibutani implies this possibility at
several points in his discussion of reference groups:
. . . the choice of definitions depends upon one's
sentiments toward the significant others who serve
aa representatives of reference groups
180
Although the evidence la far from conclusive, it
appears that many transformations of perspective are
accompanied by a displacement of significant others.
Studies of social solidarity also reveal that con
formity to group norms is related to favorable in
terpersonal relations.12
This view is further confirmed by Rosen in still
another discussion on reference groups. He says:
The reference group, as it is understood by Sherif,
Newcomb, Merton and Kitt, is a collectivity which
consists of significant others, persons of importance
to the individual and to whom he psychologically
relates himself. A major proposition of reference
group theory is that the individual’s attitudes and
inodes of behavior stem from and are related to those
of his significant others, and that this congruence
of attitudes and behavior is a function of the pro
cess of interiorization and legitimation of referents'
expectations.1^
When this definition of a reference group as only
a number of significant others is carried still further,
it can be assumed that there is no real group aspect to a
reference group as far as normative influence.1^ There is
no importance to the fact that significant others are in
12Tamotsu Shibutani, "Reference Groups and Social
Control," Human Behavior and Social Processes: An Inter
act lonlst Approach, ed. Arnold M. Rose (Boston:Houghton
Mifflin Company, 1962), pp. 140, 141, 142. Italics mine.
1 -^Bernard C. Rosen, "The Reference Group Approach
to the Parental Factor in Attitude and Behavior Formation,"
Social Forces. XXXIV (December, 1955), 138.
^Although this considerably weakens Shibutani's
definition of reference groups in terms of the normative
function« Merton's concept of the comparative function of
reference groups may still be valid for groups per se.
Supra, p. 60.
181
groups. In terms of influence over the individual there
would be no basic difference in measuring the force from
groups of significant others and merely adding up the
influence from significant others who are not in physical
or conceptual proximity. The influence of a delinquent
gang over an individual should be equal to the influence
of the same number of equally salient significant others
who might not even be aware of each other's presence.
If one assumes that groups influence the acceptance
of a value by the individual only in so far as they are re
ference groups, and then further assumes that reference
groups are only the occasional occurrence of a . clustering
of significant others in physical or conceptual proximity,
then it becomes necessary to assume there is no group
influence over individuals.
This does not mean that there would be no societal
influence over the individual entirely. The patterning
and operation of society would still limit the range of
conceptual schemes and values with which the individual
might become familiar; it would influence what things might
get results and be useful for the individual so he would
internalize the values involved, and it would affect the
possibilities of contacts with particular significant
others. However, beyond these kinds of societal influences
the demands and needs of the group would not be relevant to
what the Individual accepted as his own values.
182
Age is another factor which may relate to the lack
of group influence if as a person matures there is an in
crease in the need for interpersonal acceptance and identi
fication and a decline in loyalties to conventional group
ings, Berenda found an age difference in types of loyal
ties, although the entire age range of the sample was
younger than that of the present study. She uses the term
"group" to apply both to the school or class as a whole and
to any unit of people with which the individual identifies
(such as "we in the back of the room"). This lack of con
sistent use of the term "group" makes her findings a bit
difficult to interpret for present purposes, but in general
she concluded that the older children (ten years and older)
referred more to individuals (which includes the groupings
of individuals such as "we in the back") than they referred
to groups as a whole such as the class or school,
The results of the present study do not measure
whether a greater need to identify with specific others or
specific sub-groupings of others would lead to greater in
ternalization, The study measured only identification with
externally obvious groups. It might be that identification
with conceptualized sub-units, as in the Berenda study,
might have correlated better with internalization. This
^Ruth Berenda, The Influence of the Group on the
Judgments of Children (New York: King's Crown Press,
1950) , pp." 32, 58-59.
183
further supports the possibility that identification is
related to significant others rather than to organizational
units. If Berenda's findings apply into the adult years
one might expect to find interpersonal influences of more
importance than group influences.
For the sociologist this limitation of group influ
ence may seem a rather stringent restriction. Many psycho
logists have been trying to say this for a long time. It
relates back to the old argument that the group is merely
the sum of the individuals within it which Floyd Allport
and others have been proposing for years. Those who have
been studying small groups have bordered on this view at
several points.
In Personal Influence Katz and Lazarsfeld do not
really come out against group influence, but the following
quotation indicates their hesitancy to accept it whole
heartedly:
We sometimes talk as if people belong to only one
group; or we may imply, at other times, that only
the groups to which an individual belongs influence
his opinions .... we may talk as if we had
established beyond doubt that the individuals take
their standards only from small groups of others
with whom they are personally acquainted and nearer
from people whom they don't know personally or from
mass media. Sometimes, too, we talk as if no other
structure except the informal group exists in the
world and as if no mechanisms of control or sanctions
other than interpersonal influence and ostracism
were operative. Very often, we sound as if all
standards, judgments, values, and ideas which govern
an individual's thinking and acting originate within
the small groups in which they are "anchored"
184
(maintained and enforced). For all these, we plead
over-simplif ication.* 6
As the previous quotation indicates, however, even
when it is recognized that the group influence is not as
simple, clear-cut, or powerful as it is sometimes supposed,
there is a reluctance to say that groups per se do not
influence individuals' standards. In the light of the
results of this study and on the basis of theoretical con
siderations which will be advanced below, the paper will
consider it possible that: Groups per se have only an
external influence over the individual, and Influences over
the internal processes that lead to internalization of
values comes only from evaluation of personal experiences
and from relationships with significant others.
There is some empirical support for the above con
tention. For instance, Turner, who is quite knowledgeable
in the field of reference groups, on at least one occasion
was unable to locate their influence in regard to values
and was forced to say, "Contrary to our anticipation, only
in a small minority of cases were any of the groups rele
vant to any of the values.
Perhaps one of the conditions that has prevented
consideration of the lack of group influence over the
*^Elihu Katz and Paul F. Lazarsfeld, Personal
Influence (Glencoe, 111,: The Free Press, 1955;, p. 64.
^Turner, p. 132
185
internalized values of the individual is the common con
formity by the individuals with the group. In many instan
ces the individual will go along with the group in his
behavior, and this is often over-simplified into a belief
that the individual really holds the values of the group.
In a study by Pelz it was found that members of
high consensus groups tended to act in line with the con
sensus more often than members of low consensus groups, but
this conforming action might be either (1) carrying out the
1 8
action or (2) reporting having done so. These results
would be in line with Kelman** test for showing lack of
internalization. That is, behavior was based on the opera
tion of surveillance and saliency of individuals only and
was not on the basis of internalization of values.
There have been any number of studies indicating
that individuals do yield in many cases to pressures from
groups. The present study indicates that conformity is not
related to using the group as a generalized other, to iden
tifying with the group, or to playing an identifiable group
role. This does not mean, though, that the individual
would not conform in relation to awareness of group sanc
tions and in relation to maintaining self esteem within the
group situation. There is thus the possibility that
^Edith Bennett Pelz, "Some Factors in 'Group
Decision,'" Readings in Social Psychology, ed. E. E.
Maccoby, T. M. Newcomb, and E. L. Hartley (New York: Holt,
Rinehart and Winston, 1958), p. 217.
186
individuals will conform to group standards, but one
should be quite cautious in equating conformity with inter
nalization.
Even the fact that in many groups the individual
members not only conform to an apparent group standard but
give indications of internalizing this standard should not
necessarily encourage one to leap to the conclusion that
the group is responsible for the internalization. The
explanation for this situation could be that individuals
selectively associate with others who have internalized
previous values. It would not be necessary to account for
the internalization as a consequence of participation in
the group.
For those who feel that conformity shows more than
external group influence over the individual there is a
mediating hypothesis which might be considered at this
point. The concept of differences in private and public
attitudes and acceptance of values presents the possibility
that at the public or social level the group values are ac
cepted and fitted into the individual's conceptual scheme,
but at the private level other values are internalized.
Even authorities such as Sherif and Cantril who
hold that it is the “social" values which are dominant
within the individual's ego recognize the possibility of
187
"personal" values.^ Perhaps this distinction of social
and personal values would help account for cases where the
individual's values did match the group's, but it does not
really add much weight to the argument that the group per
se does influence the individual.
Another argument in favor of the view that groups
do not influence internalization by individuals is that
most individuals really don't know what the group's view
is. In an experiment where the subjects were asked to es
timate the group's opinion Wallen pointed to the lack of
correspondence between the estimates and the group opinion.
He was forced to conclude:
The demonstration of this degree of inaccuracy under
conditions highly favorable to accurate judgments
implies that under the ordinary conditions of limited
social interaction found in most communities we
should expect still greater errors.™
It was argued earlier in this paper that even if the views
were inaccurate there still might be an influence if the
individual thought he could perceive the group view. The
results of this study point to a lack of correlation even
between the perceptions of group views and group unity
(i.e., the generalized other) and the internalization of
19
Muzafer Sherif and Hadley Cantril, The Psychology
of Ego-Involvements (New York: John Wiley and Sons, Inc.,
1947), p. 150.
?0
Richard Wallen, "Individuals' Estimates of Group
Opinion," The Journal of Social Psychology. XVII (1943),
272.
188
value s•
If demands for action and coordination with the
group were really influencing the individual one might
expect that the individual would become very sensitive to
the group needs and to how he was fitting in with others
in the group. When individuals demonstrate, as they have
done in a number of studies, that their views of group
opinions are much closer to their own opinions than to the
actually dominant views of the group, then there is rather
great support for the opinion that members are not sensi
tive to group views. The individual is not responding to
group needs for consensus, coordination, and action. Again,
this supports the contention that groups per se have little
or no influence on individual internalization of values.
Perhaps one of the major arguments in favor of a
lack of group influence over the individual arises when the
concept of group consensus is considered. For this study
one of the major reasons for assuming that the group would
influence the individual is the underlying postulate that
groups require consensus to be able to carry out coordi
nated group action. If group action required consensus
then it would be expected that group needs would force
pressures on the individuals to bring their views in line
with group views.
One of the most interesting results of the present
experiment was that different demands for group action did
189
not produce significant differences in consensus. The
letters to the newspapers written by the section as a
whole demonstrated again and again that the group func
tioned quite adequately on a compromise basis with no real
consensus involved. The experimental situation was not
loaded by devising a situation where it was apparent that
consensus would be required for group action. Not only
was this considered to be objectionable for purposes of
fairly testing the hypotheses, but it was considered that
there really are not many situations where group members
are forced to act in complete unity.
On the basis of the results of the present study
it is strongly suggested that the concept of consensus and
its relation to group requirements for action need to be
reconsidered and re-evaluated. The common assumption of
consensus within the group on the basis of the group taking
some form of action may be very precarious if not completely
invalid.
If consensus is not necessary for the group then
the contention that a group need for consensus will create
additional "group" pressures on the individual which will
influence his internalization has very little support.
From the results of the present expeiment it can be shown
that not only did group factors not influence individual
internalization, but the whole notion that groups would be
expected to influence individuals is undermined by the lack
190
of consensus in groups where group action was demanded.
Thus the results of the study provide another theoretical
agrument for a lack of relationship between group factors
and individual internalization.
Perhaps the results of this experiment are an
excellent example of the fallacy of the oversocialized view
of man and the overintegrated view of society which Wrong
haa questioned.
The assumption of the maximization of approval from
others is the psychological complement to the socio
logical assumption of a general value consensus. And
the former is as selective and one-sided a way of
looking at motivation as Dahrendorf and others have
argued the latter to be when it determines our way of
looking at social structure. The oversocialized view
of man of the one is a counterpart to the overinte-
grated view of society of the other.
A final theoretical consideration, which will be
discussed in some detail, is that groups may not provide
a positive influence over individual internalization be
cause groups provide diversity rather than uniformity of
views and externalization rather than internalization.
There seems to be such a preoccupation among social psy
chologists and sociologists with explaining how society
brings uniformity among men, that the possibility of
society*s bringing diversity and externalization is almost
entirely overlooked.
^Dennis Wrong, "The Oversocialized Conception of
Han," American Sociological Keview. XXVI (April, 1961),
190.
191
The concept of a negative reference group is rele
vant, but it really does not handle this possibility. It
merely accounts for the fact that some people reject the
views of conceptualized groups of others and this might
tend to reinforce their acceptance and internalization of
their own values. The position being postulated here is
that factors of group membership and operation of groups
would lead to externalization of values and to greater
diversity of views among individuals in the groups.
One of the few people who have considered the pos
sibility of an externalization process and discussed its
possible operation in society is Rommetveit, He mentions
it in connection with the adolescent revolt against parents
where the boundaries set by the adults coupled with inter
nal, evnironmental, and interpersonal changes in the ado
lescent's life produce a lessening of previously accepted
values.22 ;ith this illustration the possibility of ex
ternalization becomes somewhat more apparent. The possi
bility of intragroup processes and not merely Intergroup
processes explaining divergency seems more plausible.
This possibility of the operation of society in
producing divergency has also been recognized by Brim.
Ordinarily the socialization process is used as an
11
Ragnar Rommetveit, Social Norms and Roles
(Minneapolis: University of Minnesota Press,1954),
pp. 58-59.
192
explanation of how members of a society learn to narrow
their range of behaviors into the acceptable patterns.
Brim explains it in opposite terms. Socialization is the
process that expands, not contracts, the person's varia
bility.
Socialization is successful to the extent that it
prepares individuals to perform adequately the many
roles that will be expected of them in the normal
course of their careers throughout society. It does
this by increasing a person's repertoire of behavior;
extending the range and increasing the complexity of
responses which he has at his command; freeing him
from a limited series of stereotyped responses; pro
viding him with a richer set of discriminations
between various social situations; and proliferating
the specific motives which can be switched into
action by appropriate social stimuli.
In their search for the recurrent and the predict
able in social life sociologists have fallen into what may
be a trap of believing that it is society and social orga
nization that makes man and his behavior sufficiently uni
form to formulate laws of social behavior. It may be that
society and social organization are the very things which
make man's behavior so unpredictable and difficult to for
mulate into laws of social behavior. Subhuman behavior of
animals is much simpler than man's to put into the analyti
cal machine and crank out very satisfactory theories which
hold comparatively high degrees of predictability. Man's
? 3
Orville G. Brim, Jr., "Personality Development
as Role-Learning," Personality Development in Children,
ed. Ira Iscoe and Harold V7.Stevenson(Austin: University
of Texas Press, 1960), p. 138.
193
behavior is not so, and one explanation for this has been
that man is a different sort of being, one that is somehow
innately more variable and unpredictable.
Other explanations of man's variability have em
phasized malfunctioning of the society or social disorga
nization. Wrong points to these explanations in his
counter-emphasis on man's struggle against society:
So long as most individuals are "socialized," that
is, internalize the norms and conform to them in con
duct, the Hobbesian problem is not even perceived as
a latent reality. Deviant behavior is accounted for
by special circumstances: ambiguous norms, anomie,
role conflict, or greater cultural stress on valued
goals than on the approved means for attaining them.
. . . The presence in man of motivational forces
bucking against the hold of social discipline has
over him is denied.24
Might it not also be possible that the very pro
cesses of social organization which characterize the human
environment and the operation of social interaction them
selves produce greater variability in man? Certainly the
concept of division of labor was recognized by Durkheim as
a potential for producing one sort of solidarity in society
while increasing the diversity of the members of society.
Might it not also be possible that conditions which char
acterize the group situation could operate to weaken the
internalization processes and to produce greater diversity
among the members of the group?
The possibility of externalization and nonconformity
^Wrong, p. 187
resulting from group forces should be considered. Without
being conscious of doing so a number of people have pre
sented theoretical possibilities for explaining this kind
of group influence. For instance, Newcomb discusses
attitude change in reference to a balancing of forces. He
explains that attitudes remain unchanged in some cases
because no new forces upset the balance of supporting and
opposing factors. Is it not possible that discussing the
subject of the attitude in a group, especially a group of
close friends, might upset the balance of forces and pro
duce change or externalization?
Newcomb says in other cases the attitude is un
changed because every opposing force is neutralized by
equal supporting forces. If the individual started using
the group as a generalized other, or if he were expected
to play a group role, or if he identified more with the
group might not this add rather heavy weight to one side
of the forces involved in his previous attitude balance?
In the above circumstance, where group forces pro
vided an imbalance, it might be expected that the individu
al would then fall in line with the group view (if there
were a "group" view). However, Newcomb adds another com
ment which points out why greater conformity might not occur
"Ego-defense is a particularly important kind of force
195
against attitude change.”25 Here the group may be seen as
providing forces to off-balance the individual's view, and
in response the individual feels a greater need to maintain
his own identity in opposition to the group.
In experiments which were variations on the Asch
line-matching study Gerard pointed to the greater ego in
volvement in the "group" interaction situation.
There is some evidence in the post-experimental
questionnaire • . . which indicates that compared
with the anonymous condition, the face-to-face con
dition was both more stressful and produced greater
ego involvement.^
As with the case of the adolescent resisting the pressures
of the adult group, the individual is not passive under
group influence, and the operation of group influence over
the individual may produce greater need to not-conform
than to conform.
One of the interesting factors of the Asch line
experiment that is often overlooked in the effort to ex
plain the conformity of subjects is that so many did not
conform. One might explain this in terms of the indivi
dual personality, of his having previously internalized
contrary values or his having contrary reference groups,
but it is also possible that just the group factors of the
25
Theodore M. Newcomb. Social Psychology (New
York: The Dryden Press, 1950;, p. 254.
7 f t
Harold B. Gerard, "Conformity and Commitment to
the Group," Journal of Abnormal and Social Psychology.
LXVIII (1964), 210.
196
experimental situation itself helped produce a sometimes
obstinant non-conformity in the individual. One such
factor might be the face-to-face condition of the group as
Gerard explains:
We propose that the key to understanding this ap
parent lack of an increase in the tendency to yield
to the group consensus over time lies in the nature
of face-to-face confrontation. . . . In a face-to-
face situation an avowal of a discrepant stand is a
public commitment to the group of one's stalwartness. '
In quite a different setting (group therapy), Back
points to the upsetting influence of group interaction:
In order to reduce this tension not created, but un
leashed, by the stimulating group situation, the in
dividual actively seeks supporting figures in the
group whom he will try to entice to behave in a way
that will help him to reduce the activated tension.2°
In the above quotation Back points to the prospect of the
individual seeking out supporting figures in response to
group promoted tensions. Thus it would appear that group
factors may drive the individual into a less stable, more
insecure position. They force the individual from his com
fortable uniformity and his predictable position. In oppo
sition to group diversifying pressures the individual is
forced to seek out significant others who will help him
maintain his stability. Interpersonal ties with
27lbld.. p. 209.
2®George P.. Back, "Observations on Transference and
Object Relations in the Light of Group Dynamics," The
International Journal of Group Psychotherapy. VII
(January, 1957), 67.
197
significant others which might lead to greater internali
zation thus would be in opposition to, not a result of,
group factors.
The theory of group influences producing greater
diversity and externalization has not been given complete
coverage or thorough development here. However, perhaps
enough consideration has been given to point to the possi
bility of this type of group influence over individuals.
Certainly this is one prospect that ought not to be over
looked in future research. It might provide a very fruit
ful approach to explaining why man acts as he does in
society. It might fit neatly into the explanation of a
great many aspects of society and of individuals.
Blumer has pointed to the lack of fit between
theories of organization and explanations of individual
human behavior. Maybe a theory accounting for externali-
zation and divergency within group situations could fit
into the quest offered in the following quotation:
Students of human society will have to face the ques
tion of whether their preoccupation with categories
of structure and organization can be squared with the
interpretative process by means of which human beings,
individually and collectively, act in human society.
It is the discrepancy between the two which plagues
such students in their efforts to attain scientific
propositions of the sort achieved in the physical
and biological sciences. It is this discrepancy,
further, which is chiefly responsible for their dif
ficulty in fitting hypothetical propositions to new
198
7Q
arrays of empirical data.
The present study has failed to prove its theore
tical hypotheses of a positive correlation between group
factors and individual internalization. Neither has it
proved an opposite, negative, correlation. If one can
assume that methodological problems have not distorted the
findings it becomes necessary to hold either that there is
no influence of the group per se over individual interna
lization or that the group may have opposing influences
over the individual. In the light of theoretical and some
empirical support (apart from the present study) for both
the positive and the negative relationships between groups
and individual internalization, the author prefers to hold
that the group factors operate in conflicting ways so that
their combined effect is to neutralize any results which
might appear in a study such as the present one.
Differences in the influence of the group factors
of identification and use of the generalized other in the
present research suggest that further isolation of group
factors may add great insight to the relationship between
the group and the individual. This study has been a start,
but much more needs to be done to provide a strong empiri
cal and theoretical base for knowledge in this area.
CHAPTER VII
SUMMARY AND CONCLUSIONS
The intent of this study has been to determine if
there are conditions of group need and factors which are
present in group interaction which influence internaliza
tion of group values within the individual. Specifically
it was hypothesized that increased demands for action on
the part of a group would lead to greater consensus on
values within the group, and in turn the factors within the
group as it struggled for consensus would influence the
individual to greater internalization.
Building on the assumption of a group influence on
individual internalization, it was further hypothesized
that some of the specific factors within the group which
would help produce greater conformity and internalization
would be: playing an identified group role, using the
group as a generalized other, and identifying with the
group (not just identifying with individuals). Since it
was expected that these factors would relate the group to
the individual it was also expected that the factors would
increase as the demands for group action (and the presumed
demands for consensus) increased.
199
200
To test these hypotheses four samples composed of
various kinds of on-going groups were used. The high
school, college, adult, and senior citizen samples totaled
405 participants. An experimental situation was devised
where there would be differences in demands for group
action as individuals considered the comparative values of
several qualities which might be important for getting
ahead today.
The experiment consisted of: considering the qua
lities and choosing one, writing a brief letter to an
imaginary newspaper editor about what was important for
getting ahead, and filling out a questionnaire. Each on
going group was randomly divided into sections which dif
fered in the following ways: In section one each partici
pant worked individually throughout the experiment; in
section two the participants discussed the qualities and
then worked individually, and in section three there was
discussion of the qualities and the joint writing of one
letter to the newspaper.
vhile the experimental situation created differen
ces in demands for group action and provided for the iden
tified role of group coordinator, the questionnaire was
designed to afford measurement of all the other critical
variables. These variables included: consensus, interna
lization, conformity, use of the generalized other, and
identification with the group. With a few Bfcceptions these
201
measurements were designed by the author in an attempt to
operationalize concepts which have not previously had tech
niques to measure them. Internalization of a value was
the most critical concept to measure and had not had appro
priate operational techniques developed. Three different
theoretical approaches to operationalizing internalization
were tested in the questionnaire.
Although the validity of the operational methods
used is somewhat speculative, the measurement of interna
lization of a value by the three approaches did seem to
validate some conclusions in regard to this measurement.
It was discovered that there was little relationship be
tween consistency of choice of a quality and guilt feelings
if one changed his choice. Neither was there much relation
ship between consistency of choice and certainty that this
was the correct choice of a quality for getting ahead.
There was a little more relationship between being certain
of the value of the quality and feeling guilty if one
changed to another quality that "experts" had chosen. In
general, however, it was determined that there was so
little relationship between consistency, guilt, and cer
tainty that one could not say these three factors would be
present if the individual had internalized a value.
The internalization measurements of whether one
would feel surer if he knew the views of others (i.e., if
he could take-the-role of others) and whether a person
202
would stay by his original choice in the face of contrary
expert view did show correlation. These two approaches to
internalization of a value, based on the opinions of
Parsons and Turner, did seem to have reliability on the
basis of their strength of correlation with each other.
These two approaches had been fitted together under the
concept of certainty of one's valuing a particular quality.
It was decided that this "certainty" measurement should be
used as the index of internalization in the experiment.
3y necessity the development of some of these
methodologies for testing the basic concepts of this re
search became a strong secondary purpose of the study.
Only after good techniques are devised can the theoretical
constructs be validated or revised or rejected. The
results of this study have only made a beginning in some
methodological considerations.
Although the technique for measuring consensus has
not been significantly improved by this study, the neces
sity for reconceptualization has been rather dramatically
emphasized. The need for, and the occurrence of, consensus
probably are much less important than they are described
as being in many sociolgicai and social psychological writ
ings. Certainly the assumption that consensus is necessari
ly present when groups are able to carry out coordinated
action has been challenged by the results of the present
research.
203
Apart from the measurement techniques two major
methodological problems have not been solved by this re
search and may have greatly affected the validity of the
conclusions in regard to the hypotheses. One problem in
volved the nature of the value area being used in the ex
periment. It may be that group influences on individual
values are paramount in certain value areas, particularly
those areas that are vital for the achievement of group
goals. The experiment dealt with values likely to be of
interest to individuals but not essential to the group,
and before the hypotheses are completely rejected there
should be research to test these, or similar hypotheses,
in value areas that are of direct concern to the particular
group.
A second methodological problem that should be
subjected to future research before the hypotheses of this
dissertation are completely rejected is that of the time
factors in the internalization process, Especially when
the techniques for measuring internalization are not devel
oped to the point where one could be sure that it would
distinguish fine degrees of change in internalization, it
might be essential to allow greater time for all influences
toward internalization to take hold. The present study
allowed all too brief an interval of interaction. A longer
time span would have made it virtually impossible to
control other experimental variables, but the decision to
204
compromise on the time factor may have greatly influenced
the results.
In general the results of the research indicate
that the hypotheses should be rejected. There were some
minor exceptions, however. There was an indication in the
college and adult samples that an increase in demands for
group action, which was experimentally induced, would cor
relate with greater identification with the group. In the
experiment identification was largely operationalized in
terms of enjoying the group and wanting to be with the
others in the sroup activity. Thus, enjoyment and liking
seemed to go along with more group activity. For the high
school and senior citizen samples other factors, not measur
ed in the experiment, seemed to be more closely associated
with enjoyment than increased interaction and activity,
Liven in the college and adult samples there was no
increase in internalization of value with increased demands
for activity; so the theoretical foundation behind the
study was not confirmed by this association of activity
and identification. even though participants identified
more as activity increased this did not necessarily mean
the participants internalized values to any greater degree.
When identification was correlated with internali
zation and also with conformity, the results showed an
interesting age difference even though the results were
not statistically significant. The study seemed to
205
indicate that as one's life progressed from high school
onward identification with a group would increasingly be
aseociated with conformity and would decreasingly be asso
ciated with internalization of values. This trend appeared
fairly constant when two different measured of identifica
tion with the group were used, i.e., identification with
the experimental section and identification with the larger,
on-going group.
For the high school sample the playing of an iden
tifiable group role was associated with greater internali
zation. However, it was felt that for these participants
the role probably meant closer association with the re
searcher who might have been a significant other rather than
meaning a different "role" relationship to the experimental
group. Even though the hypothesized relationship between
having a role and the internalization of a value was accep
ted for this sample it was not taken as providing much
support for the theoretical basis of the research.
Apart from these just mentioned exceptions, the
hypotheses of this study seemed not to be proved by the
results. Not only did the direct testing of the hypotheses
fail to confirm them, but more indirect considerations also
failed to provide any support and even made rejection of
the hypotheses of greater certainty. The explanation may
lie in poorly developed techniques or the problems in the
basic design of the experiment as has been discussed above.
206
However, it is just as likely that on a theoretical basis
the hypotheses ought to be rejected.
It had been assumed that because of group needs for
consensus to carry out coordinated action the group would
have some stake in producing conformity and internalization
in the individual. Conformity is not to be equated with
internalization. Still, if the group could produce confor
mity and consensus by producing uniformly internalized
values in the individuals, the group would be able to carry
out coordinated action more efficiently and without devoting
extra energy merely to keeping people in line.
Even though internalization processes are seldom
explained in terms of group organization and activity it is
common for writers to refer to "group values" which the
individual has acquired by his association with the group.
The concept of a reference group has tended to advance the
assumptions that somehow certain groups play particularly
important parts for the individual as he acquires his
system of values.
The results of the present research indicate that
the group per se does not play a great part in the indivi
dual's internalization of a value as it meets its need for
consensus. In fact, the need of the group for consensus
has apparently been highly overrated, and it appears that
at least in certain areas the group can operate very
effectively with practically no consensus.
207
If the group has few if any "common values" it
would not be expected that the individual would be taking
on the non-existent "group" values. Even in cases where
there is some evidence of a reference group the values in
volved are probably not "group" values but values common
to several individuals. It is this factor that makes re
ference groups so difficult to identify since it may be any
conceptually joined individuals whether or not there are
any group characteristics common to them.
Host explanations of internalization are based on
factors within the individual and/or factors of interper
sonal attractiveness or power among individuals. The
results of this study would lead one to believe that this
is where the explanation ends, i.e., the group interaction
does not add any further influences beyond the interper
sonal factors at work in the group setting.
As a final possibility it is thought that the lack
of support for the experimental hypotheses might be ex
plained by the operation of a reverse group influence. It
is considered that participation in group activities may
frequently have the effect of producing diversity and ex-
ternalization of values rather than consensus and interna
lization. Although this is not a widely accepted proposi
tion, there are indications in the writings of others that
support the possibility of this type of group influence.
This is certainly one area that deserve greater
208
consideration and research.
It is not felt that future study will indicate di
versity and externalization of values are the prime results
of participation in groups. Probably there are both inter
nalization and externalization, consensus-producing and
diversity-producing aspects of group influence over indi
viduals. Past research has strongly emphasized only one
side of this scale however. The results produced in this
study indicate group factors do not have a clear-cut, one-
directional influence over the individual.
It is possible that the group has no influence in
the area of internalization, but before this view is com
pletely accepted there should be further research to deter
mine if "no results" means the possibility of conflicting
influences rather then merely "no influence." The task of
formulating an adequate theory to incorporate both possible
group influences and the task of developing means to test
such a theory are left to the future. An adequate, proven
theory in this area should be of utmost value to many areas
of sociology, psychology, and social psychology. As was
anticipated in the introductory remarks in this dissertation
the present study falls far short of solving the problem,
but it is hoped that it has made some theoretical and metho
dological contribution to the beginning struggle.
A P P E N D I X E S
209
APPENDIX A
MATERIALS USED IN THE
EXPERIMENT
210
APPENDIX A
MATERIALS USED IN THE
EXPERIMENT
The Quality Check Sheet
GSM
What quality on this list do you think really gets a person
ahead today? Check one--after careful thought (and after
the »roup discussion if your section has one). Check the
one that you, individually, feel is really the most impor
tant quality for getting ahead today.
Hard work
Pleasant personality
Brains
Knowing right people
Good luck
211
212
Form for the Newspaper Letters
GSM
In the space below write a brief letter to an imaginary
newspaper telling the readers your best advice of what
will get them ahead. Assume that you are not expected to
sign the letter, but the newspaper editor has asked you
to forcefully tell his readers what you feel will get
them ahead today.
213
Questionnaire
Your answers to the following questions will be kept confi
dential* You are not even putting your name on these
papers so they can't be traced back to you except that you
were a certain member in a certain section of this experi
ment .
1. Check the item below that best tells how well you think
your section did in the experiment just conducted:
My section did a very good job.
My section did a fairly good job.
My section did all right.
My section did a slightly below average job.
My section did a fairly poor job.
My section did a very poor job.
I have no opinion.
2. Check the item below that best tells how much you en
joyed being a member of your particular section:
I enjoyed being a member very much.
I enjoyed being a member fairly well.
I somewhat enjoyed being a member.
I didn't enjoy being a member much, but it was
all right.
I didn't enjoy being a member.
1 definitely did not enjoy being a member.
3. If you were to do the experiment over would you want to
do it with the same people or with other people?
(Check one.)
other people, ___ same people, don't care
4. Once more check which quality you think really will get
a person ahead today.
Hard work
Pleasant personality
Brains
Knowing right people
Good luck
5. Check the item below that best indicates how sure you
are that this is really the most important quality:
very sure
fairly sure
somewhat sure
somewhat uncertain
fairly uncertain
very uncertain
Check the number of people in your section that you
think agree that the quality you have chosen is the
most important one:
1 (only you)
2 (you and one other person agree)
3
4
5
6
Haven't any idea
Check the item below that best tells how sure you think
the others in your section were of their choices:
very sure
fairly sure
somewhat sure
somewhat uncertain
fairly uncertain
very uncertain
They were so different I can't tell how sure most
of them were of their choices.
Haven't any idea.
Sometimes in group discussions different people take
particular roles. These roles might include such
things as: "suggestion giver," "peace maker," "ques
tioner," "disagreer," "expert," etc. Hot counting the
role of section coordinator, which was assigned, can
you identify people in your section who played roles
in your discussion? (Check one of the answers below.)
I don't think particular people played such roles
in my section.
Everyone did the same things; there were no roles.
Each person did many things so no one took a par
ticular role.
There were roles, but I can't say exactly which
person played which role.
There were roles, and I can identify some roles
that people played in our section:
_____________________played the role of___________.
(name, or some identification, e.g. color of shirt)
215
played the role of
played the role of
(If you felt there were no roles taken in your section do
not answer number 9,)
9. If you said that you felt there were roles in your
section:
What role did you play? ____________________
Were you aware of these roles during the discussion?
If you answered yes, how much did they influence you?
a little
10. For each of the paired qualities below check the one
in each pair which you think is more important for
getting ahead today.
I*m not sure what role I played.
yes no
very much, quite a bit, somewhat
A hard work
knowing rieht people
knowing right people
brains
G hard work
pleasant personality
D good luck
hard work
£ pleasant personality
good luck
F good luck
knowing right people
G pleasant personality
brains
216
H ___ brains
good luck
1 ___ knowing right people
pleasant personality
J ___ brains
hard work
The following questions are merely to get some idea of the
background of the people who are participating in this
experiment. Remember they are confidential and can not be
associated with you individually.
11. Age (check one) _____ _____ _____ _____ _____
1-17 13-29 30-44 45-59 60-over
12. Sex (check one) _____ _____
male female
13. Education (check the highest grade you have completed)
3th 12th 1 yr. 2 yr. 4 yr.
grade grade college college or more
14. '-hat is the occupation of the person who is the head
of your household:
15. What was the highest grade completed by the person who
is the head of your household:
8th 12th 1 yr. 2 yr. 4 yr.
grade grade college college or more
16. How long have you been a member of this group that is
taking part in this experiment--not the sections but
the whole group:
217
17. On the line below check how well you like being a
member of this group:
Don't like Indifferent Like it
it (Don't care)
Please read the following two stories and check the answer
you think best explains what happened in each of them.
18. Joe worked in a soap factory. He seemed smart enough
and got along well with his fellow employees. Still,
when promotions were given out two other fellows in
his department were given promotions before he was.
One day, however, the boss unexpectedly called Joe
aside and told him he would be advanced to foreman
next week.
What do you think was the most likely reason for Joe's
promotion: (Check one)
The boss thought Joe was a nice fellow all along,
but he had to promote the other fellows first
because of company rules. He promoted Joe just
as soon as he could since he liked Joe so well.
Joe's wife has a brother who had a rather high
position in a company that bought from Joe's
company. The brother-in-law probably put a
little pressure on Joe's company to promote him.
The cards finally fell just right for Joe.
The boss had not previously realized that Joe
took very few coffee breaks and even worked after
hours on his own time sometimes. When Joe's
efforts were called to the boss's attention he
promoted him immediately,
Joe probably used his head to work out a new
technique on his job that saved the company money
so they promoted him.
19. When Jane was ten she decided she wanted to be a
singer. Almost everyone she told about it explained
to her how hard it was to break into this profession.
However, by the time she was eighteen she had signed
a contract with a big record company, was making a
sizeable amount of money, and was receiving all kinds
of offers for singing engagements.
218
What do you think was the most likely reason for
Jane's success:
She probably practiced very hard since she was
ten so that her voice was easily recognized as
superior.
She probably just happened to get a couple of
good breaks that happened at the right times.
she probably has a good agent who saw that she
met the right people.
She was probably such a likeable person that
other people were glad to give her a chance to
get ahead.
She was probably clever enough to figure out and
do just the right things that would make her a
success,
20. In giving your answers to these last two stories do you
feel you would be surer of your answers if you knew
what other people felt the most likely answer was:
(Check one)
would feel much surer if I knew others' views
would feel some surer if I knew others' views
would not feel much surer if I knew others' views
would not feel any surer if I knew others' views
You might be interested in some previous results of the
experiment which was done here in your section earlier.
Previously there was no division into sections, but the
question was part of a questionnaire given to a large
number of vocational counselors and also to quite a few
successful people (all of whom were listed in '/ho's ’ ho).
It is interesting that both groups agreed that the quality
they really believed would get a person ahead today was
hard work.
If their answer didn't agree with yours, do you think you
might consider changing your answer: (Check one)
probably would change my answer
219
might change my answer
might change, but I doubt that 1 would
probably would not change my answer
Pretend that you had changed your answer after hearing that
these other groups did not agree with you. How do you
think you would feel about it: (Check one)
would not feel anything
would feel X had improved my answer by changing
would feel I was justified in changing my answer
would feel guilty if 1 changed my answer because
I would be going against what I had already
marked down
would feel guilty because I would be. changing to
what I feel is an incorrect answer
220
Alternate Form for LaBt
Page of Questionnaire
You might be interested in some previous results of the
experiment which was done here in your sections earlier*
Previously there was no division into sections, but the
question was part of a questionnaire given to a large
number of vocational counselors and also to quite a few
successful people (all of whom were listed in Who's irho).
It is interesting that both groups agreed that the quality
they really believed would get a person ahead today was
knowing the right people.
If their answer didn't agree with yours, do you think you
might consider changing your answer: (Check one)
probably would change my answer
might change my answer
might change, but I doubt that 1 would
probably would not change my answer
Pretend that you had changed your answer after hearing that
these other groups did not agree with you, Tow do you
think you would feel about it: (Check one)
would not feel anything
would feel I had improved my answer by changing
would feel a little uneasy In changing my answer
would feel guilty if I changed my answer because
I would be going against what 1 had already
marked down
would feel guilty because I would be changing to
what I feel is an incorrect answer
APPENDIX B
SUPPLEMENTARY TABLES
221
APPENDIX B
SUPPLEMENTARY TABLES
TABLE 15
CONSISTENCY-GUILT
COLLEGE SAMPLE
Guilt
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
3 15 40 2 2 62
High 0 12 13 0 1 26
Rather
High
0 9 33 2 0 44
R.ather
Low
0 11 12 3 0 26
Low
0 6 9 2 0 17
Very
Low
2 7 14 3 1 27
Total 5 60 121 12 4 202
222
223
TABLE 16
CON S1STENCY-GUILT
HIGH SCHOOL SAMPLE
Guilt
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
0 10 15 1 1 27
High 0 3 7 1 0 11
Rather
High
1 2 4 1 0 8
Rather
Low
0 2 8 3 0 13
Low 0 3 6 3 0 12
Very
Low
0 7 9 2 2 20
Total 1 27 49 11
3 91
224
TABLE 17
CONSISTENCY-GUILT
ADULT SAMPLE
Guilt
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
6 2 16 1 2 27
High 0 1 1 0 1
3
Rather
High
1 0 2 1 0 4
Rather
Low
0 2 5 1 0 8
Low 1 0 5 0 1 7
Very
Low
1 0 5 0 1 7
Total 9 5 34 3 5 56
225
TABLE 18
CONSISTENCY-GUILT
SENIOR CITIZEN SAMPLE
Guilt
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
0 4 4 1 1 10
High 0 2 1 1 0 4
Rather
High
0 1 3 0 0 4
Rather
Low
0 1 4 1 0 6
Low 0 2 6 0 0 8
Very
Low
0 2 2 1 0 5
Total 0 12 20 4 1 37
226
TABLE 19
CONSISTENCY-CERTAINTY
COLLEGE SAMPLE
Certainty
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
High 4 27 65 6 0 102
Rather
High
1 29 43 4 2 79
Average 0 4 15 3 1 23
Rather
Low
0 0 0 0 0 0
Low
0 0 0 0 0 0
Total 5 60 123 13 3 204
227
TABLE 20
CONSISTENCY-CERTAINTY
HIGH SCHOOL SAMPLE
Certainty
Consistency
High
Rather
High
Ave-
rage
Rather
Low
Low Total
High 0 3 14 2 0 19
Rather
High
1 17 00
4 4 44
Average 0 5 15 4 0 24
Rather
Low
0 2 2 1 0 5
Low 0 0 0 0 0 0
Total 1 27 49 11 4 92
228
TABLE 21
CONSISTENCY-CERTAINTY
ADULT SAMPLE
Certainty
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
High 5 1 15 1 1 23
Rather
High
4 2 15 2 2 2 5
Average 0 3 4 0 2 9
Rather
Low
0 0 0 0 0 0
Low
0 0 0 0 0 0
Total 9 6 34 3 5 57
229
TABLE 22
CONSISTENCY-CERTAINTY
SENIOR CITIZEN SAMPLE
Certainty
Consistency
High
Rather
High
Ave
rage
Rather
Low
Low Total
High 0 7 13 4 0 24
Rather
High
0 4 4 1 0 9
Average 0 0 2 0 1 3
Rather
Low
0 0 0 0 1 1
Low 0 0 0 0 0 0
Total 0 11 19 5 2 37
230
TABLE 23
CERTAINTY-GUILT
COLLEGE SAMPLE
Certainty
Guilt
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
44 17 2 0 0 63
High 13 11 1 0 0 25
Rather
High
18 21 5 0 0 44
Rather
Low
7 14 3 0 0 26
Low 5 8 3 0 0 16
Very
Low
14 7 6 0 0 27
Total
101
78 22 0 0 201
I
231
TABLE 24
CERTAINTY-GUILT
HIGH SCHOOL SAMPLE
Guilt
Certainty
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
10 15 2 0 0 27
High 3 5 2 0 0 10
Rather
High
1 3 3 1 0 8
Rather
Low
2 7
3
1 0 13
Low 0 2 7 3 0 12
Very
Low
3 11 6 0 0 20
Total 19 43 23 5 0 90
232
TABLE 25
CERTAINTY-GUILT
ADULT SAMPLE
Guilt
Certainty
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
15 8 6 0 0 29
High 2 1 0 0 0 3
Rather
High
2 2 0 0 0 4
R.ather
Low
2 5 1 0 0 8
Low 1 6 0 0 0 7
Very
Low
3 3 1 0 0 7
Total 25 25 8 0 0 58
233
TABLE 26
CERTAINTY-GUILT
SENIOR CITIZEN SAMPLE
Guilt
Certainty
High
Rather
High
Ave
rage
Rather
Low
Low Total
Very
High
7 2 1 0 0 10
High 4 0 0 0 0 4
Rather
High
2 2 0 0 0 4
Rather
Low
3 1 1 0 0 5
Low
5 2 0 0 0 7
Very
Low
3 1 1 0 0 s
Total 24 8 3 0 0 33
234
TABLE 27
CONSISTENCY BY SAMPLES
Score
Sample
Total
No.
High
School
No. %
College
No. 4
Adult
No. %
Senior
Citizen
No. %
High 1 1.1 5 2.4 2 3.4 0 0 8 2.0
27 29.0 60 29.3 13 22.0 12 29.3 112 28.1
50 53.8 .23 60.0 36 61.0 20 48.8 229 57.5
11 11.8 13 6.3 3 5.1 7 17.1 34 8.5
Low 4 4.3 4 2.0 5 8.5 2 4.9 1 5 3.8
TABLE 23
IDENTIFICATION WITH SECTION
BY SAMPLES
Score
Sample
Total
No. %
High
School
No. %
College
No. %
Adult
No.
Senior
Citizen
No.
High 2 2.2 2 1.0 0 0 0 0 4 1.0
20 21.5 31 14.9 13 22.8 2 5.4 66 16.7
42 45.2 133 63.9 25 43.9 14 37.8 214 54.2
Low 29 31.2 42 20.2 19 33.3 21 56.8 111 28.1
B I B L I O G R A P H Y
235
BIBLIOGRAPHY
Books
Berenda, Ruth. The Influence of the Group on the Judgments
of Children. New York: King's Crown Press, 1950.
An explanation of an experiment very similar to the
line experiment Asch conducted. Here smaller children
in a school setting were used,
Blalock, Hubert M., Jr. Social Statistics. New York:
McGraw-Hill Book Company, Inc., 1960.
A widely used textbook in statistical procedures
and principles. The technique for Kendall's tau from
group data used in the present research was taken
from this text.
Durkheim, Emile. The Elementary Forms of Religious Life,
trans. Joseph Vard Swain. New York: The MacMillan
Company, 1915.
An explanation and description of elementary reli
gion with emphasis on his concepts of a "collective‘l
and of a "collective representation."
Etzioni, Amitai. A Comparative Analysis of Complex Organ
izations. New York: The Free Press of Glencoe, Inc.,
1961.
An ingenious effort to present a middle range
theory of organizations. The major theme concerns
compliance, and other variables are considered in re
lation to compliance within organizations.
Freud, Sigmund. Group Psychology and Analysis of the Ego.
New York: Liveright Publishing Corporation, 1951.
(First published, 1922)
An attempt by Freud to integrate the "crowd" and
"group mind" principles from Le Bon and McDougal with
his own psychoanalytic concepts. In the light of later
thinking in both fields the book becomes an interesting
historical review of antiquated and inadequate thinking
in the area of social psychology.
Havighurst, Robert J. and Bernice L. Neugarten. Society
and Education. 2d ed. Boston: Allyn and Bacon, Inc.,
1962.
236
237
A text covering a variety of sociological views of
education in America with fairly heavy emphasis on
stratification factors. Major divisions of the book
include coverage of general characteristics of the
society and the child*s social environment, the school
as a part of the total society, and the teacher as an
agent in the society and educational institution,
Homans, George Caspar, Social Behavior: Its Elementary
Forms. New York: Harcourt, Brace and World, Inc,,
1961.
A development and discussion of sociological pro
positions concerning "subinstitutional" areas of be
havior; based on a behavioral psychology and elementary
economics approach.
Homans, George Caspar. The Human Group. New York: Har
court, Brace and World, Inc., 1950.
An attempt to build a middle range theory using a
few fundamental concepts with wide applicability.
Basic studies such as those of the Western Electric
Bank Wiring Room, the Norton Gang of "Streetcorner
Society," the Tikopian Family, etc. arc used to explore
a series of hypotheses interrelating: external and
internal systems, activity, interaction, sentiment,
norms, and rank.
Hopkins, Terence K. The Exercise of Influence in Small
Groups. Totowa, N. J.: The Bedminster Press, 1964.
A discussion of the concepts of properties of
status of a group member, i.e., rank, centrality, ob
servability, conformity, in relationship to the concept
of influence in a small group. The important terms are
discussed, their measurement considered, and then a set
of interrelated propositions using the terms in pre
sented, followed by two applications of the previous
discussion.
Katz, Elihu and Paul F. Lazarsfeld, Personal Influence.
Glencoe, 111.: The Free Press, 1955*
A comprehensive analysis of mass media and personal
influence as it directs individual behavior and atti
tudes. Chapter IV, "Norms and Small Groups: The
Shared Character of Opinions and Attitudes," was espe
cially appropriate to the present study.
Mead, George Herbert. Mind. Self and Society. Chicago:
The University of Chicago Press, 19^4.
The handbook for the symbolic interactionist ap
proach in social psychology. This compilation of notes
238
from Mead's classes by various students covers Mead's
emphasis on the social aspects of the development of
the self and the operation of society. It covers
Mead's basic concepts such as "symbolic interaction,"
"taking-the-role-of-the-other," "generalized other,"
"conversation of gestures," and "the nature of
meaning,"
Merton, Robert K, Social Theory and Social Structure.
Glencoe, 111,: The Free Press, 1957.
The basic handbook of Merton's major theoretical
contributions in sociology, including coverage of
manifest and latent functions, middle range theories,
anomie, bureaucratic structure and personality, and
self-fulfilling prophecy. Of most interest to consi
derations of internalization and group factors are his
chapters on reference group theory.
Miller, Daniel R, and Guy F. Swanson. The Changing Ameri
can Parent. Mew York: John Wiley and Sons, 1958.
A study of differences in child rearing practices
and their effects in entrepreneurial and bureaucratic
families. An internalization index based on child
rearing practices is one of the major measurement
tools.
I ewcomb, Theodore M. Social Psychology. Mew York: The
Dryden Press, 1950.
One of the basic textbooks in social psychology.
Generally a symbolic interactionist approach is used.
Attitudes are discussed in considerable detail, and
"frames of reference" play an underlying role in most
areas discussed.
L.ewcomb, Theodore M. Personality and Social Change. Mew
York: The Dryden Press, 1943.
A detailed report of the techniques and results of
Kewcomb's "Bennington Study: of political conservatism
vs liberalism in college girls during their four year
stay at the college--with particular importance attach
ed to reference groups.
Parsons, Talcott. The Social System. Glencoe., 111.: The
Free Press, 1951.
In Parsons' words: "The present volume is an
attempt to bring together, in systematic and general
ized form, the main outlines of a conceptual scheme for
the analysis of the structure and processes of social
systems."
239
Parsons, Talcott and Edward A, Shils (ed.). Toward a
General Theory of Action. New York: Harper and Row,
Publishers, 19^1.
An encompassing coverage of explaining man’s be
havior on the basis of the interrelationship between
the personality system and the cultural and social
systems, Freudian concepts of the individual are re
lated to the components and operation of the social
system. The discussion includes the concepts of
"pattern variables," value orientations, and the system
approach under an action frame of reference.
Reiss, Albert J., Jr. Occupations and Social Status.
Glencoe, 111.: The Free Press of Glencoe, 1961.
A general discussion of the problems and techni
ques of using occupation as an indicator of socioecon
omic class, with special consideration of the NORC
ratings and the development of the Socioeconomic Index
for all occupations.
Rommetveit, Ragnar. Social Norms and Roles. Minneapolis:
University of Minnesota Press, 19 54.
A very deep and explicit consideration of basic
theory of social norms and roles, leading into the
presentation of models on which basic research may be
carried out. A second section presents research that
has been done along the lines of the basic model.
Rosenberg, Morris. Occupations and Values. Glencoe, 111.:
The Free Press, 1957.
An analysis of the results of two comprehensive
surveys on occupations and values with college samples.
The questionnaire provided the basic question for the
present study.
Sargent, S. Stanfeld and Robert C. Williamson. Social
Psychology. 3d ed. New York: The Ronald Press Com
pany, 1966.
A well-known introductory text in social psychology.
Areas covered include: the social setting of person
ality, bases of social behavior, group dynamics, mass
communication and collective behavior, and the psy
chologist and the problem of change.
Sherif, Carolyn W., Muzafer Sherif, and Roger £. Neberfall.
Attitude and Attitude Change. Philadelphia: W. B
Saunders Company, 1965.
An examination of some of the psychological prop
erties of attitudes and attitude measurement. It con
siders some of the often overlooked or poorly handled
psychological aspects of attitude measurement.
240
Although the discussion relates to attitude research
the basic concepts it explores are fundamental aspects
for the development of hypotheses and theory about
attitudes*
Sherif, Muzafer. The Psychology of Social Norms. New
York: Harper & Brothers Publishers, 1936.
A readable classic in the explanation of man in
the social environment. Sherif1s famous autokinetic
experiment is discussed and analysed.
Sherif, Muzafer and Hadley Cantril. The Psychology of
Ego-Involvements. New York: John Wiley & Sons, Inc.,
19477
A rather comprehensive, but somewhat dated, cover
age of a wide variety of factors which are related to,
influence, or result from ego-involvement. The writing
is more philosophically than empirically oriented but
has a rather sound theoretical approach which is
generally still pertinent.
Shibutani, Tamotsu. Society and Personality. Englewood
Cliffs, N. J.: Prentice-Hall, Inc., 1961.
A social psychology text based on the symbolic
interactionist viewpoint. Major sections include:
social control, motivation, interpersonal relations,
and socialization.
Siegel, Sidney. Nonparametric Statistics for the Behav
ioral Sciences. New York: McGraw-Hill Book Company,
Inc., 1956.
A very useful coverage of nonparametric techniques
and their rationale. Statistical methods found in
scattered sources are pulled together in one reference,
and a number of methods seldom discussed in the common
statistics text are presented.
Articles and Periodicals
Asch, S. t., "Effects of Group Pressure upon the Modifi
cation and Distortion of Judgments," Readings in Social
Psychology, ed. E. E. Maccoby, T. M. Newcomb, and E. L.
Hartley. New York: Holt, Rinehart and Winston, Inc.,
1958, pp. 174-183.
A summary of his classic line experiment on confor
mity to erroneous judgments under group pressures.
The coverage includes a general description of the ex
periment, sub-classifications of the "independent" and
"conforming" subjects, and a brief discussion of some
experimental variations.
241
Asch, S. E. "Studies in the Principles of Judgments and
Attitudes: II Determination of Judgments by Group and
by Ego Standards," Journal of Social Psychology. XII
(1940), 433-465.
A report of two studies to discferer changes in
attitude after subjects were: (1) told of their own
first judgments, (2) told the judgments of ficticious
congenial groups, and (3) told judgments of ficticious
antagonistic groups. Generally the results showed the
subjects tended to follow the views of congenial
authorities--more than their own previous views or
antagonistic group views.
3ack, George R. "Observations on Transference and Object
Relations in the Light of Group Dynamics," The Inter
national Journal of Group Psychotherapy. VII (January,
1957), 64-76.
A consideration of processes occurring in group
therapy situations beyond those occurring in therapist-
patient interaction. Some of the group factors include
the interstimulation and "here and now" needs of the
group situation, the mutual support available at a
peer level in the group, the opportunity to react to a
variety of "symbolic triger fingers," and the oppor
tunity for role maturation within the group.
Bac’ onan, Carol, Paul F. Secord and Jerry R. Pierce.
"Resistance to Change in Self-concept as a Function of
Consensus among Significant Others," Sociometrv. XXVI
(March, 1963), 102-111.
The report of an experiment to test congruence be
tween self attitudes and interpretations of attitudes
of others. Pressures to change traits were exerted by
using false personality assessments.
Blumer, Herbert. "Society as Symbolic Interaction," Human
Behavior and Social Processes: An Interactionist
Approach, ed. Arnold M. Rose. Boston: Houghton Mifflin
Company, 1962. Pp. 179-192.
A brief summary of some basic tenents in symbolic
interaction theory followed by a comparison of the
symbolic interactionist's view of society with the
generally more prevalent views within sociology today.
Brim, Orville G., Jr. "Personality Development as Role-
Learning," Personality Development in Children, ed. Ira
Iscoe and Harold U. Stevenson. Austin: University of
Texas Press, 1960. Pp. 127-159.
A very provocative view of socialization as provid
ing for individual variability. Personality traits are
residuals of minor importance in explaining behavior
242
after socialization occurs.
Deutsch, Morton and Harold B. Gerard. "A Study of Norma
tive and Informational Social Influences upon Indivi
dual Judgment," Journal of Abnormal and Social Psycho
logy. LI (1955), 629-636.
A report of a series of experiments based on those
by Asch but with supplementary hypotheses and the use
of differing variables.
Gerard, Harold B. "Conformity and Commitment to the Group,"
Journal of Abnormal and Social Psychology. , LX/II1
(1964), 209-211.
A further discussion of why the nonconformist
sticks to his views, based on the Deutch and Gerard
variation of Asch's experiment which used both face-to-
face groups and (ficticious) anonymous groups.
Glake, Robert R. and Jane Srygley Mouton, "Conformity,
Resistance, and Conversion," Conformity and Deviation,
ed. Irwin A. Berg and Bernard K. Bass. New York:
Harper and Brothers, 1961. Pp. 1-37.
A good survey of studies in the area of conformity.
It considers: descriptions of experimental situations,
a review of conformity studies, studies of conversion
behavior, and future research possibilities and prob
lems .
Harvey, 0. J. and Conrad Consalvi. "Status and Conformity
to Pressures in Informal Groups," Journal of Abnor
mal and Social Psychology. LX (March^ 1960), 182-187.
A study of group pressures on the leader, the
second ranking member, and the lowest status member of
informal groups. The source of influence on the group
was also considered.
Hoijer, Harry. "The Sapir-Whorf Hypothesis," The American
Anthropological Association. LVI, No. 6 part 2 (Memoir
79, December, 1954), 92-105.
An explanation of the Sapir-Whorf hypothesis of the
role of language as a determinant for frames of refer
ence of individuals within a culture. Illustrations
are given from Hoijer*s study of the Navajo.
Katz, Martin R. "A Hypothesis on Anti-Negro Prejudice,"
American Journal of Sociology. LIII (September, 1947),
100-104.
A study comparing children*s own racial attitudes
with perceived attitudes of others.
243
Kelman, Herbert C. "Compliance, Identification and In
ternalization: Three Basic Processes of Attitude
Change," Journal of Conflict Resolution. II (March,
1958), 51-60.
A report of research designed to test Kelman*s
distinction between compliance, identification, and
internalization under varying conditions of surveil
lance, salience, and issue-relevance. The results
support his theoretical distinctions between the basic
concepts.
Kelman, Herbert C. "Processes of Opinion Change," Public
Opinion Quarterly. XXV (Spring, 1961), 57-78.
A rather widely referred to definition and explana
tion of differences and basic ingredients of compliance,
identification, and internalization processes.
Koslin, Bertram. "Preliminary Experiments on the Process
of Norm and Attitude Change During Collective Behavior,"
Dissertation Abstracts. XXIV (September-October, 1963),
12>0-12>1.
A report of an experiment where findings relate to
the formation of norms in the collective behavior
situation and their relation to the previous group
norms and also to the changes in norms in an alone
situation.
Lee, Alfred McClung. "Attitudinal Multivalence in Relation
to Culture and Personality," American Journal of Socio
logy. LX (November, 1954), 294-299.
A discussion of various cultural and societal situ
ations and factors which influence the individual's
shift in attitudes as he shifts roles (multivalence)*
Luchins, A. S. "Social Influences on Perception of Complex
Drawings," The Journal of Social Psychology. XXI (1945),
257-273.
A description of an experiment to determine whether
hearing a previous response by another subject would
influence the critical subject to primarily perceive
the mentioned object in complex drawings of varying
ambiguity.
Maccoby, Eleanor E. "Role-Taking in Childhood and Its
Consequence for Social Learning," Child Development.
XXX (1959), 239-252.
A theoretical statement of role training through
role-taking in relation to frequency of interaction
and power of the other over ego.
244
Mead, Margaret, "Social Change and Cultural Surrogates,"
Personality in NatureP Society, and Culture, ed, Clyde
Kluckholn and Henry A, Murray, New York: Alfred A.
Knopf, 1956, Pp. 651-662,
A comparison of American patterns of socialization
and internalization with those of other cultures and
a discussion of the relationship between these pat
terns and social change,
Mussen, Paul and Luther Distler. "Masculinity, Identifi
cation, and Father-Son Relationships," Journal of
Abnormal and Social Psychology. LIX (November, 1959),
350-356.
A report of an experiment which compared masculine
identification factors in homes where the father was
present and homes where the father was absent much of
the time.
Nelson, Harold A, "A Tentative Foundation for Reference
Group Theory," Sociology and Social Research. XLV
(April, 1961), 274-280.
An attempt to specify six "base points" for the
construction of a real theory of reference groups.
Pelz, Edith Bennett. "Some Factors in 'Group Decision,'"
Readings in Social Psychology, ed. E. E. Maccoby, T. M.
Newcomb, and £. L. Hartley. New York: Holt, Rinehart
and Winston, 1958. Pp. 212-219.
A report of an experiment to further test four
hypotheses based on earlier experiments where group
discussion versus the lecture method were compared for
effectiveness in changing behavior.
Rosen, Bernard G. "The Reference Group Approach to the
Parental Factor in Attitude and Behavior Formation,"
Social Forces. XXXIV (December, 1955), 137-143.
A report of an experiment studying the relationship
between adolescent and parent religious (Jewish) atti
tudes in terms of reference group factors in an attempt
to explain some causal aspects of the well-known fact
of similarity between attitudes of parents and children.
Shibutani, Tamotsu. "Reference Groups and Social Control,"
Human Behavior and Social Processes: An Interactionist
Approach. ed. Arnold M. Rose. 3oston: Houghton
Mifflin Company, 1962. Pp. 128-147.
A capsule version of some of the concepts and as
sumptions discussed in his social psychology text and a
re-working of an earlier reference group article.
245
Spaulding, Charles B. "Social Class and Social Percep
tion." Sociology and Social Research. XLI (Seotember-
October;'1956)? 18-26. ------
An argument for the case that most people have no
clear-cut ideas of the class system in America.
Turner, Ralph H. "Reference Groups of Future-Oriented
Men," Social Forces. XXXIV (December, 1955), 130-136.
An analysis of one section of a larger research
project to explore two aspects of reference group
influences: (1) the relevance of different reference
groups in different areas of values and (2) the inter
relationship between the importance of the group as a
reference group and participation variables.
Turner, Ralph H. "Value-Conflict in Social Disorganiza
tion," Sociology: The Progress of a Decade, ed.
Seymour M. Lipset and Neil J. Smelser. Englewood
Cliffs, N. J.: Prentice-Hall, Inc., 1961. Pp. 522-527.
Basically the article is a reconceptualization of
the concepts of value conflict as it relates to social
disorganization. Social disorganization as value
conflict does not necessarily mean logical incompati
bility of values, but that values support contradict
ory behavior patterns.
Wallen, Richard. "Individuals Estimates of Group Opinion,"
The Journal of Social Psychology. XVII (1943), 269-274.
A short account of a study of the accuracy of
students in estimating the number who would agree or
disagree with their views.
Whiting, John r. r . M. "Resource Mediation and Learning by
Identification," Personality Development in Children,
ed. Ira Iscoe and Harold W. Stevenson. Austin: Uni
versity of Texas Press, 1960. Pp. 112-126.
A theory of role learning through role-taking
similar to, but more developed than, Maccoby's
approach (see above).
Wrong, Dennis. "The Over-Socialized Conception of Man,"
American Sociological Review, XXVI (April, 1961), 185-
193.
An account of, and questioning of, how sociologists
have answered the Hobbesian question of the relation of
man to society. Wrong asks for a reconsideration of
the Freudian view that man is social but not entirely
socialized.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The Major League Professional Baseball Player: A Sociological Study

PDF

Class Consciousness And Social Mobility In A Mexican-American Community

PDF

Social Factors Related To Dentistry As A Career

PDF

A Critique Of The Concept Ethnocentrism As Set Forth In Selected Social Science Literature

PDF

Perception Of The Power Structure By Social Class In A California Community

PDF

Reference Group Theory, Selection, And The Images Of Professions

PDF

Normative values of selected law enforcement officers and adult male offenders

PDF

Time And Man-Nature Value-Orientations As Related To Sex Identity, Ethnicidentity, And Socioeconomic Status And As Predictors Of Academic Success

PDF

An Empirical Examination Of The Relationship Of Vertical Occupational Mobility And Horizontal Residential Mobility

PDF

Complementarity, Homogeneity, Heterogeneity And Marital Stability

PDF

Some Social Factors Affecting The Power Structure And Status Of A Professional Association In Reference To Social Work

PDF

Interpersonal Relations In Ethnically Mixed Small Work Groups

PDF

The Career Business Executive As A Definitive Occupational Type

PDF

Social Stratification In A Selected Community

PDF

Social Class Membership And Ethnic Prejudice In Cedar City

PDF

Organizational Capacity For Change And Adaptation: An Exploration In A Public Research And Development Organization

PDF

Familism, Suburbanization, And Residential Mobility In A Metropolis

PDF

Consensus Of Role Perceptions In A Welfare Planning Council

PDF

Middle-Class Marital Roles - Ideal And Perceived In Relation To Adjustment In Marriage

PDF

Academic Specialization And The Construction Of Social Reality: Ideologies Regarding Deviance

Asset Metadata

Creator Roberts, Martha Anne (author)

Core Title Group Factors And Individual Internalization Of A Value

Degree Doctor of Philosophy

Degree Program Sociology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,Sociology, general

Format dissertations (aat)

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Lasswell, Thomas E. (committee chair), Martin, David W. (committee member), McDonagh, Edward C. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-149004

Unique identifier UC11360129

Identifier 6713760.pdf (filename),usctheses-c18-149004 (legacy record id)

Legacy Identifier 6713760.pdf

Dmrecord 149004

Document Type Dissertation

Format dissertations (aat)

Rights Roberts, Martha Anne

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA