Social Support and Conformity: The Effect of Differentiation from the ...

DOCUMENT RESUMEED 064 658AUTHORTITLEINSTITUTIONSPONS AGENCYREPCRT NCBUREAU NOPUB DATECONTRACTNOTEEDRS PRICEDESCRIPTORS24 CG 007 507Allen, Vernon L.; Levine, John M.Social Support and Conformity: The Effect ofDifferertiation from the Group and Order ofResponding.Wisconsin Univ., Madison. Research and DevelopmentCenter for Cognitive Learning.Office of Education (DHEW), Washington, D.C.TR-72BR-5-0216Jan 69OEC-5-10-15414p.MF-$0.65 HC-$3.29*Behavior Patterns; *Conformity; Experimental Groups;Group Structure; *Interaction Process Analysis;*Interpersonal Relationship; Learning; *ResponseModeABSTRACTIn this experiment two variables relevant toexplaining the social support phenomenon were investigated. Responseposition of the partner (first vs fourth) and contact between partnerand S prior to the group situation (contact vs. no contact) weremanipulated. Male and female Ss responded to visual, information, andopinion items in a grcup pressure situation. Results indicated thatwhile social support :Ln both the first and fourth positionssignificantly reduced conformity compared to a unanimous condition,support in Pcsition 1 was significantly more effective than inPosition 4. In addition, contact with the supporter increasedconformity in the unanimous condition, but decreased conformity inboth social support conditions. Possible mechanisms underlying theposition and contact effects were discussed. (Autlxn)

STATEMENT OF FOCUSThe Wisconsin Research and Development Center for Cognitive Learningfocuses on contributing to a better understanding of cognitive learning bychildren and youth and to the improvement of related educational practices.The strategy for research and development is comprehensive. It includesbasic research to generate new knowledge about the conditions and processesof learning and about the processes of instruction, and the subsequent developmentof research-based instructional materials, many of which are designed foruse by teachers and others for use by students. These materials are tested andrefined in school settings. Throughout these operations behavioral scientists,curriculum experts, academic scholars, and school people interact, insuringthat the results of Center activities are based soundly on knowledge of subjectmatter and cognitive learning and that they are applied to the improvement ofeducational practice.This Technical Report is from the Peer Group Pressures on Learning Projectin Program 1. General objectives of the Program are to generate new knowledgeabout concept learning and cognitive skills, to synthesize existing knowledge,and to develop educational materials suggested by the prior activities. Contributingto these Program objectives, ihis project is directed toward identificationof the effects of peer group pressures on the utilization of concepts alreadylearned and on the learning of new concepts.iii

CONTENTSpageList of TablesAbstractviviiI. Introduction 1II. Method 3Design 3Subjects 3Apparatus 3Procedure 3Stimuli 4Method of Analysis 4III. Results 5Perception of Dissenter's Position 5Conformity 5Evaluation of the Supporter 6IV. Discussion 7References 84

LIST OF TABLESTable1Proportion of Ss Perceiving Another Group MemberFrequently Agreeing with ThemAnalysis of Variance on Mean Conformity Scorespage53Mean Conformity Scores as a Function of Contactin the Three Conditions 6vi

ABSTRACTIn this experiment two variables relevant to explaining th(. social supportphenomenon were investigated. Response position of the partner (first vs.fourth) and contact between partner and S prior to the group situation (contactvs. no contact) were manipulated. Male and female Ss responded to visual,information, and opinion items in a group pressure situation. Results indicatedthat while social support in both the first and fourth positions significantlyreduced conformity compared to a unanimous condition, support in PositionI was significantly more effective than in Position 4. In addition, contactwith the supporter increased conformity in the unanimous condition, but decreasr.dconformity in both social support conditions. Possible mechanismsunderlying the position and contact effects were discussed.VII

INTRODUCTIONSeveral recent investigations (Allen &Levine, 1968a, 1968b, in press) have supportedAsch's (1951) conclusion that thepresence of one person, answering before thenaive S and correctly dissenting from erroneousgroup consensus, reduces conformitysignificantly compared to a unanimous group.The present study was designed to investigatetwo unexplored variables of potentialtheoretical utility in explaining the socialsupport phenomenon. These variables are,first, response position of the partner, and,second, brief interaction between S and thepartner.In studies of the effectiveness of socialsupport in reducing conformity, typically thesocial supporter has answered fourth and theS has answered fifth (last) in a group of fivepersons (Allen & Levine, 1968a, 1968b, inpress). Whether, however, it is crucial forconformity reduction that the social supporterrespond immediately prior to S, and, hence,after the other group members, is open toquestion. Let us examine two social supportsituationspartner answering in either firstposition or fourth positionto determinefactors which may differentially affect conformityreduction. As will be seen, plausiblepredictions of relatively greater conformityreduction can be advanced for either answeringposition.The position in which the social supporterresponds may influence the "gestalt" of thesituation, as perceived by S. That is, S mayattribute varying psychological interpretationsto the relation between the supporter and thegroup, depending on the ordinal answeringposition of the partner. When the social supporteranswers first, followed by three opposingresponders, S may attribute maximumvalidity to the partner's responses. For onlyindependent assessment of the stimuli, unfetteredby group pressure, determined thepartner's judgment. Responses of the socialsupporter answering fourth, though correct orpopular, may be perceivA as partially determinedby the group, i.e., the partner may bemerely an anticonformer. Thus, S may attributegreater credibility to the supporter answeringfirst, rather than fourth. However, the partnerresponding fourth, though relatively low incredibility, has another potential advantagethat may enhance his effectiveness in reducingconformity. Subjects may attribute greatercorerage to the later supporter because hebravely stands up against three unanimouslyopposing gr)up members. This courage, inturn, may create a positive impression of thesupporter that will increase his ability to freeS from group pressure. In sum, the supporterin position four will be seen as more courageous,while the supporter in position one willbe seen as more credible. Recent data on theimportance of independent assessment of realityin social support (Allen & Levine, inpress) lead to the prediction that a partnerwho responds first will be more effective inreducing conformity than will a partner whoresponds fourth.Turning now to the second variable underinvestigation in this study, let us examinethe potential influence on conformity reductionof a brief interaction between supporter andS prior to the experiment. In the ordinary socialsupport situation, the partner is not differentiatedfrom the other group member; i.e.,S neither knows the identity of the partnernor interacts with him. However, it seemshkely that contact and mutual awareness betweenS and the supporter might produce apsychological situation for S having implicationsfor the supporter's efficacy in reducingconformity. Firt, S's usual degree of anonymity)n the group pressure situation is reducedby the presence of one person whoknows his identity (and, as will be discussedlater, his answering position). As researchhas consistently demonstrated (Deutsch &Gerard, 1955; Asch, 1956; Levy, 1960), conformityto a unanimous group varies directlywith publicness of response. That is, S'sresponses are closer to group consensus when1

ide:.tity and answers ary open to groupu iinihir veii i . th.:iefory, itmidnt bt 1;ipufnY1512l.A. tnt. if S S rt2:;po1isestif motatoled by only one gloup Inembel, Sill 1.-nd to rt,spond similarly to thy observer.coueetly dissents nom error.,nus group ednsensusS will also tend todisbent.byLulhae ul iutuil oW0ILht.:66uh(z intezaction. S and the supporter are differentiatedhoof the otherwise mutually anonymousdroup. his Cxplanation rests more onIt.ciprocol relation between S and the partnyithan does the anonymity explanation above.llety, we az y suggesting that the brief interactioncreates a "minimal" relationship betweenS and the partner which might influenceS's response in the group pressure situation.This relation might create, for instance, afelt obligation on the part of S to support theindividual who correctly dissented from thegroup. Or, perhaps, the relation might producein S a feeling of accountability or re-:31-)01:b1h1l1;y ;;I:; l'Slttlutluh. Iflib latt-1 notion 15 closelylated to thy concept of "individuation" it:Itoducedby I.-stinget. Pepitune, and :\ewt.-01:;1,(10) and recyntly investigated by Sin.rez.Brusn. and Lublin (I'lt,-)). Since S and thesupporter are ryciplocally individuated, S'sfeelings of responsibility may be heightenedby thy suppoltyr's aceyptancy 01 pk.isonalaccountability. correct dissent from elroneousgroup consensus. Thus, the oilleier,-tiation xplanation. like thy anonymity hypothesisabove, predicts reduced conformity whenS and the supporter interact prior to the grouppressure situation.In the present experiment, we shall investigatethe influence of ( I) partner's responseposition and ( 2) interaction betweenS and the partner on conformity reduction inthe social support situation. It is hypothesizedthat social support in position oneand S-partner interaction will produce thegreatest reductions in conformity.

IIMETHODDESIGNA repeated measures analysis of variancedesign was used. The between Ss factorswere: Conditions (Unanimous, Social Support4, Social Support 1), Contact (Contact, NoContact), and Sex of S (Male, Female). Thewithin Ss factor was Type of Item (Visual,Information, Opinion).In the Unanimous condition, the (simulated)Ss who answered in the first four positionsagreed on each item. On critical pressuretrials, the group gave answers placed atthe 95th percentile of responses given by astandardization group answering alone,' Onkiller trials the group gave modal or correctanswers, also obtained from the standardizationgroup. In the Social Support conditions,one simulated S. answering either first orfourth in the group of five, dissented fromthe erroneous responses of the three othersimulated Ss on critical trials by givingmodal or correct answers. On filler trials,the supporter agreed with the three simulatedSs.In the Contact groups each S interactedbriefly, prior to the experiment, with a confederatewho posed as another S. In the NoContact groups. Ss did not meet or interactwith the confederate. Male and female Sswere tested in same-sex groups in ezIch ofthe four experimental conditions describedabove. All Ss saw the same stimulus seriesconsisting of three types of items.'For information and opinion items, thestanciorclizDtion group was 300 introductorypsychology students who had filled out questionnairesin class. For visual items, thestandardization group was introductory psychologystudents tested by Tuddenham, Macbride.and Zahn (1956).SUBJECTSThe Ss were 276 undergraduates, 140males and 136 females, taking an introductorypsychology course at the University ofWisconsin. Data from 21 Ss were discardedbecause of their knowl2dge of the experimentaldeception, leaving a total of 255 Ss (123males and 132 females). The Ss receivedcredits applicable to thcir class grade forparticipating in the experiment.APPARATUSThe apparatus was a Crutchfield-typeelectrical signaling device (Crutchfield, 1955),consisting essentially of five adjacent boothscontaining signal lights and answer switches,and a master control panel in an adjoiningroom. Subjects are led to believe, by instructionsarid practice trials, that signal lights intheir booths indicate responses of other groupmembers and that one person responds in eachof t:ie five answering positions. Actually, thefirst four lights in all booths are controlledby E from the master control panel, and allSs answer last (fifth) on all trials. Thus, Ecan provide simulated group responses thatagree or disagree with Ss' private judgments.PROCEDUREContact GroupsSubjects were taken separately to privateresearch rooms where they were left to fillout adjective checklists prior to the experiment.After about 5 minutes, E returned toS's room with another S (actually, a confederate)who was told to fill out the checklisthere while F. searched for a vacant room.(The confederate had been instructed to initiateno conversation and to tactfully refrain3

from discussion initiated by S during K's absence.)The E left and zeturned about 5 nw.-utes later, stating that a room was availablefor the confederate. The E then informedboth S and the confederate of their answeringpositions in the following experimentalIn the Social Support 4 and Unanimous condi-Lions, the confederate was told to answerfouith in the group of five. In the Social Support1 condition, the confederate was instructedto respond first. The confederate always repeatedhis (her) answering position aloud, toinsure that S had noticed it. In all three conditionsS was told to answer last (fifth). Theconfederate was then taken to another S untilthe sequence had been repeated with all fournaive Ss. Finally, the four Ss and the confederatewere taken to the experimental room.No Contact GroupsFive naive Ss were taker. together directlyfrom the waiting room to the experimentalroom.After being seated in the five 1.Jooths inthe experimental room, Ss were instructed tomake accurate judgments of perceptual, information,and opinion items projected on ascreen in the front of the room. Instructionsand practice trials were utilized to familiarizeSs with use of the apparatus and to convinceSs that signal lights actually indicated lesponsesof all group members. The E thenpresented a series of 30 stimulus slides.simulating the first four responses on eachtrial. At the completion of the experiment, Ssfilled out a postexperimental questionnaire.A careful debriefing followed.STIMULIThe 30 items used in this experimentwere selected from a series developed byN ; gulledamon:lengths of nine comparison lines .31::stEdc;: ,:1 , fa number below one of tne nino panel switcnos.Information items dealt \Nun iLlativelyfactual quostions,tron:Francisco to New York. Again, these itymswen.: answered using th: numb,:tsbeneath the panel switcncs. Opinion itemsconsisted. of such statem,2nts as "I wouldnever go out of rny way to help another peisonif it meant diving u;,) some personal pl:.,:sure"and "Mo...t young peeple get too mucheducation." These items w,)re answered byusing one of Me nine labels located belowthe switches, ranging from Very StronglyAgree" to "Very Strongly Disagree."Of the 30 items used, 12 (40%) were critical,or grc-.;up pressure, itemsfour each ofvisual, information, and opinion items balancedover the series. The remaining 18 stimuli,6 of each type, were neutral filler items.METHOD OF ANALYSISFor each S a mean conformity score wascalculated separately for visual, information,and opinion items. Mean conformity scoreswere computed by summing the algebraic differencesbetween initial responses and responsesgiven in the group situation anddividing by the number of items used. (Forinformation and opinion items, initial scoreswere obtained from questionnaires Ss filledout in class several weeks prior to the experiment.For visual items, modal respoi..es ofthr -uddenham et al. (1956) standardizationgr weie used.)4

IIIRESULTSPERCEPTION OFDISSENTER'S POSITIONSu o s._; of the expeiiirn..ntal manipulauonis partially reflectue in the k:ccuracy of Ss'per(eption of tne dissenter's arr Neting position.Tble I shows the !)lopor,.ion ofeach condition who pef ceived another groupmember as frequently .1greeing with them.(These data ore based on responses fo thepostexperimental questionnaire.) Inspectionof Table 1 indicates that Ss in both the SocialSupport 4 and Social Support 1 conditionsaccurately perceived the presence and positionof the dissenter (.66 and .61, respectively),while Ss in the Unanimous condition rarelyperci wed either POIson 1 or P'_J1son 4 as frequentlyagreeing witn them (.07). Moreover,it is clear that the illsvering position of thedissenter in the two social support conditionsdid not differentially affect correct identification(.66 vs. .61). In addition, it appearsthat correct ..irentificatin of the social supporterdiffered huh_ in the Contact and NoContact conditions (.71 vs. .66 and .64 vs..55). As might be expected. however, Ss whohad interacted with the confederate (Contact)were slightly more aware of the supporter'sposition than were Ss in the No Contact condition.CONFORMITYTable 2. presents the analysis of varianceconducted on mean conformity scores. Boththe Conditions and Items main effects weresignificant at less than the .01 level. Moreover,the Conditions x Contact and Items xSex of S interactions were significant beyondthe .05 and .01 levels, respectively.Mean conformity scores in the Unanimous,Social Support 4, and Social Support 1 conditionsfor the Contact and No Contact groupsare presented in Table 3. Subsequent testson overall condition means indicate that conformityin both the Social Support 4 (.45) andSocial Support 1 (.31) conditions is significantlylower than in the Unanimous condition(.90) (t = 6.89, P< .01, and t = 8.97, p < .01,respectively). In addition, social support inposition 1 (.31) produces a significantlygreater decrease in conformity than does socialsupport given immediately before S'sresponse (.45) (t 2.08, p < .05).Regarding the significant Conditions xContact interaction, it is clear from Table 3that the effect of contact on conformity variedas a function of degree of group ananimity.That is, in the Unanimous condition Ss whointeracted with the confederate conformedmore than did those who had no contact (.99Table 1. Proportion of Ss Perceiving Another Group MemberFrequently Agreeing with ThemCondition Contact No Contact CombinedUnanimousSocial Support 4Social Support 1927984.05 .07 .07.71 .66 .66.64 .55 .61Note.Unanimous: proportion of Ss perceiving either Person 4 or Person 1frequently agreeing with themSocial Support 4: proportion of Ss perceiving Person 4 frequentlyagreeing with themSocial Support 1: proportion of Ss perceiving Person 1 frequentlyagreeing with them5

vs. ,82). lioo:,:vr. in both the Social Support4 and Social Support 1 conditions conformitywas lower in the Contact than in the No Contactgroups: ,36 vs. .54, and .25 vs. .37.Turning now to the Items main effect andthe items x Sex of S interaction, results aresomewhat less meaningful. The significantItems effect cannot be interpreted because thetwee types of items diffei in variability, difficulty.and susceptibility to group influence.The Items x Sex of S interaction is accountedfor by differential conformity of male (.26) andfemale (.54) Ss on Information items.EVALUATION OFTHE SUPPORTERRatings of thc social supporter on five 12.-point evaluative scales were obtained in thepostexperimontal questionnaire. For each scale,a separate analysis of variance was conductedusing two factors: Conditions (Social Support4, Social Sup,)ort 1) and Contact (Contact, NoContact). No significant effects were obtainedon ratings of the supporter's Intelligence. Accuracy,or Independence. Howevei, on scalesassessing liking for the supporter significantrcsults were found. Subjects expressed significantlygreater Personal Liking (I. 3.41,p .10) and Group Liking (Iforthe supporter in the No Contact than in theContact condition. Moreover, significantlygreater Croup Liking was estimated by Ss whenthe supporter answered first, rather than fourthp .05). It L;houlL.: netG.ever, tnat the mean scoles on both liking scalesin all four Contact Condition groups fell nearthe neutral point of the Cvalutotivc ulme nsluii,indicating little strong positive or negativeevaluation of tn.- supportk.r.Table 2. Analysis of Variance onMean Conformity ScoresSource df MS rConditions (A) 2 21.07Contact (B) 1 .33 .57Sex of S (C) 1 1.12 1.93A x B )- 1.97 3.40':'A x C 2 1.48 2.55B x C 1 .15 .26AxBxC 1.32 2.28Error (a) 243 .58Items (D) 2 7.29 19.18**A x D 4 .53 1.40B x D .. ) .31 .82C x D 2 1.96 5.16"AxBxD 4 .54 1.42AxCxD 4 .26 .68BxCxD 2 .50 1.32AxBxCxD 4 .59 1.55Error (b) 486 .38p < .05.01Table h Mean Conformity Scores as a Function of Contactin the Three ConditionsCondition Contact No Contact MeanUnanimous .99 .82 .90Social Support 4 .36 .54 .45Social Support 1 .25 .37 .316

IVDISCUSSIONThe purpose of the present experimentwas to investigate the influences of partner'sresponse position and interaction between Sand partner on conformity reduction in thesocial support situation. As predicted, conformitywas significantly lower when the supporteranswered in Position 1 than in Position4. Nioreover, a significant Conditions x Contactinteraction showed that while conformitywas higher in the Unanimous condition withContact than with No Contact, the inverserelationship occurred in both social supportconditions; i.e., Ss who interacted with theconfederate conformed less than did thosewho had no contact.Regarding the decreased conformity inthe Social Support 1 condition, our hypothesisof greater perceived partner credibility receiveslittle support from the postexperimentalratings of the supporter. That is, on the scaleon which a priori differential predictions mighthave been made between the two social supportconditions (Independence), statisticallysignificant results were not obtained. Thus,Ss did not perceive the partner as more independent(or credible) when he answered beforeany group pressure had been exerted (Position1). What, then, may mediate the greatereffectiveness in reducing conformity of thepartner who answered in the first position?Perhaps, Ss are maximally sensitive to thefirst answer given after the stimulus is shown,because of a desire to predict how the groupwill respond. If so, a correct first answer, inconfirming S's own judgment, may produce aninitial committment to the correct responsethat endures even though other group membersdisagree. If so, conformity should be quitelow in the Social Support 1 condition, as found.In the Social Support 4 cond:tion, on the otherhand, the initial answex is incorrect, perhapsproducing confusion in S. Thus, S would notmake initial committment to the correct answer,and conformity would be somewhat higher inthe Social Support 4 condition, as obtained.Turning now to the Contact conditions,let us examine the Conditions x Contact interactionmore carefully. As mentioned earlier,conformity was lower, in both social supportconditions, in the Contact than in the No Contactgroup. Postexperimental questionnairedata indicate, interestingly, that the effectwas not mediated by greater liking of the earlysupporter. On the contrary, Ss rated the partnerin the No Contact condition higher on thetwo indices of liking. Thus, the anonymityand differentiation explanations offered earliergain some credence, in that neither was postulatedon a liking relationship between thepartner and S. Moreover, the fact that conformityin the Unanimous condition was higherin the Contact than in the No Contact groupalso supports, though again does not give evidencewhich differentiates between, the twohypothesized mechanisms underlying the Contacteffect.7

REFERENCESAllen, V. L., & Levine, J. M. Social support,dissent, and conformity. Sociometry,1968, 31, 138-149. (a)Allen, V. L., & Levine, J. IVI. Consensus andconformity. Technical Report from theWisconsin R & D Center for CognitiveLearning, University of Wisconsin, 1968,No. 41. (b)Allen, V. L., & Levine, J. M. Social supportand conformity: The role of independentassessment of reality. Technical Report_from the Wisconsin R & D Center forCognitive Learning, University of Wisconsin,in press.Asch, S. E. Effects of group pressure uponthe modification and distortion of judgments.In H. Guetzkow (Ed.), Groups.leadership, and men. Pittsburgh: CarnegiePress, 1951.Asch, S. E. Studies of independence and conformity:I. A minority of one against aunanimous majority. PsychologicalMonographs, 1956, 70 (9, Whole No. 416).Crutchfield, R. S. Conformity and character.American Psychologist, 1955, 10, 191-198.Deutsch, M. and Gerard, H. B. A study ofnormative and informatic.nal social influencesupon individual judgment.Journal of: Abnormal and Social Psychology,1955, 51, 629-636.Festinger, L., Pepitone, A., & Newcomb, T.Some consequences of de-individuationin a group. Journal of Abnormal andSocial Psychology, 1952,47, 382-389.Levy, L. Studies in conformity behavior: Amethodological note. Journal of Psychology,1960, 50, 39-41.Singer, J. E., Brush, C. A., & Lublin, S. C.Some aspects of de-individuation: identificationand conformity. Journal ofExperimental Social Psychology, 1965,1, 356-378.Tuddenham, R. D., Macbride, P., & Zahn, V.Studies in conformity and yielding: I.Development of standard experimentalseries. Tech. Report No. 1., ONR, ContractNR 170-159, Univ. of California, 1956.8GPO a 1 1 259-2