american sociological review 1942 newcomb

8/2/2019 American Sociological Review 1942 Newcomb

1/11

Community Roles in Attitude FormationAuthor(s): Theodore M. NewcombSource: American Sociological Review, Vol. 7, No. 5 (Oct., 1942), pp. 621-630Published by: American Sociological AssociationStable URL: http://www.jstor.org/stable/2085688

Accessed: 04/11/2009 10:42

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=asa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to

American Sociological Review.

http://www.jstor.org
http://www.jstor.org/stable/2085688?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2085688?origin=JSTOR-pdf


2/11

COMMUNITY ROLES IN ATTITUDEFORMATION*THEODOREM. NEWCOMB

University of MichiganPoints of interest.The roles played by the conservatives n a progres-sive community.What kinds of persons ail to see themselvesas theyare on the scale of conservatism-progressivism?he relationsbetweenpolitical-economic ttitudes,community ooperation, nd personalitytraits established n earlieryears. [Ed.]

T IS THE thesis of this paper that the personality processes involved inattitude formation are closely related to the adaptive devices by whichan individual comes to occupy a given position in any group or groups

which take a characteristic attitude toward the issue in question. It followsthat observations or measurements of position in the group provide an es-sential kind of data to those concerned with the problem of how personalityfunctions as social attitudes are developed. If this sounds like a truism, letme remind you, first, that it is almost totally unsupported by empirical in-vestigation; and secondly, that almost all attitude research has been castin a framework of individual rather than of group and community deter-minants.The phrase "position of the individual in the group" is a vague one, butit can be rendered concrete and objective in terms of roles. The concept ofrole has been much honored in social-psychological discussions, but it hasbeen honored more in the breach than in the observance of quantitativedefinitions. It is here proposed to relate certain measured attitude changesto objectively determined changing roles in a given group.The group selected for study consisted of the entire student body at Ben-nington College during the four academic years between I935 and 1939.During each of these years there were approximately 250 women students.Many attitude scales were given repeatedly during these years, but thepresent results are drawn almost exclusively from a single attitude scale,constructed according to Likert's methods, labeled "Political and Eco-nomic Progressivism"-P. E. P. for short. It had to do primarily with issuesmade prominent by the New Deal, such as organized labor, public relief, andthe role of corporate wealth.

This particular attitude scale was chosen not because of any chance pref-erences of the investigator, but because investigation revealed that it cor-responded closely to the issues which were most alive in this community* This paper is a partial summary of the writer's Personality and Social Change,to be pub-lished in I942. The study was made possible by grants-in-aid from the Social Science ResearchCouncil.

62i


3/11

622 AMERICAN SOCIOLOGICAL REVIEWduring these years. Since the major determinants of attitude formation arehere described in community-wide terms, it is important to describe, brieflyat least, the nature of this community.

The college is situated on a hilltop four miles from the Vermont villagefrom which it takes its name. It is a very young institution; the first year ofthis study was the fourth year of the college's existence, i.e., the first yearin which there was a senior class. Its educational plan is somewhat novel,particularly its emphasis upon individual diagnosis and guidance. As theresult of the relative novelty of this plan, of the smallness of the total group,of the camaraderieand informality of faculty-student relationships, and ofthe limitations of nearby extra-college attractions the community may bedescribed as highly integrated, self-contained, and self-conscious. Anotherimportant consideration, commonly omitted in reports of attitude studies,is the fact that the majority of the faculty (including the investigator) werecommonly described as "liberal," though the term "radical" was preferredby some.This study deals with individual attitude changes projected against thebackdrop of the community setting. The first aspect of the community set-ting to be noted is that the great majority of students became less con-servative after one, two, or three years in the community (conservatism ishere defined exclusively in terms of scored responses to attitude scales).While the data here presented refer only to the P.E.P. scale, the sametrend was observed with regard to every attitude scale employed, includingsuch issues as Loyalist Spain, the C.I.O., President Roosevelt's proposal forSupreme Court reform, and Bogardus social distance. Table i indicates theextent of this trend in political terms, according to one questionnaire re-sponse.

TABLE I. PERCENT OF PREFERENCES BY STUDENTS AND THEIR PARENTS FORPRESIDENTIAL CANDIDATES IN I936

52 Freshmen 40 Sophomores 52 Juniors-SeniorsStudents Parents Students Parents Students Parents

Landon 62 66 43 69 I5 6oRoosevelt 29 26 43 22 54 35Thomas-Browder 9 7 I5 8 30 4Table 2 indicates the freshman and senior attitudes in terms of meanP.E.P. score, based upon a 26-item scale, each item of which was scored

from I to 5, high scores being conservative. Thus i30 is the most conserva-tive possible score, and 26 the least conservative possible score. This tabledoes not, of course, reveal the attitude differences between freshmen and thesame individuals, three years later, as seniors. This difference, for the oneclass which was followed throughout its entire four years in college, was ofapproximately the same magnitude as those appearing in Table 2-to be


4/11

COMMUNITY ROLES IN ATTITUDE FORMATION 623exact, i2 points. In spite of the small numbers involved, it also is a statis-tically reliable difference, the critical ratio being 5.9.

TABLE 2. MEAN P.E.P. SCORES AND CRITICAL RATIOSFreshmen Seniors Critical Ratio*

N Mn N MnFall I935 88 74.5 45 65.8 3.87Fall I936 69 75.8 27 6o. I 6.49Spring I938 86 70.6 37 59.9 4.58Spring I939 73 72.8 45 62.7 4.55

* Based upon differences between freshmen and junior-senior mean scores, because of thesmallness of the senior groups. The differences between junior and senior scores are slight.The second important aspect of the community setting is that non-con-servatism is related to prestige, measured in terms of frequency of choice asone of five students most worthy to represent the college at a hypotheticalintercollegiate gathering.* The relationship between prestige, thus meas-ured, and conservatism is shown in Table 3. The results in the followingyear were almost exactly the same. Very similar results were also obtainedfrom another question, asked at the same time, but couched in terms of per-sonal desirability as a friend.

TABLE 3. MEAN P.E.P. SCORES, CLASSIFIED ACCORDING TO FREQUENCYOF BEING CHOSEN AS REPRESENTATIVE (1938)Frequency Freshmen Sophomores Junior-Seniors Entire Collegeof Choice N Mn N Mn N Mn N Mn

40-89 3 60.3 5 50*4 8 54.II2-39 5 65.6 I5 57.6 20 59.75-II - 5 65.3 I8 62.2 23 62.72-4 IO 64.6 i8 68.6 I9 6i.6 47 65.3I I2 63.4 I7 68.6 I5 62.I 44 65.00 6i 72.8 39 7I.3 I4 69.0 II4 71.7

A third aspect of the community setting is the finding that conservatismis also related to reputation for community-cooperativeness as opposed tocommunity negativism. These results were obtained from "Guess-Who"ratings made by a group of 24 students, carefully selected so as to representevery cross-section and grouping of importance within the college. As hereemployed, this technique involved the naming by each of the 24 judges ofthree individuals from each class who were most extreme in each of 28characteristics described. One way of presenting these results is shown inTable 4, where those 13 juniors and seniors who have changed least fromfreshman scores are contrasted with those 13 whose scores have declinedmost.

* These responses were obtained, in two consecutive years, in sealed envelopes from 99%of the student body.


5/11

624 AMERICAN SOCIOLOGICALREVIEWTABLE 4. NUMBER OF LEAST AND MOST CHANGINGJUNIORS AND SENIORSWHOWERENOMINATEDMORE THAN ONCE FOR CERTAIN GUESS-WHO ITEMS*

Out of 13 Out of 13individuals not individualsdecreasing in decreasing mostP.E.P. score in P.E.P. score

9 most absorbed in social life and week-ends outside ofcollegea most absorbed in college community affairs5 I most critical of faculty or administrative staff5 I least concerned about educational policies of thecollege

I 5 anxious to hold positions of community responsibility5 o most resistant to community codes, standards, etc.2 6 most likely to be enthusiastic supporters of the college7 o least apt to continue pursuits related to college in-interests4 o most likely to lead life of sheltered leisure2 6 least likely to be deterred by family disapproval

* The top line of Table 4 is to be read as follows: nine of the thirteen individuals who failedto become less conservative in P.E.P. score are named by more than one Guess-Who rater asbeing most absorbed in social life and week-ends outside of college, while one of the thirteenwhose P.E.P. scores changed most in a non-conservative direction is thus mentioned bymore than one Guess-Who rater. The ten items selected for inclusion in this table are repre-sentative of the total 28 items, with the exception of four which showed only slight differencesbetween the two groups; for the sake of brevity the others are not here included.

So much by way of backdrop and setting. Before going on to a discussionof individual roles in such a setting, it may be pointed out that a fairly ob-vious conclusion seems to have emerged, viz., that full community assimi-lation involves the taking on of less conservative attitudes. From this itmight be deduced as a hypothesis that any sort of negativistic role in thiscommunity would not be accompanied by decreasing conservatism,whereas community-participating roles would be accompanied by decreas-ing conservatism. This is equivalent to saying that generalized anti-com-munity roles include rejection of the dominant attitude patterns, whilegeneralized pro-community roles include acceptance of them. Further in-vestigation, however, reveals that this is far too simple.The fact is that a minority of reputedly enthusiastic and community-co-operative individuals do not become less conservative in attitude, and thata minority of those whose attitudes do become less conservative are notreputedly community-cooperative. Any adequate theory of community-relationship and its bearing upon attitude formation must make room forthese minorities as well as for the majority groupings.One of the shortcomings of the too-simple theory outlined above is its as-sumption that community roles may be adequately delineated solely on the


6/11

COMMUNITY ROLES IN ATTITUDE FORMATION 625basis of observations of an individual's behavior by others. Many lines ofevidence suggest that an individual's own view of his community role has agood deal to do with the manner of his adaptation to it. This self-view, herereferred to as subjectiverole, may or may not be much like the objective ole,i.e., that based upon external behavior observations. As a matter of fact adefinition of role which includes both its objective and its subjective aspectsgoes far toward accounting for the apparent discrepancies noted above. Letus therefore turn to the question of measuring the subjective role.About half of the P.E.P. items were reproduced in somewhat abbreviatedform at the left of a sheet of paper. On the right were three columns headed,respectively, "freshman," "juniors-seniors,"and "faculty." Directions wereas follows: "First, indicate your agreement or disagreement with the state-ment. Then make an estimate, in each column, of what percent of eachgroup (freshman, faculty, etc.) agrees with the statement." Several groupfindings from these responses are of interest. For example, nearly all fresh-men consider their own class much more conservative than themselves, butconsider juniors and seniors less conservative than themselves; nearly allseniors consider freshmen much more conservative then themselves, but amajority of seniors consider themselves more conservative than most of theclass.

Of several possible ways of handling these responses for individual pur-poses, the most promising was that of calculating, for each individual, thedegree of divergence rom estimatedclass majority, i.e., divergence from herestimate of the class to which she belonged. Total conservative divergencewas taken as the sum of the estimated class percentages of non-conservativeresponse to all items answered conservatively by the subject. Total non-conservative divergence, similarly, is the sum of the estimated class per-centages of conservative response to all items answered non-conservativelyby the subject. Final score equals total conservative minus total non-con-servative divergence, divided by the number of items responded to.*Divergence score, which may have either plus or minus value, thus rep-resents the average degree to which the subject considers her own responsesto be more conservative than those of her class. In the following para-graphs, those subjects whose scores of divergence are opposite in directionto that shown by their actual attitude scores are labeled "unaware." That

* E.g., suppose that a subject's estimates of non-conservative response by her class to the4 items which she herself answers conservatively are 20, 40, 40, and 5o; her total conservativedivergence is the sum of these percentages, or i50. If her estimates of conservative response byher class to the remaining ten items, which she answers non-conservatively, are 6o, 50, 40, 40,30, 30, 20, 20, Ia, and o, her total non-conservative divergence is the sum of these, or 300.Divergence score is thus (150-300) .I4, or -I0.7. This score, which is taken to indicatethat the subject estimates her class's position as slightly more conservative than her own, issimply a weighted algebraic average of the estimated percentages of class disagreement withown response: the number times the amount of non-conservative disagreements outweighs thenumber times the amount of conservative disagreements to the average amount of I0.7.


7/11

626 AMERICAN SOCIOLOGICAL REVIEWis, if divergence score is minus for an individual whose P.E.P. score is ac-tually considerably above the class average, she has estimated her classas more conservative than herself when objective scores show the reverseto be the case; she is therefore considered unaware of her own relative con-servatism. Among the subjects discussed in the next few pages, those whoseself-estimates of divergence are rather slight are also labeled "unaware,"since their actual scores are rather extreme. Those whose self-estimateddivergence is relatively great, and in the same direction as that shown byattitude scores, are labeled "aware"; i.e., they are aware of their own atti-tude positions as being relatively conservative, or relatively non-conserva-tive, as the case may be.*An intensive study was now begun of all of the members of the two con-secutive graduating classes except those whose final P.E.P. scores werewithin one-half standard deviation of the class average. Roughly speaking,the most conservative and least conservative quarters of these two classeswere thus studied, including 43 individuals altogether. The sources for thisintensive study included all the questionnaire responses over a period ofthree or four years; the official college records, including extensive reportsby instructors on various aspects of personality, and the medical and psy-chiatric as well as the academic files; and the records of a semi-standardizedinterview held with each senior concerning her history of attitude develop-ment. The manner in which all these data were used will appear later; forpurposes of classification by role, however, only three sorts of objectivedata were used: P.E.P. score for classification as conservative or non-conservative, reputation according to Guess-Who ratings for objective role,and self-estimate of conservative divergence for subjective role.In Table 5 are shown both the mode of classifying these 43 subjects, andtheir individual scores of estimated conservative divergence. Classificationsof objective role were made upon the following basis: individuals mentionedby two or more Guess-Who raters as resistant or indifferent in two ormore items are considered "negativistic," provided they are not also con-sidered cooperative by two or more raters on more than one item. Allother conservative subjects are labeled "non-negativistic." Individualsmentioned by two or more Guess-Who raters as community-cooperative intwo or more items are labeled "cooperative," provided they are not alsonamed by two or more raters as being negativistic. All other non-conserva-tives are labeled "non-cooperative."

The rationale of this classification into four groups of conservatives and* Five subjects of the 48 whose records were otherwise suitable for inclusion in this partof the study were eliminated from consideration because their estimates of divergence were ata border-line point. This procedure is justified only because it is not claimed that the patternshere described represent all of those which are to be found. It is claimed only that certain pat-terns may be found, by such a mode of classifying, which are significant for the problem athand.


8/11

COMMUNITY ROLES IN ATTITUDE FORMATION 627TABLE 5. MEAN PERCENT ESTIMATESOF CONSERVATIVEDIVERGENCE,

FOR SELECTEDGROUPS OF SENIORS*ConservativeSubjects

Negativistic Non-Negativisticand Aware and AwareQ IO +26.8 F 32 +27.9Q I2 +23.9 Q 28 +22. IG 32 +I2.9 Q 35 +2I.4F 22 +II.I Q 78 +i8.9E 2 +IO.7 Q 73 +IO.7

Negativistic Non-negativisticand Unaware and Unaware

Q 47 --0.7 Q 68 0.0L i2 - 2.I Q 70 - I.IQ I9 - 0.7 D 22 - 6.IQ 8i - 4.2 M I2 -Io.6Q 6o -I3.6

Non-conservativeSubjectsNon-cooperative Cooperativeand Unaware and UnawareQ 25 - 7.9 Q 6I + 8.2E 22 - 8.6 Q 57 -I2.9Q 6 -ii.8 Q 72 -I3.2

Q 62 -I2. I Q 75 -I3.9D 52 -IS-7 Q 63 -I5.0E I2 -i6.7Non-cooperative Cooperative

and Aware and AwareC 32 -27-7 Q 83 -2I.4Q 7I -27.7 E 72 -2i .8Q 7 -28.6 M 42 -2I.9Q 9 -30.0 H 32 -23.2Q 2I -30.2 Q 22 -34.3K 42 -30.7 B 72 -34.7Q 43 -35 .7

* The code numbers, e.g., Q IO, F 32, refer to individual students. These code numbers arenot the same as those used during the four years in which the data were being gathered.

four groups of non-conservatives is as follows. Individuals, in such a com-munity as this, tend to be assigned labels in terms of the degree of activecommunity participation shown. Some of those assigned a given label, e.g.,negativistic, have self-assigned roles corresponding to their reputations,but the self-assigned roles of others do not correspond to the reputed roles.The personality processes by which a given degree of conservatism is ar-rived at are presumably different for these two groups alike in objective rolebut different in subjective role. More concretely, two hypotheses may bestated: (i) that the personality processes by which a given degree of con-servatism is arrived at are similar for those alike both in objective and in


9/11

628 AMERICAN SOCIOLOGICAL REVIEWsubjective role, (2) that these processes are different for those alike either inobjective or in subjective role but different for the other role.With these hypotheses in mind, the available personality data were as-sembled for each of the eight groups of subjects. Space forbids any lengthypresentation of these data. The brief summaries here reported are con-densed in such manner to show those personality characteristics applyingto all members of each group which appear to be directly involved in theattitudinal history of the subjects during their three or four years in college.Take, for example, the two groups of negativistic conservatives. Thosewho are not aware of their own relative conservatism, i.e., who believe theyare attitudinally typical, are found to be timid and socially insecure,* tohave small and limited groups of friends, and to come to college with almostno aspirations toward "social success." This latter characteristic is clearlyrelated to their almost complete failure to achieve any sort of "social suc-cess" in pre-college relationships. The negativistic conservatives who areaware of their own relative conservatism are markedly less retiring and lessinhibited, are more socially facile; they do not, like the unawares, tend to be-long to compact little friendship groups; unlike the unawares, they hadachieved a considerable degree of pre-college social success, and came tocollege with high hopes of continued success-hopes which were doomed todisappointment. In short, the unawares are insulated in tiny social groups;their negativism functions as a protective shell of indifference toward whatthey cannot cope with; hence their unawareness. The negativism of theawares is an aggressive reaction to the frustration of the ambitions whichwere at first directed at the total community; hence their awareness.Among the non-negativistic conservatives, the distinctions between theawares and the unawares are partly similar to and partly different from thepreceding distinctions. The awares have markedly greater self-confidenceand possess greater social skills; they are eager and enthusiastic, whereasthe unawares tend to be plodding and conscientious. The awares have con-siderable prestige, the unawares almost none. Both groups show more thanaverage attachment to and dependence upon parents, but the crucial dis-tinction seems to be that the unawareswere so absorbed by home and familyconflicts and allegiances as to be scarcely at all susceptible to community-wide college influences, while the awares, equally "loyal" to parents, werecapable of maintaining a divided allegiance; i.e., they yielded to all collegecommunity influences except those attitudinal ones which would havebrought conflict with parents. The unawares, unable to cope with twoworlds, participate only superficially in college community life; hence theirunawareness. The awares are capable of participating in both worlds, but

* Space forbids the presentation of the data upon the basis of which these generalizationsare made. Full documentation is to be found in the writer's forthcoming monograph.


10/11

COMMUNITY ROLES IN ATTITUDE FORMATION 629reject such college influences as would result in home conflict; hence theirawareness.Among the two groups of non-cooperative non-conservatives, both ofwhich may be described as passively conforming rather than as negativis-tic, the unawares are considered dependent upon instructors, and anxiousto please, while the awares are highly independent. The unawares are eagerand enthusiastic, while the awares are not. The awaresare more outstandingacademically, and it is clear that they have set higher standards for them-selves. The unawares believe that they would follow the majority attitudetrend in a conservative college, while the awares would not. The unawaresdescribe their major ambitions, on entering college, in terms of friendshiprather than of prestige, while the reverse is true of the awares, for whom in-tellectual prestige is particularly important. Self-interpretations, finally,show the unawares tend to think of their own attitude change as just oneaspect of being assimilated into the community; hence their unawarenessof their own relative non-conservatism. The awares, however, tend to thinkof their attitude change as an intellectual achievement in respect to whichthey have outdistanced most students, and hence their awareness.The two groups of cooperative non-conservatives have much in common.Both are composed of "substantial citizens," hard-working and conscienti-ous, though the unawares are considered more enthusiastic and the awaresmore persistent. Both groups are high in prestige; both groups had achievedconsiderable recognition before coming to college, though memories of pre-college "failures" are more acute on the part of the unawares. They differprimarily in the following respects: the unawares are more anxious to please,and need more guidance from instructors. The awares are much more com-monly described as "meticulous" or "perfectionist," and are more apt tobe intellectual leaders. The awares have come to reformulate their ambi-tions less in terms of personal success and more in terms of the success of"causes." In short, lack of awareness on the part of the one group representsloyal cooperation in respect to approved social attitudes; as "leaders" theymust, of course, be slightly "ahead" of the majority, but not too far. Theawares, on the other hand, are not only sufficiently secure that they can af-ford to go beyond the majority, but their awareness is a mark of the hard-won struggle by which they reached their non-conservative positions; henceof course, they are aware.In short, it may be said for each of the eight groups that the personalityprocesses which appear to be essentially responsible for whatever attitudeadaptation is made are also responsible for whatever degrees of awarenessis shown. Those who are conservative because they have avoided the com-munity could scarcely have an opportunity to discover that they are con-servative. Those who are conservative because they aimed at leadership,


11/11

630 AMERICAN SOCIOLOGICAL REVIEWfailed, and repudiated whatever the community stood for could scarcelyfail to be aware of their own conservatism. Those who have acquired morethan the average degree of non-conservatism because they are anxious toconform cannot be aware of their own relative extremeness, else they wouldwithdraw to a more moderate position. Those who are extremely non-con-servative because they need to excel must be aware that they are somewhatextreme, etc.Such is the rationale of the classification of individuals on the basis ofroles. Objective roles are assigned, with more or less correctness, by fellowcommunity members on the basis of observable personality characteristics.Among those assigned similar objective roles, different subjective roles areself-assigned on the basis of other personality characteristics. Those forwhom objective and subjective roles are similar, according to "objective"data, thus have many personality characteristics in common. The pointworthy of emphasis, however, is not merely that certain groups of subjectscan be objectively classified in such manner that common personality char-acteristics are found, but rather that these common personality character-istics aredirectly related to the processesof personality adaptation by whichattitudes are acquired in a given community.Such findings suggest the following conclusions as to the theory of atti-tude formation. The personality processes revealed by these aubjects mustbe conceptualized in terms which have group or community referents. Theconcepts of objective and subjective role, here defined in terms of simplemeasurements, serve the scientific purpose of delineating the conditions un-der which several individuals, with differing personal histories and personal-ity organizations, arrive at similar attitude positions by similar personalityprocesses. This channeling of unique, diverse personalities into a limitednumber of patterns, in respect to attitude formation, is the result of thelimited number of roles for which the community makes allowance. It is thecommunityrole which mediatesbetweensocial attitudes and total personalityorganization. But the processes by which individual personalities are ledto make such patterned adaptations become clear only as we conceive ofrole as a dually enforced limitation of behavior. As here employed the ob-jective limitation corresponds to degree and manner of community partici-pation which is recognized within the community; and the subjective limita-tion corresponds to the individual's perception of his own relationship in agiven context (i.e., social attitudes) to such norms as he recognizes.

american sociological review 1942 newcomb

Documents