Interpersonal Communicaion in Small Groups

7/29/2019 Interpersonal Communicaion in Small Groups

http://slidepdf.com/reader/full/interpersonal-communicaion-in-small-groups 1/8

INTERPERSONAL COMMUNICATION IN SMALL GROUPS

BY LEON FESTINGER AND JOHN THIBAUT1

Research Center jor Group Dynamics, University of Michigan

SM A L L face-to-face groups, or as they havesometimes been called, primary groups,play an important part in influencing

attitudes and opinions of their members.This important fact about social behavior has

been assumed for many years. In the past

decade experimental facts have accumulatedto substantiate this fact and to specify therelationships involved.

In summary, the following is a list of somemajo r conclusions which may be drawn from

experimental work:1. Belonging to the same group tends to

produce changes in opinions and a t t i tudesin the direction of establishing uniformity

within the group (5, 6).

2. The amount of change toward uniform-ity which the group is able to accomplish is

a direct function of how attractive belonging

to the group is for its members (i, 2).

3. Members who do not conform to the

prevailing patterns of opinion and behavior

are rejected by others in the group. Thedegree of rejection is a direct function of howattractive belonging to the group is for its

members and of the importance for the groupof the issue on which the member does not

conform (2, 7).

These facts leave unclarified the means bywhich such social influence is accomplished.The continual process of informal communi-cation among members of face-to-face groupsin part represents the attempts to influencemembers by others in the group. To under-stand completely the social influences whichgroups exert we must, then, also understandthe determinants of what does and does not

get communicated in social groups, and who

are the recipients of communications. There

are some data available. These may be sum-marized as follows:i. Persons whose social behavior is changed

by hearing something tend to relay this in-

1This study was conducted under contract with theOffice of Naval Research (N6onr—23212 NR 151—698) .It is part of a larger program of research on social com-munication and influence no w being conducted at theResearch Center fo r Group Dynamics at the Universityof Michigan.

format ion to others who are seen as likely to

be affected by it (2, 3).

2 . Persons who do not conform to the

group pattern tend to have fewer communi-cations addressed to them if they are rejectedbut tend to have more communicationaddressed to them if they are not rejected (7).

A more detailed understanding of this

process of communication and its relation to

the process of influence is the major purposeof the theories and experiments reported in

this paper.

T H E O R E T I C A L O R I E N T A T I O N

The fact that groups do exert pressures on

dieir members toward uniformity is beyonddispute. For our immediate purposes we

need not concern ourselves with the sourcesof these pressures or the reasons fo r theirexistence. We will look only at the effects of

these pressures toward uniformity on the

communication and influence process thatactually takes place in a group. A group may

be looked upon as composed of a number ofparts with each part characterized by a givens t a t e

2with respect to a certain dimension.

If the group has the property of tending

toward uniformity of state, dien any dis-crepancy among die different parts of die

group will give rise to forces which will be

exerted on parts of the group to change dieir

state in such a way as to re-establish uni-

fo rmi ty . The strength of these forces will bea function of the magnitude of the tendencytoward uniformity which the group possesses.

The force to change exerted on any par-t icular part of the group is also a direct func-

t ion of the discrepancies in state between

this part and all other parts of the group.

The larger the discrepancy between part A

and part B, the larger will be the force

exerted on A by B and on B by A, since the

2 In the experiments to be described later an individualperson is coordinated to a part of a group and anopinion concerning a certain issue to the state of theparts of the group. Cliques of people, levels in an

organization, or work groups may also be looked onas pa r t s of a group.

92



I N T E R P E R S O N A L C O M M U N I C A T I O N I N S M A L L G R O U P S

disequilibrium is greater, the greater this

discrepancy.The preceding hypotheses concerning tend-

encies toward uniformity within a group donot, of course, hold for any arbitrarily defined

collection of individua ls or parts. When dis-crepancies exist among a collection of per-sons, uniformity of any group that existswithin this collection can be achieved eitherby the exertion of forces to change variousparts of the group or, alternatively, by form-ing the group in such a way that uniformityalready exists. Redefinition of the bound-aries of the psychological group (changingthe membership composition) can, then, alsobe a response which the group ma kes to pres-

sures toward uniformity.In a group where the tendencies toward

uniformity concern an opinion about someissue, the exertion of pressures on persons tochange their opinion must of course makethemselves felt through a process of com-municat ion among them. What can we inferabout this process of communication fromthe hypotheses we have presented?

1. Within a psychological group comm uni-cations should be directed mainly toward

those m embe rs whose opinions are extrem e ascompared to the opinions of the others. This

would follow f rom our hypothesis that thestrength of the force applied on any par t ofthe group is a direct function of the dis-crepancy between the state of that part andthe states of the other parts of the group.

2. If it is possible for a group to subdivideor exclude m em bers then, as the discrepanciesin state become clear, there w ill be tendenciesto cease communicat ing to the extremes.

This would follow from a number of con-siderations that have been stated or impliedabove.

a. If it is impossible for the group to rede-fine its boundaries, then uniformity can onlybe achieved through changing others andbeing receptive to change.

b. If it is possible to redefine the bound-

aries of the group then uniformity can alsobe achieved by omitting the persons with

extreme opinions from the group.c. The perception that it is possible to

redefine the boundaries of the group should,

then, have two consequences. There should

be greater resistance to change on the part

of the members, and there should be lesscommunication to those who may be ex-

cluded from the group, namely, those with

extreme opinions.3. The less the pressure towards uniformity

in a group and/or the greater the possibilityfor the group to subdivide, the less will

be the actual accomplishment of influence.Since bodi of the factors here mentioned willaffect the readiness of members to change in

response to influence which is exerted onthem, and since possible group subdivisionwill also prevent the exertion of influence onthe most deviant members, it follows that theend result of the process of communication

will be less uniformity in the group if sub-

division is seen as possible or if the tend-encies toward uniformity are weaker.

The experiments which are describedbelow were specifically designed to test thesehypo theses. In the description of the pro-cedure we will elaborate further on the

operational definitions of the theoreticalconcepts.

E X P E R I M E N T A L P R O C E D U R E

Subjects

The subjects used in these experimentswere college undergraduates recruited fromthe various sections of the elementary psy-chology course and the elementary course ineducational psychology at the Un iversity ofMichigan. All subjects were volunteers.

General Characteristics of the GroupsFormed

Sixty-one groups were studied of which 24were composed entirely of women, 37 of

men. The size of groups ranged from 6 to14 members. Each group assembled in theexperimental room, and each member wasassigned one of a series of small tablesarranged in a circle. Each member was

identified by a letter which was printed ona 5 by 8 inch card placed in front of him so

that all others could see it.

General Setup for All Groups

Each group was given one problem to con-

sider. The problem was such that opinionsconcerning it could be placed on a prescribedseven-point continuum. Each member was

given seven 5 by 8 inch cards with numberscorresponding to diose on the seven-point



94 LEON FESTINGER AND JOHN THIBAUT

scale of opinion. The members were in-structed to consider the problem and then, allsimultaneously, to place in f ront of them thatcard which represented their tentative opin-

ion on the mat te r at issue. The experimenterthen proceeded to call attention to each per-son's decision in order both to verify it andto insure that all were fully aware of it.

Small slips of paper bearing some addi-

t ional information relevant to the problemwere then distributed at random among the

subjects. It was announced that each mem-be r of the group was receiving a differenti tem of informat ion. The purpose of thispart of the procedure was to maximize theinitial force to communicate by causing each

member to believe that he had some uniqueinformation relevant to the problem-solvingactivity. Actually, however, only two i temsof information were distr ibuted. One itemwas intended to push the member towardthe upper end of the scale, the other towardthe lower end. This device was essential to

get adequate dispersion along the scale.After th e subjects had read the new infor-

mation, each recorded directly on his infor-mation slip his identifying letter and the

scale number representing his current opin-ion. These were collected by the experi-menter and read aloud in order to makepublic the new opinions. Any membe rwhose opinion ha d changed was asked tomake the appropriate change in the num-bered card in f ront of him.

With this preliminary procedure finished,the experimenter described th e manner inwhich the problem was to be discussed.Stapled pads of paper were distributed to the

subjects. For each pad the staple was placedin a slightly different position on the page.

These differences were undetectable to thesubjects, but they allowed the experimentersubsquently to match each pad with themember to whom it had been given. Thesubjects were informed that discussion aboutthe problem had to be restricted to writingnotes to one another. The subjects were left

free to include anything they l iked in the

notes. However, a member could write anote to only one person at a time, and eachnote must bear only th e letter of the personto whom i t was directed; no reference to thesender's identity was permitted. This rule

was adopted to minimize th e chances thatany member, in the act of deciding to whomto direct a note, would be affected by aknowledge of what people had sent notes to

him. On completing a note, a sender was toraise his hand, whereupon the experimenteror his assistant would deliver it to the re-cipient. It was emphasized that if and whenany member decided to change his opinion,he should change th e numbered card in f rontof him.

At a signal f rom the experimenter the sub-jects then began to write notes. As eachnote was finished, the messenger (experi-menter or assistant) took the note, recordedon it the time in minutes and seconds from

the starting signal, and dispatched it. Arecord wa s also kept of the exact t ime of eachchange of opinion, that is, of each change inthe numbered card in f ront of a subject .The note-writ ing continued for 20 minutes .

The Discussion Problems

Two problems were used in the course ofexperimentation. A problem in footballstrategy was assigned to 31 of the groups,and a problem in evaluating a case study of

a delinquent boy was assigned to the remain-ing 30 groups.

The problem in football was concernedwith making a decision about the beststrategy for an imaginary anonymous teamwhich has the ball on the 50-yard line, firstdown, 5 minutes of play remaining, with thescore 18-18. Seven alternative types of

strategies are outlined to the subjec ts. Theserange from extremely conservative power

plays (at point i) to extremely reckless passplays (at point 7). The two items of addi-t ional information distr ibuted among th esubjects are that "our star running back hasjust been injured ..." (intended to push

th e recipient upward on the scale) and that"the opposing team has tightened up itspass defense and has caught on to our spec-tacular plays" (intended to push the recipientdownward on the scale).

The case study was a brief fictitious

account of the history of a boy who hadcaused trouble all through his life and whoha d ended in jai l . The history of the boywas deliberately made to be as ambiguousas possible in order to encourage dispersion



I N T E R P E R S O N A L C O M M U N I C A T I O N I N S M A L L G R O U P S 95

on the scale of opinion about the best pos-sible way of treating the case. The-subjectswere told that by prior decision of the socialworkers assigned to the case,, the boy was tobe put into a foster home; th e assignmentfo r the subjects was to determine the besttype of home fo r this boy. The scale ofopinion consisted of seven alternative typesof foster homes, ranging from one in whichlove and kindness were exclusively empha-sized (point i) to a home in which disciplineand punishment were exclusively used (point7). The two items of addi tional inform at ionreceived by the subjects were: (i) that for aperiod of a year his mother, acting on theadvice of a social worker, ha d tried to make

the boy's home life warm but that it did nogood, since his crim inal ac tivity increas ed(intended to push th e recipient upward onthe scale) and, (2) that the boy's old-est brother had returned home for a whileand had given the boy stern but fatherly

discipline but that the boy's delinquency onlyworsened (intended to push the recipientdownward on the scale).

The selection of these two problems was

guided by our need to create discussion situ-

ations in which there would be markedlydifferent amounts of resistance to change ofopinion. In the case study problem, it wasfelt that subjects would bring into the ex-perimental situation fairly strong predispo-sitions toward certain of the scaled opinions.These predispositions could be expected tobe quite resistant to change.

In the football problem, on the other hand,there was no expectation that strong prejudg-ments would be imported into th e si tuation.

Relative to subjects working on the casestudy problem, the subjects ought more

readily to accept the present experimentalgroup as a relevant reference group for theiropinions and hence ought to be relativelyless resistant to change.

Experimental Variations

Six experimental variations were appliedto each of the two problems. These vari-

ations were created by further instructionsover and above the general instructionsalready described. F ive groups (three maleand two female) were assigned to each of theexperimental variations in each of the prob-

lems, except fo r variation V in the footballproblem, which had six groups, fo ur male a ndtwo female.

$Instructions to Create the Hom ogeneity-

Heterogeneity Variable

In the first three variations, the intentionof the additional instructions was to createa perception that there was no basis for groupsubdivision among them. To achieve thisperception the homogeneity of the group wasemphasized as follows: "You people in thisgroup have been deliberately selected tomake up the kind of group we are interestedin observing. Y ou have been selected insuch a way that we believe you all will have

about an equal interest in this problem andabout equal knowledge about it. . . ."

In the second three variations it was in-tended that the subjects perceive the possi-

bility of group subdivision as having a basisin fact . The heterogeneity of the group wasemphasized by telling the members that theyha d been selected to be as different as pos-sible both in their interest in the problemand in their knowledge about it.

Instructions to Create th e Pressure-Taward-Uniformity Variable

Variation 1 (High pressure-homogeneity,abbreviated H-Hom). In this variation wewere interested in creating very strong pres-sures toward uniformity of opinion. Thegroup was told that th e experimenter's inter-est was in observing how a group went aboutcoming to a unanimous decision. Thus,

whatever intrinsic pressures toward uni-formity might exist in the group were

strengthened by externally induced pressures.Variation II (Medium pressure-homogene-

ity, abbreviated M-Hom). This variationwas designed to produce pressure towarduniformity of a conditional nature. The

instructions were that a body of experts ( thecoaching staff of the Universi ty of Michiganfootball team, for the football problem, andsome members of the Law School faculty fo rthe case study problem) ha d considered theproblem and had unanimously decided that

one of the seven scale points represented the"correct solution." The group was told thatit would receive a score for its performance,which would be the propor t ion of members



LEON F E S T I N G E R A N D J O H N T H I B A U T

who at the conclusion of the experiment wererecommending the "correct solution."

Variation III (Low pressure-homogeneity,abbreviated L-Hom). No external pressuretoward uniformity was applied in this vari-ation. The group was merely informed that

th e experimenter was interested in observingthe way a group went about discussing sucha problem. In this case, it was supposed thatif any pressure toward uniformity developedit would be attr ibu tab le to a need for "socialreality" within the group (2 , 4) . Accordingto this principle, there is a force on the group

member to achieve support for his point ofview; and to the extent that this point ofview is untestable by demonstration, the

member is increasingly required to acceptthe criterion of social agreement with a rele-vant reference group.

Variation IV (High pressure-heterogene-ity, abbreviated H-Het). In this var ia t ion wewere intent on establishing high pressuretoward uniformity while at the same timepermitting the fo rmat ion of subgroups. Thevariation includes instructions that the groupis composed of heterogeneous members.Otherwise it is largely a counterpar t of

Variation I (H-Hom). This t ime , however,instead of asking for a unanimous decision,the experimenter informed the group that aplurality would be sufficient. The groupwould be taken as recommending the de-cision which the greatest number of mem-bers accepted. In addition, th e subjects weretold that in such heterogeneous groups asthis, one usual ly did not find more thantwenty per cent of the members agreeing onany one alternative. These last two instruc-

t ions were made somewhat different fromthe instructions in the homogeneity condi-tions in order to allow sub-group formationto take place.

Variation V (Medium pressure-hetero-geneity, abbreviated M-Het). This variat ionwas also expected to permit subgroup forma-tion. The instructions to these groups weresubstantially the same as for variation II(M-Hom) except for the emphasis on hetero-

geneity of the members and an additionalinstruction that i t was not customarily pos-

sible fo r more than twenty per cent of thegroup to hit upon the "correct solution."

Variation VI (Low pressure-heterogeneity,

abbreviated L-Het). Except for the pretensethat th e group was heterogeneously com-posed, this variation was precisely the sameas Variation III (L-Hom).

The following tabulation is presented to

help clarify the relations among the six

experimental conditions:

P R E S S U R E T O W A R D U N I F O R M I T Y

HIGH M E D I U M Low

Homogeneous Gr oup I II IIIH eterogeneous G roup IV V VI

E X P E R I M E N T A L R E S U L TS

Hypothesis I

The volume of communication betweentwo persons is a functio n of the ma gnitudeof the discrepancy between their opinions.

Since extreme opinions are most discrepantf rom all the other opinions, we would there-fore predict that most communications shouldbe directed toward members who hold ex-t reme points of view.

\

\

CASE STUDY PRO BLE M —

(N'30 GROUPS)

FOOTBALL PROBLEM

(N*3I GROUPS)

3+OFFXTREME POSITION I OFF 2 OFF

LOCATION OF RECIPIENT

F I G . i. P A T T E R N S OF C O M M U N I C A T IO N

(First 10 min.)

Figure i summarizes the experimentalfindings relevant to this prediction in termsof the weighted number of communications.The distribution of opinions within thegroup could affect the pat tern of communi-cation. Thus, fo r example, if six members

held extreme opinions and only three mem-

bers maintained "middle" opinions, we wouldobtain a preponderance of communicationto the extremes even if members wereaddressed at random. To correct for this,

each message was weighted by the inverse



I N T E R P E R S O N A L C O M M U N I C A T IO N I N S M A L L G R O U P S 97

of the number of persons in the group in thesame relationship to the communicator asthe recipient of that particular message.Thus, a communication directed toward a

person at an extreme was divided by thenumber of persons in the group (excludingthe sender of the message) who held extremeopinions at the t ime. When the weightednumber of communications initiated duringthe first ten minutes

3of each session is

plotted against the location of the recipient

(in terms of being at an extreme position,one point away from the extreme position,etc.), the curve falls off rapidly. This rela-tionship seems to hold about equally for

groups discussing the football problem and

fo r groups discussing the case study prob-lem. Our hypothesis is confirmed—the

munications according to the location of therecipient (as in Fig. i) was computed. For

example, the mean of the distribution for thefootball problem in Figure i is .84 unitsaway from th e extreme opinion. This meanvalue is taken as an index of the tendency tocommunicate to the extremes. Low valuesof the index indicate a high proportion ofcommunicat ion to the extremes.

Table i presents these indices separatelyfor the first and second 10 minutes of dis-cussion fo r each experimental variation onthe football problem. Table 2 gives the samedata for the case study problem. In orderto examine these data from the point of viewof hypothesis II we will compare the indices

of the first 10 minutes with the indices of thelast 10 minutes in each variation. If our

TABLE i

MEAN COMMUNICATION INDICES FOR FOOTBALL PROBLEM' DISCUSSIONS

FIRST TEN M I N U T E S

HIGH MEDIUM Lo w

S E C O N D TEN MINUTES

HIGH MEDIUM Low

Horn

Het.68

• 83

.85

.83

.88

.86

H ornHet

•74•75

.63

1.30

.86

• 99

volume of communication directed toward a

group member is a function of his nearnessto the extreme of a range of opinions.

Hypothesis II

Since communication tends to be directedtoward the extremes of a psychological group,

it is predicted that where the formation ofsubgroups (redefinition of the boundaries ofthe group) is possible there will be less com-munication directed toward the extremes ofth e experimental .(arbitrarily defined) group.

Since the heterogeneity condition providedmore basis for subgroup form ation than didthe homogeneity condition, we may expectgreater decreases in communication towardthe extremes in the former as subgroups are

given time to develop.This hypothesis was tested in the following

way. For each experimental group the mean

value of the frequency curve showing thedistribution of weighted number of com-

8Exactly the same type of curve is found for the

second ten minutes of discussion. The curve is so con-sistent that only two are shown as examples.

hypothesis is correct we would expect to findthe indices increase for the heterogeneityconditions more than for the homogeneityconditions.

Examining the homogeneity conditionsfirst we find no tendency toward any changef rom the first to the second 10 minutes. For

the high pressure condition there is an ex-

tremely slight and insignificant increase forboth discussion problems. For the mediumpressure condition there is a tendency for theindex to decrease, which again does notapproach significance. For the low pressurecondition the index stays virtually the samefo r the football problem and increases insig-nificantly for the case study problem.

In the heterogeneity conditions a quite dif-ferent picture presents itself. In the highpressure condition there is no change in theindex, but in the medium and low pressure

conditions there are consistent increases inthe index from the first to the second 10m inutes . Two of these fo ur increases, themedium condition for the football problemand the low condition for the case study



L E O N F E S T I N G E R A N D J O H N T H I B A U T

problem, are significant at the 5 per cent levelof confidence. Taken together the changesin the medium and low pressure conditions

are significant at the i per cent level of

confidence.These results seem to substantiate but

qualify hypothesis II. While the homogene-ity conditions show no increase in the index,

the heterogeneity conditions show such an

increase only where the pressure toward uni-formity is sufficiently low to permit subgroupformation. In the high pressure conditionswhere strong pressures toward uniformityare exerted by the experimenter on the totalgroup, subgroup formation does not occur.

sures toward uniformity. When pressures

toward uniformity become very high, theseother forces in the situation may become lesseffective in comparison.

Hypothesis 111

As pressure toward uniformity increases,

both pressure to communicate and readinessto change also increase. Since both of thesefactors are conducive to change, there shouldbe increasing change toward unifo rm ity ofopinion as the pressure toward uniformityincreases.

In order to test this hypothesis a measureof the amount of change toward uniformity

TABLE 2

MEAN COMMUNICATION I N D I C E S F O R C A S E S T U D Y P R O B L E M D I S C U S S IO N S

F I R S T TEN M I N U T E S


S E C O N D T E N M I N U T E S


HornHct

.27 .62 .48•5° -3 i

HornH et

•3 5•3 0

.5 6

.7 2•74• 78

Where the pressure toward uniformity isweaker, subgroup formation does occurwhen a basis for it (perception of hetero-

geneity) exists.

It is also apparent from Tables i and 2

that in both the homogeneity and hetero-geneity conditions, increasing the magnitudeof pressure toward uniformity produces morecommunication toward the extremes. If wecompare the indices for the high pressureand low pressure conditions we find that inthe eight possible comparisons, the index fo r

low pressure is greater in seven instances andtied in one instance. The index fo r mediumpressure is higher than for high pressure insix of eight possible comparisons and tiedin one instance. There is no consistency inthe comparison between the m edium and lowpressure conditions.

In view of the consistency of the result wemay conclude with a high degree of confi-dence that high pressure toward uniformityresults in increased communication to theextremes. This result probably depends uponthe degree to which tendencies to communi-cate arising from other sources can competewith communications resulting from pres-

was calculated fo r each experimental group.

The index used was the quotient of the

standard deviation of opinions within thegroup by the end of the 20-minute discussion,

divided by the standard deviation within thegroup at the beginning. The lower the

index, the greater ha s been the changetoward uniformity of opinion. Thus, fo rexample, an index of 1.0 represents no changeat all, and this value may be regarded as abase line in the figure.

Figure 2 presents these indices fo r each of

the experimental variations and for each ofthe discussion problems. It can be seen thatin each instance, as the pressure toward uni-formity is decreased, the amoun t of changetoward uniformity is decreased. The trendsmay be regarded as significant well beyondthe i per cent level of confidence, since the

probability of obtaining this predicted orderof three points in four ^independent com-parisons would be by chance about onein a thousa nd. The data fully support

hypothesis III.Hypothesis IV

If subgroup format ion is seen as possible,the readiness to change when influence is



I N T E R P E R S O N A L C O M M U N I C A T IO N I N S M A L L G R O U P S 99

exerted should be less than where no sub-group formation is possible. In addition, in

the former case there is less actual exertionof influence on the extreme opinions in the

group. Both of these factors should combine

to produce less change toward uniformity inthe heterogeneity than in the homogeneityconditions.

Figure 2 shows the data relevant to thishypothesis. The difference between the

amount of change in the heterogeneity and

HIGH MEDIUM LO W

P R E S S U R E T O W A R D U N I F O R M I T Y

FIG. 2. M E A N A M O U N T S OF C H A N G E T O W A R D

U N I F O R M I T Y O F OPINION

homogeneity conditions is highly significant(beyond the i per cent level by analysis of

variance) when the football problem is dis-

cussed. There is, however, little or no differ-ence between these two conditions when the

case study problem is discussed.It will be recalled that the case study prob-

lem was selected in the belief that subjectswould bring with them fairly strong predis-positions toward certain of the opinions,which would be relatively more resistant to

change. The football problem was selectedin the belief that subjects would not bring

such predispositions intothe

experimentalsituation. This difference between the two

discussion problems is clearly reflected in themuch lower degree of change toward uni-formity in the case study problem. It is alsoprobable that this relatively high resistance

to change in the case study problem madethe added effect of the heterogeneity-homo-geneity difference relatively negligible.

We may conclude that, where strong pre-dispositions do not exist and where, conse-

quently, the group has power to changeopinions, the perception of heterogeneitywill increase resistance to change. Hypothe-sis IV , thus amended, may be considered tobe substantiated.

S U M M A R Y

The variables of (i) amount of pressuretoward uniformity existing in a group and(2) the degree to which th e members per-ceived the group as homogeneously com-

posed, were manipulated experimentally in alaboratory setting of a discussion group to

test certain hypotheses concerning the pat-tern of communication within the group and

the amount of change in opinion whichoccurs. The results strongly support the

theoretical hypotheses and may be sum-marized as follows:

1. When there is a range of opinion in thegroup, communications tend to be directedtowards those members whose opinions areat the extremes of the range.

2. The greater the pressure toward uni-formity and the greater the perception ofhomogeneous group-composition, the greateris the tendency to communicate to theseextreme opinions.

3. The greater the pressure toward uni-formity and the greater the perception ofhomogeneous group-composition, the greater

is the actual change toward uniformity which

takes place.REFERENCES

1. BACK, K. The exertion of influence through social

communication. /. abnorm. soc. PsychoL, 1951,

46, 9-23.2. FESTINGER, L. , SCHACHTER, S., AND BACK, K . Social

pressures in informal groups: A study of a housingproject. New York: Harper, 1950.

3. FESTINOER, L., CARTWRIGHT, D., ef al. A study of

a rumor: Its origin and spread. Hum. Relat.,

1 9 4 8 , i, 464-486.

4. FESTINGER, L. Informal social communication.Psychol . Rev., 1950, 57, 271-282.

5. NEWCOMB, T. Personality and social change. New

York: Dryden Press, 1943.6. SHERIF, M . Psychology of social norms. New York:

Harper, 1936.

7. SCHACHTER, S. Deviation, rejection and communica-tion. /. abnorm. soc. Psychol. , 1951, 46. Inpress.

Received January 23, 1950.

Interpersonal Communicaion in Small Groups

Documents

Transcript of Interpersonal Communicaion in Small Groups