VOL. Psychological Bulletin - Gwern.net · Psychological Bulletin ... personality assessment is...

VOL. 56, No. 4 JULY, 1959

Psychological Bulletin

A REVIEW OF THE RELATIONSHIPS BETWEEN PERSONALITYAND PERFORMANCE IN SMALL GROUPS1

RICHARD D. MANNUniversity of Michigan

A wide range of practical and theo-retical interests have found expres-sion in the study of small groups. Asthe major bibliographic sources(Hare, Borgatta, & Bales, 1955; Mc-Grath, 1957; Strodtbeck & Hare,1957) amply attest, small group re-search has proceeded along numerousindependent lines. One interest, how-ever, has been dominant for morethan 50 years. While phrased invarious ways, the relationship be-tween the personality characteristicsof the individual and his performancein the group has remained a centralconcern.

There have been at least three con-ceptual approaches to this problem.One approach considers the individ-ual as having various needs and asbeing motivated to satisfy some ofthese needs through interaction withothers; the point of interest is the re-lation between the individual's per-sonality and his goal-directed be-havior in groups. In another view,the individual is conceived of as astimulus, or set of stimuli, for theother members of the group, and the

1 The author wishes to express his gratitudeto Roger W. Heyns for his valuable sugges-tions and criticism throughout the prepara-tion of this manuscript. The survey of theliterature was carried out during the periodwhen the author held a Research TrainingFellowship granted by the Social Science Re-search Council.

relation between the individual's per-sonality and the way in which he isperceived and judged by his peers as-sumes primary importance. In thethird approach, the group is con-ceptualized as a system confrontedwith various problems, external andinternal, and attention shifts to theprocesses whereby particular individ-uals volunteer or are selected to oc-cupy various positions and performvarious roles necessary for the solu-tion of the problems. Although thesethree approaches have generatedmany nonoverlapping research ques-tions, they have produced a body ofdata which may be considered mean-ingfully as a whole.

This review attempts to summarizethe present state of knowledge aboutthe relationship of an individual'spersonality to his behavior or statusin groups. Although the independenteffects of varying the nature of thesample and history or size of thegroup upon the performance of indi-viduals are not considered, an effort ismade to determine the effect of suchsituational factors on the relation-ships observed between personalityand performance.

While the purpose of this review isto provide an adequate and accuratedescription of the present state ofknowledge in the field, its intent is tostimulate research rather than tomake a final summary. It is thought

241

242 RICHARD D. MANN

that an organized presentation of thefindings to date may help to clarifyrelationships which have been over-looked or misunderstood. Moreover,it is hoped that this summary may beused as a target and a taking-offpoint for future research, thus en-couraging publication and helping tomake knowledge in this field morecumulative.

Selection of the StudiesThe studies selected for detailed

examination meet the following fivecriteria: (a) the sample was drawnfrom a population of high school ageor older; (&) the groups studied wereface-to-face groups; (c) some assess-ment was made of the individual'spersonality; (d) some assessment wasmade of the individual's behavioror status in the group; and (e) theresults were either in correlationalform or made use of a control group,i.e., studies testing only leaders oronly social isolates are not considered.The only exception to (b) occurs inthose studies of conformity in whichthe individual believes he is inter-acting with other individuals, where-as, in fact, the experimeter has con-trolled the interaction through taperecordings or false statements aboutthe actual behavior of the others inthe group.

This review covers the availableliterature from 1900 through Octo-ber, 1957. The bibliography was col-lected by searching the most relevantjournals and published abstracts, byfollowing the network of referencesfrom article to article, and by ob-taining as much unpublished re-search as possible. In addition, theearlier reviews (Bass, 1954; Bor-gatta, 1954; Gibb, 1950, 1954; Jen-kins, 1947; Roseborough, 1953; Smith& Krueger, 1933; Stogdill, 1948)which emphasize leadership or popu-

larity to the exclusion of other as-pects of performance covered herehave been useful. No claim is madeto completeness, but no sources ofknown relevance have been delib-erately overlooked.

The Personality VariablesThe studies which meet the criteria

for selection used over 500 differentmeasures of personality. However,less than a quarter of these measuresappear in more than one study. As acommentary on the level of integra-tion within the field, this fact needslittle amplification. There is a notice-able failure throughout these studiesto resolve methodological issues in aconsistent fashion.

Clearly, it is not feasible to pres-ent each separate personality vari-able and its correlates. Some organ-ization of the measures was called for.But what organization? The field ofpersonality assessment is test richand integration poor. The 500 meas-ures all have labels, to be sure, butthey are as divergent as oral sadism,the F scale, spatial ability, adven-turous cyclothymia, hypochondria-sis, and total number of vista re-sponses. Yet all of these measureshave been used to predict somethingabout an individual's performance ingroups. In addition, there are in-numerable adjectives used for ratingsand self-descriptions. The situationrequired a set of personality factorssmall enough to remain manageableand pure enough to be meaningful,and then empirical grounds on whichto classify as many of the variablesas possible into the selected set offactors.

To arrive at a useful set of person-ality dimensions, the empirical workin the field of personality assessment,particularly the work of French(1953), Cattell (1946, 1956, 1957) and

PERSONALITY AND PERFORMANCE IN SMALL GROUPS 243

Eysenck (1953) was examined. Withone exception, the seven dimensionsor factors chosen are those frequentlyisolated in the study of personalityby factor analytic techniques, al-though two emerge only as second-order factors in some reports. A briefdescription of each personality factoris presented here.

Intelligence, This factor includesall the diverse and specific mentalabilities. Sixty-nine of the 500 differ-ent variables included in the researchreviewed are measures of this factor;of these sixty-nine measures 45 arederived from questionnaire and ob-jective tests, 24 from adjective rat-ings. The four most frequently usedmeasures of intelligence are: schoolor college grades, American Councilof Education (ACE) PsychologicalExam, Cattell's Sixteen PersonalityFactor Questionnaire (16 P.F.) Fac-tor B, and total number of responseson the Rorschach.

Adjustment. The positive end ofthis dimension has been called ad-justment, ego strength, and normal-ity, while the negative end has beencalled maladjustment, emotionality,neuroticism, psychoticism, and anx-iety. Seventy-one objective test andquestionnaire variables and 60 adjec-tives are considered as measures ofthis factor. The most frequently usedmeasures of this factor are derivedfrom standard personality inven-tories: Minnesota Multiphasic Per-sonality Inventory (MMPI), Guil-ford-Zimmerman, Bernreuter, and 16P.F.

Extroversion-introversion. Eysenck(1953) presents the fullest discussionof this dimension, although the needto integrate as many variables aspossible led to the use of a broaderdefinition. Extroversion-introversionas used in this review, more closelyresembles one of Cattell's (Cattell,

Saunders & Stice, 1951) second-orderfactors from the 16 P.P., which pullstogether the dimensions of sociabil-ity, surgency, and cyclothymia vs.schizothymia. Frequently-used meas-ures of this factor are: the BernreuterF-2 scale (self-sufficiency), MMPIHypomania scale, ratings on "so-ciable," and the relevant scales fromthe 16 P.F. (Cattell, 1956) and Guil-ford-Zimmerman (French, 1953). Atotal of 38 questionnaire and objec-tive test variables and 61 adjectiveratings were used in the studies re-viewed.

Dominance. The positive end ofthe dimension is described by domi-nance or ascendance, the negative endby submissiveness or helplessness.Seventeen objective test and ques-tionnaire variables and twelve ad-jective-ratings which have beenfound to measure dominance wereemployed in these studies.

Masculinity-femininity. This factormeasures the extent to which an in-dividual's interests or preferences re-semble those common to his own orthe opposite sex. Of the 14 question-naire and objective test variables andsix ratings, the ones most frequentlyused in these studies are the mascu-linity-femininity scales from theMMPI, Guilford-Zimmerman, andGoodenough Speed of AssociationTest.

Conservatism. The positive end ofthis dimension is defined by conserv-atism, conventionalism, or author-itarianism, the negative end byradicalism. In the studies review, themeasures of this factor include 36questionnaire and objective test vari-ables and 11 adjective-ratings. Byfar the most frequently used are theF scale and factor Qi from the 16P.F.

Interpersonal sensitivity. This fac-tor has not been found in factor ana-

244 RICHARD D. MANN

lytic studies of personality, and someauthors have questioned whether it isproper to speak of empathy and in-sight as characteristics of the individ-ual. However, it is included in thisreview because it has been related toan individual's status in groups asufficient number of times to meritseparate treatment. For the mostpart, the measures describe an in-dividual's ability to guess (a) his ownstatus in a group, (&) the statushierarchy of the entire group, as de-termined by the pooled estimates ofthe members, or (c) the opinions andattitudes of the other group members.

One hundred and fifty variablesout of the total of over 500 could notreasonably be classified into any oneof the seven factors. Some of these150 variables fall into other knownfactors or clusters, but the number ofadditional results which could be in-cluded by considering them is toosmall to justify the consequent com-plexity of the presentation. Themajority of the excluded variables,however, come from projective tests;in such cases, both the titles and theknown correlations with other per-sonality measures combined to mys-tify this reviewer as to what meaningthey might have outside the languagesystem of the particular technique.Many projective test variables donot fall into stable and identifiableclusters or factors; further, the levelof description used in projective testsmakes it difficult to bridge the gapbetween the seven aspects of person-ality examined in this review and thevarious projective measures. Exceptfor the measures of interpersonalsensitivity, the distribution of vari-ables into factors was determined bythe empirical evidence for the meas-ure's validity. Where no validitydata were found for a measure, a cal-culated risk was taken in assigning it

to a factor if the title and operationclosely resembled the set of variablesalready chosen on empirical groundsas measures of the factor; this proc-ess accounted for no more than 50of the variables classified.

The Status and Behavior VariablesIn contrast to personality vari-

ables, measures of an individual'sstatus and behavior in groups falleasily into a small number of classes.On the basis of both operations andlabels the following six dependentvariables were selected: (a) leader-ship, (b) popularity, (c) total activityrate, (d) task activity, (e) social-emotional activity, and (f) conform-ity. Leadership and popularity areconsidered to be status variables; theremaining four are considered to bebehavior variables.

Leadership has been measured infour ways: by having an observerrate the individual's attained leader-ship, by having an individual's peersrate him, by using an individual'sformal selection for office as the cri-terion of leadership, or by having theindividual rate himself. The onlymeasures included in the discussionof leadership which do not bear thatlabel are a few measures on the in-dividual's productivity and effec-tiveness. Popularity has been meas-ured by having an individual's peersrate him on such dimensions as theextent to which they like him, findhim acceptable as a friend, wouldchoose him for leisure time activities,or perceive him to be popular.

The remaining dependent variablesare based upon actual observations ofthe individual's behavior in thegroup. Activity rate has been meas-ured in terms of either the number ofacts initiated or the number of sec-onds spent talking. The distinctionbetween task activity and social-


emotional activity is made by Bales(1950). He distinguishes betweentask acts, relevant to the external-adaptive problems of the group(suggestions, opinions, orientations,and task questions), and social-emo-tional acts, relevant to the internal-integrative problems of the group(agreeing and disagreeing, showingtension release and tension, showingsolidarity and antagonism). Meas-ures of behavior other than those em-ploying the Bales categories werematched as carefully as possible tothe Bales categories and classified onthat basis. Conforming behavior in-cludes all measures of an individual'stendency to yield to the opinions orpressures of the group.

The review thus covers seven as-pects of personality and six aspectsof behavior and status. If the datawere available in sufficient quantity,we would be able to examine 42 dif-ferent relationships between person-ality and behavior or status.

Method of Presentation

One final issue, the most appro-priate unit of research, must be dis-cussed before the presentation of thefindings. The problem arises fromthe fact that a single study may con-tain, for example, more than onemeasure of leadership and more thanone measure of intelligence. On theone hand, we might consider thestudy as the unit, examining only theover-all trend of the many results.On the other hand, we might con-sider each result as the unit, examin-ing the findings from a study in asmuch detail as possible.

The advantage of using a wholestudy as the unit is that units arethen independent, and, therefore,statistical tests of the significance ofthe trends are possible. Another ap-proach is to consider as the separate

unit each result, that is, each correla-tion or measure of difference betweengroups. This can lead to overrepre-sentation of a particular sample anda particular set of measures in thetotal summary of research to date.Moreover, it is not possible to usestatistical tests to evaluate trendsbased on more than one result perstudy, since using the same subjectsand then using independent or de-pendent variables which are highlycorrelated with each other wouldviolate the assumption of independ-ence which underlies statistical tests.

If each relationship had been in-vestigated in a sufficient number ofstudies to permit statistical tests inmost cases, we would have chosenstudies as the units. Because such isnot the case, we have chosen the re-sult as the unit of research, but it isrecognized that, for the above-men-tioned reasons, any trends based onseparate results must remain as de-scriptive indications of the findingsto date. Where the number of studiesis sufficient to provide an opportu-nity to use statistical tests, the testswill be made.

There are a number of advantages,however, to using the results as units.Over 1400 results are examined inthis review. The far greater numberof results may compensate for the dis-advantages of this approach by of-fering greater stability to the trends.

The association between a person-ality variable and a status or be-havior variable is reported in one ofeight forms: (a) positive and signifi-cant, (6) positive and not signifi-cant, (c) positive but no report ofsignificance, (d) negative andsignifi-ficant, (e) negative and not signifi-cant, (f) negative but no report ofsignificance, (g) zero correlation, and(h) not significant but no report ofdirection. Throughout this review a

246 RICHARD D. MANN

positive finding refers to the associa-tion between the positive ends of thepersonality and performance vari-ables as described earlier, not to theconfirmation of an hypothesis. The.05 level is accepted as the criterionof significance.

Each relationship for which five ormore results are available and whichhas been investigated in more thanone study is examined in detail.Three summary statistics are usedthroughout to describe the findingsfor each relationship. First, the over-all direction of the results is shown bythe percentage of results which arepositive; this is calculated by divid-ing the number of positive results (a,b, and c above) by the number of re-sults which indicate direction (thetotal number minus g and h). Second,the over-all direction of the resultsis further shown by the percentageof significant results which are posi-tive; this is calculated by dividing thenumber of results which are signifi-cantly positive (a) by the total num-ber of significant results (a and d).Third, as a way of examining thesignificance of the results underlyingthe trends, the percentage of the totalnumber of results which are signifi-cantly positive or negative, depend-ing on the direction of the trend, isshown. If the trend is positive, thisis calculated by dividing the numberof significantly positive results (a)by the total number of results minusthe number which are positive butuntested (c); if the trend is negative,this is calculated by dividing thenumber of significantly negative re-sults (d) by the total number of re-sults minus the number which arenegative but untested (f).

There appears to be a general be-lief that many inconclusive and nega-tive findings are filed away into ob-scurity, doomed never to enter the

professional literature. To the extentthat this bias exists in the area re-viewed, the trends are misleading.This reviewer has succeeded in ob-taining some unpublished data anddoctoral dissertations in an effort tocounterbalance the alleged distortionin the published materials. However,it may be noted that the unpublisheddata included here are in almost per-fect agreement with the data in thejournals and monographs. Thisseems to suggest that considerationsother than the conclusiveness of theresults operate to determine whichresults will be published.

Leadership

Viewed historically, the study ofleadership has stimulated more thanits share of controversy. The traitapproach to leadership, the view thatleadership is an attribute of the in-dividual, has received the harshesttreatment throughout the years. Tohave spoken of an individual aspossessing a measurable quantity ofleadership was perhaps an unfortu-nate choice of words. The clear im-plication of such a statement is thatsince leadership is specific to the in-dividual, it will remain constant forthe individual regardless of the situa-tion in which he finds himself. In-vestigations of the actual consistencywith which an individual maintainsleadership status in different groupsand under varying conditions haveyielded results sufficiently equivocalto permit a new bifurcation of thefield. On the one hand, the trait ap-proach has been modified to implythat an individual's achieved leader-ship status is a function of his per-sonality. On the other hand, suffi-cient evidence has been accumulatedto give impetus to the situational ap-proach to leadership, which main-tains that leadership is an emergent


phenomenon, created through the in-teraction of individuals (leaders andfollowers), and that the selection andstability of any leadership pattern isa function of the task, composition,and culture of the group. From allthis work has emerged some suchsummary formulation as that an in-dividual's leadership status in groupsis a joint function of his personalityand the particular group setting.There is an interesting parallel hereto the controversy over the role ofheredity and environment in deter-mining behavior; the initial criti-cisms and intensity gave way to con-cessions that each factor sets limitsfor the operation of the other, andresearchers turned to studying therelative importance of and the inter-action between the two major fac-tors.

Table 1 presents a summary of therelationships between seven aspectsof personality and leadership. Shownthere are the number of relevant stu-dies, the number of results containedin those studies, the distribution ofresults into the various forms inwhich they are reported, and thethree summary statistics. The re-

sults are reported in eight forms: thepositive and negative associationsmay be significant (Sig.), not signifi-cant (N.S.), or untested (Unt.); theremaining two forms, zero correlation(zero) and not significant but no di-rection reported (?N.S.) are com-bined in the table. The base num-bers for the summary percentagesare enclosed in parentheses below thepercentages. The base number forthe percentage of results which arepositive (i) is the total number of re-sults which indicate direction; thebase number for the percentage ofsignificant results which are positive(j) is the total number of significantresults; the base number for the per-centage of results which are bothsignificant and in the direction of theover-all trend (&) is the total numberof results minus the positive but un-tested (c) or negative but untested(/) results, depending on the directionof the trend. A separate sectioncovers each relationship between anaspect of personality and leadership.

Intelligence. Twenty-eight of thestudies reviewed, (Arbous & Maree,1951; Bass, 1951b; Bass & Coates,1952; Bass, McGehee, Hawkins,

TABLE 1THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND LEADERSHIP

PersonalityFactors

Intelligence

Adjustment

Extroversion

Dominance

Masculinity

Conservatism

Sensitivity

No.of

Stud-ies

28

22

22

12

9

17

15

No.of

Re-sults

196

164

119

39

70

62

101

Sig.W

91

50

37

15

11

3

15

Positive

N.S.

68

55

38

9

37

18

55

Negative

Unt.M

14

14

6

3

0

0

3

tfr1

2

6

6

1

17

1

N.S.W

22

28

23

4

19

21

25

Unt.

0

0

3

0

0

3

0

Zero?N.S.(*.*)

0

15

6

2

2

0

2

Positive

8

88(196)

80(149)

72(113)

73(37)

71(68)

38(62)

74(99)

%ofsit, y)

99(92)

96(52)

85(43)

71(21)

92(12)

15(20)

94(16)

% Sig, &in Dir.

of Trend(ft)

50(182)

33(150)

33(113)

42(36)

16(68)

29(59)

15(98)

248 RICHARD D. MANN

Young, & Gebel, 1953; Bass & Wurs-ter, 1953a; Bass, Wurster, Doll, &Clair, 1953; Borgatta, 1953; Carter &Nixon, 1949; Cattell & Stice, 1954;Cobb, 1952; Cowley, 1931; Dunkerly,1940; Flemming, 1935; French, 1951;Gibb, 1949a; Gordon, 1952; Gowan,1955; Green, 1950; Howell, 1942;Hunter & Jordan, 1939; McCuen,1929; Richardson & Hanawalt, 1943;Riggs, 1953; Slater, 1955a; Stolper,1953; Sward, 1933; Wurster & Bass,1953; Zeleny, 1939) have investigatedthe association between an individ-ual's intelligence and his leadershipstatus in one or more groups. Thesestudies contain 196 results, 173 (or88%) of which indicate a positiverelationship between intelligence andleadership. Furthermore, 91 (or 99%)of the 92 significant results are in thepositive direction. Omitting those re-sults which are positive but untestedfor significance, exactly half of the re-maining 182 results are both positiveand significant at the .05 level. Con-sidering independent studies as theunits of research, the positive asso-ciation between intelligence and lead-ership is found to be highly significant(p<,0l) by the sign test. However,the magnitude of the relationship isless impressive; no correlation re-ported exceeds .50, and the medianr is roughly .25.•IIThere is some indication thatverbal intelligence is a better pre-dictor of leadership than such non-verbal factors as memory and nu-merical ability. Grades are notstrongly related to leadership in col-lege social groups, although this factmay reflect competition betweenscholastic and social activities for thestudent's time and energy.

There would seem to be littledoubt that higher intelligence is as-sociated with the attainment of lead-ership in small groups. That the null

hypothesis may be emphatically re-jected should not obscure the factthat the magnitude of the relation-ship is not high.

Adjustment. The 22 studies (Bass,McGehee et al., 1953; Bass, Wursteretal., 1953; Borgatta, 1953; Carter &Nixon, 1949; Cattell & Stice, 1954;Cowley, 1931; Dexter & Stein, 1955;Dunkerly, 1940; Flemming, 1935;French, 1951; Gibb, 1949a; Gordon,1952; Gowan, 1955; Holtzman, 1952;Hunter & Jordan, 1939; Richardson& Hanawalt, 1943, 1944, 1952;Slater, 1955a; Stolper, 1953; Wil-liamson & Hoyt, 1952; Zeleny, 1939)relating the personal adjustment ofthe individual to his leadership statusyield 164 results. The trend of theresults is clearly positive, as indi-cated by the fact that 80% of theresults are in the positive direction.If only the 52 significant results areconsidered, the proportion of positiveresults rises to 96%. One third of theresults are both positive and signifi-cant. The over-all trend withinevery study but one is positive, andthe sign test indicates that the nullhypothesis of no association may berejected at the .01 level. No singlevariable measuring adjustment is cor-related with leadership over .53, andthe median correlation appears to lieclose to .15.

Four studies using the Bernreuter(Gowan, 1955; Richardson & Han-awalt, 1943, 1944, 1952) and oneusing the 16 P.F. (Cattell & Stice,1954) present the most striking evi-dence of this positive association butthe various techniques for measuringadjustment (questionnaires, objec-tive tests, and ratings) are aboutequally productive of positive results.While no single measure of adjust-ment can be expected to be an effi-cient predictor of leadership, there isstrong evidence to indicate a positive


relationship between an individual'sadjustment and the leadership statushe is likely to attain.

Extroversion-introversion. Twen ty-two studies (Bass, McGehee et al.,1953; Bass, Wurster et al., 1953;Borgatta, 1953; Carter & Nixon,1949; Cattell & Stice, 1954; Cowley,1931; Dexter & Stein, 1955; Dun-kerly, 1940; Flemming, 1935; French,1951; Gordon, 1952; Gowan, 1955;Hunter& Jordan, 1939; Moore, 1935;Richardson & Hanawalt, 1943, 1952;Slater, 1955a; Stolper, 1953; Sward,1933; Williamson & Hoyt, 1952;Zeleny, 1939) have investigated theassociation between extroversion andleadership; 72% of the results arepositive, and 85% of the 43 signifi-cant results are positive. The non-chance character of this associationis suggested by the fact that 33% ofthe results are both significant andpositive. Finally, the sign test onthe over-all trends for the independ-ent studies reveals that the positiveassociation is significant at the .01level.

No single measure of extroversionis consistently related to leadership,with the possible exception of therelevant Guilford-Zimmerman scales.The median correlation is roughly.15, and the highest correlation re-ported is .42. Those individuals whotend to be selected as leaders aremore sociable and outgoing, al-though the process of inferring such acharacterization from the titles ofthe personality scales is a tenuousmatter at best.

Dominance. Twelve studies (Bass,McGehee et al., 1953; Bass, Wursteretal., 1953; Borgatta, 1953; Carter &Nixon, 1949; Cattell & Stice, 1954;Cobb, 1952; Cowley, 1931; Dexter &Stein, 1955; Gordon, 1952; Moore,1935; Stolper, 1953; Zeleny, 1939)have investigated whether domi-

nance, as measured by personalityscales, is associated with an individ-ual's leadership status; 73% of theresults are positive, and 71% of the21 significant results are positive.The significance of the positive as-sociation is suggested by the factthat 42% of the results are both posi-tive and significant. No correlationreported exceeds .42, and the mediancorrelation is roughly .20.

The two measures of dominancewhich yield the best evidence for apositive relationship between domi-nance and leadership are the Ascend-ence and Dominance scales from theGuilford-Zimmerman and 16 P.F.,respectively. Particularly unsuccess-ful, however, have been the attemptsto use Allport's Ascendence-Submis-sion Test. Although the trend is notvery strong, these data suggest thatdominant or ascendent individualshave a greater chance of being desig-nated leader.

Masculinity-femininity. There is aslight positive association betweenmasculinity and leadership status;71% of the results are positive. Al-though 92% of the 12 significant re-sults are positive, significant resultsare found in only two of the ninestudies. No single measure of mas-culinity relates to leadership in a con-sistently positive direction, and thecorrelations are uniformly low (Bass,Wurster et al., 1953; Bell, 1952;Carter & Nixon, 1949; Cobb, 1952;Dexter & Stein, 1955; Gordon, 1952;Slater, 1955a; Stolper, 1953; Zeleny,1939).

Conservatism. Only one measure ofthis factor displays any consistencyin its association with leadership. TheCalifornia F scale, a measure ofauthoritarian trends within the per-sonality, has been used 10 times inthe prediction of leadership. In eachcase, high-F, or authoritarian, indi-

250 RICHARD D. MANN

viduals were found to be rated loweron leadership than nonauthoritarianindividuals. In general, there is anegative association between con-servatism and leadership. This isespecially evident within the signifi-cant results, 17 out of 20 being in thenegative direction (Bass & Coates,1952; Bass, McGehee et al., 1953;Bass, Wurster et al., 1953; Carter &Nixon, 1949; Cattell & Stice, 1954;Cowley, 1931; Flemming, 1935;French, 1951; Hays, 1953; Haythorn,Couch, Haefner, Langham, & Carter,1956a; Hollander, 1954; Hunter& Jordan, 1939; Martin, Gross, &Darley, 1952; Masling, Greer, & Gil-more, 1955; Slater, 1955a, 1955b;Stolper, 1953).

Interpersonal sensitivity. Few areascovered by this review contain somuch research which builds uponprior results as this one. Unfortu-nately, few are so plagued by diffi-culties and contradictory evidence.The over-all trend of the results ispositive; in 74% of the cases leadersare found to be more accurate in esti-mating various aspects of the opin-ions of other group members thannonleaders. More impressive is thefact that 15 out of the 16 significantresults indicate greater insight amongleaders. Although two of the relevantstudies report a zero correlation be-tween interpersonal sensitivity andleadership, the trends of the resultsin the remaining 13 studies are posi-tive. It would appear that while mostresearchers have been unable to ob-tain positive results which are sta-tistically significant, they have ob-tained positive results with impres-sive consistency (Bell, 1952; Bell &Hall, 1954; Campbell, 1953; Chow-dhry, 1948; Chowdhry & Newcomb,1952; Gage & Exline, 1953; Green,1948; Greer, Galanter, & Nordlie,1954; Kites & Campbell, 1950; Nord-

lie, 1954; Smith, Jaffe, & Livingston,1955; Sprunger, 1949; Stolper, 1953;Trapp, 1955; Zeleny, 1939).

According to Campbell (1955) onepart of these results is open to a seri-ous methodological criticism. Wheninterpersonal sensitivity is measuredin terms of an individual's accuracyin guessing how his peers will ratehim on leadership, the correlation be-tween interpersonal sensitivity andleadership is spuriously positive. Ifaccuracy is measured by the dis-crepancy between an individual's ac-tual leadership status and his guessedleadership status, and if, further,there is a tendency for most indi-viduals to guess that they will berated as having fairly high status,then the higher the actual status, theless the discrepancy and the higherthe apparent interpersonal sensitiv-ity. Thus, the positive correlation be-tween actual leadership status andthis accuracy score is a statisticalartifact. The cogency of Campbell'scriticism may be reflected in the factthat 14 of the 17 correlations reportedbetween actual leadership status andaccuracy about one's own leadershipstatus are positive. Since the propor-tion of these questionable resultswhich are positive (82%) is higherthan the proportion of results remain-ing (70%) when these are eliminated,the validity of Campbell's criticismis at least suggested.

There are a number of problems ofinterpretation in this area of re-search. Gage and Cronbach (1955)have written a penetrating analysisof the difficulties in measuring inter-personal sensitivity. Among otherthings, they point out the importanceof controlling the contribution of theindividual's actual similarity to othersin measuring his empathic ability.To conclude that the leader is moreaware of group opinion is a different


matter than to conclude that hisopinion is more similar to the averageopinion. More rigorous examinationof the components of interpersonalsensitivity and their various associa-tions with leadership remains a taskfor future research.

A second problem arises when oneattempts to specify which of themany items of group opinion leadersmay be expected to estimate moreaccurately than nonleaders. Accord-ing to Chowdhry and Newcomb(Chowdhry, 1948; Chowdhry & New-comb, 1952) the item cannot be tooirrelevant to the group under studyor the leader will not have adequatedata on which to base his estimate.On the other hand, according to New-comb (1954) the item cannot be toorelevant or everyone will know theopinion of everyone else, and the dif-ference will disappear. Chowdhryand Newcomb are proposing a rangeof relevance within which accuracyin estimating group opinion will bepositively related to leadershipstatus. In the absence of an objectivedefinition of relevance, this proposi-tion, for all its attractiveness on thecommon sense level, has remained anad hoc instrument to be wieldedagainst conflicting results. An ex-amination of five studies (Campbell,1953; Gage & Exline, 1953; Greer etal., 1954; Hites & Campbell, 1950;Trapp, 1955) subsequent to Chow-dhry and Newcomb's reveals a lowpositive relationship between leader-ship and accuracy, but fluctuations inthe magnitude of the association can-not be related to the relevance of theitems because no valid scale of rele-vance can be applied across studies.

One additional fact emerges fromthe research in this area. Groupmembers believe that their leadersare more aware of their opinions andfeelings than the nonleaders of the

group (Campbell, 1953; Sprunger,1949; Zeleny, 1939). In summary,there appears to be a low but clearlypositive relationship between inter-personal sensitivity and leadership.However, methodological and con-ceptual problems remain which canbe resolved only by future research.

Techniques of measurement. Lead-ership status has been measured in atleast four ways: by observer ratings,by peer ratings, by criterion meas-ures, and by self-ratings. The lattertechnique has been used only once inthese studies, but for the three re-maining techniques it is possible toask whether different results are ob-tained when different techniques areused.

Peer ratings and criterion measuresrest upon the estimates of an indi-vidual's peers. Peer ratings are es-sentially descriptions of an indi-vidual's present leadership status,whereas criterion measures reflect thegroup's selection for future leader-ship. The peer ratings are assess-ments of the informal leadershipstructure, whereas criterion measuresreflect the formal leadership struc-ture. Numerous studies have notedthat there is seldom complete cor-respondence between the designa-tions which emerge from these twoapproaches. Observer ratings meas-ure the present informal leadershipstructure of the group, but the evalu-ation is made by someone outside thegroup, in most cases by someone notpersonally involved in the future ofthe group, and, therefore, the ob-server is not implicitly locating him-self on the status hierarchy by theact of rating. Finally, the observeris a member of a unique species ofjudging humanity, a social scientist,with special training and perhapseven special criteria of leadership.

In the cases of intelligence, ad-

252 RICHARD D. MANN

justment, and extroversion vs. lead-ership, the number of results is largeenough to permit detailed analysis ofthe relationships in terms of the dif-ferent measuring techniques em-ployed. Table 2 shows the percentageof results which are positive when thethree techniques of measuring leader-ship are related to the personalityfactors of intelligence, adjustment,and extroversion. The base numbersare shown in parentheses.

The relationship between intelli-gence and leadership appears to bequite independent of the techniquesof measuring leadership. On theother hand, there is a striking differ-ence between the way adjustmentand extroversion are related to lead-ership as the technique of measure-ment varies. Adjustment is posi-tively related to peer ratings on lead-ership in 97% of the cases, while it ispositively related to criterion meas-ures in only 76% of the cases. Extro-version is not related to peer ratingsat all, but it is consistently related tocriterion measures. It appears thatan individual's adjustment is moreimportant in determining his infor-

TABLE 2

THE RELATIONSHIP BETWEEN PERSONALITYFACTORS AND LEADERSHIP USING THREE

DIFFERENT TECHNIQUES OF MEAS-URING LEADERSHIP

Percentage of ResultsPositive

Peer Criterion ObserverRatings Measures Ratings

Intelligence

Adjustment

Extroversion

91(66)

97(31)

50(30)

85(40)

76(87)

86(58)

89(69)

76(41)

70(35)

mal leadership status (peer ratings)than his formal leadership status (cri-terion measures). In contrast, extro-verted individuals are no more likelythan introverted individuals to berated as informal leaders by theirpeers, but they are quite likely to beselected as the formal leader for thefuture. Finally, scanning the thirdcolumn of the table, it may be notedthat intelligence is more consistentlyrelated to observer ratings thaneither adjustment or extroversion.

It does appear that different as-pects of an individual's status are be-ing measured by these techniques,and that these different aspects arenot uniformly related to his person-ality. This crude division of the op-erations into peer ratings, criterionmeasures, and observer ratings sug-gests at least two dimensions of pos-sible relevance. It may be importantto differentiate between descriptionsof present or informal leadership andchoices for future or formal leader-ship; to rate on leadership and toselect for leadership may engagequite different standards on the partof the group member. Secondly, itmay be critical to know more aboutwho is doing the judging, about thejudge's training and his involvementin the group and in the outcome ofthe rating process itself.

Summary. A number of relation-ships between an individual's per-sonality and his leadership status ingroups appear to be well established.The positive relationships of intelli-gence, adjustment, and extroversionto leadership are highly significant.In addition, dominance, masculinity,and interpersonal sensitivity arefound to be positively related toleadership, while conservatism isfound to be negatively related toleadership. Finally, evidence hasbeen presented to indicate that the


TABLE 3

THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND POPULARITY

PersonalityFactors

Intelligence

Adjustment

Extroversion

Dominance

Masculinity

Conservatism

Sensitivity

No.of

Stud-ies

13

18

13

6

4

11

11

No.of

Re-sults

38

78

46

9

8

18

38

Positive

Sig.(«)

6

15

9

0

0

3

6

N.S.(&)

23

34

13

5

5

7

16

Unt.M

0

s

9

0

0

1

0

Sig.(d)

1

0

1

2

0

2

1

Negative

N.S.w6

19

S

1

3

2

13

Unt.CO

0

0

0

0

0

0

0

Zero,?N.S.Or.*)

2

5

9

1

0

3

2

Positive

%81

(36)

74(73)

84(37)

63(8)

63(8)

73(IS)

61(36)

%ofSig. 0)

86(7)

100(15)

90(10)

0(2)

—(0)

60(S)

86(7)

% Sig. &in Dir.

of Trend(ft)

17(36)

21(73)

24(37)

0(9)

0(8)

IS(17)

16(38)

relationship between personality fac-tors and leadership varies with thetechnique of measuring leadership.

PopularityThe personality determinants of

individual popularity have receivedless attention than the determinantsof leadership. At the same time, how-ever, the importance of personalityfactors has been more or less as-sumed, and the situational approachto popularity is not well developed.While less is known about the actualconsistency with which an individualmaintains his popularity in differentgroups and across changing condi-tions, there is reason to believe thatpopularity, no less than leadership,may be profitably examined in termsof both personality and situationalfactors.

Table 3 presents a summary of therelationships between seven aspectsof personality and an individual'spopularity in groups. Since thistable is constructed in a mannerparallel to Table 1, no detailed ex-planation of its form will be given.

Intelligence. Thirteen studies(Bass, Wurster et al., 1953; Bonney,

Hoblit, & Dreyer, 1953; Borgatta,1953; Burks, 1937; Cronbach, 1950;Fiedler, Doyle, Jones, & Hutchins,1957; French & Mensh, 1948; Kelly,1957; Mill, 1953; Reilly, 1947; Riggs,1953; Shapiro, 1953; Slater, 1955a)have related an individual's intelli-gence to his popularity. An exam-ination of the 38 results shows that81% are positive and 86% of theseven significant results are positive;17% of the results are both positiveand significant. The maximum cor-relation obtained is .37, and themedian correlation is no higher than.10.

College grades are more stronglyrelated to popularity than any othermeasure of intelligence. In contrast,it may be remembered that gradeswere less strongly related to leader-ship than other measures of intelli-gence. In general, there appears tobe a tendency for intelligent indi-viduals to be more popular.

Adjustment. All of the 15 signifi-cant results relating an individual'spersonal adjustment to his popularityare in the positive direction, but whenthe insignificant and untested resultsare included, the proportion falls to

254 RICHARD D. MANN

74%. Although several of the cor-relations reported are over .50, themedian is close to .10 (Bass, Wursteret al., 1953; Bonney et al., 1953;Borgatta, 1953; Burks, 1938; Cat-tell, 1934; Cohen, 1954; Cronbach,1950; Fiedler et al., 1957; French &Mensh, 1948; Guthrie, 1956; Kelly,1957; Martin et al., 1952; Mill, 1952,1953; Shapiro, 1953; Slater, 1955a;Tagiuri, 1952).

No single measure of adjustment isconvincingly related to popularity,with one exception (Guthrie, 1956),an opinion survey designed to meas-ure "satisfactory personal habits."There is some indication in these datathat more popular persons are betteradjusted.

Extroversion. When the separateresults within each study are pooledand the trend over independentstudies is assessed, it is found that 11of the 12 trends are in the positivedirection. Further indication that ex-troversion is positively associatedwith popularity comes from the 46separate results; 84% of the resultsand 90% of the significant results arepositive.

The scales measuring extroversionon the 16 P. F. are better predictors ofpopularity than the correspondingscales on the Guilford-Zimmerman,and ratings on extroversion are morehighly related to popularity thaneither. The highly chosen, popularindividual emerges from these studiesas a sociable, surgent, and emotion-ally labile person (Bass, Wurster etal., 1953; Bonney et al., 1953; Bor-gatta, 1953; Burks, 1938; Cattell,1934; Cronbach, 1950; French &Mensh, 1948; Kelly, 1957; Lemann &Solomon, 1952; Mill, 1952, 1953;Shapiro, 1953; Slater, 1955a).

Dominance. On the basis of ninecontradictory results little can besaid about the relationship between

dominance and popularity. Thetrend is positive, but the two signifi-cant results are negative (Bass,Wurster et al., 1953; Bonney et al.,1953; Borgatta, 1953; Kelly, 1957;Lemann & Solomon, 1952; Shapiro,1953).

Masculinity-femininity. None ofthe attempts to relate masculinity topopularity have yielded significantresults, and the trend, though posi-tive, is weak (Bass & Wurster, 1953a;Mill, 1953; Shapiro, 1953; Slater,1955a).

Conservatism. Conservatism ispositively associated with popularityin 73% of the results. More popularindividuals tend to be more conserva-tive, conventional, or authoritarian(Bass, Wurster et al., 1953; Bonneyet al., 1953; French & Mensh, 1948;Hays, 1953; Kelly, 1957; Martin etal., 1952; Masling et al., 1955;Rohde, 1951; Shapiro, 1953; Slater,1955a).

Interpersonal sensitivity. The rela-tionship between empathy or inter-personal sensitivity and popularityhas been investigated in 11 studies(Ausubel, 1955; Ausubel & Schiff,1955; Gage & Exline, 1953; Greer etal., 1954; Lemann & Solomon, 1952;Nordlie, 1954; Norman, 1953; Singer,1951; Taylor, 1956; Trapp, 1955;Van Zelst, 1953), but the criticismsapplied to the studies of empathyand leadership are equally relevantto this body of research. Although61% of the results are positive, thissummary statistic must be examinedmore carefully.

If Campbell's (1955) criticism isvalid, it is improper to relate an indi-vidual's popularity to his awarenessof that popularity, using the latteras the measure of interpersonal sensi-tivity. Since the accuracy score isbased in part on the person's actualstatus, this produces spuriously posi-


tive correlations. Eliminating the re-sults to which Campbell's criticismwould be directed, the proportion ofpositive results among those remain-ing is only slightly over 50%. Untilthe possibility can be discounted thata number of spurious results are em-bedded in this body of data, thedirection of this relationship cannotbe estimated with safety.

Summary. Extroversion, intelli-gence, adjustment, and conserva-tism are found to be positively re-lated to popularity. The research todate, for various reasons, provides nodefinite answer to the question of howdominance, masculinity, and inter-personal sensitivity are related topopularity.

Total Activity Rate

Only three aspects of personalityhave been related to total activityrate a sufficient number of times towarrant their inclusion in this review.Table 4 presents a summary of therelationships of intelligence, adjust-ment, and extroversion to the indi-vidual's total activity.

Intelligence. The relationship be-tween intelligence and activity ratecould hardly be clearer. All 36 re-sults are positive, and one-third ofthe results are significant. Themedian correlation is between .15and .20, the highest correlation re-

ported is .34. The data leave littledoubt that the relationship betweenintelligence and total amount of par-ticipation, although of low magni-tude, is positive (Bass, 1951b; Bass,Wurster et al., 1953; Borgatta, 1953;Brown, 1950; Slater, 1955a; Zeleny,1939).

Adjustment. Roughly three-quar-ters of the total number of results in-dicate a positive relationship be-tween adjustment and total activityrate, but few of the results reach sig-nificance. With particular consist-ency, adjustment as measured by theMM PI is positively related to thetotal amount of an individual's par-ticipation (Bass, Wurster etal., 1953;Borgatta, 1953; Brown, 1950; Cer-vin, 1956, 1957; Slater, 1955a).

Extroversion. Measures of extro-version are positively related to totalactivity rate in 11 (or 79%) of the 14results; all four significant results arepositive. Two studies (Brown, 1950;Slater, 1955a) report a positive cor-relation between the Hypomaniascale of the MM PI which indicatesgreater maladjustment among highparticipants. Since other research(French, 1953) has shown that theHypomania scale measures both ex-troversion and maladjustment, theseresults at least suggest that extrover-sion may be more strongly related tototal activity rate than adjustment(Bass, Wurster et al., 1953; Bor-

TABLE 4THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND TOTAL ACTIVITY RATE

PersonalityFactors

Intelligence

Adjustment

Extroversion

No.of

Stud-ies

6

6

5

No.of

Re-sults

36

42

14

Positive

Sig.W

12

3

4

N.S.(»)

24

28

7

Unt.(«)

0

0

0

Negative

Sig.W)

0

1

0

N.S.to0

10

3

Unt.(ft0

0

0

Zero,?N.S.(«.*)

0

0

0

Positive

% % of(») Sig. W)

100 100(36) (12)

74 75(42) (4)

79 100

% Sig. &in Dir.

of Trend(fc)

33(36)

7(42)

29(14) (4) (14)

256 RICHARD D. MANN

gatta, 1953; Brown, 1950; Slater,1955a;Zeleny, 1939).

Summary. Of the three personalitymeasures used with any frequency,intelligence stands out as the per-sonality characteristic most conclu-sively related to activity rate. Extro-version and adjustment also seem tobear a positive relationship to ac-tivity rate.

Task ActivityTask activity includes measures of

the frequency with which an indi-vidual gives suggestions, opinions,and orientations and asks questions.It is necessary, therefore, to differen-tiate between task contribution andtask questions. Unfortunately, per-sonality variables have not been re-lated to the frequency of askingquestions a sufficient number oftimes to be included here. Therefore,this section deals exclusively withtask contributions.

There is room for doubt, however,whether task activity deserves to betreated independently of total ac-tivity, since correlations as high as.93 between the number of task con-tributions and the total number ofacts initiated have been reported(Borgatta, 1953). This is hardly sur-prising in view of the fact that insome studies two-thirds or more ofthe total number of acts are taskcontributions. There is no questionthat the operation for determiningthe number of task contributions isdistinct from the operation for meas-uring total activity. The issue iswhether, in the light of the high cor-relation between these two measures,results based upon task contribu-tions are not actually misleading.The implication is that one categoryof behavior, task activity, is mean-ingfully related to some personalityvariable. But if task activity plus

nontask activity is related to the per-sonality characteristic in a nearlyidentical fashion, what is the value ofthe categorization? Total activityrate accounts for both relationshipsmore parsimoniously.

Two researchers, Borgatta (1953)and Slater (1955a) were aware ofsuch difficulties. Arguing that anindividual's task-relevant behaviorshould be considered independentlyof his total activity rate, they meas-ured task activity by taking thepercentage of his total activity whichfell within the task contribution area.It is possible to use their data toexamine the relationship betweenpersonality characteristics and taskactivity, controlling for total ac-tivity rate. In fact, the best argu-ment for including a separate sec-tion in this review devoted to taskactivity is that Borgatta and Slater'spercentage data raise a separate is-sue. Their data provide an estimateof the relation between an individ-ual's personality characteristics andthe extent to which he concentrateshis activity in the task contributionarea.

Table 5 presents a summary of theresults relating personality to taskactivity. In the three relationshipsof intelligence, adjustment, and mas-culinity to task activity the resultsbased upon percentages are shownbeneath the results based upon theraw numbers of task contributions.

Intelligence. Whereas 80% of theresults relating intelligence to theraw number of task contributions arein the positive direction, only 23% ofthe results relating intelligence to thepercentage of task contributions arepositive. Apparently, the findingbased upon raw numbers is highly de-pendent on the correlation betweenthe raw number of task contributionsand the total amount of activity. As

PERSONALITY AND PERFORMANCE IN SMALL GROUPS

TABLE 5

THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND TASK ACTIVITY

257

PersonalityFactors

Intelligenceraw

%

Adjustmentraw

%

Extroversion

Dominance

Masculinityraw

%

Conservatism

No.of

Stud-ies

4

2

4

2

4

3

1

1

4

No.of

Re-sults

45

13

19

20

19

8

21

3

12

Positive

Sig.(a)

13

0

7

1

6

'3

5

0

3

N.S.0)

23

3

10

7

9

3

12

1

3

Unt.W

0

0

0

0

0

0

0

0

0

ft10

0

0

0

0

2

0

2

Negative

N.S.

8

10

2

12

3

2

2

2

2

Unt.

0

0

0

0

0

0

0

0

0

Zero,?N.S.(S.h)

0

0

0

0

1

0

0

0

2

Positive

w80

(45)23

(13)

89(19)40

(20)

83(18)

75(8)

81(21)33(3)

60(10)

%ofSig. 0)

93(14)

(0)

100(7)

100(1)

100(6)

100(3)

71(7)

(0)

60( 5)

% Sig. &in Dir.

of Trend(ft)

29(45)

0(13)

37(19)

0(20)

32(19)

38(8)

24(21)

0(3)

30(10)

noted above, the higher the indi-vidual's intelligence, the more likelyhe is to be a high participator. Sincethe total number of acts initiated andthe number of task contributions ini-tiated are almost perfectly correlated,a positive association between intelli-gence and the raw number of taskcontributions could have been pre-dicted. However, the negative rela-tionship between intelligence andconcentration of activity in the taskarea was unexpected; 10 out of 13correlations are in the negative direc-tion. It appears that although in-telligent individuals talk more thanless intelligent individuals, they con-centrate less of their total activity inthe area of task contributions (Bor-gatta, 1953; Carter & Nixon, 1949;Cattell & Stice, 1954; Miller, 1939;Slater, 1955a).

Adjustment. The relationship be-tween personal adjustment and taskactivity depends upon the contribu-tion of total activity rate to the re-sults. When the raw number of task

contributions is employed, the rela-tionship is positive; when task ac-tivity is measured in terms of the per-centage of total activity, the relation-ship fails to hold. Actually, thetrend of the results is slightly nega-tive, but the only significant result ispositive. When the factor of totalactivity rate is controlled, the strongpositive relationship between adjust-ment and task activity is reduced toa low negative relationship (Bor-gatta, 1953; Carter & Nixon, 1949;Cattell & Stice, 1954; Miller, 1939;Slater, 1955a).

Extroversion. Extroversion is posi-tively related to the raw number oftask contributions, but it is not possi-ble to partial out the total activityfactor underlying these results. Itmay at least be suspected that therelationship would be altered if ex-troversion were related to task ac-tivity, holding activity rate constantthrough statistical controls (Bor-gatta, 1953; Carter & Nixon, 1949;Cattell & Stice, 1954; Miller, 1939).

258 RICHARD D. MANN

Dominance. The studies reviewedcontain only eight results bearing onthe relationship between dominanceand the number of task contribu-tions. Six of the results are positiveand the three significant results arepositive (Borgatta, 1953; Carter &Nixon, 1949; Cattell & Stice, 1954).

Masculinity-femininity. Masculin-ity has been related to task activity inonly two studies (Carter & Nixon,1949; Slater, 1955a). The one studywhich related masculinity to the rawnumber of task contributions found81% of the results to be positive. Onthe other hand, the study employingpercentages of total activity as themeasure of task contributions foundonly one out of three results to bepositive. The extent to which mas-culinity relates to an individual'stendency to concentrate his activityin the task area remains largely un-known.

Conservatism. Four studies (Carter& Nixon, 1949; Cattell & Stice, 1954;Haythorn et al., 1956a, 1956b) haveexamined the relation between con-servatism and the amount of task ac-tivity initiated. A slight positivetrend emerges from the data, indicat-ing that conservative or authoritar-ian individuals tend to give more taskcontributions than nonauthoritarianindividuals. However, there are toofew results to establish this trend assignificant.

Summary. A serious difficulty un-derlies the attempts to relate per-sonality variables to task activitywhen the latter is measured in termsof the raw number of task contribu-tions. The results are not independ-ent of the relationship between per-sonality variables and total activityrate. Adjustment, extroversion, mas-culinity, intelligence, dominance, andconservatism are all found to be posi-tively related to the raw number of

task contributions, but the relation-ships are reversed in the three caseswhere it is possible to control for totalactivity rate by using percentages.Intelligence, adjustment, and mascu-linity are negatively related to theproportion of a man's total activitywhich falls within the area of taskcontributions. It must be admittedthat the reversals are not uniformlyconvincing; the pattern of results foradjustment vs. task activity is mixed,and there are only three results in thecase of masculinity. On the otherhand, for all their faults, these threereversals succeed in raising the ques-tion of whether personality variablesmay not relate one way to task ac-tivity when measures of it are con-founded with the general activityfactor and quite another way to taskactivity when this confounding factoris removed.

Social-Emotional ActivityThere are two general categories of

social-emotional activity. Positivesocial-emotional activity includesshowing agreement, tension release,and solidarity; negative social-emo-tional activity includes showing dis-agreement, tension, and antagonism.They are treated separately in thissection.

Two aspects of personality, intelli-gence and adjustment, have been re-lated to social-emotional activity asufficient number of times to warrantdetailed analysis. Fortunately, theessential difficulty with the results ontask activity, the confounding of thetotal activity factor with the resultsfor a segment of an individual's totalbehavior, is not a problem here. Onthe one hand, the correlation be-tween the total number of positive ornegative social-emotional acts andthe total number of acts initiated islow. On the other hand, Borgatta


(1953) and Slater (1955a) have re-lated intelligence and adjustment tothe proportion of a man's total ac-tivity which falls within the positiveor negative social-emotional areas.In addition, Borgatta has related in-telligence to the raw number of posi-tive and negative social-emotionalacts. Table 6 presents a summary ofthe relationships of intelligence topositive and negative social-emo-tional activity, measured in terms ofraw amounts and percentages, andthen the relationship of adjustmentto positive and negative social-emo-tional activity, measured only interms of percentages of the totalamount of activity.

Intelligence. Despite the low num-ber of results, there is a trend emerg-ing from the data. Intelligence meas-ures are positively related to boththe total number of positive social-emotional acts and the percentage oftotal activity falling in this area. Onthe other hand, intelligence is nega-tively related to the two correspond-ing measures of negative social-emo-tional activity. Controlling for totalactivity by the use of percentagesdoes not disturb the trends. In com-

parison with less intelligent groupmembers, the more intelligent indi-viduals appear to concentrate moreof their behavior in the area of posi-tive social-emotional activity and lessin the area of negative social-emo-tional activity (Borgatta, 1954;Slater, 1955a).

Adjustment. The individual's per-sonal adjustment is positively relatedto the proportion of his total activitywhich is rewarding or supportive.The trend is not very strong, withonly 59% of the results in the positivedirection, but when contrasted withthe relationship between adjustmentand negative social-emotional ac-tivity, the pattern is interesting. Ad-justment is negatively related to theproportion of a man's total activityin the negative social-emotional area.On the basis of these data, it appearsthat the better adjusted the individ-ual, the more likely he is to initiatepositive social-emotional acts and theless likely he is to initiate negativesocial-emotional acts (Borgatta,1954; Slater, 1955a).

Summary. Social-emotional ac-tivity has received less attentionfrom researchers than any other as-

TABLE 6THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND SOCIAL-EMOTIONAL ACTIVITY

PersonalityFactors

Intelligencepositive

raw

%

negativeraw

%

Adjustmentpositive %

negative %

No.of

Stud-ies

1

2

1

2

2

2

No.of

Re-suits

13

13

13

13

22

22

Positive

Sig.(a)

2

0

0

0

10

N.S.(»)

10

11

6

4

12

7

Unt.(c)

0

0

0

0

0

0

Negative

Sig. N.S.W (e)

0 1

0 2

0 7

0 9

0 9

0 15

Unt.(/)

0

0

0

0

0

0

Zero,?N.S.(g, h)

0

0

0

0

0

0

Positive

I

92(13)85

(13)

46(13)31

(13)

59(22)32

(22)

%ofSig. 0)

100(2)

(0)

(0)

(0)

100(1)(0)

% Sig. &in Dir.

of Trend(W

15(13)

0(13)

0(13)

0(13)

5(22)

0(22)

260 RICHARD D. MANN

pect of an individual's performancecovered in this review. The scantyevidence available indicates that themore intelligent or the better ad-justed the individual, the morelikely he is to concentrate his ac-tivity in the positive social-emotionalarea, and the less likely he is to con-centrate his activity in the negativesocial-emotional area.

ConformityBeginning with Asch's (1951) in-

genious experiment on conformityand his suggestions about possiblepersonality differences between inde-pendent and yielding subjects, anumber of researchers have been con-cerned with the problem of relatingan individual's personality to histendency to conform to the opinionsof others. One special problem arisesin reviewing the results in this area; aconsiderable number of the resultsdepend upon personality measure-ments which ask the individual todescribe himself. While other sec-tions of this review contain resultsbased upon self-ratings, those resultshave not created any difficulty. Inthe first place, they have always beenrelatively few in number, and, in the

second place, the trends of results de-pending upon self-ratings have beenin close agreement with the trendsbased upon other techniques of meas-urement. However, self-ratings onadjustment and extroversion do notrelate to conformity in the same wayas peer ratings, questionnaires, andobjective tests. The trends for re-sults based upon self-ratings must betreated separately in order to obtaina valid summary of the findings. Afurther complication, introduced bythe use of adjective check-lists forself-rating, is the tendency forauthors, under pressure to remainbrief, to report only the adjectiveswhich differentiate conformers fromnonconformers at some specifiedlevel of significance. As a result, anyestimate of the significance of thefindings is inflated. Moreover, it isnot possible in the case of those ad-jectives which do not yield significantresults to assess the direction of therelationship. Only the trends relatingdominance and conservatism to con-formity are sufficiently free of resultsbased upon self-ratings to escapethese criticisms.

Table 7 presents a summary of therelationships of adjustment, extro-

TABLE 7THE RELATIONSHIP BETWEEN PERSONALITY FACTORS AND CONFORMITY

No.Personality of

Factors Stud-ies

Adjustmentself-rating

other techn.

Extroversionself-rating

other techn.

Dominance

Conservatism

2

8

2

5

4

6

No.of

Re-sults

18

30

16

10

8

20

Positive

Sig.(a)

13

2

10

0

0

16

N.S.(&)

1

1

1

0

0

3

Unt.(«)

0

2

0

2

2

0

Negative

Sig.(d)

1

4

1

1

2

0

N.S.(«)

3

1

4

1

1

1

Unt.(/)

0

6

0

2

2

0

Zero,?N.S.(*.*)

0

14

0

4

1

0

Positive

%(0

78(18)31

(16)

69(16)33(6)

29(7)

95

%ofSig. W)

93(14)33(6)

91(11)

0(1)

0(2)

100

% Sig. &in Dir.

of Trend(t>

72(18)17

(24)

62(16)13(8)

33(6)

80(20)


version, dominance, and conserva-tism to conforming behavior. Forthe first two relationships the resultsfor self-rating techniques are pre-sented separately from those usingother techniques of personality as-sessment.

Adjustment. Those individualswho tend to conform to group opin-ion also tend to see themselves asbetter adjusted, as indicated by thefact that 78% of the results are posi-tive. However, peer ratings and per-sonality inventories do not yield thesame results; only 31% of the resultsrelating adjustment, as measured bytechniques other than self-ratings, toconformity are positive. One way toresolve the dilemma is to assume thatself-ratings do not measure personal-ity as validly as the other measures.It may be, for example, that those in-dividuals who tend to conform togroup opinion also tend to conformto an acceptable personality charac-terization in their self-descriptions.If only the results based upon tech-niques other than self-ratings wereconsidered, it would be possible toconclude that well-adjusted indi-viduals are less likely to conform tothe opinions of others (Barren, 1953;Bray, 1950; Cervin, 1955, 1957;Hardy, 1954; Hollander, 1954; Kagan& Mussen, 1956; Kelman, 1950).

Extroversion-introversion. Extro-versionispositivelyrelatedtoconform-ity in 78% of the results employingself-ratings; those who conform tothe opinion of others describe them-selves as kind, friendly, helpful, andoptimistic. However, the results em-ploying projective and personalityinventory variables do not confirmthis relationship. While there is someevidence that extroversion is nega-tively related to conformity, a moreaccurate summary would be that thefew relevant findings are inconclu-

sive (Barren, 1953; Bray, 1950;Hardy, 1954; Hoffman, 1953; Kel-man, 1950).

Dominance. Although only eightfindings bear on the relationship be-tween dominance and conformity,the trend is negative. Only one ofthese results is based upon self-rat-ings. It might be concluded thatthose who yield to group pressuresare less dominant individuals (Bar-ron, 1953; Bray, 1950; Hoffman,1953; Kelman, 1950).

Conservatism. There has been con-siderable speculation that there is apositive association between con-servatism and conforming behavior.The burden of evidence suggests thatconservative, conventional, and au-thoritarian subjects are more likelyto yield to group pressure than radi-cal or unconventional subjects. Nosingle measure of the conservatismdimension emerges as an especiallypotent predictor of conformity in allconditions; in fact, there is a sugges-tion that it is important to controlfor a number of conditions if the rela-tionship is to hold at all. There issome indication in these data thatauthoritarian subjects are less likelyto conform to a small group of peersbut more likely to conform to eithera large group of peers or perceivedsuperiors (Barren, 1953; Bray, 1950;Cervin, 1955, 1957; Hardy, 1954;Hollander, 1954; Kagan & Mussen,1956; Kelman, 1950).

Summary. Those who are morelikely to conform to group opinionsee themselves as better adjusted andmore extroverted. Only in the pre-diction of conformity do personalitymeasures other than self-ratings con-tradict the results based on self-description. By such measures asratings by others, projective tests,and personality inventories, adjust-ment and extroversion are nega-

262 RICHARD D. MANN

tively related to conformity. In addi-tion, there is a slight indication thatdominance is negatively related toconformity. The relationship be-tween conservatism and conformityhas received considerable attention,and the data tend to confirm thehypothesis that conservative indi-viduals are more likely to conform tothe opinions of others.

Situational Factors

A number of relationships betweenpersonality variables and individualbehavior and status variables appearto be well established. It is possibleto extend the analysis one step fur-ther and inquire whether the magni-tude of these relationships variesunder certain conditions. This sec-tion examines the extent to whichsituational factors affect the relation-ships between personality character-istics and performance in groups, inan effort to understand the limitingconditions within which the relation-ships operate.

A situational factor represents acondition of research about whichsome decision must be made, butwhich, once the decision is made,may affect the generality of the re-sults. Four examples, for which ade-quate data are available in thesestudies, are selected for analysis inthis section: (a) the nature of the pop-ulation from which the sample isdrawn; (6) the sex of the group mem-bers; (c) the history of the groupprior to observation; and (d) the sizeof the group. The first example con-trasts four populations; these arehigh school students, college under-graduates, military personnel, andall other adults. The second con-trasts groups composed entirely ofmales with those entirely of females;mixed groups are omitted from theanalysis of situational factors. Thethird example compares ad hoc ex-

perimental groups with natural, on-going groups. In the majority ofcases, the former are composed ofpersons unknown to each other be-fore the period of observed interac-tion, and usually the group existsentirely for the purposes of research.In contrast, natural groups are com-posed of acquaintances and exist formany reasons other than to bestudied. The fourth example com-pares groups made up of seven orfewer members with larger groups ofeight or more. In treating each situa-tional factor the effort will be to de-termine whether the relationship be-tween the individual's personalityand his performance in the group re-mains constant across the variousconditions.

There are sufficient data to studyonly the relationships of intelligence,adjustment, and extroversion to onedependent variable, leadership. Com-parisons are made between conditionsof research by determining, for eachrelationship, the percentage of re-sults within each study which arepositive; these percentages are thenaveraged over studies employing thesame conditions of research. The ad-vantage of averaging the proportionwithin separate studies is that eachstudy is thus weighted equally in thefinal statistic, the percentage of posi-tive results. It may be rememberedthat intelligence, adjustment, and ex-troversion all bear a strong positiverelationship to leadership; the ques-tion is whether the strength of theserelationships varies with the condi-tions of research under which theyare obtained.

The nature of the population. Overtwo-thirds of studies relating intelli-gence, adjustment, or extroversionto leadership draw their groups fromthe college population. One conse-quence of such a sampling bias is thatrelatively little data can be found to


determine if, as a consequence, a dis-torted view of the relationships in-volved has been created. There ap-pears to be little variation in the rela-tion of intelligence to leadershipacross populations; the proportion ofpositive results in military groups is89% and in adult groups, 100%.However, adjustment is morestrongly related to leadership inundergraduate and adult groups thanin high school and military groups,the proportions varying from nearly90% in the former to nearly 70% inthe latter. These differences may re-flect the influence of a number ofunderlying factors; for example, itmay be that as the age, education, orsocial class of the group membersincreases, adjustment is more stronglyrelated to leadership. Finally, extro-version is less strongly related toleadership among high school stud-ents (61% positive results) thanamong the three remaining popula-tions (over 80% positive results).

The sex of the group. The numberof studies which examine all-malegroups is almost identical with thenumber examining all-female groups.There is little difference between maleand female groups in the way intelli-gence relates to leadership; 89% ofthe results are positive in the malegroups and 86% are positive in thefemale groups. Relating adjustmentto leadership, 83% of the results arepositive in male groups, 62% arepositive in female groups. In con-trast, 69% of the results are positivewhen extroversion is related to lead-ership in male groups, while 85% arepositive in female groups. Thus, ad-justment is more positively relatedto leadership in male groups than infemale groups, whereas extroversionis more positively related to leader-ship in female groups.

History of the group. When groupswhich have interacted prior to ob-

servation are compared with groupswithout prior interaction, no differ-ences of any magnitude emerge in therelationships of intelligence, adjust-ment, and extroversion to leadership.The differences between the percent-ages of positive results in experi-mental and natural groups do not ex-ceed 4%. Apparently the way thesethree aspects of personality relate toleadership status does not vary as aresult of studying either experi-mental or natural groups.

Size of the group. In studying theeffect of variations in group size, thenecessity of holding constant an-other condition of research, thehistory of the group, requires thatonly experimental groups be consid-ered here since in these studies nat-ural groups are in every case largerthan experimental groups. The ex-perimental groups range in size fromthree to 10 members. Dividing themat the median size, seven and one-half, it appears that size alone doesnot strongly affect the relationships.However, the slight trend is intri-guing. Intelligence is more stronglyrelated to leadership in smaller groupsthan in larger groups, while adjust-ment is more strongly related to lead-ership in larger groups. It may behypothesized that, at least within therange of group size considered here,as the size of the group increases, in-ternal and integrative problems be-come more important, relative to theexternal or adaptive problems of thegroup. It may be that the need foran integrative leader in the largergroups makes it more likely that awell-adjusted individual will be se-lected, whereas the greater need foran adaptive leader in the smallergroups makes the intelligent indi-vidual more likely to be chosen.These differences are only matters ofdegree, since the relationships are pos-itive in both large and small groups.

264 RICHARD D. MANN

Summary. Four decisions whichmust be made before any researchcan be undertaken concern the na-ture of the population, the sex of thegroup members, the previous his-tory and the size of the groups. Ofthese, only the previous history ap-pears to make no difference in theway intelligence, adjustment, and ex-troversion are related to leadershipstatus. To a slight extent, intelli-gence is more positively related toleadership in smaller experimentalgroups than in larger ones. Adjust-ment is more positively related toleadership in undergraduate andadult groups, in male groups, and inlarger groups, while it is less stronglyrelated to leadership in high schooland military groups, in femalegroups, and in smaller groups. Extro-version is less strongly related toleadership in high school groups thanin college, military, or adult groups,and it is more strongly related toleadership in female groups than inmale groups. However, two pointsshould be emphasized. The strongpositive relationships between thesethree aspects of personality and lead-ership status are not reversed underany variation in the conditions of re-search; the differences are only amatter of degree and many of the dif-ferences are slight.

Summary and ConclusionThis review has examined a num-

ber of relationships between the per-sonality characteristics of the indi-vidual and the way he behaves or isperceived in groups. Seven aspects ofpersonality were selected for study;all but one were chosen on thegrounds that factor analytic studiesof personality had repeatedly demon-strated their importance. Roughly350 out of over 500 different person-ality variables were then categorized

as measures of the seven dimensions.Six aspects of the behavior and statusof the individual were selected, pri-marily on the basis of the labels andoperations of the measures. The rele-vant findings on the 29 relationshipsfor which adequate data were avail-able were then examined. Finally, anexamination was made of the effect offour situational factors or conditionsof research on three of the relation-ships.

Any attempt to evaluate the con-clusiveness of this review shouldtake a number of considerations intoaccount. To the extent that contra-dictory or insignificant findings havebeen either overlooked by the re-viewer or unpublished by the re-searcher, the trends may be inaccu-rate. A considerable number of un-published data are included in thisreview, but the impact of the remain-ing unpublished data cannot beknown. Secondly, the selection offactors and especially the location ofvariables within the factors involveddecisions which future research mayprove ill-advised. Every effort wasmade to use evidence other than theauthor's original label for a variable,but the present state of knowledge inthe field of personality assessmentleaves much to be desired. Finally,the decision to examine the totalpool of results for each relationshipinvolved a risk that the statisticalinterdependence of the measureswould bias the results. Whether theproportion of results which are posi-tive for any relationship would in-crease or decrease if based upon inde-pendent samples cannot be deter-mined.

The best predictor of an individ-ual's performance in groups is intelli-gence. In order of the proportion ofpositive results, intelligence is foundto be positively related to total ac-


tivity rate, leadership, and popu-larity. In addition, it is positively re-lated to the number of task contribu-tions made by an individual, but,controlling for the total activity rate,it is negatively related to the propor-tion of his total activity falling in thetask contribution area. Intelligenceis positively related to both theamount and proportion of positivesocial-emotional activity, whereas itis negatively related to negativesocial-emotional activity.

Adjustment is found to be posi-tively related to leadership, popu-larity, and total activity rate, in thatorder. It is positively related to thetotal number of task contributions,but negatively related to the per-centage of the total number of actswhich are task contributions. Ad-justment is positively related to posi-tive social-emotional activity andnegatively related to negative social-emotional activity. Although sub-jects who tend to conform to groupopinion see themselves as better ad-justed, other measuring techniquesindicate a negative relationship be-tween adjustment and conformity.Except for the fact that intelligencehas not been related to conformity asufficient number of times to be re-viewed, adjustment is related to be-havior and status variables in muchthe same way as intelligence.

Extroversion is positively relatedto popularity, total activity rate, andleadership. Although it is positivelyrelated to the total number of taskcontributions, this result cannot beconsidered as independent of the re-lation between extroversion and totalactivity rate. While individuals whoconform more than others to groupopinion tend to see themselves asmore extroverted, techniques otherthan self-ratings fail to show any sig-nificant association.

Dominance is positively related tothe total number of task contribu-tions initiated and to leadership. Itis negatively related to an indi-vidual's tendency to conform togroup opinion.

Masculinity bears a low positiverelationship to leadership and popu-larity. A positive association is foundbetween masculinity and the totalnumber of task contributions, al-though there is some slight indica-tion that this relationship is reversedwhen the total activity factor is con-trolled.

Conservatism is negatively relatedto leadership, but positively relatedto popularity. A positive associationis found between conservatism andthe raw number of task contribu-tions. Finally, conservatism is posi-tively related to conformity; thosewho tend to conform to the opinionsof others are more conservative orauthoritarian.

The measures of interpersonalsensitivity relate positively to bothleadership and popularity. Both rela-tionships are of low magnitude, and,in the case of the latter, there is apossibility that the number of the re-sults which are spuriously positive issufficient to cast doubt on the wholetrend.

In reviewing the results for theseven aspects of personality, refer-ence has been made to the percentageof the total number of results whichare positive. In the majority of cases,when only the significant results areexamined, the strength of the trendincreases. Moreover, several rela-tionships have been examined in asufficient number of studies to permitthe use of the trend for a single studyas the unit of research; in each casethe proportion of studies yieldingpositive trends was significantly dif-ferent from chance.

266 RICHARD D. MANN

A number of conditions of researchare found to influence the relation-ship of intelligence, adjustment, andextroversion to leadership. Whendifferent techniques of measuringleadership, different populations, andgroups of different sizes are used inthe research, the relationships are al-tered. It appears to make no differ-ence to the relationship obtainedwhether the group members wereacquainted with one another beforethe period of observation and meas-urement. It should be noted that inno case was the positive relationshipof intelligence, adjustment, and ex-troversion to leadership reversed indirection.

Throughout this review the resultshave been used to examine the direc-tion of the various associations be-

tween personality characteristics andmeasures of behavior or status. Occa-sionally, however, the results con-tained a sufficient number of correla-tions to afford some estimate of themagnitude of the relationship. In nocase is the median correlation be-tween an aspect of personality cov-ered here and performance higherthan .25, and most of the median cor-relations are closer to .15.

In conclusion, it may be noted thatthe relationships reviewed are by nomeans the only ones to which atten-tion has been and should be directed.It is encouraging to note, however,that many clear and significanttrends emerge when the body of re-search on these relationships is con-sidered as a whole.

REFERENCESARBOUS, A. G., & MAREE, J. Contribution

of two group discussion techniques to a vali-dated test battery. Occup. Psychol., 1951,25, 73-89.

ASCH, S. E. Effects of group pressure upon themodification and distortion of judgments.In H. Guetzkow (Ed.), Groups, leadership,and men. Pittsburgh: Carnegie Press, 1951.Pp. 177-190.

AUSUBEL, D. P. Sociempathy as a function ofsociometric status in an adolescent group.Hum. Relat., 1955, 8, 75-84.

AUSUBEL, D. P., & SCHIFF, H. M. Some intra-personal and interpersonal determinants ofindividual differences in sociempathic abil-ity among adolescents. /. soc. Psychol.,1955, 41, 39-56.

BALES, R. F. Interaction process analysis:A method for the study of small groups. Cam-bridge, Mass.: Addison-Wesley, 1950.

BARRON, F. Some personality correlates of in-dependence of judgment. /. Pers., 1953,21, 287-297.

BASS, B. M. Situational tests: I. Individualinterviews compared with leaderless groupdiscussions. Educ. psychol. Measmt, 1951,11, 67-75. (a)

BASS, B. M. Situational tests: II. Leaderlessgroup discussion variables. Educ. psychol.Measmt, 1951, 11, 196-207. (b)

BASS, B. M. The leaderless group discussion.Psychol. Bull., 1954, 51, 465-492.

BASS, B. M., & COAXES, C. H. Situational and

personality factors in leadership in ROTC.Unpublished manuscript, Louisiana StateUniver., 1952.

BASS, B. M., GAIER, E. L., & FLINT, A. W.Attempted leadership as a function of mo-tivation interacting with amount of con-trol. Baton Rouge: Louisiana State Univer.,1956. (Tech. Rep. 7, Contract N7 ONR35609).

BASS, B. M., McGEHEE, C. R., HAWKINS,W. C., YOUNG, P. C., & GEBEL, A. S. Per-sonality variables related to leaderlessgroup discussion behavior. /. abnorm. soc.Psychol., 1953, 48, 120-128.

BASS, B. M., & WURSTER, C. R. Effects ofcompany rank on LGD performance of oilrefinery supervisors. /. appl. Psychol.,1953, 37, 100-104. (a)

BASS, B. M., & WURSTER, C. R. Effects ofthe nature of the problem on LGD perform-ance. /. appl. Psychol, 1953, 37, 96-99. (b)

BASS, B. M., WURSTER, C. R., DOLL, P. A.,& CLAIR, D. J. Situational and personalityfactors in leadership among sorority women.Psychol. Monogr., 1953, 67, No. 16 (WholeNo. 366).

BAXTER, B. A. Hope of success and fear offailure as aspects of need for affiliation. Un-published honors thesis, Univer. Michigan,1957.

BELL, G. B. The relationship between leader-ship and empathy. Unpublished doctoraldissertation, Northwestern Univer., 1952.


BELL, G. B., & HALL, H. E. The relationshipbetween leadership and empathy. J. abnorm,soc. Psychol., 1954, 49, 156-157.

BONNEY, M. E., HOBLIT, R. E., & DREYER,A. H. A study of some factors related tosociometric status in a men's dormitory.Sociometry, 1953, 16, 287-301.

BORGATTA, E. F. An analysis of three levelsof response: An approach to some relation-ships among dimensions of personality.Sociometry, 1951, 14, 267-316.

BORGATTA, E. F. Project I. Unpublishedmanuscript, Harvard Univer., 1953.

BORGATTA, E. F. Analysis of social interactionand sociometric perception. Sociometry,1954, 17, 7-31.

BORGATTA, E. F., & ESCHENBACH, A. E.Factor analysis of Rorschach variables andbehavior observations. Psychol. Rep., 1955,1, 129-136.

BOWDEN, A. 0. A study of the personality ofstudent leaders in colleges in the Unitedstates. J. abnorm. soc. Psychol., 1926, 21,149-160.

BRAY, D. W. The prediction of behavior fromtwo attitude scales. J. abnorm. soc. Psychol.,1950, 45, 64-84.

BROWN, I. S. Training university students ingroup development: An experiment. Un-published doctoral dissertation, Univer.California, 1950.

BURDICK, H. A. The relationship of attrac-tion, need achievement and certainty toconformity under conditions of a simulatedgroup atmosphere. Unpublished doctoraldissertation, Univer. Michigan, 1956.

BURKS, F. W. The relation of social intelli-gence test scores to ratings of social traits./. soc. Psychol., 1937, 8,146-153.

BURKS, F. W. Some factors relating to socialsuccess in college. /. soc. Psychol., 1938, 9,125-140.

CAMPBELL, D. T. A study of leadership amongsubmarine officers. Columbus, Ohio: OhioState Univer. Res. Found., 1953.

CAMPBELL, D. T. An error in some demonstra-tions of the superior social perceptiveness ofleaders. J. abnorm. soc. Psychol., 1955, 51,694-695.

CARTER, L. F., & NIXON, M. Ability, percep-tual, personality, and interest factors asso-ciated with different criteria of leadership.J. Psychol., 1949, 27, 377-388.

CATTELL, R. B. Friends and enemies: A psy-chological study of character and tempera-ment. Charact. & Pers., 1934, 3, 55-63.

CATTELL, R. B. Description and measurementof personality. Yonkers-on-Hudson, N. Y.:World Book, 1946.

CATTELL, R. B. Second-order personalityfactors in the questionnaire realm. J. con-

sult, Psychol, 1956, 20, 411-418.CATTELL, R. B. Personality and motivation

structure and measurement. Yonkers-on-Hudson, N. Y.: World Book, 1957.

CATTELL, R. B., SAUNDERS, D. R., & STICE,G. F. The 16 P.P. Questionnaire. Cham-paign, 111.: Inst. of Pers. and Ability Test-ing, 1951.

CATTELL, R. B., & STICE, G. F. Four formulaefor selecting leaders on the basis of per-sonality. Hum. Relat., 1954, 7, 493-507.

CERVIN, V. Experimental investigation ofbehaviour in social situations: I. Behaviourunder opposition. Canad. J. Psychol., 1955,9, 107-116.

CERVIN, V. Individual behavior in socialsituations: Its relation to anxiety, neuroti-cism, and group solidarity. /. exp. Psychol.,1956, 51, 161-168.

CERVIN, V. Relationship of ascendent-sub-missive behavior in dyadic groups of humansubjects to their emotional responsiveness./. abnorm. soc. Psychol., 1957, 54, 241-249.

CHOWDHRY, K. Leaders and their ability toevaluate group opinions. Unpublished doc-toral dissertation, Univer. Michigan, 1948.

CHOWDHRY, K., & NEWCOMB, T. M. The rel-ative abilities of leaders and nonleaders toestimate opinions of their own groups. J.abnorm. soc. Psychol., 1952, 47, 51-57.

COATES, C. H. Forecasting officer potentialusing the leaderless group discussion. Un-published master's thesis, Louisiana StateUniver., 1952.

COBB, K. Measuring leadership in collegewomen by free association. J. abnorm. soc.Psychol, 1952, 47, 126-128.

COHEN, A. R. Personality and sociometricchoice. Unpublished manuscript, Univer.Michigan, 1954.

COHEN, A. R. Experimental effects of ego-defense preference on interpersonal rela-tions. J. abnorm. soc. Psychol, 1956, 52,19-27.

COWLEY, W. H. Three distinctions in thestudy of leaders. J. abnorm. soc. Psychol.,1928, 23, 144-157.

COWLEY, W. H. The traits of face-to-faceleaders. J. abnorm. soc. Psychol, 1931, 26,304-313.

CRONBACH, L. J. Studies of the group Ror-schach in relation to success in the collegeof the University of Chicago. J. educ. Psy-chol, 1950, 41, 65-82.

CROWELL, L., KATCHER, A., & MIYAMOTO,S. F. Self-concepts of communicator skilland performance in small group discussions.Speech Monogr., 1955, 22, 20-27.

CRUTCHFIELD, R. S. Assessment of personsthrough a quasi group-interaction technique.J. abnorm. soc. Psychol, 1951, 46, 577-588.

268 RICHARD D. MANN

DEXTER, E. S., & STEIN, B. The measurementof leadership in white and negro womenstudents. /. abnorm. soc. Psychol., 1955,SI, 219-221.

DUNKERLY, M. D. A statistical study ofleadership among college women. Stud.Psychol. Sf PsycUat., 1940, 4,1-65.

ESCHENBACH, A. E, The relationship of basicRorschach scoring categories to observedthree-man-group interaction behavior. Un-published doctoral dissertation, Univer.Florida, 1955.

ESCHENBACH, A. E., & BORGATTA, E. F.Testing behavior hypotheses with theRorschach: An exploration in validation.J. consult. Psychol., 1955, 19, 267-273.

EYSENCK, H. J. The structure of human per-sonality. London: Methuen, 1953.

FIEDLER, F. E. Assumed similarity measuresas predictors of team effectiveness in sur-veying. Urbana: Univer. of 111., Coll. ofEduc., 1953. (Tech. Rep. No. 6, ContractN60RI-07135).

FIEDLER, F. E. Assumed similarity measuresas predictors of team effectiveness. J. ab-norm. soc. Psychol, 1954,49, 381-388.

FIEDLER, F. E., DOYLE, J. S., JONES, R. E., &HUTCHINS, E. B. The measurement of per-sonality adjustment and personality changein non-clinical populations. Urbana: GroupEffectiveness Res. Lab., Univer. Illinois,1957. (Interim Tech. Rep. No. 5, ContractDA-49-007-MD-569.)

FIEDLER, F. E., HARTMANN, W., & RUDIN,S. A. The relationship of interpersonal per-ception to effectiveness in basketball teams.Urbana: Univer. Illinois, Coll. of Educ.,1952. (Tech. Rep. No. 3, Contract N60RI-07135.)

FIEDLER, F. E., WARRINGTON, W. G., &BLAISDELL, F. J. Unconscious attitudes ascorrelates of sociometric choice in a socialgroup. /. abnorm. soc. Psychol., 1952, 47,790-796.

FLEMMING, E. G. A factor analysis of the per-sonality of high school leaders. /. appl.Psychol, 1935, 19, 596-605.

FRENCH, E. G. Motivation as a variable inwork-partner selection. /. abnorm. soc.Psychol, 1955, S3, 96-99.

FRENCH, J. R. P. Group productivity. In H.Guetzkow (Ed.), Groups, leadership, andmen: Research in human relations. Pitts-burgh: Carnegie Press, 1951. Pp. 44-54.

FRENCH, J. W. The description of personalitymeasurements in terms of rotated factors.Princeton, N. J.: Educational Testing Ser-vice, 1953.

FRENCH, R. L., & MENSH, I. N. Some rela-tionships between interpersonal judgmentand sociometric status in a college group.

Sociometry, 1948, 11, 335-345.GAGE, N. L., & CRONBACH, L. J. Conceptual

and methodological problems in interper-sonal perception. Psychol Rev., 1955, 62,411-422.

GAGE, N. L., & EXLINE, R. V. Social percep-tion and effectiveness in discussion groups.Hum. Relat., 1953, 6, 381-396.

GEBEL, A. S. Self-perception and leaderlessgroup discussion status. J. soc. Psychol.,1954, 40, 309-318.

GIBB, C. A. The principles and traits of lead-ership. /. abnorm. soc. Psychol., 1947, 42,267-284.

GIBB, C. A. The emergence of leadership insmall temporary groups of men. Unpub-lished doctoral dissertation, Univer. ofIllinois, 1949. (a)

GIBB, C. A. Some tentative comments con-cerning group Rorschach pointers to thepersonality traits of leaders. J. soc. Psychol.,1949, 30, 251-263. (b)

GIBB, C. A. The research backbround of aninteractional theory of leadership. Aust. J.Psychol., 1950, 2, 19-42.

GIBB, C. A. Leadership. In G. Lindzey (Ed.),Handbook of Social Psychology. Cambridge,Mass.: Addison-Wesley, 1954. Pp. 877-920.

GORDON, L. V. Personal factors in leadership./. soc. Psychol., 1952, 36, 245-248.

GOWAN, J. C. Relationship between leader-ship and personality factors. /. educ. Res.,1955, 48, 623-627.

GREEN, G. H. Insight and group adjustment./. abnorm. soc. Psychol, 1948, 43, 49-61.

GREEN, N. E. Verbal intelligence and effect-iveness of participation in group discussion.J. educ. Psychol, 1950, 41, 440-445.

GREER, F. L., GALANTER, E. H., & NORDLIE,P. G. Interpersonal knowledge and indivi-dual and group effectiveness. J. abnorm.soc. Psychol, 1954, 49, 411-414.

GROESBECK, B. L. Personality correlates ofthe achievement and affiliation motives inclinical psychology trainees. Unpublisheddoctoral dissertation, Univer. Michigan,1956.

GUTHRIE, M. R. The measurement of per-sonal factors related to success of officeworkers. J. appl. Psychol, 1956, 40, 87-90.

HANAWALT, N. G., HAMILTON, C. E., &MORRIS, M. L. Levelof aspiration in collegeleaders and non-leaders. J. abnorm. soc.Psychol, 1943, 38, 545-548.

HANAWALT, N. G., & RICHARDSON, H. M.Leadership as related to Bernreuter per-sonality measures: IV. An item analysis ofresponses of adult leaders and non-leaders./. appl. Psychol, 1944, 28, 397-411.

HANAWALT, N. G., RICHARDSON, H. M., &HAMILTON, R. J. Leadership as related to


Bernreuter personality measures: II. Anitem analysis of responses of college leadersand non-leaders. /. soc. Psychol., 1943, 17,251-267.

HARDY, K. R. The influence of affiliative mo-tivation and social support upon conformityand attitude change. Unpublished doctoraldissertation, Univer. Michigan, 1954.

HARE, A. P., BORGATTA, E. F., & BALES, R. F.Small groups: Studies in social interaction.New York: Knopf, 1955.

HARVEY, O. J. An experimental approach tothe study of status relations in informalgroups. Amer. social. Rev., 1953, 18, 357-367.

HAYS, D. G. The F-scale as a predictor of thepost-meeting reactions of small-group mem-bers. Unpublished manuscript, HarvardUniver., 1953.

HAYTHORN, W. The influence of the individ-ual group members on the behavior of co-workers and on the characteristics of thegroups. Unpublished doctoral dissertation,Univer. Rochester, 1952.

HAYTHORN, W., COUCH, A. S., HAEFNER, D.,LANGHAM, P., & CARTER, L. F. The be-havior of authoritarian and equalitarianpersonalities in small groups. Hum. Relat.,1956, 9, 57-74. (a)

HAYTHORN, W., COUCH, A. S., HAEFNER, D.,LANGHAM, P., & CARTER, L. F. The effectsof varying combinations of authoritarianand equalitarian leaders and followers. J.abnorm. soc. Psychol., 1956, S3, 210-219. (b)

KITES, R. W., & CAMPBELL, D. T. A test ofthe ability of fraternity leaders to estimategroup opinion. /. soc. Psychol., 1950, 32,95-100.

HOFFMAN, M. L. A study of some psychody-namic determinants of compulsive con-formity in a task involving linear judg-ments. Unpublished doctoral dissertation,Univer. Michigan, 1951.

HOFFMAN, M. L. Some psychodynamic fact-ors in compulsive conformity. /. abnorm.soc. Psychol, 1953, 48, 383-393.

HOLLANDER, E. P. Authoritarianism andleadership choice in a military setting. /.abnorm. soc. Psychol., 1954, 49, 365-370.

HOLTZMAN, W. H. Adjustment and leader-ship: A study of the Rorschach test. J.soc. Psychol., 1952, 36, 179-189.

HOWELL, C. E. Measurement of leadership.Sociometry, 1942, 5, 163-168.

HUNTER, E. C., & JORDAN, A. M. An analysisof qualities associated with leadershipamong college students. /. educ. Psychol.,1939, 30, 497-509.

JENKINS, W. O. A review of leadership stud-ies with particular reference to militaryproblems. Psychol. Bull., 1947, 44, 54-79.

JENNINGS, H. H. Leadership and isolation.(2nd ed.) New York: Longmans, Green,1950.

KAGAN, J., & MUSSEN, P. H. Dependencythemes on the TAT and group conformity.J. consult. Psychol, 1956, 20, 29-32.

KELLY, E. L. Medical School project. Un-published manuscript, Univer. Michigan,1957.

KELMAN, H. C. Effects of success and failureon "suggestibility" in the autokineticsituation. /. abnorm. soc. Psychol, 1950,45, 267-285.

KIDD, J. W. Personality traits as barriers toacceptability in a college men's residencehall. J. soc. Psychol, 1953, 38, 127-130.

LEMANN, T. B., & SOLOMON, R. L. Groupcharacteristics as revealed in sociometricpatterns and personality ratings. Socio-metry, 1952, IS, 7-90.

MAAS, H. S. Personal and group factors inleaders' social perception. J. abnorm. soc.Psychol, 1950, 45, 54-63.

McCuEN, T. L. Leadership and intelligenceEducation, 1929, SO, 89-95.

McGRATH, J. E. A framework for integrationof small group research studies. Arlington,Va.: Psychol. Res. Associates, 1957.

MARTIN, W. E., GROSS, N., & DARLEY, J. G.Studies of group behavior: Leaders, follow-ers, and isolates in small organized groups./. abnorm. soc. Psychol, 1952, 47, 838-842.

MASLING, J., GREER, F. L., & GILMORE, R.Status, authoritarianism, and sociometricchoice. J. soc. Psychol, 1955, 41, 297-310.

MIDDLETON, W. C. Personality qualities pre-dominant in campus leaders. /. soc. Psy-chol, 1941, 13, 199-201.

MILL, C. R. Personality patterns of sociallyselected and socially rejected male collegestudents. Unpublished doctoral disserta-tion, Michigan State Coll., 1952.

MILL, C. R. Personality patterns of socio-metrically selected and sociometricallyrejected male college students. Sociometry,1953, 16, 151-167.

MILLER, D. C. An experiment in the meas-urement of social interaction in group dis-cussion. Amer. social. Rev., 1939,4,341-351.

MOORE, L. H. Leadership traits of collegewomen. Social & Soc. Res., 1935, 20,136-139.

NEWCOMB, T. M. Personal communication,cited in Greer, F. L., Galanter, E. H., &Nordlie, P. G. Interpersonal knowledgeand individual and group effectiveness./. abnorm. soc. Psychol, 1954, 49, 413.

NORDLIE, P. G. The performance of smallmilitary groups as a function of intragroupknowledge. Unpublished master's thesis,Univer. Maryland, 1954.

270 RICHARD D. MANN

NORMAN, R. D. The interrelationshipsamong acceptance-rejection, self-otheridentity, insight into self, and realistic per-ception of others. J, soc. Psychol., 1953,37, 20S-23S.

PAGE, D. P. Measurement and predictionof leadership. Amer. J. Social., 1935, 41,31-43.

REILLY, J. W., & ROBINSON, F. P. Studiesof popularity in college: I. Can popularityof freshmen be predicted. Educ. psychol.Measmt, 1947, 7, 67-72.

RICHARDSON, H. M., & HANAWALT, N. G.Leadership as related to Bernreuter per-sonality measures: I. College leadershipin extra-curricular activities. J, soc.Psychol., 1943, 17, 237-249.

RICHARDSON, H, M., & HANAWALT, N. G.Leadership as related to the Bernreuterpersonality measures: III. Leadershipamong adult men in vocational and socialactivities. J. appl. Psychol., 1944, 28,308-317.

RICHARDSON, H. M., & HANAWALT, N. G.Leadership as related to the Bernreuterpersonality measures: V. Leadershipamong adult women in social activities.J. soc. Psychol., 1952, 36, 141-154.

RIGGS, M. M. An investigation of the natureand generality of three new personalityvariables: Part II. Related behavior. /.Pen., 1953,21,411-440.

ROHDE, K. The relation of authoritarianismof the aircrew member to his acceptanceby the airplane commander. Amer.Psychol, 1951, 6, 323.

ROSEBOROUGH, M. E. Experimental studiesof small groups. Psychol. Bull, 1953, SO,275-303.

SHAPIRO, D. Psychological factors in friend-ship choice and rejection. Unpublisheddoctoral dissertation, Univer. Mich. 1953.

SINGER, E. An investigation of some aspectsof empathic behavior. Amer. Psychol.,1951, 6, 309-310.

SLATER, P. E. Psychological factors in rolespecialization. Unpublished doctoral dis-sertation, Harvard Univer., 1955. (a)

SLATER, P. E. Role differentiation in smallgroups. Amer. social. Rev., 1955, 20,300-310. (b)

SMITH, A. J., JAFFE, J., & LIVINGSTON, D. G.Consonance of interpersonal perceptionand individual effectiveness. Hum. Relat.,1955, 8, 385-397.

SMITH, M., & KRUEGER, L, M. A brief sum-mary of literature on leadership. Bull. Sch.Educ. Indiana U., 1933, 9, No. 4.

SPRUNGER, J. A. The relationship of groupmorale estimates to other measures ofgroup and leader effectiveness. Unpub-

lished master's thesis, Ohio State Univer.,1949.

STOGDILL, R. M. Personal factors associatedwith leadership: A survey of the literature.J. Psychol., 1948, 25, 35-71.

STOLPER, R. An investigation of some hy-potheses concerning empathy. Unpublishedmaster's thesis, Louisiana State Univer.,1953.

STRODTBECK, F. L., & HARE, A. P. Bibliog-raphy of small group research: (From1900 through 1953). Sociometry, 1954, 17,107-178.

SWANSON, G. E. Some effects of member ob-ject-relations on small groups. Hum.Relat., 1951, 4, 355-380.

SWARD, K. Temperament and direction ofachievement. /. soc. Psychol., 1933, 4,406-429.

TAFT, R. Use of the "group selection ob-servation" method in the selection oftrainee executives. J. appl. Psychol.,1948, 32, 587-594.

TAGIURI, R. Relational analysis: An exten-sion of sociometric method with emphasisupon sociometric perception. Sociometry,1952, 15, 91-104.

TALLAND, G. A. The assessment of groupopinion by leaders, and their influence onits formation. J. abnorm. soc. Psychol.,1954, 49, 431-434.

TAYLOR, F. K. Awareness of one's socialappeal. Hum. Relat., 1956, 9, 47-56.

TRAPP, E. P. Leadership and popularity asa function of behavioral predictions. J.abnorm. soc. Psychol., 1955, 51, 452-457.

U. S. OFFICE OF STRATEGIC SERVICES. As-sessment of men. New York: Rinehart, 1948.

VAN ZELST, R. H. Validation evidence onthe empathy test. Educ. Psychol. Measmt.,1953, 13, 474-477.

WATSON, J. Some social psychological corre-lates of personality: A study of the useful-ness of psychoanalytic theory in predict-ing social behavior. Unpublished doctoraldissertation, Univer. Michigan, 1952.

WILLIAMSON, E. G., & HOYT, D. Measuredpersonality characteristics of studentleaders. Educ. psychol. Measmt., 1952, 12,65-78.

WURSTER, C. R., & BASS, B. M. Situationaltests: IV. Validity of leaderless group dis-cussions among strangers. Educ. psychol.Measmt., 1953, 13, 122-132.

YOUNG, N., & GAIER, E. L. A preliminaryinvestigation into the prediction of sug-gestibility from selected personality vari-ables. /. soc. Psychol., 1953, 37, 53-60.

ZELENY, L. D. Characteristics of group lead-ers. Social, soc. Res., 1939, 24, 140-149.

Received August 25,1958.

VOL. Psychological Bulletin - Gwern.net · Psychological Bulletin ... personality assessment is...

Documents

Transcript of VOL. Psychological Bulletin - Gwern.net · Psychological Bulletin ... personality assessment is...