'v'wWwwr—VVVVvaw—w—y'vv— -——-—‘—V‘.-‘v ‘ T

 

_ .v- .1 ‘sw-Q
I, a A ~m“mm‘.¢”c'0'OOO-oovao Q \ 0’.-Q..--‘.“..""' If“. 1 v
. -0 0.0.0.‘ , _
‘ ‘ '0. ’0‘. m‘. .“,‘.
r. .' ”MA, ‘ A

 

 

 

 

r ’ Thesis for the Degree. of M._
MICHIGAN my: umvansm
-3 on NA momma mewoon

 

   
  
 
 
  
   
  

 

 

 

 

 

 

 

     

l
- ‘0 .
x-
. .
.fi ‘ . . .
' . o o
'0
V - o O
a. . . -.
v
A-
.-. .
A
- o
’ -
u . O '
-.‘.
V‘ n
‘. p v I -
-‘ - . .
o - * ' -
5 ‘ . . . . I o . ‘
O-.- - v , _ -
" . - , ‘ . I - .
“- , r . u - . ... O
o . ' -
. l - I , "’ _ .
" ‘ - -. . _ 7 ~
. V . , ' — , . .- -
.. - .. 1 .
. ' r , . ' - I _
J.. _ o , ._- o a -
. .
o , , . , - . , .. , _ -_- -
V . . o . .o 0' _ ' _ o I .
. . -o - ' ‘
“‘ . ‘ ', . . .. . v . ~uo. .1, v
(f‘ n o ‘ __.. _' ' - ,- . r .. .o o
‘ § . c _ . _ .1 r. .. ’ - ' ' ’ I
A. _ . .- - - . -
u . . . - ... -~ - V.- o
' . o .. 4 .‘ . ' .o - y‘.
"'~ . . ‘ , , ,. . . _ . ,., .. ...r‘
Y ' . . . ‘ . . . . , _ . . v , - a
. I ' . - ._ . —- a ‘ -‘ ' - O
I. ' ' ‘ . - ' . ... . a- . o ' ' ---a .
I . -. o . . -. , . . . . - . o0 . b- . -.. . ..
‘ - - r ,. ' -, .. 1 , .
f." ’ - ' ‘- . .., .,,. , v-7 .1. v . 0- ‘ - 0.
.-'\ -. . _ ._ ‘ . v ' .1 . _ . . ._
“‘5 ’ . . . . . - . ,. .., .... . . ., . - ‘ - . - . o .
w‘ ' ' 00 1-; o ' ‘- . - - , ‘ .( , . . o. .- . .
.‘-. . '. ' - . . - . - 0",97‘ 0‘ a -0 "~ 0' " ‘
v a . - - . .. . . I... -. .. . -v—GI ' .
‘ ‘- » - - ’ . . . o oo. ‘ — y'. a. '0' .- '0
- . - . .. . .o - v~. o O I
F: ’ ’ - - - . o ‘ .Ovt ..-... 00‘ II'I" . I' ---0‘ I
I ' v ' . - ... a -- . ro- , . . O ,, o . o q -
‘ ' ' " - ' . u o .. - - , _ 00.40.1‘ " . -l-. 0 ~-
. ’ ' ' . ’ " a a ~ ' v. o - I a ,.vfl’. o, . '.. l ‘ -
. , ,1. ,. .¢ 0 . . o 4 u. o o. . _ , .—. A-O'OV‘I " 'l""-
. ’ “ - ‘ ' ' - 0' - -4‘--- ' ,,..oc-'. . .-;1, - v, '- - "". '
u - . .‘ o . . _ . ..,_ ..' ‘ . . ., o ."." _ .0..—-OO .-
I ‘ 4! ’ ' .o-‘ 0-0 . .9 ,_._-... ..,./ ,. ‘-’- I f- v ‘ " '-
. .. _ . .4 . a v .vro. , . .. - , ,,- . ,, - . 10,. .n....v.-oo.v-v - o ’-o , O
. . _.. a o v . - .V ;, ...4 .- ..~ .. 'oo ... - O. -.’.’ ”.m' '. prvalp’ - o'pOoOr.polrr0. .
‘ ‘ _ a. . ,_, " . . . ' _ ., I. '..... 1. -. .'
- a . . . - - , ’ , ,
- - - - v.. . .. . 't"---- - ' - 1
'0" ' ‘ ‘ 'lu . ,‘_, ,rl. -' ’ . , o'cc'uf ‘a. ‘0’.
.. . .- . ‘ 'I. .- o . .,.. . z N44. voyO‘l-Woo . 11¢: .r-‘o - -. u'l-I
‘ ' ' "" " "" ‘ " " o or'l I I’airplfpllt'f‘ - "" ’ "'f--" “ " "\V
' _'l o o ' "O ' - p , . - l- o c .f.....,, >'.," H, ’,.. ‘(t'WPQO‘CI‘:""’ ,‘I '4 y- o . ’
"" '00.; 0 II to v‘ ‘ .,-. . . _:‘T".,, '0-0- 4 v V o Y ‘-oo'r.a‘vr‘alr-‘O" ’
. '0 . . . . — I r 0-.Otu. - v ,1... ‘,".":r.:c;.'v’,p,"o-o-.’d
.; - . . >.. . r”: ' _,,‘ V . 0". “p-‘ J-O'f ' ‘
, v. . . , ’ "". ' r"_ .».-/ .‘M."'.‘d -- . o-n
-" l . o u - .' - . y - ,'. .' ol-‘I; od'CF- - oog‘l
, o.- 5": — a , . . - , . _.. --, .o , . ."",'_-.'>.”"".W _' c- g".‘. ..(I
’ " v p—- " .¢ I). .o,. ..v.q .0. 7. ' -vx'r,>loO’I-> '
‘ - . I . , . ~0.- . V .0.‘
' o o g . A ' ' - -
‘ ' v'-.r -. ‘.
v -- r .. ..a. co .
.
., .. . .. r ... oc-- .-10v
0 . . . o -o v. .I.’... , . '4 .
.,o -- . . . .. 1 -.ol ‘-"l-
'I.. - o _. 9..
o . a- . , '1’ o .
u . l ‘ n . ._
. . .- _ . o ‘ I
. , - . , - D -t¢l-
‘ '- ' -- - ' ... . ._-_._¢
. . . . _'r- Jl-“' 'Ill‘dltln -
. -- . - -

     

 

    
     

  

LEE:- ZARY

Michib an State
Univmsit)’

HillllllllljllzlllfllllllllHHIIHUIIHIIIIIHUIlllllllllW

310381 5126

 

 

 

 

 

ABSTRACT

COGNITIVE DIFFERENTIATION AND TASK LEADERSHIP

BY

Donna Antoinette Lingwood

The primary purpose of this study has been to test
the hypothesis, based upon Kelly's Theory of Personal
Constructs, that cognitive differentiation is positively
correlated with task leadership in small groups. Some
assumptions underlying Kelly's theory are also examined.
It is hypothesized that ratings of others are influenced
by familiarity and the desire of the rater to be socially
agreeable. In addition, Kelly's assumption that construc-
tive ability, defined here in terms of Bieri's concept of
cognitive differentiation or the number of dimensions
the individual has in his repertoire for description of
other peOple, is independent of the type of the dimensions
is examined.

The hypotheses were tested in groups consisting of
a resident advisor and resident assistants on each floor
of the Michigan State University resident halls. The
number of groups, including male and female groups, totaled

thirty-three. Findings of the study show that the

Donna Antionette Lingwood

hypothesis relating coqnitive differentiation, measured by
a method which approximates Tripodi and Bieri's modifica-
tion of the Kelly Role Construct Repertory Test, with task
leadership, measured by the number of choices a person
receives for having best ideas, is not supported.
Familiarity is shown to positively correlate with cog-
nitive differentiation, and cognitive differentiation is
shown to be dependent upon type of dimensions. The lack
of support for the cognitive differentiation-task leader-
ship hypothesis is evaluated with suggestions for an

improved test of the hypothesis.

COGNITIVE DIFFERENTIATION AND TASK LEADERSHIP

BY

Donna Antoinette Lingwood

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

Master of Arts

Department of Sociology

1973

ACKNOWLEDGMENTS

I would like to eXpress my appreciation for the
time, effort, and many interesting suggestions given by
Dr. Bernard Finifter. I would also like to thank Dr. Bo
Anderson and Dr. William Crano for their constructive

criticisms and comments.

ii

TABLE

LIST OF FIGURES . . . .

Chapter
I. INTRODUCTION . .
II. THEORY . . . .
III. METHODOLOGY . . .
IV. RESULTS . . . .
V. DISCUSSION . . .
VI. SUMMARY . . . .

REFERENCES. . . . . .

OF

iii;

CONTENTS

Page

iv

25
42
46
53

56

LIST OF FIGURES

Figure Page
1. Bales 1970 Questionnaire . . . . . . . 32

2. Bieri and Tripodi 1963 Measure of
Cognitive Differentiation . . . . . . 33

3. Measure of Cognitive Differentiation . . . 37

iv

CHAPTER I

INTRODUCTION

The general focus in this paper is on personality
and position in small groups. The approach to personality
and social structure relations taken by such authors as
Inkeles (1963), Levinson (1963), Thomas (1968) and Heine
(1971) provides the broad framework for this thesis. In
this perspective, personality and social structure are
viewed as interacting, yet separate systems where knowledge
of one system can aid in the analysis of the other. Social
systems can be studied more fruitfully through inclusion of
psychological theory in the analysis. Inkeles (1963: 321)
argues for this approach, contending that "sociological
research has suffered from the failure to use psychological
theory and established knowledge about personality as one
element in the analysis."

A common meeting ground for psychological and
sociological theory is small group research. In particular,
the relationship of personality to the role position of
leadership in small groups has been the subject of numerous

research efforts (e.g., Mann, 1959; Geier, 1970). This

paper is a report of an investigation of the relationship

of the personality trait of cognitive differentiation to
task leadership in small groups. Specifically, drawing

upon G. A. Kelly's (1955) theory of personal constructs, it
is hypothesized that cognitive differentiation will be
positively correlated with the number of sociometric choices
received on best ideas, in a group of individuals with
variation in levels of differentiation.

A review of the literature on cognitive differentia-
tion raised questions regarding the validity of the concept
(Vannoy, 1965; Wicker, 1969; Mitchell, 1970; and Crano and
Bettinghaus, 1970). Because the validity of the concept
and Operationalization is believed to be a critical issue
in correlating differentiation with task leadership, much
of the research effort was directed toward this area. A
substantial portion of the present paper is a description
of the examination of the validity issues, with particular
reference to the relationship of agreeableness to cognitive
differentiation.

There are several concepts of cognitive differentia-
tion in the literature (Bieri, 1966; Crockett, 1965; Harvey,
Hunt and Schroder, 1961; and Schroder, Driver and Streufert,
1966). The definitions of these concepts are generally
found in discussions of the larger concept of cognitive com-

plexity. Cognitive complexity is the larger concept in the

sense that cognitive differentiation and cognitive inte-
gration together comprise cognitive complexity.

The definition of cognitive differentiation to be
used in this project is the number of independent dimensions
a person has available for ordering a set of stimuli (Bieri,
1966). -These dimensions together form a cognitive system
in the loose sense of a group of interrelated elements.
Cognitive complexity is a term referring to particular
aSpects of the system of constructs or dimensions, including
the number of constructs in the system, their independence
of each other, and their hierarchal relationships to
each other (Kelly, 1955).

The number of dimensions and the independence of
these dimensions is cognitive differentiation, while the
hierarchal relationships among the dimensions is cognitive
integration. Independence of dimensions is considered an
important aspect of cognitive differentiation by some
authors (Bieri, 1966; Crockett, 1965) while writers such
as Schroder, Driver and Streufert (1967) do not discuss
independence of dimensions. Cognitive integration is
referred to by various authors as the interconnection
(Schroder, Driver and Streufert, 1967) or hierarchal
integration (Crockett, 1965) of the elements of the cog-
nitive system. The distinctions between cognitive differ-
entiation, cognitive integration, and cognitive complexity
are not always apparent in the literature which necessi-

tates careful attention to the above authors' works.

As stated above, cognitive differentiation is the
personality variable to be related to task leadership in
this study, using Kelly's theoretical framework and
Bieri's (1966) definition of cognitive differentiation,
which stems from Kelly's work. Kelly's theory can be
criticized for the vagueness of many of its terms. Conse-
quently, due to a lack of clarity of the terms, some
problems are likely in attempting to conduct research based
on the theory. However, Kelly does formulate an inter-
esting hypothesis relating constructive ability to leader-
ship and provides a measure of his concept of constructive
ability as well as suggestions for how he would operation-
alize leadership. As Bieri has clarified (perhaps with
some conceptual slippage) Kelly's concept of constructive
ability, and Bieri and Tripodi have refined Kelly's
measure of constructive ability, Kelly's hypothesis can be
adequately tested.

The measure of Bieri's concept of cognitive differ-
entiation is Tripodi and Bieri's 1963 modification of
the Kelly Role Construct Repertory Test. In addition to
origination in Kelly's work, selection of Tripodi and
Bieri's measure was based upon the secondary analysis
nature of this project. As the data were collected for
other purposes, this measure of cognitive differentiation

is the only measure applicable to the data. Other measures

including Crockett's method (1965) where subjects generate
their own constructs, and Harvey, Hunt and Schroder's
sentence completion test (1961), cannot be applied to the

data.

CHAPTER II

THEORY

The theoretical framework for this thesis is
Kelly's Theory of Personal Constructs (1955). Kelly's
theory is not a formal theory. Kelly himself notes
(p. 46) that there are limitations in his theory-building
efforts. He constructs his theory in terms of one
"fundamental postulate" and several corollaries. The
corollaries are described as little more than clarifica-
tion of what is implicit in his fundamental postulate.
Kelly comments further that he does not see his theoretical
system as logic-tight, but that "rather, we shall strive
to make our theoretical position provocative, and hence
fertile, rather than legalistic."

The basis of Kelly's theoretical formulations is
in his image of man in relation to his environment and his
views on epistomoloqy. Kelly's image of man is "man-the-
scientist [p. 4]," whose ultimate aim is to predict and
control his environment. Man-the-scientist is seen as
referring to all mankind and not the class of men who are

scientists.

To Kelly, the environment exists independently of
the person or perceiver. This position is similar to what
Garner (1973) terms an epistemolOgical realist position
where the concern is with a real world, rather than a
world of peeple's thoughts (p. 6).

Prediction of the real world, the central motive
in Kelly's image of man, is based upon man's ability to
represent this world cognitively. The creative capacity
of man to represent the environment, not merely respond
to it, is emphasized. Representation of the world allows
the person to place alternative constructions upon it.
Kelly terms this process constructive alternativism, the
central assumption in his theory.

Representing the environment cognitively, or
construing the environment is described in the following
terms: "Man looks at his world through transparent pat-
terns or templets which he creates and then attempts to
fit over the realities of which the world is composed
[p. 9]." Kelly gives the name constructs to the patterns
that are tried for fit to the world. Different constructs
are tried, incorrect ones eliminated, and new ones added
according to their effectiveness in predicting the
environment.

The fundamental postulate in Kelly's theory is
stated as: A person's processes are psychologically

channelized by the ways in which he anticipates events.

Kelly's discussion of this postulate indicates that, more
simply, this postulate states that a person's actions are
guided by how he anticipates events. Anticipation of
events is, in turn, determined by the construct system.

Kelly presents what he terms eleven corollaries
to his fundamental postulate. Relevant to this paper in
terms of the hypothesized relationship between cognitive
differentiation and task leadership are his individuality
and sociality corollaries. To clarify Kelly's term
"construing," an integral part of these two corollaries,
his construction corollary will be briefly discussed.

The construction corollary regards how peOple
anticipate events in the environment. A person anticipates
events by construing these events. The term construing,
which is the critical term in Kelly's framework, refers to
"placing an interpretation"; a person places an inter-
pretation upon the environment. He erects a structure,
within the framework of which the substance takes shape
or assumes meaning (p. 50). By "substance" Kelly refers
to the environment and "structure" refers to the system
of constructs.

The individuality corollary states that individuals
differ from each other in their construction of events.
This corollary establishes Kelly's idea of individual
differences in construction ability. Sharing of exper-

ience is possible as each person make construe likeness

and difference between the events in which he is involved,
together with those in which another person is involved.
While there are individual differences in the construc-
tion of events, people can find common ground through
construing the experience of others along with their own.

The sociality corollary is where Kelly develops
his social psychology and hypothesizes a relationship
between the ability to construe other pe0ple's cognitive
processes and leadership. The sociality corollary states:
To the extent that one person construes the construction
processes of another, he may play a role in a social
process involving the other person.

In clarifying the corollary, Kelly comments that
playing a constructive role in relation to another person
involves seeing eye to eye with him and an acceptance of
him and of his way of seeing things. In other words, the
person must construe the other's viewpoint, not his
actions, in order to relate to him.

Kelly argues that commonality of vieWpoints is
less important to playing a role with another person than
is construing or subsuming or understanding or inter-
preting (terms which are used interchangeably) the view-
point or construction system of the other person.
Commonality between construction systems can increase the
likelihood that one construct system can subsume another.

However, commonality is incidental to role-playing in

10

cases where roles are played between peOple who have
similar construction systems.

The emphasis on the construing or understanding
of the other person's construction system provides the
basis for Kelly's social psychology. Understanding other
peOple's construing efforts is the basis for social
interaction. In this reSpect, Kelly's position is similar
to symbolic interactionism. To Kelly, social psychology
"must be a psychology of interpersonal understandings,
not merely a psychology of common understandings [p. 95]."

In addition to his emphasis on the construing
process in social interaction, Kelly stresses his concept
of role. He defines role as a "psychological process based
upon the role player's construction of aspects of the con-
struction systems of those with whom he attempts to join
in a social enterprise [p. 97]." A role is seen to be an
ongoing pattern of behavior that follows from a person's
understanding of how others who are associated with him
in his task are construing the situation.

Kelly elaborates his definition of role. Like
other behavior, it is assumed to be tied to the person's
construct system. Role is anchored in the construction
system of the role player and does not necessarily follow
from his relationship to other members of a group. Role

is "a pattern of behavior emerging from the person's own

11

construction system rather than primarily out of his
social circumstances [p. 98]."

Kelly's emphasis on cognition in defining role is
prevalent in the social psychological literature
(Levinson, 1963; Turner, 1969) as well as the literature
with an ethnomethodological perspective (e.g., Hayek,
1942). Levinson discusses role-definition as Opposed to
structurally given role demands. He defines personal
role-definition in terms of the individual's conceptual-
ization of external reality.

Turner describes (pp. 216-17) the function of
cognition in role-playing. Cognition is responsible for
shaping the person's world into roles. The shaping of the
phenomenal world into roles is, to Turner, the key to
role-taking as a core process in interaction. Role is an
"ideal conception which constrains peOple to render any
action situation into more or less explicit collections
of interacting roles [p. 216]."

Turner notes that an initial distinction is
necessary between taking the existence of distinct and
identifiable roles as the starting point in role theory,
and postulating the conceptualizing of roles as the
orienting process in interactive behavior. The latter
approach, or Turner's approach, "has less interest in

determining the exact roles in a group and the specific

12

content of each role than observing the basic tendency for
actors to behave as if there were roles [p. 216]."

Another process of construction of roles similar to
Kelly is described by Berger and Luckmann (1966: pp. 56-
57). A person constructs the reality of social situations.
He attributes to the other motives and viewpoints and
plays a role vis-a-vis the other taking into account these
motives and vieWpoints. The cognitive element in role
playing is also emphasized by Hayek, who points out that
in studying social relations, the data must be man and the
world as it appears to man.

Kelly utilizes this sociality corollary to relate
individual differences in construing ability to leadership
in small groups, forming the basis of the hypothesis in
this study. Leadership is defined sociometrically in
terms of what criterion the group members use for nominat-
ing a group leader. Nominators may make their selections
of leaders differently, depending upon their understanding
of what the situation demands. The person selected would
be one who shows ability to meet the criterion for c0ping
with situational demands. A leader is one who performs
any of the jobs which are recognized as leadership (p. 101).

This definition of leadership is similar to the
definition of leadership given by small group theorists
and others, who define leadership in terms of a combina-

tion of personality and situational factors. Gibb (1970:

13

369) contends that leadership is always relevant to the
situation as a certain kind of situation is required before
leadership appears; and the particular set of circumstances
existing determines which attributes of personality are
important for leadership.

Selection of an individual to the leadership role
is seen by Kelly to be dependent upon the group and upon
the capacity of the individual to aid in the achievement
of the goal. Heine (1971: 103) takes a similar position
stating personal qualities are relevant to leadership
selection, when these qualities meet the demands of the
followers. The personality-social structure perspective
can also be found in the conceptualization of leadership
offered by Cartwright and Zander (1968) who note that
traits of the leader necessary and effective in one group
may be quite different from those required in a different
group. The authors conclude that selection of a leader
must take into account the type of functions to be per-
formed in a given situation.

Bales (1950) categorizes these functions into task
and socioemotional types. Task leadership relates to
those behaviors in a group which are directed toward com-
pletion of the task while socioemotional leadership is
conceptualized as behaviors in a group directed towards

integration and maintenance of the group. Both types of

l4

leadership are found in the group and the same person does
not usually occupy both leadership roles.

Kelly discusses a type of leader which he terms
the mobilizing type of leader. A mobilizing leader is
briefly mentioned as one who accelerates the group's
social prOgress. He does not clearly define what he means
by the term "mobilization." Mobilization implies gathering
forces with the purpose of achieving a goal. In a group,
mobilization would seem to refer to moving the group
towards attainment of the group goal. As working toward a
goal is the primary characteristic of task groups, mobil-
izing leadership seems to essentially imply leadership in
a task group. However, mobilizing could also involve moving
a group towards better interpersonal relations, thus
involving socioemotional leadership.

Mobilizing a group is used by other authors to
refer to either a task or a socioemotional process. Carter
(1954) refers to a similar process relating to carrying
out the task of the group as group goal facilitation.
Bosers and Seashore (1967) refer to mobilization processes
under the terms interaction facilitation (socioemotional),
work facilitation (task), and goal emphasis (task).

As noted above, Kelly does not clearly define
mobilizing leadership as referring to task or socio-
emotional leadership and the term "mobilizing" is appli-

cable to either task or socioemotional leadership

15

processes. Therefore, Kelly's hypothesized relationship
between cognitive differentiation and mobilizing leader-
ship can be specified as a relationship between cognitive
differentiation and task leadership or between cognitive
differentiation and socioemotional leadership.

In this paper, Kelly's mobilizing leadership will
be used as referring to task leadership, or aspects of
leadership related to task completion or goal attainment.*
This decision is based upon research findings which provide
some indication that cognitive differentiation is not
associated with a preference for balance in social rela-
tions. Press and Crockett (1969) found that, when con-
fronted with learning the social relations among people in
a group, low cognitively complex peOple relied upon a
balance principle while high cognitively complex people
did not rely upon a balance scheme as a tool for organizing
social relationships. Scott (1963) found high cognitively
complex persons prefer unbalanced liking relations and
liking is a less central concern for complex persons.

In the personality-social structure perspective
on leadership offered by Kelly, the personality aspect is
the quality of the individual relevant to his selection

by the group to the role of mobilizing leader. The social

 

*
Cognitive differentiation will also be correlated

with socioemotional leadership, measured by the number of
votes a person receives on likeability.

16

structural element is the mobilizing leader role. The
personality quality to which Kelly refers is the person's
ability to construe other pe0ple's construction systems.
One person may understand others better than they under-
stand him. He may construe more of others' vieWpoints.
These individual differences in construing ability or
differentiation, stated in the individuality corollary,
are referred to by Bieri as the number of constructs a
person has available for construing or interpreting or
differentiating others' construction systems.

The greater the number of constructs a person
possesses, the more likely he is to be able to construe or
understand how various group members are construing a
situation. Construing other peopples' construction systems
involves interpreting their systems in terms of one's own
constructs; therefore a large repertoire of constructs
makes it more likely one can accurately construe the view-
points of others' and consequently predict their behavior.
The ability to understand various other group members'
construction systems allows this person to better play
roles vis-a-vis the group members. Construction of others'
ideas and accurate prediction of others' behavior is seen
as necessary for mobilizing a group toward a goal.

Kelly's hypothesized relationship between con-
struction ability or differentiation and task leadership

in a small group includes the following elements.

17

Initially, several assumptions are made: (a) that members
of the group perceive the task orientation of the group;
(b) members perceive the relevance of mobilization ability
or task leadership ability for attaining the group's goal;
(c) members perceive who in the group has the best ideas;
and (d) members are utilizing the same criterion of best
ideas for nominations of the person with best ideas. The
greater the number of constructs in an individual's cog-
nitive system, or the greater his differentiation ability,
the greater is his ability to (a) See others' ideas,

(b) predict others' behavior, and (c) play roles vis-a-vis
others in the group, all of which are to Kelly components
of mobilizing, or task, leadership. Therefore, the greater
the cognitive differentiation of a group member, the
greater is his task abilities.

A further assumption is necessary to relate cog-
nitive differentiation to task leadership. The high cog-
nitive differentiator not only construes or differen-
tiates the various ideas of others. He also is assumed to
integrate these ideas and therefore develop the best
ideas in the group as to how to accomplish the task. In
support of the assumption, Streufert and Schroder (1965)
have found differentiation to correlate positively with
integration.

Best ideas is the component of task leadership

measured in this study. Given the above assumptions, the

18

higher the cognitive differentiation ability, the more
likely this individual will give the best ideas in the
group regarding task completion, and the more likely he

is to be selected as group task leader. Specifically,
constructive ability, or cognitive differentiation, is
hypothesized to be positively correlated with task leader-
ship, i.e., the number of nominations a group member
receives for having the best ideas.

Several studies, using various measures of cog-
nitive differentiation, have investigated aspects of
Kelly's hypothesized relationship of differentiation and
leadership. However, the relationship of differentiation
and task leadership as specified above has not been
examined. The relationship between differentiation and
the prediction component of leadership has been examined
by Bieri (1955). High differentiators were found to be
more accurate in prediction of others' responses.
Differentiation correlated positively .35 at the .05 level
of significance with accuracy of prediction of others'
behavior. Leventhal (1957) reports finding high differ-
entiation to be correlated with accurate prediction of
others' behavior but at a non-significant level.

A study by Mann (1959) linked accuracy of inter-
personal perception, assumed to be associated with accuracy
of prediction of others' behavior, to leadership. A

significant, positive correlation was obtained.

l9

Differentiation has also been correlated significantly
with extroversion (Bieri and Messerly, 1961). Therefore,
the high differentiator would have the force to get his
ideas across to other group members. This factor would
increase the likelihood of his selection as having best
ideas in the group, or as task leader.

Other studies using small groups have related
levels of cognitive complexity to information search be-
havior, communications, decision-making and productivity
within the group (Streufert, Suedfeld, and Driver, 1965;
Streufert and Schroder, 1965; Schroder, Driver and
Streufert, 1967). Tuckman (1967) has correlated level of
cognitive complexity in the group with task performance
and role differentiation.

In testing the cognitive differentiation-task
leadership hypothesis an important consideration is the
degree of formal structure of the group. The hypothesized
relationship would be expected to hold for groups lacking
a designated leader and given role assignments, i.e., a
formal structure. Variation in role behavior attributable
to coqnitive differences would be greatest in unstructured
groups. Levinson (1963: 438) states that in formal organ-
izations role-requirements can be narrowly defined and
social control mechanisms are powerful. Turner (1969: 216)
concurs with this view, stating that the "formal regula-
tion system restricts the free Operation of the role-

making process. . . ."

20

Leadership in unstructured groups is "emergent,"
as opposed to formally designated leadership. It is this
type of leadership that is to be examined, although it is
the special case of emergent leadership in a group where
formal leadership exists. The idea of leadership as
dependent upon a criterion held by the group is an emergent
leadership concept. A criterion of the group is much less
relevant to formally designated leaders. By taking this
definition, Kelly is implicitly assuming an emergent type
of leadership.

Implicit in Kelly's theory are certain assumptions
regarding differentiation of other peeple. Kelly seems
to assume a differentiation of people in the abstract.
What relationship the people have in relation to the dif-
ferentiator is not seen as relevant, nor is how familiar
the person is with the differentiator considered. Here
all peOple are in the relationship of group member to one
another; however perceived familiarity of persons to each
other is quite variable.

Familiarity with others logically seems to be the
result of amount of interaction. Wicker (1969) has shown
indirectly that the amount of interaction with another is
significantly, positively correlated with the degree of
differentiation of him. If Newcomb's (1961) theory that
interaction leads to increased information is correct,

then a correlation between familiarity and differentiation

21

could be expected in the same way that a correlation
between interaction and complexity is found. Increased
interaction with another would mean increased information
about the other or greater familiarity with him.. This
greater familiarity would be associated with increased
differentiation of the other person. Therefore, it is
hypothesized that the greater the familiarity with another,

the more the differentiation of this person.

 

A further assumption made by Kelly and Bieri is
that in the process of construing or differentiating other
peOple's frames of reference, all dimensions are equally
likely to be construed. That is, content of dimensions
does not affect differentiation. Types of dimensions
such as task and socioemotional would be equally differ-
entiated astflmeprocess of construction or differentiation
is seen by Kelly and Bieri to be independent of type of
item.

It appears likely, as peOple seem to habitually
use some dimensions more than others, that people would
construe or differentiate some types of items more than
others. If this position is taken, this assumption becomes
questionable. Therefore, it is hypothesized that the con-
tent of dimensions affects the degree of differentiation

of these dimensions. Task and socioemotional dimensions

 

will not be differentiated to the same degree. That is,

22

a low positive correlation is expected between the amount
of differentiation on task and socioemotional dimensions.

A final assumption of Kelly's regards the relation-
ship of motives to differentiation. The motive Kelly
employs to account for differentiation is prediction of
the other's behavior. This motive leads to accurate con-
struction or differentiation of the other. Another motive
that could be involved in the process of differentiation '
of others is the need for approval of others, or "the
approval motive" (Crowne and Marlowe, 1964). According to
Crowne and Marlowe, peOple who are motivated by a need for
approval are more likely to respond in a socially desirable
manner. Approval is believed to be attained by responding
in culturally accepted ways.

Responding in socially approved ways includes
evaluative behavior, or the rating of others on dimensions.
Crowne and Marlowe see agreeableness in the rating of
others as designed to obtain social approval. Researchers
studying rating behavior have dealt with an agreeing
phenomenon under the label of acquiescence or yea-saying
(Couch and Keniston, 1960). Couch and Keniston see what
is generally viewed as a non-personality related response
set of acquiescence, usually conceived in terms of an
artifact in measurement, as a behavioral disposition of
agreeableness. Acquiescence, or agreeableness, to Couch

and Keniston is reflected in saying yes to all items on

23

an evaluative test regardless Of content Of item. Saying
yes to socially devalued items is counted in Couch and
Keniston's scoring of the number of yes reSponses to items.
If yea-saying is viewed in terms Of a behavioral
diSposition Of agreeableness, which serves the motive of
desire for social approval, then agreeableness could inter-
fere with differentiation Of others in the following way.
A person who shows a strong behavioral disposition of
agreeableness would show less differentiation of others.
High agreeableness would indicate that the rater, rather
than considering each dimension independently to see if it
applies to the person rated (differentiation), would agree
(or disagree) that all dimensions aptly describe the
person (lack of differentiation). Therefore, i£;i§_
hypothesized that agreeableness is negative1y_associated

with differentiation.

 

Since agreeableness is conceptually Opposed to
differentiation, and differentiation is hypothesized to
correlate positively with task leadership, agreeableness
is expected to correlate negatively with task leadership.
Empirical support for this hypothesis is offered by Bales
(1950), who found giving agreement is negatively associated
with task leadership. A significant negative correlation
between agreeableness and task leadership will, in addi-

tion, establish yea-saying as behavioral disposition rather

24

than as a measurement artifact for yea-saying will relate

to a theoretically relevant variable.

CHAPTER III

METHODOLOGY

The task groups in this project are groups formed
of the resident hall advisor and resident assistants on
each floor of the residence halls at Michigan State Uni—
versity. Advisors from each hall, the total number Of
advisors being 312, are grouped into units ranging in size
from 6-16 persons. The total population of groups consists
Of 36 groups. Three groups refused to participate, leaving
a total of 33 groups. Fifteen Of these groups are male,

16 are female, and 2 are combined male-female groups.*

The test of the differentiation-leadership hypo-
thesis in this study depends upon establishing these groups
as task groups. In the literature, studies Of task groups
frequently refers to groups similar in nature to Bales task
groups. A Bales (1950) task group is a group whose pur-

pose is the solution Of a task through joint solving

 

*Questionnaires employed in the study were given
to each person in the resident hall groups (when the
groups were assembled in regular weekly meeting). The
response rate is moderately high with a great deal Of
inter-group variation in response rates.

25

26

efforts Of members of the group. A group where the aim or
main emphasis is on expressive personal interaction, such
as therapy groups, meeting of friends, and play groups, is
not considered to be a task group.

The groups used in this study are not as clearly
task groups as are the Bales task groups. However, the
groups can be shown to sufficiently conform to the nature
of the Bales task groups so as to provide an adequate test
Of the hypothesis. One way in which these groups differ
from Bales task groups lies in the emphasis on task-
solution activities. The task in these groups is to dis-
cuss and solve mutual problems encountered in their roles
of resident advisors. The groups are given this task by
the Residence Hall Office and this task is the Official
function of the groups.

It would be expected that some activity in the
groups would relate to dealing with members' feelings
regarding handling Of their roles Of resident assistants.
These activities would be Of an expressive personal inter-
action nature which, according tO Bales, are not task
groups activities. However, Bales disqualifies a group
from being a task group not on the basis of whether some
activity Of the group is of an expressive personal inter-
action nature, but rather if the aim or main emphasis of
the group is on expressive personal interaction. TO

reiterate, the legitimate aim or function Of these groups

27

is solving problems relating to methods Of dealing with
behavior of residents in the hall encountered in the role
of resident assistant, not on members expressing their
feelings regarding their performance in this role.

A conceivable line Of argument against establishing
these groups as task groups would concern the nature of
the task Of the groups. It would be argued that as the
task of these groups concerns behavior, the task is socio-
emotional and therefore the groups are socioemotional
rather than task groups. In refutation Of this argument,
it can be noted that Bales (1968: 391) considers a group
in which the task concerns behavior as a clear example of
a task group. Specifically, he refers to a diagnostic
council, where the task is for members to come to a
decision regarding the patient.

Certain characteristics of these groups are rele-
vant to the central hypothesis of this study. The resident
advisor in each group, who usually has one or two graduate
assistants, is the official leader Of each group. Thus
the groups possess a degree of formal structure. Within
the groups, one or two leaders, sometimes more, generally
emerge among the members Of the group, selected according
to the criterion of best ideas. However, in some groups,
no one has high scores on number Of choices on best ideas.

The amount of task ideas contributed by the Offi—

cial leader could relate to whether a leader emerges from

28

among the group members or resident assistants. The fre-
quency of contribution Of task-related ideas by the formal
leader is a relevant condition for emergence Of a task
leader from the group. The formal leader may be a light
or heavy contributor Of good ideas. If the formal leader
were to give a great many good ideas, it is conceivable
that choices for best ideas would be fairly evenly dis-
tributed across all people in the group, indicating lack of
emergence Of a leader from among the group members.

Regarding this issue, Bowers and Seashore (1967)
report that peer leadership (defined as leadership by peers
in a group where a formally designated leader exists, or a
type of emergent leadership) appears to be equal in
strength to formally designated leadership. Consequently,
the formally designated leader would not be expected to be
the only group source Of best ideas, and emergent leader-
ship could develop in those groups where the formal leader
contributed many task ideas, or where strong formal leader-
ship as defined by contribution of task ideas exists.
Factors other than strong formal leadership would be held
by Bowers and Seashore to be responsible for lack Of varia-
tion in choices on the criterion of best ideas, or emer—
gence of leadership from the group.

Another relevant characteristic Of these groups
is the nature Of the role of resident advisor. As differ-

entiation will be measured in the interpersonal domain

29

(as Opposed to differentiation Of Objects, for example)
these subjects, due to the interpersonal nature Of their
role, would be expected to have more experience in this
domain and therefore would have higher differentiation
scores. Therefore, a smaller range in differentiation
scores would occur among the resident advisors than among
a non-resident advisor college population. Crockett's
(1965) discussion of experience and level Of differentia-
tion lends support to this prediction.

A review Of the literature on studies concerning
cognitive differentiation indicates no article has reported
the distribution of differentiation scores for the subjects
studied. Thus no distribution is available for comparison
with the present pOpulation to determine how these subjects
differ from other subjects.

In this study, for reasOns given above, differen-
tiation is measured using Tripodi and Bieri's 1963 modi—
fication of the Kelly Role Construct Repertory Test. This
test can fit the data with some modifications which will
be presented following a description Of the 1963 test.

The modified Repertory Test (1963) measures the
degree to which an individual differentiates traits. This
test provides the subject with a 10 x 10 matrix, consisting
Of ten columns Of role types (mother, father, boss, etc.)
and ten rows Of bipolar traits (outgoing-shy, calm-

excitable). The subject is asked to rate the individual

30

who correSponds to each role type for him on each set Of
bipolar traits, using a six-point Likert scale ranging from
-3 to +3, so that each of the one hundred boxes in the
derived matrix receives a score.

Total scores are calculated by comparing the simi-
larity of ratings in every row of traits with the ratings
of every other row of constructs. Every row is compared
with every other row and a score Of l is given to every
instance of exact agreement between the two ratings on any
one person, and these are added. Each rating is compared
with every other rating once, i.e., the rating in row 1
(trait 1), column 1 (person 1) is compared tO row 4,
column 1 but not vice-versa. Rows are compared with each
other on one person at a time so that there are subtotals
for each person rated.

The highest possible score (if each person rated
receives the same rating--subtotals would equal 45, times
10 peOple rated equals 450) is 450 and the lowest, assum-
ing no omissions, is 40 (since there are six alternatives
and 10 ratings to be made in each row, four must be
repeated). Within this range, lower scores indicate
greater differentiation.

In this manner, the degree to which an individual
differentiates traits is measured by the test. ~If an
individual were to always rate persons identically on two

different traits, one could assume that he was making no

31

differentiation at all between these two dimensions.
Therefore, the test examines the extent to which an
individual uses dimensions discriminantly. Chance agree-
ment of ratings on two different traits is not considered
in this test. Previous research employing this measure
has not considered the effect of chance agreement on
scores Obtained by this method of measurement.

In the study reported here, the comparisons—Of-
pairs-Of-responses aspect Of this measure is applied to a
questionnaire consisting Of one sheet per person rated
containing 26 items (Bales, 1970), each requiring a rating
of the person on a trait (see Figure 1). For purposes of
this study, the terms trait, dimension, construct and item
will be considered equivalent terms. Bale's items are
designed to measure personality types associated with
participation in groups. His items are, in general, Of a
task or socioemotional nature compared to Bieri's mainly
non-task, interpersonal items.

Overlapping items in the two measures include,
listing Bieri's (see Figure 2) and then Bales's items
reSpectively, independent-feels independence is important,
calm-calm and understanding, outgoing-extrovert, and to
some degree interested in others-personally involved in
the group and considerate-believes equality and humani-
tarian concern for others is important. A close examina-

tion of the items on the two tests reveals a fair degree

32

FIGURE 1

BALES 1970 QUESTIONNAIRE

MEMBER CODE NUMBER PHASE II CIBBA STUDY

INSTRUCTIONS: IN MAKING THESE RATINGS, PLEASE CHECK TO MAKE SURE THAT YOU ARE
THINKING ONLY ABOUT THE PERSON WHOSE CODE NUMBER APPEARS AT THE TOP OF THIS
SHEET. CIRCLE THE BEST ANSWER:

 

1. How well do you feel you know this person?
INTIMATELY VERY WELL FAIRLY WELL NOT VERY WELL HARDLY AT ALL

2. How physically attractive do you feel this person is?
EXTREMELY ATTRACTIVE QUITE ATTRACTIVE AVERAGE UNATTRACTIVE

. Is his (her) rate Of participation generally high? YES NO
. Does he (she) seem personally involved in the group? YES NO
. Does he (she) seem likely to be rated highly on "leadership?" YES NO

3
4
5
6. Is his (her) rate of giving suggestions on group tasks high? YES NO
7. Does he (she) make inhibitory demands and want to enforce discipline? YES NO
8. Does he (she) seem dominating? YES NO
9 Does he (she) receive a lot Of laughter? YES NO
10. Does he (she) seem very extroverted? YES NO
11. Does he (she) seem warm and personal? YES NO
12. Does he (she) arouse your admiration? YES NO
13. Is his (her) rate of giving agreement generally high? YES NO

14. Does he (she) seem to stand for the most conservative ideas and beliefs of
the group? YES NO

15. Does he (she) always seem to try to Speak objectively? YES NO

16. Does he (she) seem to feel that his (her) individual independence is very
important? YES NO

17. Does he (she) seem to feel that others are generally too conforming to
conventional social expectations? YES NO

18. Does he (she) seem to reject religious belief generally? YES NO
19. Does he (she) seem to make others feel they are entertaining, warm? YES NO
20. Does he (she) seem calm, understanding? YES NO

21. Does he (she) seem to believe that equality and humanitarian concern for
Others is important? YES NO

22. Does he (she) seem very introverted, serious, shy, introspective? YES NO
23. Does he (she) seem to feel anxious, fearful of not conforming? YES NO

24. Does he (she) seem to participate only when others ask him (her) for his (her)
Opinion? YES NO

25. Does he (she) tend to believe that others dislike him (her)? YES NO

26. Does he (she) show many signs Of tension and passive resistance? YES NO
27. Does he (she) seem to be appealing for understanding? YES NO

28. Does he (she) tend to devaluate himself (herself)? YES NO

E4-

2+

1+

BIERI AND TRIPODI 1963
MEASURE OF COGNITIVE DIFFERENTIATION

33

FIGURE 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l. Yourself
2. Person you
dislike
3. Mother
4. Person you'd like
tO help
5. Father
6. Friend of same sex
7. Friend Of Opposite sex
(or Spouse)
8. Person with whom you
feel most uncomfortable
9. Boss
10. Person difficult to
understand
4.-
H- w
s |
n
w. P- m
s s o H H
n a O m o m
m m s m s' m
H 'U m '8 m n a. O O
m m w. m m m m s
w s a s H a. o ~4. n
r? O: (D m H: O P' £3 £0 +
P' m H P' C P' 9’ In W 0 N
s s m U‘ P‘ n gt P- n w-
m n n H < m a
m m a. m n. a
O
H
m
+
H
I
H
w- m
P- P- F‘ m a
a a H H ,1 P-
m o H H. m g. H
O O m 5 x m
a m a m g. n m a
G a. m 'U P- o .4. m
H P- o m n P; c 5 u
F‘ O a. a H 0 g, m m -< N
a o m m H P- n
n H P- a. 8‘ H- < m
91 0‘ m O O»
n F‘ O.
O m
I
w

 

34

of comparability; however, it should be noted that to Kelly
and Bieri content Of items is unimportant in the measure-
ment of cognitive differentiation.

The number of response alternatives differs on the
Bieri and Tripodi measure and the Bales-based measure
employed in this study. The alternatives of Bieri and
Tripodi are continuous, while the alternatives on the.
Bales-based measure (the measure employed here) are
dichotomous. The response alternatives of Bieri and
Tripodi do fall into positive and negative categories
(+3, +2, +1, -1, -2, -3) and in this way the response
categories could form a dichotomy if grouped by sign.
Positive corresponds to "yes" and negative corresponds
to "no" in the present test.

One could logically expect that to the rater, the
polar positive or negative choice would precede the choice
of a gradation number within the + or - side of the re-
sponse scale. This would lead him to treat the alterna-
tive initially as a positive-negative dichotomy. Scale
logic, however, assumes the choice is made on a continuous
basis. .

Further, in comparing two versus six response
alternatives in measuring cognitive differentiation, there
is no one right way. The important part Of the measure is
taken by Bieri and Tripodi as the comparison Of pairs Of

reSponses, used to determine independence of dimensions.

35

This procedure is equally applicable to two or six re-
Sponse alternatives. Because there are two alternatives on
the present measure instead of six, similarity Of two item
responses in rating a person is much more likely to occur,
and therefore higher differentiation scores will be
obtained.

Other differences between the present measure of
COgnitive differentiation and Bieri and Tripodi's measure
include the number of dimensions upon which a person is
rated (26 versus 10), number Of people rated (a number
ranging from 6-16 versus 10) necessitating equalizing for
the effect of number of people rated on the differentiation
score, and the identity Of peOple rated (peOple in peer
roles versus people in authoritative roles). Another dif-
ference is in the method Of calculating the differentiation
score. Rather than comparing each item by each other item
for a person rated to determine response similarities as
is done in the Bieri test, the number in each response
category is determined and entered into the formula
n(n-l)/2 to determine the number of co-occurrences Of

similar responses, or pairs.*

 

*In addition to calculating differentiation scores
by comparing across items for each person rated, a
differentiation score was computed based upon responses
given to one item compared across all peOple rated. The
characteristics of this measure have not been explored.

36

The two possible response alternatives are yes
and no, and a yes is coded l and a no is coded 0. The
formula gives the total possible pairs among n Objects,
in a symmetrical matrix. One-half of the symmetrical
matrix and the main diagonal are excluded. N here refers
to the number Of ones and the number of zeroes, each
entered separately into the formula so that the number of
pairs Of "1" responses and the number of pairs Of "0"
responses are determined separately (see Figure 3).

The number of "1" and "0" pairs in each column
(for each person rated) is calculated and these subtotals
are added and divided by the number of people rated to get
a mean differentiation score for the subject. The higher
the score, the less the differentiation.

The lowest number of combined pairs (number Of
one and zero pairs combined) occurs when there are half
ones and half zeroes on the items--13 ones and 13 zeroes
on 26 items. In this case the number of one pairs is
13-1 x 13/2 or 78 and the number Of zero pairs equals the
same. The combined total becomes 156 for the lowest pos-
sible differentiation score. Results using this method
and results employing Bieri and Tripodi's method derived
from the same hypothetical data matrix were found to be
identical. The same number Of pairs was Obtained through
using the n(n—l)/2 method and using Bieri and Tripodi's

method of item-by-item comparison. Thus the step of

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

37

FIGURE 3

MEASURE OF COGNITIVE DIFFERENTIATION

High rate of participation? Yes No
Personally involved in group?

Is leader?

Gives suggestions?

Makes inhibitory demands and wants discipline?
Dominating?

Receives laughter?

Extroverted?

Warm and personal?

Arouses admiration?

Gives agreement?

Conservative?

Speaks objectively?

Independent?

Feels others are too conforming?

Rejects religious belief?

Makes others feel they are warm, entertaining?
Calm and understanding?

Believes in equality and humanitarian concern?
Introverted?

Anxious, fearful of not conforming?
Participates only when asked?

Believes others dislike him (her)?

Tense and passively resistive?

Appeals for understanding?

Devalues self?

I-‘
N
w
A SUOSIad

0‘ 93393

7 8 9 10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

0

 

l3 ones (13 x 12)/2=78
l3 zeroes (13 x 12)/2=78
subtotal- 156

 

38

item—by-item comparison is unnecessary, simplifying
computation of the differentiation score through use Of
the n(n-l)/2 formula.

In the present measure, the number Of pairs Of
one responses and the number of pairs of zero responses
in the rating Of a person are negatively correlated
(r=-.80) because the number of zeroes going into the
n(n-1)/2 formula is determined by the number Of ones going
into the formula and vice-versa. The number of pairs of
a type Of response is determined by the number of that
type Of response. The negative correlation occurs because
ones and zeroes (yeses and nos) are the only possible re-
Sponse alternatives other than no response (blanks). Since
there are 26 items, the number of ones and zeroes must sum
to 26. Therefore, the more the ones, the fewer the zeroes
and vice-versa.

As mentioned above, task leadership is measured
sociometrically in terms Of the number of choices on the
item best ideas a person receives. Persons in the group
were asked to respond to the question "which three members
usually come up with best suggestions and ideas?" Number
Of choices (including first, second and third) for each
person is then divided by the total number Of votes cast
in the group to standardize for differences in group size

to allow for cross-group comparison.

39

Familiarity is measured by a question asking "how
well do you feel you know this person?" The range Of
familiarity ratings is from one to five. The familiarity
scores for each person rated is recorded.

Agreeableness (or disagreeableness) can be
measured from the same questionnaire items used to measure
cognitive differentiation; therefore the cognitive differ-
entiation test is taken to measure more than the variable
Of cognitive differentiation. A "yes" response is taken
to indicate agreeableness, regardless of item content (as
previously stated, yea-saying and nay-saying are used by
Couch and Keniston to measure agreeableness and dis-
agreeableness). The ones and zeroes (yeses and nos) that
go into the calculation Of the differentiation score also
are calculated as a measure of agreeableness and disagree-
ableness.

To measure agreeableness and disagreeableness, the
number of pairs Of ones and zeroes (which are directly
determined by the number Of ones and zeroes given by the
subject scores all people he rates) are calculated sep-
arately and not combined into one score as they are in
measuring differentiation. The number Of ones is the
measure Of agreeableness, while the number Of zeroes is
the measure Of disagreeableness. The larger the number,

the more agreeable or disagreeable is the person.

40

TO determine if content Of dimensions affects
differentiation Of these items, that is, if cognitive
structure exists independently of content, it is first
necessary to establish clearly which items are Of task
or a socioemotional nature. A factor analysis (principal
components, rotated with eigenvalue cut-off)* was con-
ducted on the 26 items. Items loading high on factors
interpreted as task and socioemotional were selected
(4 of each).

Task items include rate Of participation in the
group, personal involvement in the group, leadership, and
giving suggestions. Socioemotional items selected are
"is warm and personal, arouses admiration, makes others
feel they are warm and entertaining, and is calm and
understanding." Differentiation scores were calculated
separately for each Of these sets Of items in order to
determine if both sets Of items are equally differentiated.

Pearson product moment correlations were computed
between differentiation and task leadership, differentia-
tion and familiarity, differentiation scores on the two
item sets, and agreeableness and task leadership. The
distributions for the variables of familiarity and
differentiation on the task set approximate normality.

The distribution of differentiation approximates normality

 

*
Factors with eigenvalues greater than 1.0 were

used, employing Varimax rotation.

41

and 90% of the subjects lie within a lO-pOint range on a
scale of 156 to 325. Distributions for agreeableness and
differentiation on the socioemotional set of items are
slightly skewed (positively), while the distribution for
best ideas or task leadership is very positively skewed,
which would attentuate somewhat the correlation between
differentiation and task leadership, and agreeableness

and task leadership.

CHAPTER IV

RESULTS

Differentiation correlates .07 (non-significantly)
with task leadership. Thus the hypothesis that differen-
tiation correlates positively to task leadership is not
supported.*

The hypothesis that differentiation is influenced
by content of items rather than being solely a structural
concept is supported. The low, positive correlation
expected between differentiation scores on task and socio-
emotional item sets is found. The amount Of differentia-
tion on task content items correlates .11 with the degree
Of differentiation Of socioemotional items. In other

words, degree of differentiation on task items has little

 

*

Differentiation correlates .05 (non-significant)
with socioemotional leadership. The second measure Of
differentiation (computed across people) correlates -.22
(significant) with task leadership and -.12 with socio-
emotional leadership, thus more strongly supporting the
differentiation-task leadership hypothesis. It should be
noted that a high differentiation score means low dif-
ferentiation. Therefore the negative correlation between
differentiation and task leadership supports the
hypothesis.

42

43

relation to degree Of differentiation on socioemotional
items.

Familiarity was found to correlate -.23 with the
co—occurrence score, thus supporting the hypothesis that
familiarity correlates positively (read as follows) with
differentiation. This correlation reads the higher the
familiarity, the lower the co-occurrence score (the number
of pairs of similar responses). A low co-occurrence score
indicates high differentiation. Thus, the greater the
familiarity, the greater the differentiation. Correlation
ratio or eta is not significantly different from E for
these two variables, nor for any oftfluaother variable
combinations. Thus, there is no significant departure
from linearity among these variables.

Agreeableness can be shown mathematically to be an
inverse function of differentiation. The number of ones
(yeses) goes directly into the computations that determine
the differentiation score. The higher the number of ones
(the more the agreeableness as measured here), the higher
the number of pairs of ones, one Of the two components of
the differentiation score. Therefore, a higher co-
occurrence score results, or lower differentiation. It
will be recalled that the lowest co-occurrence score, or
the highest differentiation, occurs when there are 13 ones

and 13 zeroes. In this case there are 78 pairs of ones

44

and 78 pairs Of zeroes, or a co-occurrence score of 156
pairs for one person rated.

In a case where a rater gives 20 ones, indicating
he scores high on agreeableness, and 6 zeroes, there would
be 190 pairs of ones and 15 pairs of zeroes for a co-
occurrence score of 220, or low differentiation. A similar
score would be obtained if the rater scored high on dis-
agreeableness, or gave 20 zeroes. Thus the more agreeable
or disagreeable a person is, the less he differentiates
others.

The discussion of the relationship between agree-
ableness and differentiation leads to the conclusion,
implicit in the foregoing discussion of the measurement
of differentiation, that this differentiation measure is a
measure Of variance in reSponse alternatives chosen by
the rater. It is apparent that the greater the varia-
bility of response, the lower the co-occurrence score and
the higher the differentiation. Therefore, the present
differentiation measure was correlated with a measure of
differentiation (Crano and Bettinghaus, 1970) that is
computed in terms of variance of responses. Computing
Crano and Bettinghaus's measure in terms of the formula
for binominal variance, or the average pq/n across all
peOple rated by the subject (where p equals the number of
ones, q equals the number of zeroes, and n equals the

total possible responses or 26), a correlation Of -.94 was

45

obtained. The binominal variance formula was employed
to fit the measure to the two reSponse alternative mode
of this data.

As noted, a low co-occurrence score is reached
through greater variability Of responses. Thus a negative
correlation (the higher the co-occurrence score, or lower
differentiation, the lower the variance differentiation
score, or lower differentiation) indicates both measures
are variance measures. Both measures Operationalize
differentiation in terms of response variability. For
these measures, continually using all dimensions to
describe an object indicates lack of re3ponse variability
or low cognitive differentiation.

In a test Of the final hypothesis Of this study,
agreeableness was found to correlate -.16 with task
leadership, significant at the .01 level. Therefore, the
hypothesis that agreeableness correlates negatively with
task leadership is supported, replicating Bales (1950)
finding that giving agreement is negatively associated
with task leadership. Further, this finding supports the
use Of yea-saying as a behavioral diSposition rather than

as a measurement artifact.

CHAPTER V

DISCUSSION

These results have several implications for Kelly's
theory of personal constructs and the measurement of
cognitive differentiation. First, Kelly's view Of the
process by which people perceive others is simplistic. The
finding that differentiation of dimensions is influenced
by the content of the dimensions indicates that people are
more likely to perceive others in terms Of certain kinds
Of dimensions, i.e., all dimensions are not equally rele-
vant in perceiving others. Here, cultural or situational
factors present in the group can influence what kinds Of
dimensions group members consider relevant.

Further, perception of others changes over time.
Since differentiation is positively correlated with famil-
iarity, which is a function of time, the longer one knows
a person the more differentiated would be the perception
of the person. Time would be assumed to represent inter-
action, and greater interaction would result in increased
information about other persons. Greater familiarity

represents increased information, thus leading to greater

46

47

differentiation of the other based upon accumulation Of
information.

Kelly's one-motive image Of man is questionable
in view of the finding that agreeableness, a behavioral
disposition designed to gain social approval, is inversely
related to differentiation. The approval motive conflicts
with Kelly's motive to accurately predict others' behavior
and leads to a lack of differentiation upon which pre-
diction is based. Ample evidence of an approval motive
exists in the literature, meaning this motive's effect on
differentiation others is highly probable.

Finding that behavior designed to gain social
approval, i.e., agreeableness, can be measured in the same
test constructed to measure differentiation provides
grounds for questioning the validity of this measure.
Cronbach and Meehl (1967: 254) note that degree of con-
struct validity can be stated as the proportion of variance
in scores attributable to the construct variable. Since
this test clearly measures at least one other variable, the
prOportion Of variance in scores attributable to differen-
tiation cannot be very high. Further, because the test
measures variables other than differentiation, correlating
scores on this test with best ideas inka test Of the
hypothesis relating differentiation to task leadership

becomes problematic.

48

The importance of these measurement problems for
other Operationalizations Of differentiation, or for a
test Of differentiation such as Bieri and Tripodi's,
depends upon the degree to which agreeableness is also
included in the measure. Couch and Keniston (1960) state
that if agreeableness is a general personality variable, it
should permeate tests that use Likert-scale response
alternatives and those that use response categories that
are similar in meaning to agreement and disagreement,
i.e., two reSponse alternatives of yes and no, as used in
this study, or Of true or falée. Empirical support is
Offered showing that agreeableness permeates all types Of
reSponse categories.

Couch and Keniston also show that in a Likert-
scale response alternative situation, agreeableness is
uniquely highly associated with the response categories of
agree and disagree. Therefore, the effects Of agreeable-
ness and disagreeableness would be represented through an
increase in these two response categories, rather than
being distributed over all re5ponse categories. This would
have the effect of numerically increasing either the agree
or the disagree category relative to all other categories,
resulting in a greater similarity of overall responses.
The higher number Of similar responses-~agree reaponse-—

would result in more pairs of these reSponses, thus

49

raising the co-occurrence score calculated on the basis Of
Likert-scale response alternatives.

In this way, agreeableness would affect co-
occurrence scores computed in Bieri and Tripodi's test.
Thus the same validity problems would appear, although
probably not to the same degree.

It would seem likely that familiarity and content
Of items would influence differentiation measured by Bieri
and Tripodi's test in the same way as the measure in this
study is influenced by these factors. As the peOple rated
in Bieri and Tripodi's test are in roles Of varying social
distance from the rater, familiarity would not be a
constant. As noted above, the dimensions employed in
Tripodi and Bieri's measure are of varying content.

In conducting this study, several conditions would
have been ideally preferred. Groups in the general popu-
lation of college students or groups in the community
would have been preferred over the specialized pOpulation
of college resident hall advisors. A wider range Of co-
occurrence or differentiation scores could be Obtained in
the preferred populations as members would be expected, if
the co-occurrence measure is reliable, to have greater
differences in amount of interpersonal experience, thus
facilitating correlation of differentiation and best

ideas.

50

Groups with no official leader, i.e., totally
unstructured groups, would have better suited the
differentiation-task leadership hypothesis. Emergent
leadership would be more easily studied. Under conditions
where there is no Official leader, leadership is more
likely to emerge from the group. The distribution on the
variable of best ideas (strong, positive skew), while
representing all of the groups, indicates many peOple in
the groups received few choices. This distribution could
be interpreted as reflecting lack of emergent leadership
as everyone receives about the same number of votes for
best ideas--perhaps partly due to the presence of formal
leadership.

An alternative to studying groups without formal
leadership would be to test Kelly's hypothesis with the
formal leaders Of these groups by measuring the level Of
cognitive differentiation Of the formal leaders. It
could be argued that formal leaders occupy their position
due to their level of cognitive differentiation. It
could also be argued that the role of formal leader leads
to increased cognitive differentiation of the occupant.
Therefore in a design to test the first line of argument,
level of cognitive differentiation should be measured
prior to assuming the role of formal leader.

Using groups which are clearly task groups and

groups which are strongly socioemotional groups would

51

provide a better test Of Kelly's hypothesis relating
cognitive differentiation with mobilizing leadership. As
previously noted, Kelly's concept of mobilizing leadership
seems to apply to either task or socioemotional leadership.
The results Of this study indicate cognitive differentia-
tion as measured with Bieri's method does not correlate
with either task or socioemotional leadership (see
footnote, p. 42). However, in testing the differentiation-
task leadership hypothesis Of this study, it should be
noted that the groups, while predominately task groups, do
have some characteristics of socioemotional groups.
Relating c0gnitive differentiation with leadership in
straight task and socioemotional groups would provide a
better test of Kelly's hypothesis.

Utilizing straight task and socioemotional groups
would improve the clarity Of the assumptions in Kelly's
theory. First, members of the group would more clearly
perceive the task or socioemotional nature of the group,
which would lead to increased clarity as to the criterion
utilized by group members for assessing who has best
ideas. In the present groups where both task and socio-
emotional activities occur, whether the criterion for best
ideas refers to ideas relates to task or socioemotional
activity is less clear.

In general, the findings of this study provide

grounds for questioning the current conceptualization and

52

Operationalization Of cognitive differentiation. It can
be argued, based upon these data, that cognitive differen-
tiation is not a cognitive structural variable that can be
measured independently of the content of the cognitive
system. Further, current measures of cognitive differen-
tiation appear to tap more than the concept Of cognitive
differentiation, rendering problematic the testing of
hypotheses involving the concept of cognitive
differentiation.

Thus, the testing of the hypothesis relating cog-
nitive differentiation to task leadership is not definitive.
The test Of this hypothesis is a test Of the broader
personality-social structural theoretical framework in the
sense that cognitive differentiation is the personality
variable and the leadership role, defined sociometrically
as the number of nominations a person receives for having
best ideas is the social structural variable. As the
testing is not definitive, an evaluation Of the broader
personality-social structure theoretical framework in
terms of the demonstrated lack Of relationship between
differentiation and task leadership in small groups is

not feasible.

CHAPTER VI

SUMMARY

The primary Objective of this study is to test the
hypothesis, originating in Kelly's Theory of Personal
Constructs, that constructive ability correlates positively
with task leadership in small groups. Constructive
ability, the central term in Kelly's theory, is defined in
accordance with Bieri's use Of the term. Bieri, who has
tested Kelly's theory in research, defines constructive
ability in terms of cognitive differentiation, or the
number of independent dimensions an individual utilizes in
describing others. Cognitive differentiation is measured
here with an approximation of Tripodi and Bieri's modifica-
tion Of the Kelly Role Construct Repertory Test. Task
leadership is defined, with reference to Bales, as having
the best ideas in the group regarding task accomplishment.

The connection between constructive ability or
cognitive differentiation and task leadership posited by
Kelly is taken as the basis for the hypothesis. Using a
personality-social structure framework similar to current

theorists, Kelly sees constructive ability as the

53

54

individual personality variable which leads to selection

by the grOup of the individual to the role, or social
structural variable, of what Kelly terms the mobilizing
leader. Better constructive ability enables the individual
to, among other things, better perceive the ideas of
members Of the group, leading to his selection as mobil-
izing leader.

Certain assumptions underlying Kelly's theory are
also examined. Kelly and Bieri assume that cognitive
differentiation is not influenced by characteristics of
the interpersonal relationship. It is hypothesized that
cognitive differentiation is positively correlated with
degree of familiarity with the person and a desire Of the
differentiator to be socially agreeable. In addition,
Kelly's assumption that cognitive differentiation is
independent Of type Of dimensions is tested.

The groups utilized in testing these hypotheses
are groups composed Of a resident advisor and resident
assistants on each floor Of the Michigan State University
resident halls. The groups, including male and female
groups range in size from 6-16 persons. The total number
of groups is 33.

The findings of the study do not support the
cognitive differentiation-task leadership hypothesis.
This lack of support could be due to the formal leader-

ship in the groups as the hypothesis examines cognitive

55

differentiation in relation to leadership emerging from
among the group members, or to the fact that the groups
are not straight Bales-type task groups. Familiarity

is shown to correlate positively with cognitive differen-
tiation, supporting this hypothesis. The desire to be
socially agreeable is demonstrated to have an effect on
the measurement Of cognitive differentiation. Finally,
coqnitive differentiation is shown to be affected by the

type of dimensions in the person's cognitive system.

REFERENCES

Bales, R. F.

1950

1970

Bales, R. F.

1968

REFERENCES

Interaction Process Analysis. Reading,
Massachusetts: Addison-Wesley.

Personality and Interpersonal Behavior.
New York: Holt, Rinehart and Winston.

and F. L. Strodtbeck

"Phases in Group Problem-Solving." Pp.
389-398 in D. Cartwright and A. Zander (eds.),
Group Dynamics. New York: Harper & Row.

Berger, P. L. and T. Luckmann

1966

Bieri, J.

1955

1966

The Social Construction of Reality. New
York: Doubleday.

"Cognitive complexity-simplicity and
predictive behavior." Journal of Abnormal
and Social Psychology, 51: 263-268.

"Cognitive complexity and personality
develOpment." Pp. 13-39 in O. J. Harvey (ed.),
Experience, Structure and Adpatability.

New York: Springer.

Bieri, J. and D. Messerly

1961

Reported in J. Bieri, "Complexity-simplicity
as a personality variable in cognitive and
preferential behavior." Pp. 355-379 in

D. W. Fiske and S. R. Maddi (eds.), Functions
Of Varied Experience. Homewood, Illinois.

56

57

Bowers, D. G. and S. E. Seashore

1967 "Peer leadership within work groups."
Personnel Administration, 30: 45-50.

Carter, L. F.

1954 "Recording and evaluating the performance Of
individuals as members of small groups."
Personnel Psychology, 7: 477-484.

Cartwright, D. and A. Zander
1968 Group Dynamics. New York: Harper & Row.
Couch, A. and K. Keniston

1960 "Yeasayers and naysayers: agreeing response
set as a personality variable." Journal of
Abnormal and Social Psychology, 60: 151-
174.

Crano, W. D. and C. O. Bettinghaus

1970 "The generality of evaluative differentiation
across diverse attitude domains." Paper
presented at the American Psychological
Association annual convention.

Crockett, W. H.

1965 "Cognitive complexity and impression
formation." Pp. 47-90 in B. A. Maher (ed.),
Progress in Experimental Personality Research,
Vol. 2. New York: Academic.

Cronbach, L. J. and P. E. Meehl

1967 "Construct validity in psychological tests."
Pp. 243-270 in W. A. Mehrens and R. L. Ebel
(eds.), Principles Of Educational and
Psychological Measurement. Chicago: Rand
McNally.

Crowne, D. P. and D. Marlowe

1964 The Approval Motive. New York: John Wiley &
Sons.

58

Garner, W. R.

1973 Dimensions of cognitive structure. Guest
Lecture at the University of Michigan,
Ann Arbor.

Geier, J. G.

1970 "A trait approach to the study Of leadership
in small groups." Pp. 409-416 in R. S.
Cathcart and L. A. Samovar (eds.), Small
Group Communication. Dubuque, Iowa:
William C. Brown.

Gibb, C. A.

1970 "The principles and traits Of leadership."
Pp. 366-375 in R. S. Cathcart and L. A.
Samovar (eds.), Small Group Communication.
Dubuque, Iowa: William C. Brown.

Harvey, O. J., Hunt, D. E. and H. M. Schroder

1961 Conceptual Systems and Personality DevelOp-
ment. New York: John Wiley & Sons.

Hayek, F. A.

1942 "Scientism and the study of society."
Economica, 9: 267-291.

Heine, P. J.

1971 Personality in Social Theory. Chicago:
Aldine.

Inkeles, A.

1963 "Sociology and psychology," Pp. 317-387 in
S. Koch (ed.), Psychology: A Study of a
Science: Vol. 6. Hightstown, New Jersey:
McGraw-Hill.

Kelly, G. A.

1955 The Psychology of Personal Constructs. New
York: W. W. Norton.

Levanthal, H.

1957 "Cognitive processes and interpersonal
prediction." Journal of Abnormal and Social
Psychology, 55: 176-180.

59

Levinson, D. J.

1963 "Role, personality and social structure in
the organization setting." Pp. 428-440 in
N. J. Smelser and W. T. Smelser (eds.),
Personality and Social Systems. New York:
John Wiley & Sons.

Mann, R. D.

1959 "A review of the relationships between
personality and performance in small groups."
PsycholOgical Bulletin, 56: 241-270.

Mitchell, T. R.

1970 "Leader complexity and leadership style."
Journal of Personality and Social Psychology,
16: 166-174.

Newcomb, T. M.

1961 The Acquaintance Process. New York: Holt,
Rinehart and Winston.

Press, A. N. and W. H. Crockett

1969 - "Cognitive complexity and the learning Of
balanced and unbalanced social structures."
Journal of Personality, 4: 540-553.

Schroder, H. M., Driver, M. J. and S. Streufert

1967 Human Information Processing. New York:
Holt, Rinehart and Winston.

Scott, W. A.

1963 "Cognitive complexity and cognitive balance."
Sociometry, 26: 66-74.

Streufert, S. and H. M. Schroder

1965 "Conceptual structure, environmental
complexity, and task performance." Journal
of Experimental Research in Personality,

1: 132-137.

Streufert, S., Suedfeld, P. and M. Driver

1965 "Conceptual structure, information search
and information utilization." Journal of
Personality and Social Psychology, 2: 736-
740.

60

Thomas, E. J.

1968 "Role theory, personality, and the
individual." Pp. 629-727 in E. F. Borgatta
and W. W. Lambert (eds.), Handbook of
Personality Theory and Research. Chicago:
Rand McNally & Company.

Tripodi, T. and J. Bieri

1963 "Cognitive complexity as a function of own
and provided constructs." Psychological
Reports, 13: 26.

Tuckman, B. W.

1967 "Group composition and group performance of
structured and unstructured tasks." Journal
of Experimental Social Psychology, 3: 25-40.

Turner, R. H.

1969 "Role-taking: process versus conformity."
Pp. 215-230 in A. R. Lindesmith and A. L.
Strauss (eds.), Readings in Social
Psychology. New York: Holt, Rinehart and
Winston.

Vannoy, J. S.

1965 "Generality Of cognitive complexity-
simplicity as a personality construct."
Journal of Personality and Social
Psychology, 2: 385-396.

Wicker, A.

1969 "Cognitive complexity, school size, and
participation in school behavior settings."
Journal of Educational Psychology, 60:
200-203.

HICH IES

\\\11\l)fllfl(fi(fi)i)u\@i§@fl \\

 

31