In terms of the self, examine your personal background and experiences. In an essay (500-750 words), address the following prompts. Cite two to four scholarly sources to support your claims: Explain h

The Journal of Social Psychology, 1968, 74, 251-258.

EXPERIMENTS ON FACTORS RELATED TO SELF-CONCEPT CHANGE* Department of Sociology, San Francisco State College JOHN W. KINCH A. THEORETICAL FRAMEWORK The Cooky-Mead formulations concerning the self-concept are primarily focused on the relationships between other persons' responses to an individual and his conception of himself.

More specifically, they argue that a person's conception of himself comes about as a result of the way he perceives the re- sponses of others toward him [Cooley's "looking-glass self" (5) and Mead's "taking-the-role-of-the-other" (14)].

This, they aver, is particularly impor- tant, since the way the individual sees himself (his self-concept) has the func- tion of directing, or influencing, the way he will behave.

This approach has had wide acceptance among social psychologists, but not without some reservations. The vague concepts and general propositions make its application very difficult. However, when the basic ideas are used as a gen- eral framework from which more specific statements are developed, the theory IS most fruitful. These extensions or elaborations are not necessarily derived from the theory in a logical sense, but are intervening claims that reduce the generality or abstractness of the framework and bring it conceptually closer to the specific phenomena it purports to explain.

In the study reported here, previous work on one of the propositions is used as a framework from which specific hypotheses are developed to be tested. The proposition may be stated as follows: The individual's conception of himself is based on his perception of the way others are responding to hi?n.

There is am- ple research evidence supporting this general relationship between individuals' perceptions of others and their self-concepts. However, few writers in this tra- dition have paid specific attention to how this process works. It is obvious that the theorists are not arguing that every time a new response is directed toward an individual there will be a corresponding change in his self-concept. What is needed is an expansion of this general proposition so that the relationships are made explicit. The basic question seems to be: Under what conditions do the • Received in the Editorial Office, Provincetown, Massachusetts, on December 14 1966.

251 252 JOURNAL OF SOCIAL PSYCHOLOGY perceptions of the others' responses have an effect on the individual's self- conception?

The theoretical framework and what evidence there is available in the liter- ature suggest that the effect of perceived responses on the self-concept is a function of a series of factors involved in interpersonal contacts. They are (a) the frequency of responses in the course of these interpersonal contacts, {b) the perceived importance of the contacts, (c) the temporal proximity of the contacts, and (d) the consistency of the responses resulting from the contacts.

It should be understood that each of the hypotheses that follows is preceded by an implied "other-factors-being-equal." 1.

Hypothesis 1 {Frequency) The more frequently the individual perceives others as responding toward him in a particular way, the more likely he is to align his self-concept with the perceived responses.

Although very little attention is given to the direct test of this notion, there are many studies that implicitly do so. For example, experimental studies have for the most part given the subjects specific ratings or evaluations which, in effect, increase the frequency of these responses, thus indirectly testing this hypothesis (4, 9, 13, 17).

Hypothesis 2 (Importance) The more important the individual perceives the contact between himself and the others to be, the more likely it is that the individual's perceptions of the responses of the others will be used in defining his self-image. It is widely accepted that contacts with "significant others" are required before the indi- vidual's self-concept will be affected. These "significant others" may take the form of prestigeful persons (experts) or of personal acquaintances (friends).

However, the research evidence available is not completely consistent on this subject (4, 10, 11).

Hypothesis 3 (Temporal Proximity) The individual's concept of himself is a function of (a) the earliest evalua- tions he receives on a particular attribute and (b) the most immediate evalua- tions. This hypothesis has two parts. First, it is contended that the first con- tacts which the individual perceives as favoring a particular self-concept are of extreme importance. Self-conceptions that develop in early childhood are likely to persist throughout life. No direct evidence is available on this hypothesis, but there is considerable indirect evidence when one considers the selective as- JOHN W.

KINCH 253 pect of early evaluations.

If tbe author's basic contention is true that tbe re- sponses of others are used in the original formation of the self-concept, tbere is ample evidence in the literature to suggest that persons choose friends and join groups which they perceive as evaluating them congruently witb their percep- tion of themselves (2, 3, 8, 15).

Therefore, once an individual develops a con- ception of himself, he will interact as much as possible witb others wbo will reinforce this conception for him. The second part of Hypothesis 3 suggests that those most immediate contacts are important in understanding tbe indi- vidual's self-concept at any given time. Almost without exception those studies which have compared perceived responses of others to self-concepts have dealt with responses of others in tbe immediate situation (7, 16). In one study that directly confronted this issue, it was found tbat the most exaggerated changes were observed immediately following the experiment. However, some signifi- cant change still remained six weeks following the experiment, whicb suggests some lasting effects (9).

Hypothesis 4 {Consistency) The more the individual perceives a consistent pattern in the responses of others, the more likely he is to let this affect his self-concept.

RESEARCH DESIGN The researcb presented here empirically evaluates certain aspects of the hy- potheses suggested above.

It employs a series of experiments—each following the classical design with before and after tests on both experimental and con- trol groups.

The dependent variable concerns changes in one aspect of an in- dividual's self-concept and the independent, or experimental, variables consist of the "factors" (frequency, importance, temporal proximity, and consistency) wbicb are varied, one at a time, in the experimental situations.

In the present study tbe notion of the self-concept is defined as the organiza- tion of qualities which the individual attributes to himself.

Although this self- concept is "organized," it cannot be measured on a single continuum of self- regard (as many social psychologists have used the concept), but must be analyzed in terms of individual attributes or clusters of attributes. There seems to be substantial evidence for this contention (1, 6, 13, 17). In the pres- ent study, subjects were required to evaluate themselves on several descriptive objectives by use of a seven-point scale. Rather than attempt to accumulate these self-evaluations, the investigation concentrated on one adjective "leader- ship." As in other experimental studies of the self, the subjects were asked to eval- 254 JOURNAL OF SOCIAL PSYCHOLOGY uate themselves before and after they were involved in an experimental situa- tion. In the set of experiments, conditions were varied in order to demonstrate the effect of those factors which the hypotheses suggest should influence self- concept change. In order to specify the procedure more clearly, a detailed de- scription follows: Each experiment followed a design whereby all participants, with one exception, were confederates of the investigator. During the activity in the experimental setting, the one naive student was assigned a position of leadership and was required to direct the others. On completion of this activity the confederates were presented to the naive subject as students who were ex- perts in organizational dynamics, all having taken advanced courses on the subject. They were then asked by the investigator to rate the naive subject's performance as a leader. This was done according to a prearranged system, independent of the subject's actual performance. The ratings were communi- cated by the verbal response of each confederate and were marked on a scale on the blackboard in front of all the participants. The confederates were pre- instructed to perform the task reasonably well regardless of the naive subject's orders, so that the prearranged ratings did not appear out of line.

Four variations in the confederates' ratings made up what were called "ex- perimental conditions" and were designed to test several of the hypotheses. In Experimental Condition 1, each subject was rated favorably by all experts with very little variation in ratings. In the second condition (E.C. 2) each subject was involved in two sessions, thus doubling the number of evaluations which they received. Experimental Conditions 3 and 4 followed the same pat- tern as E.C. 1, except that in E.C. 3 only five confederates were used and a sixth rating was given by the investigator conducting the experiment, and in E.C. 4 the six ratings were more dispersed and less consistent than in E.C. 1, although the average ratings were the same for both. These categories were compared with a control group: a matched category of students who did not participate in the experiment.

All the prearranged ratings were higher than the individual had rated him- self before the experiment. The question of the consequence of lower ratings was not examined. In a limited experiment in which negative ratings were given a number of subjects, the results seemed to suggest that these ratings were even more influential in changing self-concepts than were the positive ratings. However, for ethical reasons this study was limited to only a few stu- dents and was not continued after the first responses were considered (12).

About half of the naive students were interviewed and questioned about their impressions of the experiment. Although there were several reasons for these interviews, one major concern was with establishing some idea of the subjects' JOHN W. KINCH 255 perception of the ratings they were given. Since the theory suggests that it is the way the individual perceives the responses of others that changes his self- concept, it was felt that the investigators must make certain that the favorable ratings which were given the subjects were perceived as such.

The difficulty of setting up an experiment that would test the hypotheses of this theory is apparent. The experimental situation had to be somewhat artifi- cial. The ratings were rather formal compared with the ratings that one gets in everyday life. The student participation in the experiment took a relatively short time (10 minutes). Four to five weeks elapsed between the before and after tests. During that time these freshmen in college were experiencing many other situations relevant to their self-concept. For these reasons the sensitivity of the experiment and its significance for the respondents was seen as one of the major problems to be dealt with in the research presented here. Several de- vices were employed to increase the significance of the situation for the sub- jects.

The subjects were required to work with the other members of the ex- periment "as a team," and it was this "team" that rated them. In other experiments of this type, the subjects were rated by someone (expert, stranger, or friend) who was not engaged with them in the experiment. In the inter- views, the subjects indicated considerable concern over what ratings they re- ceived and considered the ratings as favorable. This suggests that the others in the experiment were significant to them.

The problem of sample size is particularly difficult in this type of design, since, for each subject, there must be a separate experiment, plus complete be- fore and after test information. When the total number of cases is divided into four experimental conditions and a control group, the number of cases in any one condition is likely to be cut seriously low. This was the case in the present study, which started with 105 subjects, 21 assigned to each of the four experi- mental conditions and 21 to the control group. Some of the disadvantages of the small numbers were reduced by precision matching and randomization of assignments.

C. RESULTS By comparing before-after changes in self-concept ratings on leadership for the subjects in each of the conditions, evidence is brought to bear on at least part of three of the four hypotheses mentioned above. The complete results are reported in Table 1.

There are several ways in which these data may be applied to Hypothesis 1.

In this hypothesis it is suggested that the more frequently a person perceives a particular type of response directed toward him, the more likely he is to use 256 JOURNAL OF SOCIAL PSYCHOLOGY TABLE 1 MEAN SELF-RATINGS ON LEADERSHIP BEFORE AND AFTER EXPERIMENTAL EVALUATIONS BY EXPERIMENTAL CONDITIONS Experimental condition E,C, 1 (Standard) E,C, 2 (Repeat) E,C, 3 (Importance) E,C.

4 (Dispersed) All experimental conditions Control group N 19 18 19 15 71 21 Before M 4,26 4,28 4,26 4,20 4,25 4,29 SD 1,02 1,04 1.12 1,11 1,07 1,16 After M 4,63 5,28 4,89 4,80 4,90 4,52 SD ,93 ,73 ,73 .83 .83 .96 Before-after differences M .37 1,00 ,63 .60 .65 .23 mu .98 1,15 .74 .92 .99 .75 t 1.59 3,57** 3,62** 2,36* 5,45** 1.42 Note: All tests are one-tailed. The t's above refer to the difference between the before and after means of dependent samples, » p< .05, ** # < ,01.

that response in changing his self-concept. Since all those taking part in the experiments had a particular type of response (favorable ratings on leader- ship) directed toward them, we can assume that "on the average" this group has had more of this type of response than the control group. Therefore, one test of the hypothesis is a simple comparison of those in the experimental groups with those in the control group.

The results are in the expected direc- tion with mean changes of +.65 of a unit on the rating scale for those in the experimental groups (2V = 71) and only -\-.23 of a unit for the control group {N = 21). The difiEerences between the two changes are statistically signifi- cant at the .05 level of significance {t = 1.76, difference between independent mean test).

A more refined test on this hypothesis involves comparing those in E.C. 2 with those in E.C. 1 and the control group, since the subjects in E.C. 2 had more of a particular type of response directed toward them than did the sub- jects in E.C. 1. The results support the hypothesis. Those who performed under E.C. 2 showed a mean change in their self-conceptions of one complete unit on the rating scale, while those in E.C. 1 changed only .37 of a unit.

This difference is statistically significant {t = 2.29, difference between inde- pendent mean test).

From Hypothesis 2 (Importance) it was felt that the added prestige of the investigator's rating in E.C. 3 would lead to greater changes than in the com- parable E.C. 1.

Here the results are in the expected direction; however, the difference between the two experimental groups was small and not statistically significant (means of .37 and .63 for E.C. 1 and 3, respectively—t = .93).

Hypothesis 4 suggests that the consistency of responses is important and JOHN W.

KINCH 257 would lead to the expectation of a greater before-after difference in E.C.

1 than in E.C.

Here the results show just the opposite. Those students who were given about the same ratings by all the raters changed less than those subjects who were given more dispersed ratings (mean changes of .37 in E.C.

1 as compared with a change of .60 in E.C.

4).

SUMMARY The study described here is one in a series of studies proposed by the author designed to vary systematically factors relevant to changes in self-conceptions.

Although experimental studies of this type are hampered by the several diffi- culties apparent in this report, the valuable manipulative power that the inves- tigator has over his variables allows crucial tests which could not be accom- plished by other methods.

The need for systematic empirical support or investigation of social psychological theory is recognized by all.

This study has attempted to provide that tj'pe of support for the Mead-Cooley notions about the self-concept.

REFERENCES 1.

AKERET, R. U.

Interrelationships among various dimensions of the self concept.

Counsel. PsychoL, 1959, 6, 199-201.

BACKMAN, C. W., & SECORD, P. F. The effect of perceived liking on interpersonal attraction.

Hum.

Relat., 1959, 12, 379-384.

Liking, selective interaction, and misperception in congruent interper- sonal relations. Sociometry, 1962, 25, 321-335.

4. BERGIN, A.

The effect of dissonant persuasive communications upon changes in a self-referring attitude.

Personal., 1962, 30, 423-438.

CooLEY, C. H.

Human Nature and the Social Order.

New York: Scribner, 1902.

COUCH, C.

Family role specialization and self-attitudes in children. Sociolog.

Quart., 1962, 3, 115-122.

7. DAVIDSON, H.

H., & LANG, G.

Children's perceptions of their teachers' feelings toward them related to self-perception, school achievement and behavior.

Expcr. Educ, 1960, 29, 107-118.

8. DiTTES, J. E.

Attractiveness of group as a function of self-esteem and acceptance by group.

Abn. & Soc. PsychoL, 1959, 59, 77-82.

HAAS, H. L., & MOEHR, M. L.

TWO experiments on the concept of self and the reactions of others.

Personal.

Sf Soc.

PsychoL, 1965, 1, 100-105.

10.

HARVEY, O. J., KELLEY, H. H., & SHAPIRO, M. M. Reactions to unfavorable eval- uations of the self-made by other persons.

Personal., 1957, 25, 393-411.

11.

KENNEDY, J. L., & LASSWELL, H. D. A cross-cultural test of self-image. Hum.

Organization, 1958, 17, 41-43.

12. KINCH, J. W. The manipulation of subjects in experiments. Unpublished paper presented at the Pacific Sociological Association Meetings, Vancouver, British Columbia, Canada, 1966.

13.

MAEHR, M., MENSING, J., & NAFZGER, S. Concept of self and the reaction of others. Sociometry, 1962, 25, 353-357.

14. MEAD, G. H.

Mind, Self, and Society. Chicago, III.:

Univ. Chicago Press, 1934.

258 JOURNAL OF SOCIAL PSYCHOLOGY 15.

REESE, H, W. Relationships between self-acceptance and sociometric choices, J. Abn. Sf Soc. Psychot., 1961, 62, 472-474.

16.

RosENGREN, W. K, The self in the emotionally disturbed, Amer. J. SocioL, 1961, 66, 454-462.

17.

ViDEBECK, R. Self-conception and the reaction of others. Sociometry, 1960, 23, 351-359.

Department of Sociology San Francisco State College 1600 Holloway Avenue San Francisco, California 94132