self-esteem: a self-social construct

tournat of Consulting and Clinical Psychology1969, Vol. 33, No. 1, 84-95

SELF-ESTEEM:

A SELF-SOCIAL CONSTRUCT1

ROBERT C. ZILLER, JOAN HAGEY, MARY DELL C. SMITH

University oj Oregon

AND BARBARA H. LONG

Goucher College

Self-esteem is defined within a context of self-other orientation, and an instru-ment is described which is designed to measure the evaluation of the self inrelation to significant others using topological representations of self and othersand involving limited verbal demands. The results of a program of research aredescribed which emanate from an evolving theory of social self-esteem. Theresults suggest that self-acceptance and social acceptance are inextricablycombined and raise serious doubts about the meaning of earlier results con-cerning self-esteem which were based upon verbal self-reports.

Self-esteem is usually defined as the indi-vidual's perception of his worth. In evaluat-ing the self, however, few physical cues areavailable which provide a reliable basis for anestimate. Still, the individual has recourse topaired comparisons of the self and significantothers; that is, self -evaluation evolves interms of social reality (Festinger, 1954). Self-evaluation, then, emerges largely within a so-cial frame of reference.

According to this definition of self-esteem,if the social environment changes, a corres-ponding change in self-esteem may be antici-pated. It is now proposed that the person'sresponse to the social environment is a func-tion of self-esteem. Self-esteem mediates so-cial stimuli and response (Social stimuli— >Self-esteem — > Response) . It is proposed herethat self-esteem is a component of the selfsystem which regulates the extent to whichthe self system is maintained under conditionsof strain, such as during the processing ofnew information concerning the self. Thus, forexample, evaluations of either a positive ornegative nature do not evoke immediate, cor-

research program from which this reportemanated was supported in part by a grant to thesenior author by the National Science Foundationand in part by the United States Office of Educa-tion through a contract to the Center for the Ad-vanced Study of Educational Administration, Uni-versity of Oregon.

2 Requests for reprints should be sent to RobertC. Ziller, Department of Psychology, University ofOregon, Eugene, Oregon 97403.

responding action by the individual with highself-esteem. New information is examined onthe basis of its relevance and meaning for theself system and is disregarded if its meaningtends to be tangential. In this way the or-ganism is somewhat insulated from the en-vironment or is not completely subject to mo-mentary environmental contingencies.

Persons with low self-esteem, on the otherhand, do not possess a well-developed con-ceptual buffer for evaluative stimuli. In Wit-kin's terms (Witkin, Dyk, Foterson, Good-enough, & Karp, 1962), the person with lowself-esteem is field dependent; that is, hetends to passively conform to the influence ofthe prevailing field or context. Since the indi-vidual's behavior is directly linked to immedi-ate environmental circumstances and is notmediated or differentiated and integrated bythe self-concept, he is thereby inclined towardinconsistency.

Thus, the concept of self-esteem as de-scribed here is linked to the concept of person-ality integration as used by Lewin (1935).Development was described by Lewin as in-cluding an increase in the number of the rela-tively independent subparts of the person(differentiation) and increasing the unity ofthe person (integration or organization). Simi-larly, Piaget (1947) presents the concepts ofassimilation and accommodation. It is antici-pated that the behavior of persons with highself-esteem is more integrated and that theircognitive processes are characterized by a

84

SELF-ESTEEM: A SELF-SOCIAL CONSTRUCT

selective consideration of relevant social ele-ments (stimulus control).

COMMUNICATION AND THE MEASUREMENT OFSELF-ESTEEM

Previous research concerning self-esteem hasnot emphasized sufficiently the social nature ofthe self system. The failure to incorporate andweight social factors within the self-evalua-tion framework may have contributed, in partat least, to the disppointing state of the inves-tigation of self-esteem.

A second shortcoming of earlier studies inthis area is their descriptive nature, which,coupled with the serious shortcomings of themeasurement techniques, have left the areaat a low level of theoretical development.

Finally, and most seriously, previous re-search has largely involved a verbal self-report measure of self-esteem. Kelly (19SS)qualifies his assertions continuously by point-ing to the most tentative of his assumptions,that S's word labels for his constructs meanwhat the examiner thinks they mean. Finally,Kelly suggests that if a testcan be arranged to produce a kind of protocol whichcan be subjected to a meaningful analysis, indepen-dent of words, we shall have made progress towarda better understanding of the client's personalconstructs [p. 268].

The approach used here involves a methodof communcation with limited verbal demandsand relies primarily upon rudimentary formsof abstraction which are assumed to predateverbal communication systems.

The approach is a confluence of the ap-proaches of DeSoto and Kuethe (19S9), Kelly(19SS), and an evolving theory of self-otherorientation. The measurement approach as-sumes that the human organism finds itexpedient to order and categorize or tostructure generally the multitude of self-surrounding stimuli. The processes used bythe respondents are expected to be somewhatidiosyncratic, but owning to commonalityamong human experience, sensory processes,and classification systems, the evolving ab-straction systems possess sufficient common-ality that the basis of a communicationsystem exists. Some of these processes includeextent of separation between objects (Kuethe,1962), number of objects in a category, and

ordering of objects (DeSoto, London, &Handel, 196S).

The most relevant ordering process withregard to self-esteem is what DeSoto, London,and Handel refer to as "spatial paralogic"and "linear ordering." It is observed thatpeople are prone to place elements in a linearordering to the exclusion of other structures,and that they handle linear ordering moreeasily than most other structures (Coombs,Raiffa, & Thrall, 19S4; DeSoto, London, &Handel, 196S). Indeed, DeSoto, London, andHandel note that serial ordering proceedsmore readily in a rightward direction thanin a leftward direction. The tendency toattribute greater importance to the objectplaced at the extreme left position in a hori-zontal display has been noted by Morgan(1944).

The measure of self-esteem developed hereutilized the serial ordering predilection of 5swithin a social context. (See Figure 1.)

The measure involves presenting a horizon-tal array of circles and a list of significantothers (including the self) such as those usedby Kelly (1955). The task requires 5 to as-sign each person to a circle. The score is theweighted position of the self. In accordancewith the cultural norm, positions to the leftare assumed to be associated with higherself-esteem.

The item in Figure 1 is one of six items inthe student form of the instrument. The otherfive self-esteem items involve the followingsets of significant others: (fl) doctor, father,friend, mother, yourself, teacher; (b) some-

DIRECTIONS. The circles below stand for people.Mark each circle with the letter standing for one ofthe people in the list. Do this in any way you like,but use each person only once and do not omitanyone.

F-someone who is flunkingH-the happiest person you knowK-someone you know who is kindS-yourself

Su-someone you know who is successfulSt-the strongest person you know

ooooooFIG. 1. Measure of self-esteem.

86 ZILLER, HAGEY, SMITH, AND LONG

one you know who is a good athlete, someoneyou know who is a good dancer, someone youknow who is funny, someone you know whogets good grades, yourself, someone you knowwho is unhappy; (c) an actor, your brotheror someone who is most like a brother, yourbest friend, a dean of students, yourself, asalesman; (d) someone you know who iscruel, your grandmother, a housewife, a po-liceman, yourself, your sister or someone whois most like a sister; (e) doctor, father,friend, nurse, yourself, someone you knowwho is unsuccessful.

In a study involving 75 randomly selectedstudents from Grades 7 through 12, the split-half reliability (odd-even) was .80 correctedfor length (Long, Ziller, & Henderson, inpress). Split-half reliability (odd-even) forthe adult form3 was .85, uncorrected forlength, in a study (Mossman & Ziller, 1968)involving 60 neuropsychiatric patients. Test-retest reliability for 86 sixth and seventhgraders was .54 for the student form.

VALIDATION

The measure of self-esteem proposed hereis assumed to involve social reasoning anda norm of hierarchical ordering of social ob-jects in a horizontal line from left to right.This assumption was examined in a series ofseparate studies.

In the first of these (Ziller, Megas, &DeCencio, 1964), 45 patients in an acuteneuropsychiatric treatment ward were pre-sented with seven circular pieces of whitefelt cloth 2 inches in diameter. The circles

3 The six sets of social objects included in theadult form of the instrument are: (a) doctor,father, a friend, a nurse, yourself, someone you knowwho is unsuccessful; (6) doctor, father, friend, poli-tician, yourself, an employer; (c) someone you knowwho is a good athlete, someone you know who ispopular, someone you know who is funny, someonewho knows a great deal, yourself, someone youknow who is unhappy; (d) an actor, your brother orsomeone who is most like a brother, your bestfriend, yourself, a salesman, a politically active per-son; (e) someone you know who is cruel, a judge,a housewife, a policeman, yourself, your sister orsomeone who is most like a sister; (/) a defeatedlegislative candidate, the happiest person you know,someone you know who is kind, yourself, someoneyou know who is successful, the strongest person youknow.

were marked by symbols indicating the personthey represented. These persons includednurse (N), nurses' aide (NA), other patientsin the ward (0), psychiatrist (Pi), psycholo-gist (Po), social worker (SW), and yourself(Y). A list of the symbols and their referentswere placed on a table in alphabetical orderfor the S's information. These were also readto the Ss. The 5s were instructed to arrangethe circles on a black felt board, 2 X 1\yards, in any way that they wished.

In support of the serial ordering tendenciesproposed by DeSoto, London, and Handel, themajority of 5s arranged the symbolic circlesin a straight line from left to right. Byassigning weights to the left-right positionsand calculating the mean weighting of theseven symbols, the resulting order of thesymbolized positions was: psychiatrist, psy-chologist, social worker, nurse, nurses' aide,other patients, and yourself. It is apparentthat a left to right status hierarchy of thesocial objects emerges.

A similar analysis was made of the left toright location of a low-status other personby college students using the student form(see Table 1). It was noted that the"unhappy" person was placed in the lastposition to the right 48% of the time;"someone you know who is unsuccessful,"56%; "someone you know who is cruel,"64%; and "someone you know who is flunk-ing," 59%.

Further support for the assumption of aleft to right paralogic was found in the associ-ation of the self-esteem score as previouslyderived and a second technique of scoring.

TABLE 1

DISTRIBUTION OF THE LOCATION OF THE "NEGATIVESIGNIFICANT OTHERS" IN FOUR ITEMS OF THE SELF-

ESTEEM MEASURES, HORIZONTAL ARRANGE-MENT POSITIONS

Item

UnhappyUnsuccessfulCruelFlunking

6

16212731

S

7321

4

7542

3

7913

2

15534

1

48566459

Note.—For Unhappy, W = 150; for Unsuccessful, ff = 147;for Cruel, AT = 154; for Flunking, N = 172. Percentage signsare omitted.

SELF-ESTEEM: A SELF-SOCIAL CONSTRUCT 87

This involved the identification of the mostnegative, significant other for each set ofsignificant others and calculating the distance,in number of circles, between the low-statusother and the self. (Only four items whichinvolved a clearly differentiated low-statusother were included in this analysis. Theseitems were a, b, d, and e.) This method ofscoring was suggested when it was noted inTable 1 that the low-status other was some-times located in the left position, indicatingthat arrangement of the self and others maybe based on other than a left to right hier-archical ordering in some cases.

The correlation between the scores derivedby these methods was .33, p < .05 (N = 163male and female college students). Only theresults with regard to males were statisticallysignificant, however (r = .46, N = 61, p <.001; females, r - .14, N = 102, p < .10).

A similar analysis of the two scoringmethods was made using the adult form withmale neuropsychiatric patients (Mossman &Ziller, 1968). The results corroborate thefindings with regard to the male collegesample (r = .56, N = 60, p < .001).

In yet another study of the left-right serialordering phenomenon, a children's form of theself-esteem measure was used (Henderson,Long, & Ziller, 1965). The analysis involvedthe responses of 48 boys and girls ranging inage from 7 to 14 who had applied for cor-rective training at a reading study center, plus48 controls matched for age, sex, and generalintelligence. The 5s were given a paper witha long horizontal line. They were next pre-sented in random order six circles with pic-tures representing self, friend, and a "smart,""dumb," "funny," and "bad" classmate. Thechildren were told to paste these symboliccircles in a row on the line. It was found thatchildren placed the "smart" classmate to theleft and a "bad" classmate to the right to asignificant degree.

Evidence that the left-right serial orderingis not a phenomenon limited to persons withinthe United States is found in the analysis ofthe location of "someone you know who isunsuccessful" in the student form of Item e.With regard to 92 boys and girls from Form Iof the M.V.D.M. High School in Vesakapat-nam, Andhra, South India, the frequency with

which the "unsuccessful" person was locatedin the positions from right to left was 74, 9,4, 0, 1, and 4. With regard to an Americansample of 94 boys and girls, the correspondingfrequencies were 53, 17, 15, 3, 3, and 3.

Another test of the left to right hierarchicalordering assumption was the association be-tween the weighted position of "yourself"among five others including "someone youknow who is flunking," "the happiest personyou know," "someone you know who is kind,""someone you know who is successful," and"the strongest person you know" under condi-tions where the social objects were to be ar-ranged horizontally as in Figure 1 and verti-cally. In the vertical display, the higherposition of the self is assumed to representhigher self-esteem. The correlation betweenthese two measures was .50 (N — 82,P< .05).

An analysis of the location of the loweststatus other, "someone you know who isflunking," indicates (see Table 2) that thenumber of reversals in the placement of thelow-status other is reduced in the verticalarrangement (9% vs. 31%). The verticalarrangement may introduce greater itemvisibility, however.

A third approach to the validation of thesocial self-esteem (SSE) measure was a cor-relational analysis of SSE with existing mea-sures of the construct. The measures selectedfor comparison were those most frequentlyreferenced in the literature (Wylie, 1961) anda more recent device developed for researchpurposes by Cutick (1962) and used byDiggory and her collaborators (Diggory-Farnham, 1964). With the exception of the

TABLE 2DISTRIBUTION OF THE LOCATION OF "SOMEONE WHO

is FLUNKING" IN Two IDENTICAL SELF-ESTEEMITEMS (VERTICAL vs. HORIZONTAL

ARRANGEMENTS)

Locationarrangement

VerticalHorizontal

6

931

5

.061

4

.062

3

33

2

34

1

8459

Note.—-In the vertical arrangement location "6" was thefirst or top position in the hierarchy. In the horizontal ar-rangement location "6" is the first position in the left-righthierarchy. Percentage signs are omitted.

ZILLER, HAGEY, SMITH, AND LONG

SSE, all the measures were based on self-reports. Thus, the Bills-Vance-McLean(1951) Index of Adjustment and Valuesrequired S to rate himself with reference toeach of 49 adjectives as to how often he was"this sort of person." Six-week test-retestreliability was .90 (N = 100).

Diggory's Self-Evaluation Questionnaireasks 5 the percentage of time that he ex-pected to succeed in eight given situations.The reliability is not reported.

Coopersmith's (1959) Self-Esteem Inven-tory contains 54 items concerned with S'sperceptions in four areas: peers, parents,school, and self. The form was modifiedslightly to make it more appropriate for acollege population. The Self-Esteem score istwice the sum of the high self-esteem items(as agreed upon by five psychologists.) marked"like me" and low self-esteem items marked"unlike me." Reported test-retest reliabilityafter 5 weeks was .88.

In an earlier test of the relationship be-tween the SSE and a single item, overall self-evaluation (Ridgeway, 1965), a negative butnot statistically significant relationship hadbeen found (r - -.15, N = 100, p < .5).Given the different theoretical frameworksand method of communication upon whichthe measures are based, a significant relation-ship was not anticipated in the present study.The purpose of this study in the programwas to establish the independence of the SSEmore systematically.

The correlation matrix for these threescales and the SSE for each sex is shown inTable 3. None of the correlations with SSE

were statistically significant. Once again, sexdifferences are quite apparent in the inter-correlations among the measures. For maleSs (N = 33), the highest correlation, r = 60,was between Diggory's and Bills' measures.Significant correlations with regard to maleSs were also found between Diggory's andCoopersmith's measures (r = .37), and Bills'and Coopersmith's (r = .46). The only signifi-cant correlation found for female 5s (N = 53)was between Diggory's and Bills' measures(r = .29). These results are worrisome, eventhough they were anticipated. Yet, the resultsmay be interpreted to indicate that the SSEand the other measures of self-esteem are indifferent psychological domains. The SSE incontrast to the other devices is a nonverbal,"low visibility" instrument, and also incorpo-rates a social frame of reference.

One of the universal criticisms of the mostfrequently used measures of self-acceptanceis that they are about equally correlated withsocially desirable responses as they are witheach other (Crowne, Stephens, & Kelly,1961). For example, the greater the tendencyto give socially desirable responses, the lessthe reported discrepancy between self andideal self.

Using the Crowne-Marlowe (1964) measureof socially desirable response tendencies, andrelating it to the SSE as well as Diggory'smeasure of self-evaluation, correlations of-.36 and .65 (N = 24, p < .05 for both)were found for sophomore female volunteersfor an experiment. Higher self-esteem as mea-sured by the Diggory device was associatedwith a tendency to give socially desirable re-

TABLE 3INTERCORRELATION MATRIX OP FOUR MEASURES OE SELF-ESTEEM

Bills-Vance McLean Index (1)Coopersmith's Self-Esteem (2)Diggory's Self-Evaluation (3)Social Self-Esteem (4)

(2)

Males

.46**

Females

(3)

Males

.17 .60***.37*

Females

(4)

Males

.29* -.10

.23 .02-.09

Females

-.14.04.21

Note.—For males, N* p < .05.

**p < .01.***p <.001.

• 33; for females, N = 53.


sponses. The opposite relationship was foundusing the SSE.

SELF-ESTEEM AND SOCIAL ACCEPTANCE

Turning now to construct validation pro-cedures, one of the earliest studies in theseries examined the frequently hypothesizedrelationship between acceptance of self andacceptance by others (Mann, 1959; Wylie,1961). In one of the reported studies, Cooper-smith (19S9) found that fourth, fifth, andsixth graders showed a significant positivecorrelation (.34) between self-esteem andpopularity. The rationale for the relationshipis often tautological (see Rogers, 1951, p.520) but the findings are nevertheless con-sistent. Within the present framework, self-acceptance and the acceptance by others areperceived as inextricable components of socialself-esteem.

The Ss in this study (Ziller, Alexander, &Long, 1964) were 321 sixth-grade students in11 classrooms from four elementary schools.The Ss were all white, and the compositionof the classes remained unchanged throughoutthe school day. All 5s completed a socio-metric item asking them to name the five chil-dren with whom they would most like to play.Twenty-five children (17 boys and 8 girls)who were unchosen and 25 children (17 boysand 8 girls) who were most highly chosenfrom the same classes as the unchosen wereadministered one item of the social self-esteemmeasure. The social set included "doctor,""father," "friend," "the person with whomyou are most happy," "mother," "yourself,""the most successful person you know," and"the person with whom you are most comfort-able." The directions were read to 5s. Themean position of the popular children (1 beingthe left position) was 3.8 and of the un-popular 5.7 (t = 3.87, p < .005).

SUCCESS AND FAILURE OF POLITICALCANDIDACY AND SELF-ESTEEM

One of the difficulties of studying changesin the self-concept is that conditions associ-ated with changes in the self-concept are notreadily generated and are rarely encounteredunder circumstances amenable to statisticalanalysis. A political election, however, pro-vides an exceptional opportunity to study

TABLE 4

CHANGES IN SELF-ESTEEM or WINNING AND LOSINGPOLITICAL CANDIDATES

Candidates

WinnersLosers

Direction of change in self-esteem

Increased

IS8

Decreased

411

No change

42

X" - 6.01 (p < .05).

changes in the self-concept associated withwinning as opposed to losing the election.

One month prior to the 1967 state electionin Oregon, 44 candidates for the state legis-lature were administered the first four itemsof the adult Form II of the SSE. The sameitems were again administered to the samecandidates approximately 1 month after theelection. Each candidate was approached indi-vidually and completed the form in the pres-ence of the data collector. The results areprovided in Table 4. Fifteen of the 23 winningcandidates increased in self-esteem as opposedto 8 of the 21 losing candidates. Moreover,11 of the 21 losing candidates decreased inself-esteem as opposed to 4 of the 23 winners.The results are significant at the .05 level ofconfidence (x2 = 5.99).

SELF-ESTEEM AND CONSISTENCY OFSOCIAL PARTICIPATION

In an experiment by Mossman and Ziller(1968), it was hypothesized that self-esteemis: (a) positively related to frequency ofparticipation in group discussion, and (b) isassociated with the organism's consistency ofresponse. Relative to the first hypothesis, ithas already been demonstrated that highself-esteem is associated with social accept-ance. This finding suggests that in a groupdiscussion the individual with high self-esteemwill receive and expects to receive verbal andnonverbal cues from the other members whichinvite or support his bid for participation.In addition, the high self-esteem—high soci-ally accepted member may be expected toreceive more social reinforcement for his par-ticipation. Thus, the high self-esteem—highsocially accepted member is assumed to re-ceive more self-reinforcement and social rein-


forcement for participation in group discus-sion, which results in a higher level ofparticipation.

Here, high self-esteem is assumed to beassociated with a higher potential for self-reinforcement and a higher probability ofsocial reinforcement. To some extent self-reinforcement and social reinforcement maybe complementary. When social reinforcementis withheld, the individual with high self-esteem has recourse to self-reinforcement. Theindividual with low self-esteem is moredependent upon social reinforcement, lead-ing, as we stated at the outset, to less stableparticipation.

Borgatta (1962, p. 256) demonstrated asignificant correlation between Cattell's GuiltProneness versus Confident Adequacy sub-tests and total activity in a group discussion(r = —.34, the more guilt, the less activity)and between Edwards' Abasement subtest andtotal activity (r = —.30, the less abasement,the more activity). The 5s in the study were76 neuropsychiatric patients who were mem-bers of four "autonomous problem-solvinggroups" similar in purpose to those discussedby Fairweather (1964, p. 171). Each of thefour groups was observed during one sessiona week over a 3-week period. The observerrecorded the total amount of interaction unitsper individual. The adult forms of the SSEwere administered at the end of the thirdsession.

The self-esteem scores were ordered asanticipated with regard to levels of verbalparticipation: high interactors (those whocontributed more than 5% of the interactionunits across the three sessions), 21.6 (thesum of six ratings); low interactors (thosewho contributed between 2% and 5% of theinteraction units, 19.8; and the noninteractors(those who contributed 1% or less of theinteraction units across the three sessions),17.4. Furthermore, the results were sta-tistically significant (F = 3.37, df = 1/57,p < .05).

In order to test the second hypothesis, thevariance of the relative frequency of inter-action units across the three sessions was usedas the measure of consistency of social be-havior for each individual. A highly consistentinteractor was denned as a group member

whose relative frequency of interaction vari-ance was .03 or less. The .03 point of divisioncreated two equal-size categories of Ss withregard to consistency of verbal participation.However, only the consistency of high inter-actors was analyzed since consistency amonglow interactors and noninteractors would bea statistical artifact stemming from a ceilingeffect. As hypothesized, there was a signifi-cant difference (p < .05) in self-esteem scoresbetween the high interactors-low consistency(SSE = 15.90, N = 10) and high interactors-high consistency (SSE = 26.73, N = 11) cate-gories of Ss.

SELF-ESTEEM AND SOCIOECONOMIC STATUS

Although Wylie's review of the literaturecontains no reference to an analysis of arelationship between socioeconomic statusand self-esteem, more recently two studies(Coopersmith, 1967; Rosenberg, 1965) indi-cate a positive relationship, although only theresults of the first are statistically significant.The rationale for the relationship is thatsocial status is one of the most strikingindexes of prestige and success. Personshigher in the social system have more prestigi-ous occupations, have higher incomes, andtend to live in larger and more luxurioushouses located in more desirable neighbor-hoods. These persons are perceived as moresuccessful and tend to receive material andcultural benefits that might lead them tobelieve that they are generally more worthythan others.

Coopersmith (1967) points out, however,that children's social position emanates fromexperiences in school and the neighborhoodrather than in an occupational context. Theseattenuating considerations notwithstanding, itis proposed children from higher status fami-lies are more apt to have ego enhancingmaterial reinforcements and social reinforce-ments. Social self-esteem, then, is presumed tobe a general evaluation of the self in relationto significant others, and socioeconomic statusis but one component of social self-esteem.

A reanalysis of the results of a study byLong, Ziller, and Henderson (in press) in-volved an equal number of boys and girls ofnormative age for grade in each Grade 6 to12 in four schools in Queen Anne's County,


Maryland. The 5s were white and lived in arural area on the eastern shore of Maryland.The 5s had completed six items of the studentform of the SSE. Hollingshead's OccupationalScale based upon head of household's occupa-tion (Hollingshead & Redlick, 1953) wasused. Four classifications evolved which pro-vided classes with maximally equivalentnumbers of 5s per class. Thus, 83 5s whosefathers' occupation was professional, business,office worker, or salesman constituted Class 1;66 5s whose fathers were classified as skilledlabor constituted Class 2; 77 5s whose fatherswere farmers constituted Class 3; and 69 5swhose fathers were semiskilled or unskilledlabor constituted Class 4. The mean self-esteem scores of these 5s were 23.0, 22.4, 20.4,and 20.5, respectively (F = 2.54, df - 3/292,p about .05). Self-esteem and socioeconomicstatus are positively associated in this Ameri-can sample of 5s.

SELF-ESTEEM AND CULTURE

In an analysis of cultural shaping of con-ception of the self, Hallowell (1955) assumesthat the individual's self-image and his in-terpretations of his own experience cannot beseparated from the concept of the self that isnormative in his society.

A similar analysis to the foregoing wasmade possible by the availability of the casteof Indian 5s in reanalysis of a cross-culturalstudy (Long, Ziller, Ramana, & Reddy, 1966).The 5s consisted of 50 boys and 50 girls fromForm I of the M.V.D.M. High School inVesakapatnam, Andhra, South India. Thechildren ranged in age from 10 to 14 with amedian age of 12. Six items of the studentform of the SSE were used. The instructionswere read aloud in Telugu, the native lan-guage of the children, by one of the Indian Es.

The 5s were found to be members of fourcastes: (a) Caste 1 (Brahmin), N = 39, self-esteem = 27.1; (b) Caste 3 (Visya), N = 9,self-esteem = 25.1; (c) Caste 4 (included 18types such as Satani and Najara which are allassociated with crafts), N = 48, self-esteem= 29.3; (d) Caste 5 (Harijan and Relly), N= 4, self-esteem = 30.3. Because of the ex-tremely small number of 5s in two of thecates, Castes 1 and 3 were combined and com-pared with Castes 4 and 5. The results were

statistically significant at the .05 level of con-fidence (F= 5.6, df= 1/98).

The results were in opposition to expecta-tions based upon the results of the previousstudy. In this sample of Indian 5s the highestself-esteem was expressed by students in thelowest caste (in the results above, the self-esteem scores were simply reversed; highscores represent high self-esteem). However,school was attended within this age group andin the region by only 14% of the population.The children from the lowest castes, then, maybe using other children of their caste who donot attend school as points of reference, andwithin this frame of reference their ownstatus appears extremely high. Thus, as indi-cated in the initial self-esteem framework, thefield of comparison becomes crucial. Self-esteem is defined with regard to significantothers.

In this same study of Indian children itwas possible to compare the self-esteem of acomparable sample of boys and girls from theQueen Anne's County sample mentionedearlier. Here again the high score representshigh self-esteem. The Indian and Americanmeans were 28.1 and 20.6 respectively (F —65.6, df = 1/196, p < .005). Again, however,the privilege of school attendance may be thecrucial variable. The Indian student who per-ceives himself as being a member of a selectgroup by virtue of school attendance mayhave higher self-esteem than American chil-dren, who all attend school as a matter ofcourse. In terms of the discussion of the as-sociation between self-esteem and socioeco-nomic status, it is proffered that the schoolenvironment is the most salient status variablefor the Indian children.

SELF-ESTEEM AND CONFORMITY

In a summary of the research concerningself-esteem and conformity, Wylie (1961)acknowledges that there is a trend indicatingan inverse relationship; individuals with lowself-esteem tend to be more persuasible. Inthe present study 41 high school seniors wereadministered the six-item student form of theSSE and then were placed in the classic Aschconformity situation (Asch, 1956). Under thenine extreme conditions where the unanimousmajority (of four in this study) chose the line


which deviated most from the standard line,the biserial correlation between conformityand the SSE was .32, p < .05. Higher self-esteem was associated with higher conformity.

As stated at the outset, previous researchwas equivocal. Nevertheless, it is compellingto view these results as negative. In view ofthe strong social component inherent in thepresent measure of self-esteem, however, it isstill possible to interpret the results as indi-cating that the person with high self-esteemwithin a social context may not perceive con-forming behavior under low cost conditions asdamaging to the self system.

SELF-ESTEEM AND THE NEUROTICPERSONALITY

Ausubel (1952) regards self-esteem as theoutcome of achieving a status commensuratewith one's conception of self-importance. Heproposes that a devaluation of the self-con-cept is necessary in the face of reality andin order to avoid severe injury to self-esteem.Trauma to self-esteem may result if ego im-portance is devalued extremely. On the otherhand, personality disorders may evolve fromuntenable notions of omnipotence to whichthe child is subject.

Results of experiments which have investi-gated the relationship between adjustment andself-regard (Wylie, 1961) are equivocal.Again, however, the results may simply re-flect the shortcomings of verbal self-reportmeasures of self-esteem, particularly whenused with neuropsychiatric patients.

In the first and second of three studies in-volving the SSE (Ziller, Megas, & DeCencio,1964), the felt circle approach describedearlier was employed. In the first study in-volving a set of significant other members ofan acute neuropsychiatric treatment ward(psychiatrist, psychologist, social worker,nurse, nurses' aide, other patients in the ward,and yourself), patients who had been admin-istered electroconvulsive shock therapy (N =10) from 1 to 6 weeks prior to completing theSSE as compared with those patients who hadnot received this treatment (N = 35) placedthe circle representing themselves in the lastposition to the right more frequently (x2 =12.34, p < .01). These results appear to vali-date the psychiatric screening techniques for

depression since electroconvulsive shock treat-ment was only recommended for acutely de-pressed patients.

In the second experiment in the series (Zil-ler, Megas, & DeCencio, 1964), 25 patientsfrom the neuropsychiatric ward of an acutetreatment center and 23 volunteers from thehospital staff, including personnel from severallevels of the occupational hierarchy, arrangedin a horizontal straight line on a black feltfield 10 significant social elements representedby symbols on a piece of felt 2 inches in di-ameter. The elements included "mother,""father," "your wife or girl friend," "the mostsuccessful person you know," "the happiestperson you know," "the person with whomyou are most uncomfortable," "employer,""your doctor," and "your friend." The circleswere placed in a random order on a table infront of the black felt field. On each circlethere was a one-word description of the personwhom the circle represented. A list of the so-cial elements that the circles represented waspresented in alphabetical order on the sametable. The Ss were asked to "place them allon the board in a straight line according tosome relationship that you decide upon."Normals were found to have placed the "self"in a higher position than the patients in theassumed left-right hierarchy (t = 4.57, p <.001).

In the third study in the series (Ziller &Grossman, 1967), 90 male, acute neuropsychi-atric patients and 87 male employees of thesame hospital served as Ss. Two self-esteemitems were administered to the patients dur-ing the first week of admission to the hos-pital. In the first item, S was asked to choosea circle to represent himself from 10 circlesarranged in a vertical column. Circles wereweighted from 1 to 10, with a higher scoreassociated with a higher position.

The second measure of self-esteem pre-sented a horizontal array of circles. The 5swere also presented with the same list of 10significant other persons (including the self)used in the preceding study. Here the usualleft-right hierarchy was assumed.

The results corroborate earlier findings inthe series. Neuropsychiatric patients in com-parison with normals show lower self-esteem


on both measures (horizontal, p < .10; verti-cal, p< .OS).

OVERVIEW

The series of studies described here repre-sents the first phase in a program of researchconcerning self-other orientation. The programof research involves the integration of a theoryof personality involving self-other perceptionsand an instrument designed to measure theevaluation of the self in relation to significantothers using topological representations of selfand others. The outcome of the present ap-proach, although balanced with regard to theemphasis on theory, instruments, and re-search, rests largely upon the utility of themeasures involved. Measurement remains themissing link in personality research. With re-gard to self-social constructs, measurement isdependent upon the method of communicatingbetween 5 and the scientist. Here we haveproposed that there are some distinct advan-tages to avoiding the usual verbal self-reportapproach and substituting for it a topologicalapproach with limited verbal demands.

The results tend to support the validityand utility of the approach to the measure ofself-esteem. Social objects with greater valuetend to be placed to the left in the horizontaldisplay; the absolute difference between loca-tion of self and a low-status social object issignificantly associated with the left to rightlocation of the self; left-right location of theself is significantly associated with the up-down location of the self. Higher self-esteemwas found to be associated with social ac-ceptance, social participation, socioeconomicstatus (only in an American sample), identi-fication with parents, consistency of social be-havior, and the normal as opposed to theneurotic personality. Finally, winning politi-cal candidates for state legislative offices rosein self-esteem, whereas losing candidatesdropped in self-esteem.

The results of three studies require recon-sideration. In one of these negative or verylow correlations were found between the SSEand three frequently used measures of self-esteem. The results are consistent with thoseof an earlier study, however, and stronglysuggest that the SSE measures an aspect of

self-evaluation which is in a different factorregion.

The second result requiring reconsidera-tion concerns the tendency of low caste ascompared with high caste Asian Indian stu-dents to place the symbol representing theself more to the left. By way of explanation,it was observed that only a select few of thelow caste as opposed to a high percentage ofthe high caste attended public schools, and asa consequence, the high status of the selectedlow caste student in comparison with the non-selected low caste student was reflected intheir self-evaluations on the SSE. Finally,persons with high social self-esteem were foundto be more conforming in the Asch situation.It is difficult to discount this finding, eventhough previous studies concerning a relation-ship between self-esteem and conformity havebeen equivocal. Against the background ofthe directly supporting results, the three re-sults requiring qualification do not appear torequire a reexamination of the validity of theconcept of social self-esteem.

The social context of self-esteem has beenemphasized in the present approach, and, in-deed, the results of most of the studies areconcerned with social behavior (i.e., popu-larity, frequency of participation in groupdiscussions, parent-child relationships, con-formity). An attempt was made to describethe concept used here in terms of its socialcorrelates. Thus, the term "social self-esteem,"was used. The meaning of social self-esteemas it evolves against the background of itssocial correlates suggests social acceptance orperhaps self-other confidence. The individualwho is assured of his high self-evaluationwithin a social context is more consistent insocial behavior and more accepted by others.As stated at the outset, self-acceptance andsocial acceptance are intrinsically interdepen-dent.

If, then, the traditional self-report measureof self-esteem is unrelated to social self-esteem, a reanalysis of the meaning of thetraditional measures is indicated. It is nowsuggested that the self-report measures indi-cate a socially desirable self-esteem, an evalu-ation of the self that the reporter is willing toreveal, or that he desires the other to accept.

Aside from the question of the social con-

94 ZILLEE, HAGEY, SMITH, AND LONG

text of self-esteem, the limited verbal demandsof the present approach recommend it as amost useful measure in cross-cultural research,developmental research, and research in gen-eral where there may be some question of thecomparable verbal ability of the 5s in rela-tion to the experimental tasks.

The fundamental assumption of the pro-posed measure of self-esteem is the proclivity,in the Western culture, for left to right linearordering of objects. The findings of the pres-ent series of studies along with a long historyof research appears to support this assump-tion. As has already been noted, the tendencyto attribute greater importance to the objectplaced at the extreme left position in a hori-zontal display was recorded by Morgan in1944. Introspectionists such as Lashley (1961)and Inhelder and Piaget (1958, p. 252) havealso noted the use of spatial imagery in think-ing about nonspatial orderings.

The most systematic work in this area hasbeen conducted by DeSoto (1960, 1961; De-Soto, London, & Handel, 1965). For example,in spatial imagery associated with syllogisms,the results suggest that the left end is thepreferred starting point for a horizontal order-ing (DeSoto, London, & Handel, 1965).

The studies presented here have attemptedto extend this work to the development of auniversal communication system; a sign lan-guage, if you will, for describing self-otherorientations. On this assumption, a number ofself-other configurations have been developed(see Ziller & Grossman, 1967; Ziller & Long,1966) using other spatial arrangements ofsymbols representing self and significant oth-ers which are designed to measure self-otherpower orientation, marginality, social interest,identification, identification with the major-ity, self-centeredness, social inclusion, andopenness.

REFERENCES

ASCH, S. E. Studies of independence and conformity:A minority of one against a unanimous majority.Psychological Monographs, 1956, 70(9, Whole No.416).

AUSUBEL, D. P. Ego development and the person-ality disorders. New York: Grune & Straton, 1952.

BILLS, R. E., VANCE, E. L., & MCLEAN, O. S. An in-dex of adjustment and values. Journal of Con-sulting Psychology, 1951, 15, 257-261.

BORGATTA, E. F. A systematic study of interactionprocess scores, peer and self assessment, person-ality and other variables. Genetic PsychologicalMonographs, 1962, 65, 219-291.

COOMBS, C. H., RAUTA, H., & THRALL, R. M. Someviews on mathematical models and measurementtheory. Psychological Review, 1954, 61, 132-144.

COOPERSMITH, S. A method for determining types ofself-esteem. Journal of Abnormal and Social Psy-chology, 1959, 59, 187-194.

COOPERSMITH, S. The antecedents of self-esteem. SanFrancisco: Freeman, 1967.

CROWNE, D. P., & MARLOWE, D. The approval mo-tive. New York: Wiley, 1964.

CROWNE, D. P., STEPHENS, M. W., & KELLY, R. Thevalidity and equivalence of tests of self acceptance.Journal of Psychology, 1961, 51, 101-112.

CUTICK, R. A. Self-evaluation of capacities as afunction of self-esteem and the characteristics of amodel. Unpublished doctoral dissertation, Univer-sity of Pennsylvania, 1962.

DESOTO, C. B. Learning a social structure. Journal ofAbnormal and Social Psychology, 1960, 60, 417-421.

DESOTO, C. B. The predilection for single ordering.Journal of Abnormal and Social Psychology, 1961,62, 16-23.

DESOTO, C. B., & KUETHE, J. L. Subjective proba-bilities of interpersonal relationships. Journal ofAbnormal and Social Psychology, 1959, 59, 290-294.

DESOTO, C. B., LONDON, M., & HANDEL, S. Socialreasoning and spatial paralogic. Journal of Person-ality and Social Psychology, 1965, 2, 513-521.

DIGGORY-FARNHAM, S. Self-evaluation and subjectivelife expectancy among suicidal and nonsuicidalpsychotic males. Journal of Abnormal and SocialPsychology, 1964, 69, 628-634.

FAIRWEATHER, G. W. Social psychology in treatingmental illness. New York: Wiley, 1964.

FESTINGER, L. A theory of social comparison proc-esses. Human Relations, 1954, 7, 117-140.

HALLOWELL, A. I. Culture and experience. Phila-delphia: University of Pennsylvania Press, 1955.

HENDERSON, E. H., LONG, B. H., & ZILLER, R. C. Self-social constructs of achieving and non-achievingreaders. The Reading Teacher, 1965, 19, 114-118.

HOLLINGSHEAD, A. B., & REDLicx, F. C. Social strati-fication and psychotic disorders. American Socio-logical Review, 1953, 18, 163-169.

INHELDER, B., & PIAGET, J. The growth of logicalthinking. New York: Basic Books, 1958.

KELLY, G. A. The psychology of personal constructs.New York: Norton, 1955.

KUETHE, J. L. Social schemas. Journal of Abnormaland Social Psychology, 1962, 64, 31-36.

LASHLEY, K. S. The problem of serial order in be-havior. In S. Saporta (Ed.), Psycholinguistics. NewYork: Holt, Rinehart & Winston, 1961.

LEWIN, K. A dynamic theory of personality. NewYork: McGraw, 1935.

LONG, B. H., ZILLER, R. C., & HENDERSON, E. H.Developmental changes in the self-concept duringmiddle childhood. School Review (in press).


LONG, B. H., ZHXER, R. C., RAMANA, K. V., & REDDY,J. E. Self-social orientations of Indian and Ameri-can children. Unpublished manuscript, GoucherCollege, 1966.

MANN, R. D. A review of the relationship betweenpersonality and performance in small groups. Psy-chologkal Bulletin, 1959, 56, 241-270.

MORGAN, J. J. B. Effects of non-rational factors ininductive reasoning. Journal of ExperimentalPsychology, 1944, 34, 159-168.

MOSSMAN, B. M., & ZHXER, R. C. Self-esteem andconsistency of social behavior. Journal of AbnormalPsychology, 1968, 73, 363-367.

PIAGET, J. The psychology of intelligence. (Trans, byM. Piercy & D. E. Berlyne) London: Routledge& Kegan Paul, 1947.

RIDGEWAY, S. The relationship among three measuresof self concept. Unpublished master's thesis, Uni-versity of Maryland, 1965.

ROGERS, C. R. Client-centered therapy. Boston:Houghton Mifflin, 1951.

ROSENBERG, M. Society and the adolescent self-image.Princeton: Princeton University Press, 1965.

WITKIN, H. A., DYK, R. B., FOTERSON, H. F., GOOD-ENOUGH, D. R., & KARP, S. A. Psychological dif-ferentiation. New York: Wiley, 1962.

WYLIE, R. C. The self concept. Lincoln: Universityof Nebraska Press, 1961

ZHXER, R. C., ALEXANDER, M., & LONG, B. H. Self-social constructs and social desirability. Unpub-lished manuscript, University of Delaware, 1964.

ZnxER, R. C., & GROSSMAN, S. A. A developmentalstudy of the self-social constructs of normals andthe neurotic personality. Journal of Clinical Psy-chology, 1967, 23, 15-21.

ZrtLER, R. C., & LONG, B. H. Self-social constructs:Theory and instruments. Unpublished manuscript,University of Oregon, 1966.

ZJXLER, R. C., MEOAS, J., & DECENCIO, 0. Self-socialconstructs of normals and acute neuropsychiatricpatients. Journal of Consulting Psychology, 1964,28, 50-63.

(Received March 8, 1968)

(Continued from page 77)

Comparative Study of the Wechsler Preschooland Primary Scale of Intelligence and theStanford-Binet Intelligence Scale, FormL-M, among Culturally Deprived Children:Allan Barclay* and Allan C. Yater: Cardi-nal Gennon Memorial Hospital for Chil-dren, 146S South Grand Boulevard, St.Louis, Missouri 63104.

Measurement of Social-Evaluative Anxiety:David Watson* and Ronald Friend: Psy-chology Department, University of Hawaii,Honolulu, Hawaii 96822.

Effect of Physical and Cognitive SomaticArousal on Rorschach Responses: An Ex-perimental Test of the Assumption thatBody Image Influences the Perceptual Or-ganization of Unstructured Stimuli: StevenN. Van De Mark and Charles Neuringer*:Department of Psychology, University ofKansas, Lawrence, Kansas 66044.

Object-Discrimination Learning Set Acquisi-tion in Young Cerebral Palsied Children in

* Asterisk indicates author for whom address issupplied.

Relation to Tested and Rated Intelligence:Douglas Pearson*: California State Collegeat Long Beach, 6101 E. Seventh Street,Long Beach, California 90801.

Conditioning and Transfer of Affective SelfReferences in a Role Played CounselingInterview: Robert J. Hoffnung*: Counsel-ing Service, University of Cincinnati, Cin-cinnati, Ohio 45221.

Revised Suicide Potential Scale: R. W.Miskimins and Lowell T. Wilson*: MentalHealth and Manpower Project, 3520 WestOxford Avenue, Denver, Colorado 80236.

Different Psychological Effects of LateralizedBrain Damage: Oscar A. Parsons*, ArthurVega, Jr., and Julian Burn: The Universityof Oklahoma Medical Center, 800 North-east Thirteenth Street, Oklahoma City,Oklahoma 73104.

Predicting the Success of Peace Corps Com-munity Development Workers: CharlesDicken*: Department of Psychology, SanDiego State College, San Diego, California92115.

self-esteem: a self-social construct

Documents