SOCIAL SELF AND THE SOCIAL DESIRABILITY
MACK R. HICKS
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The writer wishes to express his appreciation to
Dr. Audrey S. Schumacher, Dr. Ben Barger, Dr. Hugh C. Davis, Jr.,
Dr. John R. Barry, and Professor Mell H. Atchley for serving on
his supervisory committee.
Dr. Harry C. Anderson, Jr. and Dr. Lorna S. Benjamin
offered valuable suggestions for the design and statistical
analysis of the study. Special thanks also to Dr. Norman S.
Greenfield for arranging for the use of subjects at the
University of Wisconsin. The writer is indebted to the
Wisconsin Psychiatric Institute for the preparation of the
manuscript and the statistical analysis of data.
The writer is particularly indebted to Dr. Schumacher,
chairman of the committee, who devoted considerable time in
helping the writer to develop and clarify theoretical concepts.
TABLE OF CONTENTS
ACKNOWLEDGMENTS. . . . . . . . ... ... iii
LIST OF FIGURES . . . . . . . .... . vi
LIST OF TABLES . . . . . . . . .. ... . vii
INTRODUCTION . . . . . . . . . . . .
BACKGROUND . . . . . . . . . . . .
-Tie Concept of Self . . . . . . . .
Theory . . . . . . . . . . .
Research . . . . . . . .... . 6
Tr,e Concept of Social S If . . . . . . .
Theory . . . . . . . . ... .. .
Research . . . . . . .. . . 12
Anxiety and Dynamic Structure of the Self. ... 13
Stages of Development. . . . . . .. 1
A Theory of D velopment. . . . . . .. 14
Concept of Social Desirability . . . . .. 24
SUMMARY AND STATEMENT OF HYPOTHESES. . . . . .. 0
METHOD . ..... ... ... . . . . . . .
Subjects . . . . . . . .... . .. 7
Procedure and Instruments. . . . . . . .
Tne Self-Rating Muthodology. . . . . ... 8
Position and Sequence. . . . . . ... 40
The Social Situation . . . . . . . 40
Te Private Situation. . . . . . ... .
The Picture Identification Test. . . . ... 41
Statistical Methodology. . . . . . .. 43
Experimental H potheses. . . . . . . . 4
RESULTS . . . . . . . . . . . . 47
The Analysis . . . . . . . .. .. . 47
Findings . . . . . . . . . . . 417
High tied for Social Approval Group. . . . 52
Moderate Need for Social Approval Group. ... 53
Lcw N ,ed for Social Approval Group . . . 54
Convergent Versus Divergent S.If-Ratings . . 55
Discussion of Findings ............... 57
High Need for Social Approval Group . . . 57
Moderate Need for Social Approval Group . 59
Low Need for Social Approval Group . . .. 62
Convergent Versus Divergent Self-Ratings .. 63
Questions and Inferences Related to these Findings 64
SUMMARY . . . . . . . . .. . 70
BIBLIOGRAPHY . . . . . . . . . 73
APPENDIX ..... . . ........... . 76
BIOGRAPHICAL SKETCH . . . 79
LIST OF FIGURES
1. Topographical model of the self-system representing the
absolutism level of self-development. . . . . ... 19
2. Topographical model of the self-system representing the
relativism level of self-development. . . . . ... 21
3. Topographical model of the self-system representing the
inconsistent level of self-development. . . . .. 23
LIST OF TABLES
1. Mean scores and standard deviations for subjects on
the Marlowe-Crowne and Edwards Social Desirability
Scales .................. .. . 25
2. Correlations between scores on the Marlowa-Crowne
and Edwards Social Desirability Scales and scales
on the Minnesota Multiphasic Personality Inventory 25
3. Correlations of scores on two Social Desirability
Scales and the Barron Conformity Scale ...... 27
4. Analysis of variance . . . . . . . . 48
5. Ratings of pictures (as to likeability in social
and private situations) by subjects grouped accord-
ing to need for social approval. . . . .. 49
6. Statistical analysis of experimental predictions .50
7. Average scores of all groups on the Marlowe-
Crowne Social Desirability Scale . . . ... 52
8. Average divergence between self-ratings in social
and private situations for discrepant convergent
groups . . . . . . . . . . . 5
9. Average ajg and education of subjects grouped
according to scores on the Marlow3-Crowne Social
Desirability Scale . . . . . . . ...
10. Average age End education of subjects. . . . 7
11. Average salf-avaluations of subjects in the social
and private situations . . . . . . . 7
12. Average scores of subjects on the Harlowo-Crowne
Social Desirability Scale. . . . . . . 75
13. Number of Wisconsin and Florida subjects in
social desirability groups .......... . 77
14. Average self-ratings of Florida and Wisconsin
subjects grouped according to need for social
approval, . . . . . . . . .. .77
15. Self-ratings of all subjects grouped according
to need for social approval and situation . . .78
There is little question that the self-concept has been a
useful construct in clinical service and in the construction of
personality theories. The writings of various psychological and
social psychological theorists attest to this fact (29,22,19,20,15,9).
There has also been a large quantity of experimental research in
recent years (32).
While the evidence indicates that this is a profitable area
of research, unequivocal support for the utility of this construct
In the prediction of human behavior has not been established
experimentally. It is felt by some (32) that these Inconsistent
findings have resulted because of premature attempts to relate
the self-concept variable to other complex variables, without first
exploring and rigorously defining the structure of the self-concept.
The present study Is directed towards the solution of this
problem. It constitutes an attempt to explore experimentally the
relationship between two important aspects of the self-structure.
The social aspect of the self has intrigued theorists for a
number of years (17). Mead (19) and Piaget (6) have attempted to
relate the social aspect of the self and, to a lesser extent, the
self-concept in general to the incorporation of sociocultural
attitudes and roles. For this reason, the social aspect of self
would seem to be a logical point of entry into the study of the
relationship between personality and cultural attitudes. Sullivan
(29) has utilized the construct of interpersonal anxiety in
explaining how these attitudes are incorporated into the individual.
An attempt will be made, in this paper, to relate the constructs
of sociocultural roles and interpersonal anxieties to aspects of
More specifically, the social aspect of the self-structure
and the private phenomenal aspect of the self-structure will be
related to the social-desirability variable. The degree of
divergence between these two aspects of self, the absolute dif-
ferences between the two, the direction of differences, and
individual variability within the two aspects will all be related
to degree of need for social approval.
The question is: will the individual who finds it necessary
to play inflexible and conflicting social roles, as inferred from
a strong need to behave in a socially approved manner in many
different situations, have a greater divergence between the private
phenomenal aspect and the social aspect of his self-structure than
the individual who does not find it necessary to adhere absolutely
to these roles.
In addition, an attempt will be made to relate the divergence
and convergence of these aspects of self to psychological health,
with the need for social approval held constant.
Since the concept of self provides the general framework
for this study, theoretical formulations and research findings in
this area will be considered. The other main topics of Interest,
which will be discussed in the remaining sections of this chapter,
are: the social aspect of the self-structure, the concept of
social desirability, and the dynamic structure of the self as
it develops through experiences with anxiety. This review will
consider theory and research in these four areas and move towards
the theoretical formulation which underlies the present research.
The concept of self
Psychologists are funny fellows. They have before
them at the heart of their science, a fact of perfect
certainty, the one warrant for the being of all other
things and yet they pay no attention to it (2).
This statement, according to Allport (2), is the kind of thing
a layman might ask the psychologist about the concept of self. It
is true that until recently there has been little work done on
investigation of the self.
The concept of self has been of great concern since early In
the history of philosophy and psychology. More recently, Herbart
(14) stated that man conceives of himself as part of the subjective
world and depending on the situation may see himself as figure
or ground. Theodore Lipps (14) proposed that man has knowledge
of things, own self, and other selves. Lipps is best known
for his theory of empathy. A more recent point of departure,
and one that is often referred to by current theorists in this
area, is found in the writings of William James (15).
In his book, The Principles of Psychology (15), James devoted
a chapter to the consideration of self. James proposed four
constituents of the self: the material self, the social self,
the spiritual self, and pure ego.
That there are many different conceptions of the self is
evidenced by Allport's "The Ego in Contemporary Psychology" (2).
In this paper, Allport considers eight different conceptions of
the self: first, the ego may be considered as a knower. As he
points out, this concept was given the coup de grace by James
in The Principles of Psychology. Second, the ego may be viewed
as an object of knowledge. This Is concerned with the problem of
the nature of our experience of the self. Third, the ego is viewed
as primitive selfishness. Fourth, the ego may be viewed in terms
of dominance, drive, status, and recognition. Fifth, the ego may be
viewed as a passive organization of mental processes which con-
ciliates warring forces (i.e., Freud's usage). The ego may be
viewed as a fighter for ends. This view is represented by McDougall,
Koffka, and Goldstein. Also, the self has been considered to be a
behavioral system by Koffka and Lewin. Finally, the ego may reflect
the subjective organization of culture as proposed by Sherif and
Cantril. This last concept eradicates the distinction by Freud
of ego and superego.
Symond's (30) concept of self-as-process develops and executes
a plan of action for attaining satisfaction In response to inner
drives. The self-as-object Is how a person perceives himself;
what he thinks of himself; how he values himself; and how he
attempts to enhance or define himself.
Snygg and Combs (8) do not differentiate between the self as
an object and doer. The phenomenal self is differentiated out of
the phenomenal field and is all those parts of the field which the
individual experiences as part or characteristic of himself.
Sherif and Cantril (27) also propose no construct of ego but
have a self-as-object which they call the ego. It is a constellation
of attitudes which somehow motivate behavior.
Sarbin (25) presents no ego, but presents a self which is a
cognitive structure about various aspects of the individual's
being: conceptions of the body (somatic-self), conceptions of
the sense organs and musculature (receptor-effector-self), and
social behavior (social-self).
Hilgard (14) conceptualizes no ego and the self is defined as N
one's image of himself. It cannot be studied by asking the subject
because of unconscious factors. Rather, he suggests using projective
techniques to get a picture of the inferred self. Self-as-process,
he says, comes from the layman's mistaken belief that his behavior
is completely determined by him.
G. H. Mead (19) conceives of no ego but states that many
selves may develop, each of which presents a more or less separate
group of responses acquired from different social groups. At <
first there is no self because the person cannot enter his own
experiences directly. He becomes a self insofar as he can take
the attitude of another and act towards himself as others act.
Hall and Lindzey (14) feel that the term self has come to
have two distinct meanings: self-as-object and self-as-process.
Self-as-object is defined as the person's attitudes, feelings,
perceptions, and evaluations of himself as an object; what a
person thinks of himself. They call this the self. They define
self-as-process as a doer. It consists of an active group of
processes such as thinking, remembering, and perceiving. They
call this the ego.
The self-concept as defined In this paper is similar to Hall
and Lindzey's construct of self-as-object (14). It is a hypothetical
construct which represents the individual's conception, perception,
feeling (affect), and evaluation of his own psychophysical processes
as an object.
Research. -- An enormous amount of self-concept research has
been reported in recent years (2). One might wonder why psychologists
have felt the need to postulate inferred variables referring to the
For one thing, psychologists of a number of schools
of thought have noted that antecedent conditions, defined
in terms of interexperimenter agreement, are not sufficient
to predict either group trends or Individual differences
in human behavior. They have suggested that one could
increase the accuracy of predictions of behavior if one
found out what the subject perceives, knows, or feels
about the "objective" situation, including his own
characteristics. Personality theorists have also pointed
out that general behavior theorists, for purposes of their
own, have thus far delimited their theories in such a
way that they are unable to account for some of the
behaviors one can observe In the clinic, in school, and
in other "every day life" situations. In addition,
many personality theorists have felt that the organizational
or configurational properties of human functioning are
not subsumed by the constructs of most present day general
behavior theories (32, p. 318).
Wylie (32) states that while constructs concerning the self may
be needed for the above reasons, the way they have been used creates
another problem. That is, these self-constructs have been stretched
to cover so many cognitive and motivational processes that their
utility has been greatly reduced.
The result Is a good deal of ambiguity In the Interpretations
of various studies and considerable apparent contradictions of results.
"in short, the total accumulation of substantive findings is dis-
appointing, especially in proportion to the great amount of effort
which obviously has been expended" (32, p. 317).
Two main approaches, which may offer a solution to this
dilemma, are suggested by Wylie (32, p. 319). One is to improve
the constructs and hypotheses through the use of more carefully
delineated definitions of variables. The second approach, which
Is also suggested as an alternative to abandoning self theories, is
to improve their predictability by the addition of more variables.
The present study is directed towards the solution of this
problem. By exploring the relationship between two of the aspects
of the self-structure, the writer hopes to reduce the molar quality
of the processes which are studied and, at the same time, increase
predictability by the consideration of more than one inferred sub-
structure. The study of the self-structure, then, focuses on a
specific aspect of personality, but one which is meaningfully
related to other dimensions of personality.
The concept of social self
Theory. -- In the previous discussion, some of the problems
in current self-concept research were outlined, and the type of
research which may aid in solving this dilemma has been suggested.
The remaining discussion will give the background of variables
which are more directly pertinent to this particular study. The
first of these concerns the inferred levels of self with special
emphasis on the social-self level.
Recently, some theoretical consideration has been given to
the structure of the self (30,25,17). Leary's (17) proposals
are probably most relevant to this study. Leary presents five
levels of interpersonal relationships, which are viewed by the
present writer as characterizing five levels of self:
Level I comprises automatic role relationships in
standard institutional situations.
Level II Is the conscious verbal report of the individual
and it reflects how he chooses to present himself and his
view of the world.
Level III concerns those preconscious themes which the
individual expresses in a projective testing situation.
Level IV relates those unconscious interpersonal themes
which the individual consistently, significantly, and specifically
omits in the other three levels.
Level V represents the values which the individual
Other levels could have undoubtedly been used, but for the purposes
of this paper, Leary's primary contribution appears to be the
consideration of "psychological pressures among different levels
of personality" (14, p. vi).
The present writer's conceptualization of aspects of self
overlaps, in many cases, with those of Leary. It is, however,
more closely tied to the previously presented definition of self-
concept, which emphasizes the individual's perception of his own
processes. Several aspects of self are hypothesized, each of which
can be inferred from the individual's verbal and expressive behaviors:
1. Private Phenomenal Aspect What the person believes are his
own characteristics (self-concept) under conditions of minimal
social pressure. The degree of distortion from the person's
characteristics as they "really" exist depends upon the degree
of previous anxiety to which the individual has been subjected
and his success in coping with it. Unconscious seif-enhancement
may be found. Since present measurement techniques usually
involve the subject's realization that he is an experimental
subject, some conscious self-enhancement will probably be
measured, even though it is not included as part of this
2. Social Aspect What the person believes are his own
characteristics (self-concept) within the current interaction
of a social situation. This depends upon:
(1) The Individual's ability to recognize various social
roles and situations.
(2) The ability to enact and respond to these roles.
(3) The motivation (conscious and unconscious) to respond
in accordance with these roles.
3. Public Aspect What the person communicates verbally about
himself. This differs frcm his private phenomenal view of
himself and he realizes that it differs. It mey also differ
from his real self, as inferred by others.
4. Real-Self These are the "true" characteristics of the individual
which are not distorted by the defensive perception of aspects
one and two, or the protective corrjnunicat ons of aspect three.
This aspect includes accurate self-perceptions (conscious and
unconscious). As with other aspects, the real self must be
inferred from the statements and behavior of the individual.
Since the social aspect of the self-structure is of primary
Interest here, It might help to clarify the definition of this aspect
by examining related constructs of other theorists.
The present definition of social-self resembles Leary's level 1,
public communication (17). This level includes automatic responses
in standard institutional situations. These are subtle, ubiquitous,
automatic role relationships, which to a considerable extent, function
to minimize anxiety.
Most everyone assumes automatic role responses which
he automatically assumes in the presence of each significant
"other" in his life. These roles are probability tendencies
to express certain Interpersonal purposes with significantly
higher frequency. The individual may be quite unaware of
these spontaneous tendencies (17, p. 109).
Leary compares this level to Jung's "Persona" and Reich's
Reich, as reviewed by Leary (17), defines character armor
as a chronic alteration of the ego which has rigidity. The degree
of character mobility constitutes the difference between healthy
and neurotic character structures. Reich finds that analysis of
different characters shows them all to be merely different forms
of armoring tf the ego against the dangers threatening from the
outer world and from repressed Inner impulses. They represent
different reactions to anxiety. From Reich's frame of reference,
the establishment of character armoring solves a repression
problem: it either makes the process of repression unnecessary,
or it changes the repression into a relatively rigid, ego accepted
Jung, as reviewed by Leary (17), saw the persona as a compromise
between the Individual end society as to the kind of scmblance to
adopt, or as we might say, those aspects of the ego which are
concerned with adaptation to social roles. The persona in its
way Is a necessity:
Society expects, and indeed must expect, every individual
to play the part assigned to him as perfectly as possible, so
that a m n who is a person must not only carry out his official
functions objectively, but must ?t all times and in all cir-
cumsta'nces play the role of ~arson in : flawless ma-ner...each
must stand at his post, here a cobbler, there a poet. No ran
is expected to be both (17, p. 4) ).
Research. -- As far as the present writer can determine, there
have been no investigations of the social aspect of self, as presented
here, reported in the literature to date.
In one Investigation, Brownfaln (5) Investigated the "social-
self" which was operationally defined as the subject's rating of
himself as he believes other people in the group see him. This
measurement approach, however, does not get at many elements of
social-self as defined in this paper. First, the interpersonal
pressure of the current social situation was not present. Thus,
certain elements of the real-self, which might be called out under
this pressure, would not be measured. Also; different social
situations call for differing role qualities (I.e., modesty,
bravery, assertiveness, etc.). Actually, one would expect to
tap elements from the private phenomenal level and the public
level through the use of this procedure.
How, then, can the social aspect be tapped for measurement?
The method to be used here Involves obtaining an individual's
evaluation of self in a particular social situation. In this way,
in addition to the social aspect, elements of the public aspect
and of the real-self aspect can be inferred. Also, a definite
role behavior will be called for.
The individual's evaluation of self In a private anonymous
situation will also be used. Not only the private phenomenal
aspect, but certain elements of the public aspect and of the
real-self aspect can be obtained In this manner.
With this brief description of the social aspect of self, it
might be appropriate here to take a closer look at how the dynamic
structure of the self develops.
Anxiety and dynamic structure of the self
Stages of development. -- Three basic stages have been differ-
entiated by Piaget (20) in describing how children come to apply
social norms to their behavior. Newcomb's (20) account of the
development of the self, based upon Piaget's thinking, will be
The first stage, called the stage of autism, is the early
period In infancy when the wishes and demands of the infant are
met without reslstence on the part of the significant adult.
The second stage, absolutism, is marked by four phenomena:
first, there is some resistance to the demands of the child. This
results in the child's learning to inhibit some impulses and learning
to perform certain acts in order to have his needs satisfied. Second,
because of developing intellectual capacities, the child is able to
make discrimination, especially among people. Third, the child
Interiorizes social norms as a part of the universe (absolutism).
Finally, he learns to play the roles of others, often in rapid
succession. In this way he learns to respond to the anticipated
behavior of others. He is now able to do this because he is
capable of taking longer time perspectives, and because of the
development of language, the child is able to take a dual role
when he speaks.
The third stage in the development of the self is the stage
of relativism. Now the child recognizes that there are different
perspectives in the world and that they are not absolute. The
mother is more variable than the child's fixed anticipations of
her so he anticipates her actions in order to make his life less
chaotic. He discovers that he will be more successful in his
anticipations of her if he puts himself in his mother's place.
This is how he comes to identify with her. Then he learns that
all of his mother's motives are not necessarily of Importance to
him. This is the fact of different perspectives.
A theory of development. -- It Is the writer's view, utilizing
Piaget's concepts of absolutism and relativism, that to the extent
the individual is able to graduate from the absolutism level to the
relativism level, he will have an autonomous self-identity in which
the various aspects of the self are highly congruent. To the extent
that he is unable to make this step, the Individual will be characterized
by self-diffusion. He will play many rigid, Inflexible roles, and
this absolute construction of the world will Interfere with the
individual's ability to perceive and experience accurately new sltua-
What is it that prevents some individuals from successfully
making the transition? Speaking primarily from a Sulllvanlan (29)
framework, one might hypothesize that anxiety (attack on the then
established self-structure) will stifle the individual initiative
and industry which is necessary for relinquishment of the safe but
non-functional and inflexible roles.
From a more positive point of view, two qualities are necessary:
positive and approving feedback from others, and consistent feedback
from others. As Erickson has stated:
The form of ego identity is more than the sum of the
childhood identifications. A sense of ego identity (ego
synthesis) is the accrued confidence that one's ability
to maintain inner sameness and continuity is matched by
the sameness and continuity of one's meaning for others.
The child must recognize that his way of recognizing
reality and mastering experience is a successful variant *
of the way other people around him master experience (13
Also pertinent to this discussion are the viewpoints of several
Kelly (16) sees the psychological disorder as resulting from
any personal construction which is used repeatedly in spite of con-
sistent Invalidation. He presents a constructive alternativism
which emphasizes "the creative capacity of the living thing to
represent the environment, and not merely to respond to It" (16, p. 3).
Rogers (22) presents a process view of adjustment in which the
movement towards adjustment is away from fixity towards an integrated
fluidity of experiencing and feeling.
Leary explains how different human beings develop different, rigid,
self-defeating techniques of adjustment.
* Underline added.
In the adjusted, well-functioning individual, the
entire repertoire of interpersonal reflexes is operating
spontaneously, flexibly, and appropriately. Human societies,
however, tend not to be too well balanced. They tend to
put a premium on certain interpersonal responses--compet!-
tiveness or slavish submission for example. To survive
and flourish, human beings must tailor their responses to
the demands of such imbalanced cultures.
Even In the rost heterogeneous and tolerant society,
the developing personality interacts with so many inflexible
pressures (e.g., parents' personalities, subculture demands)
that 6 hierarchy of preferred reflexes develops. To say
that human personality is varied and different, is to say--at
this level--that most everyone tends to overemphasize certain
automatic Interpersonal responses and to underemphasize
others (17, p. 31).
Leary emphasizes the Sullivanian framework in explaining why the
individual is willing to tailor his responses to the demands of
significant others and their cultures. He employs these interpersonal
responses In order to reduce anxiety, ward off disapproval, and
maintain self-esteem. The individual discovers, as he develops,
that certain interpersonal responses bring danger, while other
responses bring a narrow, uncomfortable, but certain security.
Taking these theoretical viewpoints concerning the development
and structure of the self into consideration, the present writer
theorizes two types of individuals who behave in ways which are
consistent with the absolutism and relativism levels of self
More specifically, individuals who behave In a way which is
characteristic of the absolutism level of self-development have
indiscriminately incorporated the roles of significant others as a
means of warding off anxiety. These roles or rules of the game are
seen as absolute.
The boundary of the self-structure Is permeable and stimuli
are primarily incoming. The Individual reacts to the environment.
He is dependent on the field and oversensitive to the expectations
of others. Solf-csteem Is based almost entirely on the reactions
of others. Rigid barriers separate roles within the self-structure,
resulting in a lack of consistency within the self.
Because of the individual's dependence and oversensitivity,
behavior In a particular role situation will be highly predictable.
It will conform with the social demands of the situation. Since
rigid barriers separate roles within the self-system, behavior in
one role is, to a large degree, Isolated from the rest of the self,
but consistently so. Individuals who behave in a way which is
characteristic of the relativism and inconsistent stages of self-
development are less predictable.
The public aspect of self will be emphasized ac the expense
of other aspects of self because of the individual's need to conform
to the expectations of others (oversensitivlty) and the ability
to recognize behavior which is socially desirable. High anxiety is
expected in this group because of conflicting roles within the
self-structure arid lessened ability to act on the environment. The
individual perceives the world (and himself) through previously
assimilated rigid constructs. This results in a lessened ability
to act spontaneously in new situations.
Fewer intricate roles are developed because the perception of
new roles must be made to conform to old constructs. This brings a
narrow and uncomfortable, but certain security.
The private phenomenal self-evaluation should be lower than
that of individuals who have progressed to the relative stage because
of conflict and anxiety leading to lowered self-esteem. It should
be higher than that of individuals who, for some reason or other,
have been unable to assimilate meaningful and useful roles into
their self-structure. This is because of the greater role play-
ing ability in the absolutism group and the positive, if limited,
social feedback and communication which results from this con-
formity. Also, it Is socially unacceptable in this culture to
give ones self very low self-ratings in any situation.
The relatlvlsim stage of development represents individuals
who were able, because of minimal anxiety in their environment, to
be discriminating in the acceptance of roles. These roles are
seen as relative Instruments which are useful In achieving goals.
The boundary of the self-structure is less permeable than that
of the absolute stage Individual, but is not rigid. This Is
because of the greater autonomy of the individual. Stimuli are
both incoming and outgoing. The Individual reacts to and acts
upon the environment. He Is sensitive to the field, but not
overly so. Self-esteem Is based upon reactions from others and
a relatively permanent and consistent self-structure.
Barriers within the self-structure are more permeable than
in the absolute stage individual. This results in greater
consistency within the self. Because of the Individual's greater
autonomy, behavior in a particular role situation should be less
Figure I.-- Topographical model of the self-system representing
the absolutism level of self-development.
OUTER PERIMETER OF CIRCLE REPRESENTS BOUNDARY
OF SELF :S M.
NESS D MESION OF PFRIME'ER
PERMEABIL T OF 5. BOUDARY
GRID # REPRESENTS NUMBER OF INCRP. RATD ROLES AND
A TTTUDES NHICH COhPSE SELF -S'EM
THICKNESS- DIMENSION OF R LINES REPRESENTS
FLEXIBILITY OF ROLES AND AT TUDES T A HIN LHE SELF -SYSTEM.
ARROWS REPRESENT DIRECTION AND Q ANTITY
OF STIM LAT ON.
>E CIRCLES REPRESENT SI(
= THE SELF-SYSTEM.
FILLED IN AREA REPRESENTS THE SOCIAL SITUATION
AS USED IN THIS STUDY.
predictable than in individuals who represent the absolute stage of
development. He has learned the utility of social roles, however,
and will act in conformance with them unless this produces conflict
within the established self-structure.
The public aspect of self should come into play, but since the
individual feels autonomous as an individual and consistent within
his self-structore, this factor should not be as influential as
it is in the absolute stage individual. Low anxiety is expected
in this group because of the congruence of rjles within the self-
More intricate roles are developed because of the capacity
to perceive and experience new roles spontaneously. The private
phenomenal self-evaluation should be higher than in individuals
representing the other stages of development because of the greater
self-esteem in this group.
A logical extension of this theorizing allows the writer to
hypothesize a third type of individual. This individual is incon-
sistent in his behavior and insensitive to the roles and attitudes
of the culture.
Individuals who behave in a manner which is characteristic of
this inconsistent stage have been unable to asslilate meaningful
and useful roles into their self-structure because of Inconsistent
and negative feedbacks from significant others.
The boundary of the self-structure is rigid and while the
individual does act upon the environment it is often in a role
Figure 2.-- Topographical model of the self-system representing
the relativism level of self-development.
OUTER PERIMETER OF CIRCLE REPRESENTS BOUNDARY
THICKNESS-THINNESS DIMENSION OF OUTER PERIMETER
REPRESENTS PERMEABILITY OF SELF-BOUNDARY.
GRID # REPRESENTS NUMBER OF INCORPORATED ROLES AND
ATTITUDES WHICH COMPRISE .. F-SYSTEM.
THICKNESS-THINNESS DIMENSION OF GRID LINES REPRESENTS
FLEXIBILITY OF ROLES AND ATTITUDES WITHIN THE SELF-SYSTEM.
ARROWS REPRESENT DIRECTION AND QUANTITY
SMALL OUTSIDE CIRCLES REPRESENT SIGNIFICANT OTHERS
OUTSIDE OF THE SELF-SYSTEM.
FILLED IN AREA REPRESENTS THE SOCIAL SITUATION
AS USED IN THIS STUDY.
The individual is largely independeLnt of the field and insensitive
to the expectations of others. Sarbin (25) has found that persons
who Frequently make nonconfoi-mlng perceptual responses are charac-
terized by social behavior which is socially invalid. Individuals
who were high on the F scale of the 4ilinesota Multiphasic Personality
Inventory were low in perceptual conformity.
Because c. the individual's Independence, lack of sensitivity,
and limited ability to perceive social situations, behavior in a par-
ticular situation will be highly unpredlctab.e, both within the
group and within the Individual.
The public aspect will be less influential than it is in the
self-structure of individuals in tlie otiter two stages because of
the individual's lack of awareness of the feelings of others.
There will not be conflicting roles within the self-structure which
would produce anxiety, but the negative feedback due to socially
deviant performance ray bring it about.
Fewer intricate roles are developed and new situations are
often not accurately perceived. The private phenomenal self-
evaluation should be low because of negative feedback from others,
but this will be tempered by lack of insight.
In this section, stages in the development of the self have
been presented, and it is hypothesized that Individuals whose
behavior is characteristic of these various stages will differ in
regard to their self-structures.
In the remaining sections, a rationale for selecting individuals
who represent these various stages of development and a methodology
for testing differences in the self-structure will be presented.
Figure 3.-- Topographical model of the self-system representing
the inconsistent level of self-development.
OUTER PERIMrE" I OF CIRCLE REPRESENTS BOUNDARY
OF SELF SI'M.
THICKNESS- THNNESS DIVENSON OF TER PERIMETER
REPRESENTS PERMEABL ITY OF SEL-BOUNDARY
GRID # REPRESENTS NUMBER OF INCORPORATED ROLES AND
ATTITUDES WHICH COMPRISE SELF -SYSTEM.
THICKN SS THINNESS DIMENSION OF GRID LINES REPRESENTS
FLE ABILITY OF ROLES AND AT TITUDES /THIN THE S L-SYSTEM.
ARROWS REPRESENT DIRECTION AND QUANTITY
SMALL OUTSIDE CIRCLES REPRESENT SIGN FIANT OTHERS
OUTSIDE OF THE SELF-SYSTEM.
FILLED IN AREA REPRESENTS THE SOCIAL SITUATION
AS USED IN THIS STUDY.
Concept of social desirability
One variable which Is directly related to the need for
conformity to the rules and regulations of society Is the social
desirability variable. The social desirability construct, which
has gained wide attention in recent years both as a response set
and as a motivational variable, gives promise of providing a
classification of individuals according to the hypothesized stages
The investigation of social desirability as a motivational
variable which is relevant to non-test situations has been
initiated. Marlowe and Crowne define social desirability as
"a need for social approval and acceptance and the belief that
this can be obtained by means of culturally acceptable and
appropriate behaviors (11, p. 109).
In developing a scale to measure this motive, the authors
hoped to overcome the limitations of the statistical deviation
model drawn from the Minnesota Muitiphasic Personality Inventory
by Edwards (12). When subjects given the Edwards Scale deny,
for example, that their sleep is fitful and disturbed, it cannot
be determined whether these responses are attributable to social
desirability or to a genuine absence of such symptoms. In a
college population, high social desirability scores may reflect
the low frequency of pathological symptoms in this population
and not the needs of the subjects to present themselves In a
In their scale, Crowne and Marlowe drew from a population of
Items which are culturally approved, but improbable of occurrence.
Fifty items meeting this criterion were submitted to ten judges
for social desirability ratings. Judges were asked to score each
item in the socially desirable direction from the point of view
of college students, using true and false response categories.
Unanimous agreement was obtained on 36 items and 90 per cent
agreement was obtained on II additional items. The scale was
then reduced to 39 Items by the elimination of Items with content
relevant to pathology.
An item analysis showed 33 items that discriminated at the
.05 level or better between high and low total scores. Eighteen
items are keyed true and 15 false to control for response set.
Reliability is .88 using the Kuder-Richardson 20 to test for
reliability. A one month test-retest correlation of .89 was
obtained. The correlation between this scale and the Edwards
Scale is .35. The distribution of scores on this scale closely
approximates a normal one, while negative skewness is found on
the Edwards Scale (Tables 1-3 from Crowne and Marlowe (11)).
Table I. --- Mean scores and standard deviations for subjects
on the Marlcwe-Crowne and Edwards Social
Desirability Scales (11).
Table 2. --- Correlations between scores on the Marlowe-Crowne
and Edwards Social Desirability Scales, and scales
on the Minnesota Hultlphasic Personality Inventory (11).
HHPI Hlar low-Crowne Edwards SOS
K .40 .65
L .54 .22
F -.36 -.61
HS -.30 -.52
D -.27 -.72
Hy .15 .03
Pd -.i4 -.73
Pa .21 -.02
Pt -.30 -.80
Sc -.40 -.77
Ma -.24 -.42
Prejudice Gough 1951 Pr -.27 -.5G
Status Gough 1948 St .16 .14
Ego Strength ES .17 .46
Anxiety Taylor MAS -.25 -.75
Anxiety Welsh A -.23 -.61
Repression Welsh R .28 .07
Crowne and Marlowe administered the scale to subjects .t two
universities and then had the subjects perform a boring task for
25 minutes. They found that subjects with a strong need for social
approval (as measured by their scale) expressed significantly more
favorable attitudes towards the experiment than subjects with a weak
need for social approval(ll).
The high 5-0 (social desirability) group differed fror the low
S-0 group at the .01 level on all of the following attitudes: how
enjoyable the tasks were, how much they learned, scientific importance,
and willingness to participate in a similar experiment.
High and low groups on Barron's Conformity Scales were discriminated
only by question two, how much they learned (.05 one-tailed). But
response to the ether questions were In the same direction as the
Groups divided on the basis of the Edwards S-D Scale did not
differ significantly on these questions.
Table 3. --- Correlations of scores on tuo Social Desirability
Scales and the Barrcn Conformity S-cle (11).
Edwards S-D Scale Mar lowe-Crowne
S- D Solle
Confornlty -.12 (R-1 ) -.54** (N-5?)
Crowne S-0 .56* (r )
As a result of this analysis, Crowne and Marlowe define the
need for social approval as a iotivatrinal variable, while conformity
refers to a clsss of behaviors. They stat, that high need for social
approval is a characteritic of the Individual who gives responses on
the birron scale which are indicative of a relative lack Independence
of Judgment. Since the EAwirds Scade was unrelated to Barron's Con-
fcrmity scale, and since it correlates highly with various Minnesota
Hultiphesic Personality Inventory scales and the Toylor Menife;t
Anxiety Scale, the authors cf thli study Inferred tht it measures
the extent to WhLch an individual is w 1t(n to admit syr.ptorsn
indicativet of e.inac'l.rr.wni. Therefore, they expect relationships
between the Edwards Scale and other measures where there is a
corresponding over ep in Item content, particularly related to
Strickland and Crowne (28) used 85 outpatients with primarily
neurotic and characterological diagnoses, who were treated at the
Columbus, Ohio Psychiatric Clinic, and found that individuals with
high need for social approval terminate psychotherapy prematurely
(p <.005) and are rated by the therapist as more defensive and
less improved than those with less need for social approval. The
authors concluded that Individuals with high need for social
approval are "disorganized" and "concerned with protecting and
maintaining a vulnerable self-image" (28, p. 99).
Crowne and Liverant (10) tested the hypothesis that conformity
Is related to low expectations of success in socially evaluative
situations and is consequently accompanied by defensive processes.
The authors saw this as lending support to Rotter's social learning
theory (23). They concluded that individuals with high need for
social approval engage in defensive personal enhancement.
Allison and Hunt (1) studied the relationship between the
Edwards Scale and expression, as measured by a paper and pencil
technique, in a situational frustration test. They found that
subjects who were high on the Edwards Social Desirability Scale
expressed less aggression when the intention of the frustrating source
was not specified and, therefore, the culturally approved response
was not Implicitly defined.
It would appear, on the basis of these studies, that subjects
with very high needs for social approval resemble those individuals
who have been less successful in graduating from the absolute stage
of self-development. Individuals with a high need for social approval
lack independence of judgement, are disorganized and defensive, and
attempt to obtain social acceptance through culturally appropriate
behaviors In many situations. This conforming behavior in many
different social situations should result in the assimilation of
many conflicting roles.
Sarbin (24), writing about Piaget's stage of absolutism, stated
that "Fixation at this level leads to behavior which is primarily
'other' oriented. That Is to say, actions must first be considered
in terms of possible approval or disapproval, reward or punishment"
(24, p. 242).
Those individuals with very low needs for social approval, on
the other hand, appear to resemble Individuals who have found It
necessary to reject many of the roles of significant others. Low
social approval Individuals are not responsive to culturally approved
forms of behavior.
Individuals with moderate needs for social approval have apparently
found It useful to incorporate some culturally approved behaviors, but
are not slavishly following the cultural dictates of every situation.
In this respect, they are similar to individuals who have achieved
the developmental stage of relativism.
Following the above reasoning, individuals with high, medium and
low scores on the Marlowe-Crowne Social Desirability Scale can be assumed
to represent individuals whose behavior is characteristic of absolutism,
relativism, and inconsistent stages of self-development.
SUMMARY AND STATEMENT OF HYPOTHESES
A good deal of ambiguity has resulted from self-concept
research because of premature attempts to relate self-constructs
to other complex phenomena. Before this type of research can be
meaningful, studies must first be undertaken to explore and define
the structure of the self. The present study is directed towards
the solution of this problem.
Based upon the writings of various sociological (19,20) and
psychological theorists (29,22), the self is defined here as a
hypothetical construct which represents the individual's conception,
perception, feeling (affect), and evaluation of his own psycho-
physical processes as an object.
Drawing upon the writings of Leary (17) and others (32,2), the
present writer has conceptualized four aspects which comprise the
self-structure. Two of these aspects, the private phenomenal aspect
and the social aspect, are of primary interest in this study. The
private phenomenal aspect is defined as the Individual's evaluation
of himself with a minimal amount of social pressure. The social
aspect Is defined as the individual's evaluation of himself within
a social situation.
In conceptualizing the development of the self, the writings of
Piaget (20) and Sarbin (25) were reviewed. Three stages of development
During the autism stage of development, the individual has
been unable to incorporate the roles and acts of others into his
self-structure. The stage of absolutism is marked by the indis-
criminate incorporation of the roles of others, and these roles
are seen as absolute. In the third stage of development, the
roles of others are useful to him. lie discovers that roles
represent relative rather than absolute rules and he becomes
discriminating in his acceptance of them.
It is the general hypothesis of this study that individual
self-systems will vary, depending upon the individual's ability
to incorporate meaningful roles and acts of others into his
self-concept. In terms of Uerner's theory of mental development
(31), it Is hypothesized that differentiation of the levels of
the self is associated with the stage of absolutism, while
integration of self levels is associated with the stage of
relativism. More specifically, it is hypothesized that individuals
who are at the absolute stage will have a greater divergence between
their private phenomenal and social aspects of self than will individ-
uals who have made the transition to the stage of relativism.
In order to test this hypothesis, individuals who represent the
three stages of self-development are needed. Based upon a review
of the literature, it was determined that the need for social
approval was related to the need to accept and incorporate role
models. The Marlowe-Crowne Social Desirability Scale appeared to
be the instrument which would best measure this variable, and it
was decided that subjects would be selected to represent the three
stages of self-development based upon their scores on the Marlowe-
Crowne Social Desirability Scale.
More specific hypotheses are as follows:
1. Assuming that individuals with high scores on the
Marlowe-Crowne Social Desirability Scale represent individuals
whose behavior is characteristic of the absolutism stage of
development, the following relationships can be expected.
First, because of rigid barriers separating roles within
the self-structure, these individuals should have a greater
divergence between the private and social aspects of the self
than individuals at either the relativism or inconsistent
stage of self-development.
Second, because the social aspect of self has been molded
rigidly in accordance with various social roles and situations,
this group should show less individual variation as regards the
social aspect of self than individuals at either the relativism
or inconsistent stage of self-development.
Finally, individuals in this group should have lower self-
esteem than individuals at the relativism stage of self-development.
This is because of the limited social feedback and communication
which results from over-conformity. At the same time, however,
individuals In this group have assimilated some useful and
meaningful roles into their self-structure. This results In
some degree of positive feedback from significant others. For
this reason, self-esteem should be higher in this group than
in the group representing the inconsistent stage of self-
2. Assuming that individuals with moderate scores on the
Marlowe-Crowne Social Desirability Scale represent individuals
whose behavior is characteristic of the relativism stage of
development, the following relationships can be expected.
First, because of the highly permeable barriers separating
roles within the self-structure, these individuals should have
Fless divergence between the private and social aspects of the
self than individuals at either the absolutism or inconsistent
stage of self-development.
Second, since the social aspect of self has been molded
flexibly through the discriminate incorporation of social roles,
this group should show greater individual variation as regards
the social aspect of the self than individuals at the
absolutism stage of develcprent. On the other hand, since
the social aspect of self has developed from val'd and useful
social perceptions, less individual variation will be found
here than in the group representing the inconsistent stage of
Finally, the positive social feedback and communication,
along with congruence of roles within the self-structure, should
result In higher self-esteem in this group than in groups
representing the absolutism end inconsistent stages of self-
3. Assuming that individuals with low scores on the Marlowe-
Crowne Social Desirability Scale represent individuals whose
behavior is characteristic of the inconsistent stage of
development, the following relationships can be expected.
First, the barriers separating roles within the self-
structure of this group are considered to be more permeable
than those of individuals in the absolutism stage of
development. On tie other hand, these barriers are not
as permweble as those found in the individual at the relativism
stage of development. For this reason, these individuals
should have less divergence between the private and social
aspects of self than individuals in the absolutism stage,
but more divergence than individuals in the relativism stage.
Second, since the social self has been molded on the basis
of inconsistent and negative feedbacks from significant others,
the highest degree of individual variation can be expected in
this group. Socially invalid social perceptions are anticipated.
Finally, the negative and inconsistent social feedback
experienced by this group should result in lower self-esteem
than in groups representing the relativism and absolutism
stages of self-development.
4. Conflict within the self has been considered by various
writers (17,22,24,29) to be conducive to anxiety and lowered
psychological adjustment. On this basis, It is expected that
those individuals with a w-:dc divergence between the private
and social aspects of self, regardless of the stage of
development which they represent, will have lower psychological
adjustment than individuals with a narrow divergence between
the private and social aspects of self.
The procedure to employ, in attempting to elicit social factors
which influence the individual's self-concept, presents many
difficulties, as discussed previously. If one assumes that behavior
is not a product of self (as a doer), but rather a product of a
complex of stimuli, of which the individual is only partly aware,
then any procedure which is devised will tap various levels of
self, Focial and private, conscious and unconscious.
In a projective test type situation, for example, in addition
to tapping unconscious self-feelings, the responses will be a product
of both the individual's private feelings about himself and his
defensive presentations of himself as an individual in a social
situation. It has been demonstrated that the presence of the
examiner in the administration of the Thematic Apperception Test
has an inhibiting effect upon strongly emotional material (4).
Similarly, in a private self-evaluation situation, various
unconscious and socially defensive factors will probably be present.
It appears that unconscious, private, and social factors will be
tapped in any situation, but the degree to which each is elicited
should depend largely upon the situation itself. Self-evaluative
response in a social situation should be more influenced by social
factors than private self-evaluative factors. The reverse should be
true in the private, anonymous, self-evaluation.
Several steps will be taken to overcome these design problems.
Subjects will be grouped according to high, moderate, and low scores
on the Marlowe-Crowne Social Desirability Scale. Then they will
rate the pictures of 12 coeds as to likeability, relative to
themselves. in this way, an inference about each subject's
feelings of self-likeability can be inferred. Half of the
ratings will be made in a private anonymous situation, and
the other half in a clearly defined social situation which calls
for some degree of modesty. Since the 12 pictures will have
already been selected by each subject as being neither liked
nor disliked in terms of likeability, any differences found should
be a product of the treatment situation (social or private).
To control for position and sequence effect half of the subjects
will proceed from the social to the private situation and the
other half from the private to the social situation.
Forty-six female students who were enrolled in basic psychology
courses at the University of Florida and 44 female students who were
enrolled in basic sociology courses at the University of Wisconsin
were used. Most of the girls were freshmen or sophomores and were
18 to 20 years of age. The average age, education, and results
on the various measures used were very similar (see Tables 10,11,12,
13 and 14 in the Appendix). When this was established, both groups
were combined to make up one pool of subjects.
The first 44 subjects were students at the University of
Florida, and they participated in the experiment during the summer
of 1963. Thirteen of these subjects fell in the high social-
desirability category, 10 in the moderate category, and 21 in the
The remaining 46 subjects were from the University of
Wisconsin and participated in the fall of 1963. These subjects
were selected so that each of the three categories (high, moderate,
and low) would be brought up to a total of ;0. Therefore, the
Wisconsin and Florida samples tend to approach a significant
difference in social desirability scores (x2 = 3.56, df=l, p .10
two-tailed) (see Table 13 in the Appendix).
Subjects from the two samples did not differ, however, in the
direction of their self-ratings in the private and social situations.
Differences in the magnitudes of self-ratings were evidenced (see
Table 14 in the Appendix).
Procedure and instruments
A group administration of Part A of the Picture Identification
Test (6) and the Marlowe-Crowne Social Desirability Scale (11) was
accomplished approximately one week before the individual self-
evaluations were made. All instructions were standard and the
tests were administered in the order mentioned above.
The self-rating methodology. --After the group administration
of tests, the experimenter placed the 30 subjects with the highest
scores on the Social-Desirability Scale into one group, the 30
subjects with the lowest scores into a second group, and the 30 sub-
jects with scores clustering around the median into a third group.
For each of these subjects, the experimenter selected the 12
pictures which each subject had not placed into the most likeable
or least likeable categories on the Picture Identification Test.
Six of the pictures were used for the subject's private evaluation
and the other six for her social self-evaluation. Since these 12
pictures had previously been selected by the subject as neither
liked nor disliked, it was assumed that any differences In the
subject's evaluation of herself relative to them, from the private
situation to the social situation, would be a function of the
situation itself rather than any absolute difference in stimulus
quality (attractiveness) of the girls pictured or of the subject
herself. In this way, the experimenter feels that he is controlled
for a methodological weakness which has been evidenced in many
experiments of this type. For example, in one study (21), a subject
was presented with her own photograph (in a tachistoscopic procedure)
along with other photographs which had been Judged by the experimenter
of average attractiveness. This judgement, in the writer's opinion,
may have varied widely from the subject's phenomenal view of the
Neither is it satisfactory, in the writer's opinion, for the
subject to rate the pictures of others as equal to himself and
then to present these pictures to the subject In a different
procedure (i.e., tachistoscopic) (3). In this case, the initial
evaluation may be influenced by social and unconscious factors, and
may not represent the subject's private opinion.
Position and sequence. -- Half of the high social approval
group went from the social self-evaluation situation to the private
self-evaluation situation and the other half from the private
situation to the social one. The same counter-balance procedure
was used for the middle and low social approval groups.
The social situation. -- The social self-evaluation situation
consisted of the subject, two female peers (confederates of the
experimenter) and the experimenter. After the subject was seated,
the experimenter handed the subject and the confederates six
pictures each and read the following instructions:
"In front of you there is a measurement scale which runs from
(-) minus three to (+) plus three. Your job will be to rate these
people In terms of likeability. Use yourself as a reference point,
or zero on the scale. If a person is more likeable than you, rate
her on the plus side. If a person is less likeable than you, rate
her on the minus side. The one, two, and three represent the degree
to which she is more or less likeable than yourself. Plus three is
quite a bit more likeable than yourself. Plus one is a little more
likeable than yourself. Minus three is quite a bit less likeable
than yourself. Minus one is a little less likeable than yourself.
Tell me your ratings and I will write them down. Do not use zero.
Are there any questions?" Then the experimenter will look at the
subject and say,-----"(name), you go first".
After each rating, the experimenter restated out loud the position
of the pictured girl on the scale. For example, "She is a little more
likeable than yourself." When the subject was finished, she was excused
and taken directly to the private situation, or excused from the
experiment If she had been assigned to the private situation
immediately before the social one.
The private situation. -- The private rating situation took
place In a different room. After the subject was seated, the
experimenter gave her the other six pictures, a slip of paper
marked from one to six, and a coded envelope. The instructions
were the same as in the social situation, with the exception that
in the place of saying "Tell me your ratings", the experimenter
read the following instructions:
'Put down a number from minus three to plus three for each per-
son on this slip of paper. Then put the slip of paper in an envelope
with the letter (letter assigned to the subject) on it. This letter
is used so that your rating will be held confidential. No one
will know how you rated these people. When you have finished
and placed your ratings In the envelope, leave the envelope here
and you may leave." Then the experimenter left the subject alone
in the room. When finished, the subject was taken directly to
the social situation or dismissed from the experiment, depending
upon the order to which she had been assigned.
The Picture Identification Test. -- The full Picture Identification
Test (6) was administered to each subject approximately one week
before the individual self-evaluation was made.
The Picture identification Test is a projective test which
requires the subject to make judgements about photographs of people
of the same sex as the subject. The subject selects the pictures
of people he likes best and of the people he likes least. He also
selects the people who, in his judgement, best fit descriptions of
21 needs of the Murray Need System. He receives a Judgement, an
Attitude, and an Association score for each of the 21 Murray needs.
Since the overall Association index score was used as P criterion
of adjustment, it is of primary interest here. An Association Index
is computed for each need. This score is based on the degree to which
the subject conforms to a norm group in associating needs by attributing
them to the same person. For example, it has been quite common in norm
groups to associate need-dominance and need-aggression by attributing
these two needs to the same person, whereas the need-dominance and need-
deference were seldom associated in this manner. A high Association
Index for a need is interpreted by the authors of the test as meaning that
the subject knows how to fit the need in well by the system of other needs
which help him to satisfy the needs in an acceptable wa It means that
the individual knows that the need does not combine well with certain
other needs so that ie wouldn't try to satisfy it and these other needs
at the same time.
In one study (18), 94 male prisoners from a correctional institu-
tion and 94 male students from a technical and vocational high school
were administered a group form of the Picture Identification Test, A
procedure was established to test the hypothesis that the profiles
adequately represented various types of prisoners and non-prisoners, so
that a new subject could be correctly classified according to whether
his profile was most similar to the prisoner or non-prisoner group.
T..e results showed that 68 per cent of the subjects were correctly!
classified by a median cutting point. It was noted that while the
prisoner profile showed low association scores, there were no particu-
larly low association scores in the student profiles.
In another study (7), using college students as subjects, the
Association Index was again found to be positively related to
Statistical methodology. --Since a counter-balance procedure was
used to obtain the self-ratings in the two situations, it was decided
that a 2 x 2 counter-balanced analysis of variance would be the most
appropriate statistical methodology. The most important advantage to
this methodology is that subjects are orthogonal to treatments and
each subject therefore serves as his own control. This feature provides
that the usually large between subjects variance is removed from the
error term used for testing treatments, and thus the design is extremely
sensitive for treatments. Te assumptions of a normal distribution of
the population, random and independent observations, and similar popu-
lation variances, seemed reasonable for this type of research. T.e
sample is one which is often used in research of this type. The n of
90 is large enough to enable the use of parametric methods of analysis.
The variances were not found to differ significantly.
Combining the previous theoretical hypotheses with this methodology,
several specific experimental hypotheses can now be discussed:
I. Individuals at the absolutism stage of self-development can be
expected to have a greater divergence between the social and
private aspects of self than individuals at either the relativism
or inconsistent stages. It can be expected than: a. That the
high need for social approval group will have a greater divergence
between self-ratings in the private and social situations than will
the moderate need for social approval group. b. Tuat the high need
for social approval group will have a greater divergence between
self-ratings in the private and social situations than will the
low need for social approval group.
Because of the oversensitivity of individuals at the absolutism
stage of self-development to social expectations, it can be
expected: c. Ti.at the high need for social approval group will
have higher self-ratings in the private situation than in the
Finally, since individuals at the absolutism stage of self-
development can be expected to have higher self-esteem and to
respond more appropriately to social expectations than individuals
at the inconsistent stage of development, it can be expected that
c. The high need for social approval group will have higher self-
ratings in the private situation than will the low need for social
2. Individuals at the relativism stage of self-development are not
expected to have a great divergence between the social and private
"spects of self. It can be expected then: a. That the mocerite
need for social approval group will not have a great divergence
between self-ratings in the private and social situations.
b. Toat the moderate need for social approval group will not have
a greater divergence between self-ratings in the private and social
situation than the low need for social approval group.
Individuals at the relativism stage of self-develo7ment are
expected to have a high degree of self-esteem and the ability to
make appropriate, yet discriminating, evaluations of social
situations. It can be expected than: c. That the moderate need
for social approval group will have higher self-ratings in the
social situation than will the igh need for social approval
group. d. That the moderate need for social approval group will
have higher self-ratings in the private situation than will the low
need for social
approval group. e.
T :at the
4"erate need for
social approval group vill have higher combined self-ratings in
both the social and private situations than will the low need for
social approval group.
Individuals at the inconsistent stPge of self-development are
not expected to have a great .r bet-ween the social and
-iv te aspects of self. It can be expected then: a. T.at the
low need for social approval group will not have a great i r
between self-ratings in the private and social situations.
Individuals at the inconsistent stage of self-development are
expected to have a low degree of self-esteem and to make inappro-
priate and invalid social perceptions. It can be expected then:
b. Trat the low need for social approval group will have loi r
self-ratings in the private situation than will the moderate need
for social approval group. (See hypothesis 2 d.) c. That the
low need for social approval group will have lower self-ratings
in the private situation than i il the high need for social
approval group. (S
For social approval
e hypothesis I d.) d. That the low need
group will have lower combined self-ratings
in both the social and private situations than will the moderate
need for social approval group. (See hypothesis 2 e.)
Since the social aspect of self in this group has developed
from Inconsistent communications from others, individual
variations in reactions to social cues are expected. It can be
expected thc-: e. Trat the low need for social approval group
will have higher individual variations in self-ratings in the
social and private situations than the moderate need for social
4. Individuals ;ith s wide divergence between the private and
social aspects of self are expected to be lower in their psycho-
logical adjustment. It can be expected, ti, that those indivi-
duals with a wide divergence between self-ratings in the private
and social situations will have lower Association Index scores on
LhC Picture Icencification than will individuals with convergent
self-ratings in the two situations (with the need for social
approval factor held constant).
In the general statistical analysis, en analysis of variance was
used to compare the two treatments (private and social) and the three
groups (.igh, mAderate, and low need for social approval). It allowed
for the simultaneous control of individual differences and temporal order
of treatments. Each individual received each treatment once and different
groups of individuals received treatments in the two possible different
orders (social to private and private to social). Each treatment appears
once and onl, once in each row (group of individuals) and in each column
T. subjects x sequence ; groups and subjects position x groups
x sequence interactions were used as error terms for groups and groups
x treatments effects respectively. Error terms not including all of
these interactions could lead to a positive F test bias (Type I error).
For example, the use of the groups x sequence mean square as the error
term would have resulted in a significant F for groups. T is error term
would not have contained the replicated subjects variance.
Tne group times treatments interaction was found to be significant
at Ahe .05 level. Once this was determined, t tests were used for each
variable between the two groups. Finally, t tests for correlated means
were used to test between variables for each group. In addition, the
nonparametric median test (26) was used to test hypothesis number 12
involving two small groups of subjects.
While the variances were not found to be significantly different, a
slight correlation between means and variances was observed. This was
not considered serious enough to prevent the analysis of differences
in terms of means, however.
The analysis of variance showed the groups times treatments effect
to be significant at the .05 level. The groups effect represents the
high, moderate, and low need for social approval groups. The treat-
ments effect represents the private and social situations. The F for
the groups times treatments effect was 4.01. At 84 and 2 degrees of
freedom, the critical value needed for significance at the .05 level was
Tie groups effect alone was not quite significant at the .05 level.
The F for the groups effect alone was 3.02. With 84 and 2 degrees of
freed-m, the critical value for significance at the .05 level was .11.
The first and most general finding then was that female
students, grouped according to scores on thae Marlowe-Crowne Social
D sirability Scale and making their self-ev 'utions in private and
social situations, would show significant variation.
Table 4. ---Analysis of Variance.
Source of Variatic
Groups x Sequence
S's x Sequence x Group
Groups x Position
Groups x Treatments
S's x Pos x Grps x
Sum of Square? df Mean Square F
24, .5654 17-
At DF = 84 and 2, the critical value is F = 3.11
At DF = and 2, tiie critical value is F = ,.ll
Table 5. ---Ratings of pictures (as to
and private situations) by
according to need for soci
High need for social
Moderate need for
social approval group.
Low need for social
likeability in social
PRIVATE SITUATION SOCIAL SITUATION
X = X 2 = 283*
( X2= 1289 ( X) = 193
X = = .914
x = X = 10.93
X 2 = 24 X 2 = 251
( X) =59049 ( X)2 = 60959
X = 2213 X = 2469
X = 8.10 = 8.36
X = ;11 X 2 = 285
(X)2 = 96721 ( X)2 = 10006
X = 3/15 X = 175
X = 10.37 X = 9.50
*R w score represents the subjects' rating of the pictures and
not a self-rating. Tuus, high raw score = low self-rating.
o OL . L L-
L0 a 0 L 1- CL l- l-0
O 0l 0 0 '-3-
S 3 8 8 8 S3
X 0. 0
o0 ijL( 0 o LA LA-I I '- r '-ALA
a s o 0 00 0 0 00
'VAA v V A A V VA
r 3 a 1 4 0 4A
OR 0 L 0 IU' &0v
SZ *000 1 0 0 a 0 0 0
.-0-->. VVV V V V V VV
0 j <
0 c o
(I J 1- c I 0
S- o0- -3- A - 'U~
S I -I -I LI L.
L 1- 0 a4Q Co O L Li
fi) t3 000 OO QU 3
w- UJ I ____
4.4 z Un 3 0) o
(- LI uj 00 0 z -
i >U 0 o 0 0 o0 Lo o n u
I C&lU..JWC aJ uWi .Q .jjW
u U UJ-) U 00 007
-W -- -- 1 1= %n C
(U0)0I00. ) )
_i -u -
s -q Co
o. a x a 0 -
0 -- O
Oi >u Au
4A LI L CL
C L 0 -*
o w 04 3 ad C
33 -C 3 L
- I u- 5 0 (1
LI < 2!s Z wJOU- au
-e -I c*-a Wo mC
S-l CE Q )-. 0-q 3
< 0O O O
-C *-- u!a
Wta A A A r n- Cai
a. a a M-
-am t E m a- Y o f e
l < S: S SC MC -
5 -)C (548 S .......3 u
Table 7. ---Average scores of all groups on the Marlowe-
Crowne Social Desirability Scale.
Average Score Standard Deviation
High group 20.87 2.80
Moderate group 14.20 1.65
Low group 7.97 2.43
High need for social approval group. --Hypothesis I a., that
the high need for social approval group would have a greater divergence
between self-ratings than the moderate need for social approval group,
was correct (t = 2.94, df = 58, p .01 one-tailed). A t test for
uncorrelated means was used to compare the difference between the
private and social situations for the high group with the difference
scores between the private and social situations for the moderate group.
The sign (or direction) of the difference scores was token into
account (see Table 6).
Hypothesis I b., that the high need for social approval group
would have a greater divergence between self-ratings than the low
need for social approval group, was incorrect (t = 1.38, df = 58,
p .10 one-tailed). A t test for uncorrelated means was used to
compare the difference scores between the private and social situa-
tions for the high group with the difference scores between the
private and social situations for the low group. The sign (or direction)
of the difference scores was taken into account (see Table 6).
Hypothesis I c., that the high need for social approval group
would have higher self-ratings in the private situation than in the
social situation, was supported (t = 4.03, df = 29, p .001 two-
tailed). The statistical test used here was the t test for correlated
means (see Table 6).
Hypothesis 1 d., that the high need for social approval
group would have higher self-ratings in the private situation than
would the low need for social approval group, was not supported
(t = 1.22, df = 58, p .10 one-tailed). An inspection of the data
reveal that the direction was correctly predicted here, however,
since the high need for social approval group did tend to have higher self-
ratings in private situations than the low need for social approval group.
The statistical test used for this analysis was the t test for uncorre-
lated means (see Table 6).
Moderate need for social approval group. --Hypothesis 2 a.,
that the moderate need for social approval group would not have
different ratings in the two situations, which were statistically
significant, was supported (t = .38, df = 29, p .10 two-tailed).
Here again, while one cannot prove the null hypothesis, it is
interesting to note that significantly different ratings in the two
situations were not expected. The statistical test for this analysis
was the t test for correlated means (see Table 6).
Hypothesis 2 b., that the low need for social approval group
and the moderate need for social approval group would not show a
significant statistical difference In their self-ratings in the two
situations, was supported (t = .54, df = 58, p .10 one-tailed).
W.ile one cannot prove the null hypothesis, it is interesting
to note that the smallest difference was expected, and found,
between these two groups (see Table 6).
Hypothesis 2 c., that the moderate need for social approval
group would have higher self-rating in the social situation than
the high need for social approval group, was supported (t = 1.88,
df = 58, p .05 one-tailed). The statistical test used for this
analysis was the t test for uncorrelated means (see Table 6).
Hypothesis 2 d., that the moderate need for social approval
group would have higher self-ratings in the private situation than
would the low need for social approval, was supported (t = 2.46, df = 58,
p .01 one-tailed). The statistical test used in this analysis was
the t test for uncorrelated means (see Table 6).
Hypothesis 2 e., that the moderate need for social approval
group would have higher combined self-ratings in both the social
and private situations than would the low need for social approval
group, was supported (t = 2.54, df = 118, p .05 two-tailed).
The statistical test used for this analysis was the t test for
uncorrelated means (see Table 6).
Low need for social approval group. --Hypothesis 3 a., that
the low need for social approval group would not have ratings
in the two situations which were significantly different, was
supported (t = .96, df = 29, p .10 two-tailed). H-re again, while
one cannot prove the null hypothesis, it is interesting
to note that no statistically significant differences were expected.
T.e statistical test used for this analysis was the t test for
correlated means (two-tailed) (see Table 6).
Hypothesis 3 e., that the low need for social approval group
would have higher individual variations in ratings in the two
situations than would the moderate need for social approval group,
was not supported (t = .87, df = 118, p .10 two-tailed). The
statistical test used in this analysis was the t test for uncorrelated
means comparing the difference scores of individuals in the low
need for social approval group, with the difference scores of
individuals in the 'o rate need for social approval group. The
sign (or direction) was not considered in these difference scores
(between social and private situations). Here again, while not
statistically significant, the trend was in the direction predicted
(see Table 6).
Convergent versus divergent self-ratings. --Hypothesis 4, that
individuals, disregarding social desirability scores, with con-
vergent ratings between the private and social situations would
have higher association index scores on the Picture Identification
Test than individuals with divergent ratings, was incorrect (t = 4.16,
df = 1, p .02 one-tailed). In fact, the result here was in the
opposite direction to that predicted. Individuals with divergent
ratings between the social and private situations had significantly
higher association index scores. There were 15 subjects in the
convergent group representing the high, moderate, and low need for
social approval groups. The discrepant ratings group was comprised
of 14 subjects also from the high, moderate, and low need for social
approval groups (see Table 6).
Table 8. ---Average divergence between self-ratings in
social and private situations for discrepant
Discrepant groups 6.4
Convergent groups 1.9
Of the 12 experimental predictions made, 8 were correct and
4 were incorrect. The correct direction was predicted in 11 of
the 12 experimental hypotheses. Of the 9 hypotheses where statisti-
cal differences were expected, statistical significance was achieved
in 6 of the 9 hypotheses. As was expected, no statistical difference
was found in 3 hypotheses, but the correct direction was predicted
for these 3 hypotheses.
Perhaps one reason for the lack of statistically significant
differences, in some of these findings, is the homogeneous population
which was used. In the pilot study the top and bottom 10 scores
out of a total of 50 Social Desirability Records were used. This
allowed for a considerable divergence in the scores of the two
groups. In the final study, the top, middle, and lowest 30 scores
were used out of a total population of only 120 records.
Discussion of findings
The most general finding, and perhaps the one with the broadest
theoretical applications, was that individuals grouped according
to need for social approval would differ significantly in the
pattern of their self-ratings In private and social situations.
This finding lends support to those theorists who see the self-
concept as developing from a subjective organization of the
culture (27,25,14,19). Certainly, the fact that group differences
can be expected, in specific aspects of self, based on the knowledge
of general social attitudes, is provocative.
High need for social approval group. -- The first hypothesis,
that the high need for social approval group would have a greater
divergence between self-ratings than the moderate group lends
support to the notion that individuals who have found It necessary
to indiscriminately incorporate the roles of significant others,
will be characterized by self-diffusion. Rigid barriers separate
roles within the self-structure, resulting in a lack of consistency
within the self. The moderate approval group, on the other hand,
showed relatively less divergence, and this confirmed the expectation
of this group as having more permeable barriers within the self-
structure and greater consistency within the self.
It Is important to note, however, in analyzing this finding,
that the statistical test which was used took into consideration
both amount of difference between ratings in the two situations
and direction of difference. In ter.s of Individual variation,
vglugo regard to directlon, the group tra is reversed. The
high need for approval group has less Individual variation without
regard to direction than does the moderate need for approval group.
While this nay appear to be contradictory to the major hypothesis
at first glance, It Is actually consistent with the theoretical
formulation proposed previously. Because in the high need
group, while rigid barriers separate roles within the self-system,
behavior In one role is isolated from the rest of the self, but
consistently so. Because of the -Woindeoce end ovor-,ansitivity,
behavior ii a particular role situation is highly predictable.
The moderate need group, on the other hand, while having greater
consistency within the self is less predictable Individually
because of their greater autonomy. One would expect the low need
group to heve even greater Individuel variation without regard to
direction than the moderate need group. This is because of their
lack of sensitivity and limited ability to perceive social situations.
There was a trend In this direction but it did nctreach significance.
The second hypcthesis, that the high need group would have a
greater divergence than the low need group, wa: n;t supported at the
.05 level of significance, but did reach the .10 level of significance.
One would expect a less clear cut difference between these two
groups than between the high and moderate groups because, like
the high need group, they have assimilated few roles and these
are not consistently In*egrated. But unlike the high need group,
the reason for this is inconsistent and negative feedback from
others; not slavish compliance to a few safe roles.
The third hypothesis was correct and lends support to the
theoretical expectation that the high need group is dependent
on the field and overly sensitive to the expectations of others.
These individuals will conform to the social demands of the situation.
In this instance, modesty was built i;ito the social self-rating
situation, and this group responded with lower self-ratings in this
situation relative to the private one.
The fourth hypothesis, that ths high need for social approval
group would have higher self-ratings In the private situation than
would the low need for social approval group, was not supported
statistically. With more subjects and a wider range of social-desirability
records, this might have reached a significant level.
Moderate need for social approval aroup. -- In the first
hypothesis, no statistically significant difference was expected
between the moderate need group's ratings in the two situations.
This was because this group was not expected to be overly sensitive
to the expectations of others. On the other hand, this group has
recognized the utility of social roles and will usually act in
conformance with them. An Inspection of the raw data shows slightly
lower self-ratings in the social situation in conformance with the
The second hypothesis, that the moderate and low need groups
would not show a significant difference in their self-ratings in the
two situations, was made to point up the curvilinear quality of the
need for social approval variable. As stated in the previous
discussion, these two groups would be expected to show certain
similarities as regards their self-concepts, but for different
reasons. While one cannot prove the null hypothesis, these "no
difference" hypotheses may have some meaning in the total context
of successful predictions of differences between other groups.
Findings related to the third hypothesis lend support to
two theoretical expectations. The first, and most obvious reason,
for the moderate group to have a significantly higher self-rating
in the social situations, is that the moderate group is not overly
sensitive to the social response of modesty which is inherent in
the situation. If this were the only reason for this finding,
however, one would also expect the low need group to have signifi-
cantly higher ratings in the social situation relative to the high
need group. This is because the low need group is insensitive to
these social pressures. The low need group did not have significantly
higher ratings than the high need group, however. Therefore, another
causal factor must account for the moderate need group's higher self-
ratings. This factor is probably the greater overall self-esteem
of this group due to the positive and consistent feedback from significant
The findings of the fourth hypothesis also offer support for
two theoretical expectations. The first reason why the moderate
group can be expected to have higher self-ratings in the private
situation than the low need group is that the low need group makes
invalid social perceptions resulting in higher self-ratings in a
situation calling for modesty and lower self-ratings in a situation
not calling for modesty. The second reason for this finding is the
greater overall self-esteem of the moderate need group.
In fact, both of the previous findings lend support to the
theoretical expectation that the moderate need group has greater
overall self-esteem. The findings of the fourth prediction could
be explained by the deviant role reversal of the low need group.
If this were the onli causal factor, however, then one would expect
the high need group to also have significantly higher self-ratings
in the private situation than the low need group. This was not the
case since prediction Id. was incorrect. Therefore, the moderate
need group's higher self-ratings probably relates to their overall
These findings offer some tentative support for the curvilinearity
of the social desirability factor as an indication of greater self-
esteem. This is contrary to the viewpoint that a high degree of
conformance is equivalent to a high degree of self-esteem and adjust-
The fifth experimental prediction was supported statistically,
and this lends support Lo the greater self-esteem of the moderate need
group. Where levels of significance were not achieved in this study,
one might be able to explain this in terms of the limited number of
Social Desirability Records from which extreme scores could be selected.
Therefore, the range of subjects may have been too houogenoous.
Another possibility would be that the Harlowa-Crowne Social
Desirability Scale da s not allow for a sufficient range of
responses. A greater rarnt of responses would allow for the
selection of more entree groups.
Low need for sc ral a-,rov'pl orcr., -- The first hypothesis
was made because It was not expected that the low need group
would have statistically significant differences In ratings In
the two situations. This is because the low need group Is
largely independent of Ote field and insensitive to the expecta-
tions of others. Not only was the null hypothesis of no dif-
ferences supported, but inspection of the data shows this group
to have made higher self-ratlngs in the social situation thar In
the private one. This is in direct opposition to the quality of
modesty which was called for in the social situation. This gives
qualitative support for the invalidity of the social behavior of
The theoretical meanings involved in the second experimental
prediction were discussed previously within the context of prediction
a1 The low need group was expected to have greater individual
variation without regard to whether the social or private self-ratings
were higher than the moderate need group. This is because of their
lack of sensitivity and limited ability to perceive social situations.
In turn, the moderate need group Is expected to show greater individual
variation than the high need group. The reader is asked to compare
the outer perimeters of the drawing of the self-systems for the three
groups: iow noed -- inconsistent stage, r,iodcrate need riclet.vmsn
stLge, and high ned.d absuiutism stage. While this hypothesis wes
not supported statistically, once again the trend was in the expected
Convererit vers',s divergent self-rratrins. -- The final hypothesis
was made because one would expect to fiic iowered adjustment In
individuals who have conflict between the private arnd social aspects
of their self-systems. The results were the opposite of those
expected. Individuals with wide discrepancies between the two
aspects of self had significantly higher association index scores.
One possible explanation for this result is that the Picture
Identification Test is, in a sense, a test of attitudinal conformity.
That is, high scores are given to those individuals whose attitudes
are in conformance with the majority. Therefore, while individuals
representing the discrepant and convergent groups were selected
equally from all three social desirability groups to eliminate
the social desirability factor, the discrepant group itself may
represent those individuals who are more sensitive to what is
socially desirable than is the convergent group. If this is true,
then it would nrt be surprising to find this group having high
scores on a test which is strongly influenced by one's ability to
recognize that which is socially desirable.
A further analysis of these data shows that the higher associ-
ation index scores of the divergent group are not a function of the
direction of their self-ratings, however. They do not give themselves
iore low self-ratings in the social situation relative to the private
situOtion than eots the convergent group. Seven waaber of the
divergent group gave themselves iowr self-ratings In the social
situation relative to t.he private situation, while seven mfwbers
gave thcrseIves higher ratings. Lieven iuabers of the convergent
grcup gave themselves lower salt-ratings In the social situation
reltlve to the private situation, while 'for members gave them-
seivies higher ratings. This evidence would seem to contradict
the attitudinal owntormity hypothesis.
Questions and inerrienes rglitei to these finalnq4
These findings can be Interpreted as ending support to tae
major .ypothesis; that- inalvlduais wro hece been force, as a
way of aviloing eAceisivt 4nAlety, to iaoiscriminately essimliate
aony social roies will have a graster oevergi et wiciuin their self-
structu.re tlha ,Eiiv6uis t who have been alsoweu to be more dia-
crimititing In titeir acceptance of social roles.
Toe dats coua. e expiainea without necessarily accepting the
uniirsyinrg theory relating to anxiety anJ the iccrporstion of
sociia roles. t ne tt.gni -say that v etever the reason for differences
in e-dj fiai socicu izpproval, individuals with a high need, as
metasurct by the scale, will respond ln the socially approved manner
in a au,.vi-Vail s ,tLon. it .i.i ds Vjil'ji;y i4 ~ ;ie hI iaoow~e-
Crtwn SoCd Cl Desliubi ity SC4ise. f' ui, ilTisurprliitati Wa i'lS Lt
acitui..: i t t. o'i do f ; 4a y.i;upi s Iigwar ratings oi iouif i i the
Irlv1e 2LtutLi.. i 4;-l tOhd u unad group, !cwavrJr. it su aitunld
not account for the tendency (not significant) of the moderate need
group to give itself higher self-ratings in the social situation
than the low need group. One would expect, rather, that the lower
the need for social approval, the higher the self-ratings in the
social situation. The data show a curvilinear effect which is
more congruent with the anxiety, role assimilation hypothesis.
It is interesting to note, as regards the importance of anxiety
in this study, that subjects usually appeared to respond with varying
degrees of anxiety to the social situation. Some were hesitant in
complying with the task after only hearing the instructions, while others
appeared to comprehend the subtle pressures of the situation at some
point during the rating task.
Another interpretation of these data would relate to the age and
education of the subjects. Slightly older and better educated students
are exposed to the university climate which reinforces a constructive
criticism of the expectations of society. The older, more sophisticated
student also has a different test taking response set. He learns to
reject as false those test items which are couched in terms requiring the
acceptance of extremes (i.e., always, never, etc.). This group may also
be more intelligent since these students have managed to stay in school
longer. In this way, one could account for lower scores on the Marlowe-
Crowne Social Desirability Scale and lower self-ratings in the social
situation. H re again, however, this interpretation is not compatible
with the curvilinearity of self-ratings in the social situation by
the three groups. Aiso, differences in education and age are very
Table 9. --- Average age and education of subjects grouped
according to scores on the Marlowe-Crowne
Social Desirability Scale.
Age Years of completed college work
High S-0 group 18.55 1.04
Moderate S-D group 18.27 1.14
Low S-D group 18.96 1.27
As mentioned previously, these findings lend support to those
theorists who have emphasized the contributions of culture In the
development of the personality. The measurement of broad cultural
attitudes has enabled the prediction of behavior in specific tasks
related to self-esteem.
Certainly a replication of this work is necessary, and the use
of other populations would be of Interest. It would be interesting
to conduct this experiment utilizing males; various socio-economic
groups; individuals of various ages end degrees of education; and
various pathological groups (i.e., schizophrenics who theoretically
have fragmented self-concepts and difficulty In recognizing socially
approved behavior). It would be of great interest to examine the
longitudinal development of need for social approval in young children,
particularly during the development of the self-concept. The addition
of other aspects of the self-structure might also prove fruitful;
the unconscious and public aspects might be examined alone or in
conjunction with the private and social aspects. The reliability
of the self-rating scales should be checked, and their relationship
to some behavioral measure would be of interest.
These findings offer some support for the validity of both
the Marlowe-Crowne Social Desirability Scale and the need for social
approval as a motivational construct. One future area of exploration
would be the expansion of the social-desirability scale to provide
a greater range of items and to examine possible conflicts between
need for social approval in various areas. There is no reason to
consider Lhis ds a unitary construct. Rather, an individual may vary
in his need for social approval according to the area or experience
which is Involved (i.e., he may have a high need for social approval
in the area of sexual relationships and a low need for social
approval in the area of dominance, etc.).
Tie possible curvilinearity of the social desirability factor
as an indicator of adjustment leads one to question the validity of
many psychological instruments which purport to measure adjustment
(i.e., Minnesota Multiphasic Personality Inventory) and which may
be measuring need for social approval or the ability to recognize
that which is socially approved. While individuals with low scores
on these Instruments are less well-adjusted, those with high scares
may be less adjusted than their scores would indicate.
At a more general level, these findings add construct validity
to the value of the self-concept in the Investigation of personality.
They also stress the necessity for considering the dynamic interaction
of structures and levels within the personality.
Several general Implications can be derived from the findings
of this study:
In terms of personality theory, some support has been offered
to those theorists who have emphasized levels of self (i.e., Leary
(17)) and role constructs (1.e., Sarbin (25) and Kelly (16)). Further
investigation into the development of social roles and their incor-
poration into the self-picture are needed. Role-role conflicts and
self-role c-nflicts must also be studied more extensively if
sophisticated theories of personality are to be developed. The
incorporation of social roles and norms into the Individual per-
sonality appears to provide a natural bridge between psychology
and disciplines concerned with group and cultural variables
(i.e., Sociology and Anthropology).
In the area of diagnosis, an emphasis on the client's self-
organization and his exposure to and assimilation of conflicting
roles, offers a new perspective to the armamentarium of the
clinician. Kelly's Aole Reperatory Tests (16) may be of value
in researching this area. The recognition by the clinician of the
various aspects of self which will be tapped by various measurement
instruments should also have the effect of increasing the precision
and rigor of these diagnoses. The growing acceptance of the social
desirability variable as a motive, as well as a response set, should
serve as a caution to diagnosticians who propose to correct for it
as if it were irrelevant to the individual's personality.
In terms of the prevention and treatment of psychological
disorders, this study offers support for those clinicians who have
emphasized role relationships, role playing, and role flexibility
in their treatment programs. The approaches of Kelly (16) and Rogers
(22) are particularly relevant. It appears that personality dis-
orders may be associated with the assialiation to too many roles,
thus resulting in self-structure dlscrepances, or the assimilation
of too few roles, resulting In an ability to recog:ize many roles,
and thus tie Inabillty to profit from their utilization.
ECplideological studies might be directed towards uncovering
role conflicts within the cownunity and examining areas where
exposure to various roles and norms Is minimal. Lack of role
assimilettion may be related to a lack of knowledge concerning
roles, the Inability to absorb these roles, or the lack of motiva-
tion to learn these roles.
On the level of family adjustment, one i;ht look for less
cohesiveness in families which Indiscriminately adopt conflicting
roles as a unit. Family congruence, on the other hand, might be
found In families which have failed to aocpt important social roles,
but one would expect these families to be less effective as e unit.
This study attempted to throw some light on the organization
of the self-concept within the individual by relating the social
and private aspects of self to need for social approval.
Ninety coeds were asked to evaluate pictures of six coeds as
to likeability, relative to themselves, in the presence of a male
graduate student and two female peers. This comprised the social
aspect of the self-concept. These same coeds were then asked to
evaluate six other pictures as to likeability In a private
anonymous situation. This comprised the private aspect of the
self-concept. A counterbalanced assignment to situations was
used. Previous to the evaluations, the subjects had been
assigned to high, moderate, or low need for social approval groups,
based upon their scores on the Crowne-Marlowe Social Desirability
A theory of the development of the self-concept within the
individual was presented. This theory stressed the importance
of anxiety and the assimilation of social roles in the development
of the self-concept. It was felt that individuals who had
indiscriminately assimilated many social roles from significant
others would have a wide divergence between the two aspects of
the self-concept. Individuals with high need for social approval
scores were used to represent this group. It was felt that individuals
who had not ndiscriira nately ass iilated many social roles would
have greater convere&rnce tetweer, the two aspects of the self-concept.
Individuals with moderate need fcr social approval scores were used
to represent this gro'p. Finally, those individuals who had been
unable tc esi ii ilte i 'tan u-eful ac' :erningful rcls oI .uld I! ve
great in-kl\ijiI veriWillty ir. the dere-e of dlversence between
aspects cf the self-ccncept. Inrivic'luels with loc' naed for social
approval scores weri. usei to represent this group.
Ecme interesting relationships were found which offer tentative
support for tt.e gererl thcoretic.l framework. The which need for
Cppicvea group hat .a reatter circrrparry between self-ratings in the
two situEtions then the other two groups; it was intermediate in
self-c-ste-e relAtive to the other grstps.; tnd the grcup Las a whole
conforQied tc the detrrc:s (S the scociE'l situt ilon.
The moder te rcecd troup hao 'Tes rlicrepincy in self-ratings
In the two situatit (;: ; the t.ighE. t d(ree of self-cstemr; arnd
r.; d degreee of conforriance to the der.ands of the scctal situation.
The lov need group had the lowest degree of self-esteen; the
greatest individual variability ii, self-ratings ;n the two situations;
and tended to respond opposite to the dcnmands (f the social sl'uation.
The hypothesis thac inrdiruovci's with a wide degree of divergence
between self-rating s in the two situations vculi have lower association
index scores on the Picture identification Test than individuals with
a high degree of convergence was not supported. In fact, the opposite
relationship was found.
because of tne number of correct statistical predictions and
tile numnler uf correct predictions of direction where s.atiisical
significance was noiu ochieveu, it is fe.t that tnis tIuay ienos
tentacive support to icre rajor nypotiesis; cnat indiviiauas wno
lave ueeln tfrcec, ds a way or avoudial excCssive d..xilty, to
indiscriunately assica late erany sucidi roles wi ij neve a greater
divergence. witulil their self-structure tiian inoividuais who ihve
been allowed to be more jiscril1inatini- in iceir acceptance of
This stuoy also lends tentative support co tne methodology
useu and to ie inportaiice or tia social approval construct in
the scudy of tne sei --cncepc.
i. Allison, J. and Hunt, D. E. Social Desirability and the
Expression of Aggression under Varying Conditions of
Frustration. J. Consult. Psychol. 1959, 23, 528-532.
2. Allport, Gordon. The Ego in Contemporary Ps.holoay.
Psychol. Review. 1943, 50. 45.
3. Beloff, Haila and Beloff, J. Unconscious Self-Evaluation
Using a Stereoscope. J. Abnorm. Soc. Psychol. 1959, 59
4. Bernstein, L. The Examiner as an Inhibiting Factor in Clinical
Testing. J. Consult. Psychol. 1956, 20, 287-290.
5. Brownfain, J. J. Stability of the Self-Concept as a Dimension
ot Personality. J. Abnorm. Soc. Psycnol. 1952, 4 597-606.
6. Chambers, J. L. and Lieberman, Lewis R. Picture Identification
Test Manual for Subjects. Charles L. Mix Memorial Fund, Inc.
Americus Georgia, 1960.
7. Chambers, J. L. Trait Juagment of Photographs and Adjustment of
College Students. J. Consult. Psychol. 1961, 2, 433-435.
8. Combs, A. W. and Snygg, D. Individual Behavlor (rev. ed.) Harper
and Brothers, New York, 1959.
9. Cooley, Charles H. Human Nature and the Social Oraer. Scrlbner's,
New York, 1922.
10. Crowne, C. P. and Liverant, S. Conformity Under Conditions of
Personal Commitment. J. Abnorm. Soc. Psychol. 1963, bt. 547-
11. Crowne, D. P. and Marlcwe, D. A New Scale of Social Desirability
independent of Psychopdthology. J. Consult. Psychol. 1960, 24,
12. Edwards, A. L. The Social Desirabilitv Variable In Personcllty
Assessment and Research. Dryden, New York, 1957.
13. Erickson, Erick H. Childhood and Society W.W. Norton and
Company, Inc., New York, 1947.
14. Hall, C. S. and Lindzey, Gardner. Theories of Personality.
John Wiley and Sons, Inc., New York, 1957.
15. Ja:mes, Wiliar.. The Principles of Psycholcqy. Dover Pub.
Co., New York, 1950.
16. Kelly, Georye A. The Psychology of Personal Constructs. 1,
.V'. Norton end Co., Inc., Nz'; York, 1355.
17. Leary, T. inter3crsonal Diarncsis of Pcrsonality. Ronsld,
New York, 1957.
18. Lieben an, Lewis R. and Chambers, J. L. Differences Between
Prisoners and Trade Schgol Stud'Entr n the Picture lientifl-
cation Test. Based on a paper read at the Southeastern
Psycholoc;ical Associatior Corvzntion in Miami, Florlda, 1963.
15. Mead, G. H. Min.c, 3Slf sn Scciety. Univc-sity of C-.i;ca~ Press,
20. Newcomb, T. M. Social Psycholoqy. Holt, Rinehart and Winston,
Hew York, 13S0.
21. Rogers, A. H. :nd l,- lsh, T. M. Dcfensiveress enc' Ur'nwittin3
Self-Evaluation. J. Clin. Psychol. 1958, Ijj 302-304.
22. Rogers, C. R. On Becoming a Person. Houghton-Mifflin Co.,
New York, 1961.
23. Rotter, J. B. Social Ler-.irng d li.i:csl PSycholo'1.j Prentice
Hall Co., New York, 1954.
24. Sarbin, T. R. A Preface to a Psychological Analysis of the
Self. ?5syc.l. Rev. 1332, 2 1!-22.
25. Sorbir., T. P.. and Hair-yck, Curt;z C. "Co.,for:nnce in Role
Perception as a Personality Variable". In T. R. Sarbin (ed.)
:t:c 1: 32:hsvicr Path:lc~ hclt, RIe-.ar't :r.d Winston,
New York, 1961.
26. Seigel, Sidney. Nonparametric Statistics for the B~hevioral
W :Cy, s. I rc.. i, 'c I ;.7..
WI l'y, New York, i9-7.
28. Strickland, B"nnie R. and Crowne, Douglas P. Need for Anproval
and the Prerature Termination of Psychotherapy. J. Consult.
?PXschl. 1963, 2Z, 476-485.
29. Sullivan, H. S. The Interpersonal Theory of Psychiatry. Perry,
H. S. and Gawel, M. L. (eds.) W.W. Norton Co., New York, 1953.
30. Synonds, P. '. Thle Ego and The Seif. Appieton-Century-Crofts,
New York, i9r1.
31. Werner, Heinz. Conarative P-cholqcr of mental Development.
international Universities Press, Inc., New York, i357.
32. Wylie, Rutn. Thie Self Concet. University of Nebraska Press,
Lincoin, Nebraska, 1961.
Table 10. -- Average age and education of subjects.
Age Years of completed college work
Florida 19.15 1.39
Wisconsin 18.50 1.02
Table 11. ---Average self-evaluations of subjects in
the social and private situation.
Private self-ratings Social self-ratlngs
Florida 9.09 9.98
Wisconsin 9.15 9.24
Combined 9.12 9.60
*Rw score represents the subject's ratings of the pictures
and not self-ratings. Thus, high raw scores are equivalent
to low self-ratings.
Table 12. --- Average scores of subjects on the HMrlowe-
Crowne Social Desirability Scale.
Table 13. --- Number of Wisconsr.n and Florida subjects
in social desirability groups.
ilgh S-D Moderate S-D Low S-D
Florida 13 10 21
Wisconsin 17 20 9
Comb ine 30 0 30
Table 14, -- Average self-ratings of Florida and Wisconsin
subjects grotped according to need for social
Need for Socis Approva Florida
Private Situation Social Situation
High 7.33 10.31
ttoderate X 7.50 x 8.30
Low 10.91 10.57
The Florida sample had higher self-ratings in their high group than
the Wisconsin sample (t = 2.05, df = 58, p< .05 two-tailed, t test for
The Florida and Wisconsin samples did not differ in self-ratings for
the moderate groups (t = .56, df = 58, p> .10 two-tailed, t test for
The Florida s~cple had lower self-ratings In their low group ti.
the Wisconsin sample (t = 2.05, df = < tw-tiled, t test
for uncorrelated means) .
Tible 15. ---Self-ratings of 11 sujects grouped accur
to need for social approval and situation.
Subject Private Social Private Social Private Social
-nuiber Situation Situat ion i ion Stuation Situation Situatution
Rating Pating Rztlny Rating --Lii atlnrg
2. 13 5 7
3,. I. 1I 7 i
*. iO 10 5 I Id i
6. 12 7 g 13 10
7. 8 S & 3 12 I
6. 11 3 I1 V; 13 10
10. il 1; 7 S 10
11. i 13 3
12. 10 3 1
13. 5 11 1 10 0
I. 0 10 3 t0 3 2
17. 11 il 5 I I1 5
1l. 9 12 13 13 2
18. 1 9 13
20. i 17 9 O
21. 8 12
22. 10 Y 3 ,
23. 4 2 11 12 5 12
24. j 7 7 5
2. 10 10 5 0
26. 1014 5 12
27. ii 13 5 it 12 K
2. 9 13
30.6 It 17 12
Mack R. Hicks was born April 16, 1935, in Tulsa, Oklahoma.
In 1953 he was graduated from Watertown High School, Watertown,
Wisconsin. In June, 1957, he received the degree of Bachelor of
Science from Notre Dame University. From 1958 until 1960 he
served as a supply officer in the United States Air Force and
was stationed in England. Following his discharge from the
Air Force in 1960, he enrolled in the Graduate School of the
University of Florida. He worked as a graduate assistant in
the Department of Psychology in 1961, and received the degree
of Master of Arts in 1962. From February, 1962, until the
present time, he has pursued his work toward the degree of
Doctor of Philosophy.
Mack R. Hicks is married to the former Kathleen Susan Raue
and Is the father of three children. He currently holds a
Florida State Mental Health Grant and is a member of Psi Chi
Fraternity and the Southeastern Psychological Association.
Tiis dissertation wos prepared under the direction of
the chairman of the candidate's supervisory co~rittee and has
been approved by all ennbers of that conxittee. It waa sub-
nitted to the Decn of the Colleye of Arts jnd Sciences and to
the Graduate Council, and was approved as partial fulfillment
of the re uireents for the decree of Doctor of Philosophy.
SAugust U, Sd&4
Dean, Colle /kf Ars & Sciences
Dean, Graduate Scinol
C: ia i man