*k";*>w

>■„

^a^rola

/

{ "

/

i-W^^ oy vwiTöMOi^oo-'r • tusrrsnBDRJKpnr o*5 s^sucN-oifä • UTR-äAI^A. ix*..

A COHSIDERATfC3^ ©»• TWO ÄSSUmPTIONS ÜNDSilYtNO BEDLEi'S CONTINGINCY MOÖEI FOi THS PR^ICriON 0£ LEADiRSHJP IPFiaiVENESS

MÄtTIN HSHESIN, i A L^OY. **© GSAC5 HATCH UNivERsrrv OF laiNots

TECHNfCM mmt- NO. 52 ifff-m MAY. ^^7 - " r • -

hsftt' Jn?fWt«{ fey Ss» Mnmc»« itwaKfc ^«{»d* Ageöty, ASCA Ofi«r {%. AS* y«d«f C*!* el Neni^ i«t»HjA Coalwa Hi V7~t72, Hew «3M{34}

I

DISTRISUTJOH OF THIS POCmMt IS UriUMITfD

RECEIVED

JIJU71967

CFSTi

D D n

i •V'

BEST AVAILABLE COPY

GROUP EI-i^CTIVfiNESS RESEARCH LABORATORY

DEPARTMENT OP PSYCHOLOGY

UNIVERSITY OF ILLINOIS

URSANA, ILLINOIS

A Consideration of Two Assumptions Underlying

Fiedler's Contingency Model for the Prediction

ot Leadership Effectiveness

Msrtin Fishbein, Rya landy

a»d

Grace Hatch

University of Illinois

Technical Report No. S2 (67-8)

May, 1967

.

CotBRunication, Cooperation, and Negotiation in Culturally Hstsr^gsmeous Groups

Project Supported hy the

Advanced Research Projects Agency, ARPA Ordar No. 4S4 Under Office of Naval Research Contract NR 177-472, Nonr 1854(36)

Fred E. Fiedler, Lawrence M. Stclurow, «sn4 Harry C. Triandis Principal Investigators

DISTRIBUTION OF THIS DOCUMENT IS UNLIMITED

A Consideration of Two Assumptions Underlying

Fiedler's Contingency Model for the Prediction

of Leadership Lffectiveness

Martin Fishbein, Eva Landy, and Grace Hatch university of Illinois

ABSTRACT

The present paper c«>r. best be viewed as an attempt to exploT'e two of the

basic assumpticiis underlying Fie<'.ler's (1964, 1965) Continf»8ncy Model:

(1) the assumption that different group-task situations "require a different

leader-group member interact ion>', i.e., "demand" differenc types cf leader-

snip behaviors; and (2) the assumption that these "demands" will covary

systematically with the three dimensions of the group-task situation specified

by the Contingency Model. One hundred forty-one male undergraduates rated

the way they believed the "Host Effective Leader " (MEL) ihould perform in

each of eight gi."'jp-task situations on a Behavioral Description Questionnaire.

The results indicated that although these ratings of the MEL's behavior did

vary across the different group-task situations, the ratings were significantly

influenced by only two of the three gioup-task dimensions isolated by Fiedler,

namely, the leader-member relations and the position power dimensions.

Additional hypotheses related to the Contingency Model were also investigated

and discussed.

This study v/as supported by the Advanced Research Projects Agency, under ARPA Order 454, Contract NR 177-472, Nonr 1834(36), Fred E. Fiedler, Lawrence M. Stolurow, and Harry C. Triandis, Principal Investigatcrs. The present study represents Technical Report No. 52 (67-8) of the contract. The authors are indebted to Professors Fred E Fiedler, Stanley M. Nealey, and Gordon O'Brien for their critics] readings of an earlier draft of this paper.

A Oonsideration of Two Assumptions Underlying Fiedler's Contingency Hodel for the Prediction

of Leadership Effectiveness*

Martin Fishbein, F-va Landy, and Grace Hatch Uni\'ersity of Illinois

In a recent series of papers, Fiedler (196U, 1965) has presented a model for

the prediction of group performance. As Fiedler points out, "The model is pre-

dicted on the assumption that the type of leadership behavior required for good

rroup performance is contingent upon favorableness of the group-ta.sk situation for

the leader," More specifically, Fiedler identifies three major dimensions of the

group-task situation; (i) the nature of the affective leader-member relations;

(2) the task structure; and (3) the leader's power position in the group.

The dimension of affective leader-group relations refers to the personal

"'ationship between the leader and key members of his group. According to Fiedler,

.s dimension reflects what "is probably the most important single determinant of

group processes which affect team performance. The liked and respected leader can

obtain compliance from his group under circumstances which, in the case of a

disliked or distrusted leader, ^ould lead to open revolt."

The task structure dimension refers to the clarity or ambiguity of the task.

Here one may distinguish between highly structured, unambiguous tasks where the

leader and his group members kno'v? exactly what needs to be done and the way to dc

it (e.g., a missile crew performing a count-down) and unstructured, ambiguous tasks

where neither the leader nor the members can readily specify the manner in which

such a task is to be executed (e<.g., planning a program for a picnld), Fiedler

views the task structure dimension as the second most important determinant affect-

ing team performance.

The final dimension of "leader position power" refers to the degree of formal

or informal power inherent in the leadership position. Thus, a leader with high

-2..

position power is one who can utilize rewnrds and sanctions, ai>d »ho has authority

over hia men that is supported by the organization within which the group operates;

& leader with low oosition power is one who essentially is restricted to using

persuasion and other indirect means of influence»

By dichotomizing each of the^e dimensions, eight distinct types of group-task

situations can be identified as follows:

Affective Leader-Herabej' Task Leader's

Relations Structure Position Power

1 Good Structured High 2 Good Structured Low 3 Good Unstructured High

U Good Unstructured Low S Poor Structured High 6 Poor Structured Low 7 Poor Unstructured High 3 Poor Unstructured Low

Two points about these eight group-task situations should be noteds

(1) According to Fiedler, tnese situations vary ^long a continuum of favor-

ableness to unfavorableness for tne leader. That is, these situations differ with

respect to the degree to which they permit the leader to "influence and control his

group members." Hare specifically, the leader is seen as having maximum influence

over, and control of, his group members in group-task situation 1, and mini.nal

influence and control in situation 8.

(2) Each of these group-task situations i,my "require a different leader-group

member interaction." (Fiedler, 196ii). That is, these different group-task situ-

ations may well demand different types of leadership behavior in order for the

group to operate ?t maximum efficiency. Thus, for example, one type of si*uation

may require a permissive, .iOn-directive, considerate type of leader, while another

group-task situation may require a controlling, managing, directive type of leader.

Fiedler (196u) has argued that such is indeed 'he case.

-3-

One cf the major variables investigated by Fiedlor and his associates has been

"the leader's esteem for his least preferred coworker (LPC)," There is a consider-

able amount of evidence indicating that High LPC leaders a^e more effective than

Low LPC loaders in certain group-task situations; while Low LPC leaders are more

effective than High LPC leaders in efferent types of group-task situations. In

his earlier writings, Fiedler (e.g., 1958) viewed the High LPC leader as a person

who is permissivej non-directive, and considerate} and the Low LPC leader as a per-

son who is controlling, managing, and directive. More recently however, Fiedler

(196U, 1966b)has emphasized a motivational, rather than a behavioral basis for

distinguishing between (or describing) High and Low LPC leaders, itore specifically,

Fiedler now views the High LPC leader as a person who "obtains need satisfaction or

reinforcement as a consequence of having experienced success in interpersonal rela^-

tionsj" while the Low LPC leader is viewed as an individual who "obtains his need

satisfaction or reinforcement through his achievement.(or participation) in assigned

group tasks." Thus, although a High LPC will, under normal conditions, tend to be

more porndssive 'p.d considerate than a Low LPC leader, this is not necessarily the

case in all situations» That is, in any given situation, the High LPC leader's

motivation for achieving satisfying interpersonal relations may "cause him to

behave" in a more directive, Controlling way than a Low LPC leader. Similarly,

the Lew LPC leader's motivational structure »uay "cause him to act" in a permissive,

considerate manner. Thus, although High and Low LPC leaders are still viewed as

behaving differentially in a given situation, the specific types of behaviors they

display will vary as a function of their underlying motivational orientations.

To summarize briefly, the contintency model nay be seen as an attempt ta tie

together, and integrate, the findings of twelve years of research which sugg&s^

that different types cf leaders (i.e.. High LPC vs. Low LPC) are differentially

effective in different types of group-task situations. In particular, Fiedler

-u- hypotheaizes that in aitufitiona that are very favorable or very unfavorable for the

leader, a Low LPC leader will be most effective; while in situations ^hich are

nsodcratoly favorable or moderately unfavorable for the leader, a High LPC leader

will be raost effective. Several validation studies (e,g., Fiedler, 1966a, Shaw and

BIIM, 1966) have provided support for this as well as other hypotheses generated

by the contingency model.

The present paper can best be viewed as an attempt to «"-xplors two of the basic

aastaqptions underlying the contingency model, Itore specifically, Fiedler's concept-

ualization assumes that High and Low LPC leaders are differentially effective in

different group-task nituations because (a) thede situations call for different

kinds of leadership behaviors for maximally effic'^nt group performance, and (b)

High and Low LPC leaders differ with respect to leadership styles and orientations,

and thus one type of leader (e.g,, a High LPC leader) will better meet the ,!deniandc"

of a gi^en situation than will another type of leader (erg,, 3 Low LPC leader),

PurtheTj it should be recalled that from Fiedler's point of view, these situational

demands should covary with the threa dimensions of the contingency model (i.e.,

affective loader-member relations, task structure, and leader's position power},

lore specifically, Fiedler has hypothesized thäc these three dimensions are of dif-

ferential iiiportance. That is, the kinds of leadershir behaviors that are required

for maximal group effectiveness should vary most with the nature of the affective

ieader-mamber relations, next most w:.th the task structure, and least ,d.th the

leader's position power.

Thus the major purpose of the present paper is to investigate the following two

assumptions:

1« The assumption that different group task situations "require a different

leader-group member interaction," i.e., "de.mnd" different types of leadership

behaviors; and

-5-

2. The assumption tha^ these "demands" vd.ll covary systeinatically with the

three dimenDiona of the contingency model.

In addition, an attempt will be made to explore one possible reason for the

differential behaviors of High and Low LPC leaders. That is, although Fi8dler(l966b)

has presented evidence that High and Low LPC leaders do indeed behas; ^ differenti-

ally in given situations, ihs determinants of this differential behavior have not

Seen .'lade explicit, l/hilo Fiedler has argued that the basis of this differential

behavior is the different motivationa.'' orientations of High and Low LPC leaders,

one may raise the question of whether these motivational differences are directly

reflected in behavior or whether they operate through an intervening variable.

That is, it may vr&ll be that High r-nd Low LPC leaders behave differently because

(a) they perceive the "demands" of a given situation differently and (b,v they a«t

in accordance with their perceptions. Alternatively, High and Low LPC leaders

could (a) perceive the '"demands'* of a given situation similarly, but (b) have

characteristically different modes of responding to the same sitaation.

Thusj a second nurpoes of the present study is to investigate the ways in

which subjects who vary in LPC perceive the "demands" of the eight group-task

situations. Although the second alternative (i.e., that subjects differiug in LPC

perceive the same "demands") aeems more consistent with Fiedler's position, it

should be noted that both of these alternatives are consistent with the general

nction that "igh and Low LPC leaders differ with respect to leadership styles and

orientations. Further, it should be noted that these two alternatives are not

mutually exclusive, and it is popsible that both occur simultaneously.

To summarize briefly, the purpose of the present paper is twofold: (l) to

investigate two assumptions underlying Fiedler's contingency Model and (2) to

explore the possibility that subjects differing in LPC differentially perceive the

-6- ,ldallnand5,, of £ given group-task situation, S^nce the specific hypotheses to be

tested in the praseut paper will be more clearly understood ii the reader is famil-

iar witn the type of data that were obtained, let us first, turn to a consideration

of the methods and procedures used.

Hethoda end Procedures

The subjects wore lUl male undergraduates who participated in the experiment

aa part of e course requirement. All subjects were teated simultaneously during a

one hour session. 'Jpon entering the experiments] room (M large pnditorium), sub-

jects were randomly assigned to seats. After all the subjects were seated,

questionnsira booklets were passed out. Each booklet contained the following!

1. A standard form of Fiedler's LPC Scale, This instrument consists of 2$,

eight-place bipolar adjective scalus, in the Semantic Differential format. The

subjects are asked to

"thirK of your least preferred coworker~that is, think of tho one person you have had the most difficulty working with, and rate that person on the following scales. Remember, we are not necessarily asking you to think of ohe person you likea the least, but the one person you have had the most difficulty working with,"

2, A handout dascribing the three dimersions of the contingency model, with

examples representing the endpoints of each dimension. The handout was not

attached to the booklet, and thus the subjects could refer to it at all timee. The

handout is reproduced in Table 1.

3« Each of the following eight pages of the booklet contained a Behavioral

Description Questionnaire (BDQ). The BDQ is based directly ^n Bales' (1951) method

of Interaction Process Analysis. Each questionnaire consists of twelve items, each

representing one of Bales' interaction process categories. That is, each item

describes c different type of behavior (e.g., "helps to clarify the situation by

providing useful information") and the subjects are asked to indicate the degree to

6a TABI£ 1

Jescriptions of Group Task Situations

On each of the following pa?es, you will be given a description of a certain group task situation, These sroup tayk situations will be described in terms of three dimensions, ntmrl/i (1) the affective leader-group relations;! (2) the task structure; and (3) the power position of the leader. Follovdng is a list of short explanations of the term we have US'KL

AFFECTIVE LnAPER-GROTT RMATIOrS. Refers to the personal relationships between the leader and the me:ibers of his group.

Good affective r^lations refers to a 3i+nation where the leader feelü, and is, accepted and TlTTad by the members of his group.

Bad affective relations refers to a situation where the leader feels, ^nd is, rejected and "disliked by the members of his group.

TASK STRUCTURE. Refers to the degree of clerity or ambiguity of the task the group is wording on,

A highly structured task refers to a situation where the task the group is working on is a specific one — the goal is clear,, and it can be reached by a definite procedure. That is, there is one correct solution to the problem the group is working on, and there is one correct way of reaching the solution« Further, once a solution has been reached, or the task has been completed, the correctness of the solution can easily be seen.

A hignly unstructured task refers to a situation where the task and goal of the group are vague and unspecific. There is more than one correct solution, and more than one way to reach a solution. That is, there is no specific manner in which to execute such a task. Further, once the task has been cenw pleted, there is no precise way of knowing whether the method used, or the solution itself, was the best one.

POWER POSITION OF THE LEADH?. Hafers to the degree of control and power that the leader has over his members»

The leader has a high power position when he has a high degree of authority over his group," Further, this authority is supported by the organization within which the group works. His role as the group leader is independent of the members. He is appointed by f larger organization, ^nd is acknowledged as the leader by the members. Further, group members have to obey his instructions. He is expected to supervise and evaluate the work of the group members. He can punish or reward membere at his own discretion. For example, he can effect a promotion or fire or penalize a group member»

^e iSfäSE ^as i low power position when he has no rewards or sanctions at his Sisposai. " Re""has to influence the group mainly by persuasion. Further,, his role as the group leader is dependent upon the members. That is, the members could replace the leader if bh:/ so desired. Thus, he has relatively little direct authority over his group.

On each of the following pages, one group task situation will be described to you. For exa iple, one group task situation might be described as follows: The leader has a high power position, good affective relations, and a highly unstructured task, "

-7-

whlch a given stimulus person engages in the particular bohsvior. ilore specifically,

subjects respond to each item by checking an eight-place scale ranging from "very

trva" (8) +- ^svy untrue" (1). The BDG is reproduced in Table 2.

At the top of each page, one of the eiglt group-task situations was described

(e.g., "Given a group-task situation where the loader has GOOD AFFECTIVE RELATIONS,

where the group is working on a HIGHLY STRUCTURED TASK, and where the leader has a

LOW PUW35R POSITION"), and the cubjects were asked to describe the "Ibst Effective

Leader" in that situation, 'lore specifically, dch subject was told to

"think of the kind of situation described at the top of each page, and than indicate the way you think the most effective leader would behave in that situation. That is, in that situation, what kind of leadership behaviors do you think are necessary if the grcup is to ooerate at maxiffium efficiency and be maxiiually productive. Reiüember, we v.-ant you to tell us the way you believe the most effective leader would behave and not necessarily the way that a person you would like would behave."

The eight pages (one for each of the group-task situations) were randomized

within each booklet, and two orders of presentation were us^d to describe the group

task sitaatioi.3. For hali the subjects, the group-tadk situation was described in

terms of (1) affective relations (2) task structure, and (3) position powsr. The

order was reversed for the remaining subjects. Since neither the order effect nor

any of its iateraetionj were sig lificant in any of the ana?.yses conducted, it will

not be re.rerred zo in the present paper.

To summarize briefly, an LPC score and twelve judgments of the I-bst Effective

Leader' (IZL) in each of the eight group-task si^ lations was obtained from each sub-

ject. Rather than considering each of the twelve behavioral judgments independently,

four scores were computed for each group task situation. Following the work of

Bales (19$1). the twp'vt specific behaviors were viewed as measures of two types

of interpersonal behaviors (positive -nd negative s<~ ioemtional behaviors) and two

types of task oriented be'..aviors ("giving answers" and "askfng quos .ens").-. Thus

7a

TABLE 2

The Behavioral Description Questionnaire

H'ven a grouo-task situation where the leader iias a HIGH PCWift POSITION, and whe: were are A HIGHLY UNSTRUCTURED TASK AND GOOD AFFECTIVE RELATION'S,

THE xiOST EFFECTIVE LEADER:

1. Would help and encourage the other group members.

very truet ; ; : : :_ : : ; very untrue ~T~ 7 ~5 5 T" 3 2 1

2. Would laugh and joke a lot.

very truf ; : : : : : : :_ i very untrue

3» Would tend to agree wit1 other members' ideas and su^- 'Stions.

very true: : ; : ; : ; : ; very untrue ' 8 7 6 S h 3 2 1

k» Would give many useful suggestions to get the job done.

very true: :_ : : ; _:_ : ; : very untrue 676 5U321

5. Would freely express his own personal feelings and opinions.

very true:_ : : ; j_ :__ : : very untrue 876$li321

6. Would help to clarify the situation Ly providing useful information.

very true; ; :__ :_ : : : . ; very untrue ~T~ 7 " 6 "~$ IT" 3 2 1

7. Would often ask for more information about the task,

ve^y ixrue; : : : _:_ __; __:_^ „: very untrue ~T~ 7 ~Z T TT 3 2 1

8. Would often ask for th« opinions and feelings of others.

very true: : : : : : : : : very untrue ~n 7 T~ ~T"* "~r" "T"" "2 r~

9* Wculd often ask for suggestions.

very true; ; : : : : __: : ; very untrue "T" 7 ~T~ 1 T" 3 2 1

10. Would often disagree with the other members' ideas and suggestions,

very true: : : : : ; ; : ; very untrue T~ 7 T 5 IT" 3 2 1

11. Would often appear to be anxious and tense.

very true: : i : : : : : very untrue T~' 7 ~Z ? TT" 3 2

12.. Would often be antaßonistic and aggressive toward othe. members.

very trrs: : : ; __: :_ ; : : very untrtio -g— 7 —3 5— ■ h 3 2 !

-8-

the basic data obtained .frow each subject v/as his LPC score and eight sets of four

behavioral expectations. Since eaoh of the behavic, al expectations was obtained

by sunsning over three Judgnsnts, the ccores could range frojn 3 (it is completely

nntrue that the MEL would engage in this type of behavior) to 2h (it is completely

true that the HEL would engage in this type of behavior). These data were then

used to test the following hypotheses

1. Different group-task situations "deinand1' different kinds of leadership

behaviors for maximal group effectiveness, iiore specifically, it was hypothesized

that subjoctc' ratings of the way the most effective leader should behave would

vary across group task situations;

2. The different situational demands (i.e., ratings of the way the MEL should

behave) will covary with the three dimensions oi the contingency model. Specifi-

cally, these "denands" should vary most with affective leader-member relations,

next, most with task structure^ and least with leader position power; and

3c Subjects differing in LFC will differentially perceive the demands of the

various grovp tack situations, i.e., the xatings of the way the i-IEL should behave

will vary as a functions of the subject's LPC score.

On the basis of their LPC scores, the subjects were divided into three equal

groups of l|7 subjects each (i.e., a high, middle, und low LPC group. The basic

experimental design was a mixed analysis of variance with LPC serving as a "between1'

subject" main effect, and the three dimensions of the contingency model serving as

"within subject" main effects, A separate analysis was conducted for each of the

four types of behaviors»

Although the analysis of variance methodology i& often used only to deteirdr0

the presence or absence of a significant difference between groups or between

different levels of a given variable, it can also be used to estinute the strength

-9-

of obtained relatiDnships between independen' and dependent variables (e.g., see

Bolles and ilessick, 19$8j Fishbein, 1963} ililler, 1961; Triandis and Triandis,

I960, I96S'). ilore specifically, the analysis of variance technique is a procedure

that partitions the total variance (in the dependent variable) into Its component

parts. Thus, one can determine the per cent of the total variance that is associ-

ated with (i.e., under the control of) each of the independent variables and their-

Inters'cions. The more of the variance controlled by a variable, the »nore iinpor-

tant is the variable. However, since a corsiderable portic 1 of the total variance

is often a f'-mction of individual differences between subjects (including treatment

by subject interactions), errors of measurement, and other uncontrolled sources of

variation, a clearer picture of the relative importance of each independent variable

can be obtained if one considers the proportion of variance a-isociated with each of

the variables after individual differences and other uncontrolled sources of error

have been eliminated. That is, one may look at the per cent of controlled variance

associated with each independent variable as well as the per cent of the total

variance accounted for by each of these variables. Since the second hypothesis

presented above is primarily concerned with investigating the relative importance

of the three dimensions of the continf-.,ncy model as determinants of "situational

demands" placuJ on the .iost Effective Leader, the sunuaary analyses of variance

tables that will be presented will contain eitL.iaves of both the total and con-

trolled variance accomted for by each variable and its interactions.

Results and Discussion

The results of U-e four analyses of variance are presented in Tables 3 and U«

Mere specifically, t'ne analyses of ratings of the way the MEL should behave in a

positive and negative socio-emotional manner are presented in Table 3* while

-10-

analyses of ratings of the MEL's answer giving and question asking behaviors are

presented in Table l^t In these tables, it can be seen that, coisidered together,

Tables 3 and h about here

the four major variables and all their interactions account for between 7.6 per

cent and 17.8 per cent of the total variance in subjects' beliefs about the v:ay

"The Most Effective Leader" should behave. More specifically, LPC and the three

dimensions of ths contingency model control 17.6 per cent of the variance in

beliefs about the MEL's positive socio-emotional behavior, 11.1 per cent of the

variance in beliefs about his negative socio-smotional behavior, Q.h per cent of

the variance in beliefs about the MEL's engaging in "asking questions," and 7.6 per

cent of the variance in beliefs about his "answer giving" behavior. Although this

is clearly a significant and meaningful amoun* of variance to account for, it should

be noted that between 80 per cent and 90 per cent of t,he variance in subjects'

Judgments about the way the MEL should behave is essentially left unexplained.

Since only 2$ per cent to 35 per cent of this variance can be directly attributable

to individual di ferences, this clearly implies that other variables, relevant to

the phenomena under consideration, still have to be isolated» Keeping this in

mind, let us now turn to a consideration of the four major variables (i.e., LPC

and the three dimensions of the contingency model) and taeir interactions.

L LPC - Contrary to the third hopothesis, it can be seen that subjects dif-

fering in LPC do not have differential expectations about the way the MUL should

act in any of the eight group task situations. Indeed, LPC and all its inter-

actions with the three dimensions of the cotitingenc.v model account for less than

1 1/2 per cent of the total variance arid less '■-hat 12 per cent of the controlled

variance in all four analyses. In no case, does LPC or its interactions with other

10a

oj o I c o

0)

A ■H O

8 >

•H -P n5 oD V

«Ml

"I

pq

-Ct HO\{*-H-=rONOt-l

COUN H H

<5 iH vO -3cO r-4 <M CO O >- O O

XAOf-CMf^OJiHO

i-t C\( r-< UN O OJ O U

o

8 r-l

rn r^\ -O^—'

c~fn">OvfNCvjcMr-(i-(

OJ r- eg en

cgojoj^OOcviOO

HfJoOOOOO CVJCOCwOOJlAt^COO

IA-:' iXa^S^ H

r-CO CM C^- iH O r-( r-l vO CM CO ON , • • • • r-4 iH H r^ (SI

v V

r^ifnr-l NO OVAO ococo o o -^ \r\ir\,0 CNVO i-H H0O ONO O •-< *—' ^-'CJ OJ r^ -^

CMriCMH-^CNOCM OcnOHr^iAiHO r-i oo CM NO CJ

VfN f~( i-l ^J O NO C- ^ • ••••••< rJOOr-iXAö&t-iOOr*- r-l r-< H r-l rM

O

*

Q

f-Ot— ONCMi-tHCOONOOr-l -ScHr^v Or-t7N_3r- \A t"- P- _arO CM o^ r^\ C~- t*- rN CM H N—' .HN-«

ON^OOCMNOHHO NO -cr -^ CM NA CM H

NONOO NOHXAOOCOOOOO ^i-4JCMr-r-0>_3 H CM CNr-IIA^AvO "^OJ

OOVAr-flA^-VN^OT '*- NOONC^C^OONO -A 0-3tr\r-<HVf\CMOH ^OJr-IOJr-4r^r-:Hr^

i-t CM

fA M

W -P W w

Si o

•H

cd V5,1

H c o , -P

0)

> •H ♦^ •H . en co o OT| a,

CO OOcAPACOOOOCM-rr

r-l _3fA rAJ O CM O O ON NO rt H

CO O C^ CM CM

OUNrH

^O rH VN Vf\t»-CO

NO S

-5-aCMUNIA-^CM CM • ••••••• r^HOOOOOC o

2 O C^NO O (^ f^i vO NO O r^ O H NOo^OHr-iHOO XTv O. ^ r\ \A NO f^-t PA O

CMNOHOCMOOJOO O O O O O O o _i CM-ar\ANONO-3-3 o

^111 rA ON NO CO xrv iH A- H '»A

CM O t^ «A _d CM iH Ai u>

AJ

'—«NO CM HrAr-)(-lO(HHi-' ^ '4 ^ V J V W ./

i-4AjrAl>-r-tONOHC^ OOVAAJt^-HONOO ^JNOAJOOIAOO

NOAjr-corAOCMNor- (H rA CO A.' CO O r-l ^-'\A-^i-<UN \i\

N-'OJ

ONCOCAOOCOCACM N—CNI iH rH

cMOt*-_3oor>-*0 H CM CM r-i H

rAOrA NOXArAP--HONOr-tr- r-IOOiAr<->. CMCOV\ OIAIH^JCMC^NOUNON

XAXAf-cO fAO '.\INO t>- VA CM CO A) CO O -M CO CN-.d- r-l tA VA \A fA A'

-i NO C^ O rH vO t"- VN fAVA CM CM r-l

NO (A-^C^-IA^O CMNO H ON_^ONr-l-^CONO^Jr-l ONTO CA_a-a^AONCf\0 r-4 CM CA r-l r-l CM 1-4 CM

'•H O Al« CN. f-iHiHHHiHHH ^^ CM A» AJ CM CM CM NQ<=OCO0O00TOCOaOC>-

NO CAfArArAPAfnrAAJ ■d -^ fA CO Sm*

i~l --I '-> ^—^ O^HHHHHHHH ■—* ^—' N-" (OH

X c^cJScS •rt c_> •p • 0) H rf rH V, S ;3 OJ 3 5 -P OJ +> o

0 jg O x K x ^ 3 "■5 o o X

Pi co to n HCOCOCOOOO x

a ■H u »■ n « 3 _ o K ^3 ^ ä o M X X X Ü 3 -3 o

ü tl >— > f< C

0

to » -H +J O w W C5 > rj) « X M K nooo«XXX n CO X n

W +» CO -H Ho +> m o o W

Q) p.. t-, JJH ^ >_) ^.-1 »-I OJ C O 03 .-läääftiüüm

t/i co ä a) j«; -H rO ,0 coo « «H to w X X X H rt X X X X X X X rtXXXXXXX

Ü c c

•H H nj O J- ^ E-4 ß, < «T^ 04 <;

•rH CQ o <; «; oa ■a; ^<oao<!«3;cQ<j;H

1 5 5 ■p > > s ?. o

EH

10b

I

U u n

II H\ m

^ CO O O H f- O O

Cgt--HCOiHCMO\.J3

8

<5 M r^^r<n r"V3 Ov O f^-<-H vO O O s0r-IOrHrH--(OO rr. 00 M3 -3 (^ fV CM Ov o Jl x^i • - • « * • -.■-*♦ V • vS *•#•••• • • |5 a -SO-^ Utr~._3 o n o o o o oooooooo P— r«! CO CM vO sO sO rn. n

r^ ro vO"^-'' s»-y \A rH rH o M o -—■ ■«-»• **^ H

nted

an

d ^-4 p n a? Al i i J&sO o

H O 0> H iH r-l

rH O •

•H 0) 1 ̂ 1 t-t sa »-N CM rH rHrHrHrHOi-HHr-l O.

°^ M '\' • • VcNrS^V V v/ *■ 'V^H'V^ *

AS 5 »-< PH N^»

! S nj M

so ^s /-V »->. *^N ■—. H c<> <;

11 -^H ^ co r-cvjoojHf-iOi-i r^vO t- Ov 0-J-3r^ co CMP-VACOOCO O 8 *

g| i

• • <l

o cy O OO 00 VfNOVAVAO CO VD • •♦• ••«•

CM O iH O i-O .3 O IA rH CM CM rH r-i rH

sOOOr-U-NCMroCVCO c—vO r- 1-4 CKCO H0O CM P- H CM rH CM rH rH rH

1^ n ^^ —^ CM-r^ -if rW s^^ v—*

f\

wrJ

»-^/-N ^-s i OcvKO

« • '• * • • CO-d^H H O H fNrHf^cc OsCOCOsO C— r-- Os r^ -'■, r-i iTu O f-

■a t. Si \om\AO CMSO CM rH f^ rH VA CO O O • ••••«•

»O sO O _3 Os rH ro CM •

C

^^ \A<M CJ JV 5» CO

IAO f\ CO CM i-<

r-4CO CV4 f— ts- _rt _-f CM CM CM H sOOT^fr-lsOOssOO CVJOO^Vrit^-P-f—rH

CO OO r-H f- -3 <H rH r-x

.o o\ o\ ^—-* sOfnCMf^rHiHrHrH 00 * V v_^ ,H v^ rH CM

nS w *—'

|§ n^s y^.

rt O o r^\OHOM-lr-iOVA r-CMCu\sOfnNOiA o ^S ^1 •

l-f OOOOCOC\JOV\PH OrHrH-SOOCMO •

8 . «a: 00 \A CVi rH

> V XM* ^ -4

•H O ^■x ^^^^ *-\ *-s -P c es m SI

-^ r-H -cf NOcOO\ A4*

O CVt CM O .J H ^-^H^Hrr^OOCMO O CM o\\t\^iO\o VA o

Eff

e(

Var

L

IAOXA NO''»—'

O CM O O O O oododdod XArH g

« ^-^ '—* rH

■P -o 3i^ oo o

^-^ W 0) » r-HrHr-l-CtrHHrHH

$1 n <Hl r-( xacNj rH UNO r-l tAVA c • • • • , • •

o ■<< \' vfJCO •' CM r^v as r^\ rH Csl CM o u M^ e»> •*-*• Ä -p u P c c

o •H ,—-v *<—S *^N s~^ ,'~*. p Ü cd

> SI

vr. OM-< O CNH H O C O-S O OsC^rHOOrHCOUN COCOU\sO CMAVAO

ÄT3 \0 vo r^ CM CO as i-l O O rH r-t r-l dcOr-CM-äcMOOf^ OaOOCMOsor-sO ■P E -3 -TJ r-4 r-H r-l f^ C— CM rH r^ rH H H rH rH rH

(0 i —■ -3 V—i^ Sw*

^H < v-^

C « »—\ y—^^^ <-N • U\CO r^ COvO rH rH O O 0_30 OfACMOOsCMUNOs CM^JlA_3<^-3r~-XA (»N

Osr^iA * • <»

r", (M 00 • •••••

iH O O rH r-4 rH CM t^-^r-?c^-^r^so UNi-H rHsO r^

• • • •<*••• ^fOsCMDcnmiArH X/NOv-^r^MDlAi^CM

• "0 51 O r-< OS lAIAH rH ro C^- (\/ ^D 9 0) VTi ^ CO CVJ C^ -3 rH .j\r\_3t— CljOCO (->

•■a jr sO r^H ^^ O CM rH H i-4 rH rH oq P •^-^ r-4»-' rH

<H s»-/ %^ O 5H

O rf^N ^■■^^•■^ ^-N r-v

* . ^1 O c\iCO r^-r-rl H H r-( rH rH M _SC\iCMCMCMCMCMCM »OcOcoooaOooooco sörarnror^fAri-vr^

r- y +* 3 rn ao^^ C\J C C rH ON s-> CT\ rH rH H rH rH H rH H :i) o vv-^ V-'^^N v_^ n r-H

•H Ü n <~> ^-S'—s n co 3 t.

c *: CQ O w o — NM^SW^ co co co X

> 4> •H r_^ « oo to CO % lit P • Q) ^ O S g-s II O XXX

a. to co w a

0} P p « P O ^ L' 8^ ^ x

X 3 nj 3 o

>-) to CO CO O ü o X ■rl co a

Ü OH xxx x SS ä o

n bfl w > >H C >> c a) n w ^ ■PC X coCiOOXXXX M O O O XX X X

!0 CO ai P CO -H

p m o o en ^t-ja^caou« o a, OH PL, rH J f-H ►.) CQ O O (D

t: ^ w CO X) OJ >: -H JQ J3 «J: H

c^ c rt v. w to x « « « rtXXXXXXX rt X X X X X X X

Ü 0) o o

5^ rt o •H •rl

^s a fn d, <i; *< m ■< ^<4a3ü-<<i;m<!; H ^ CQ ü <; «a; CQ < 3 4i P > > > ts 5 ( ^

-11-

variables approach the ,01 level of significance. Thus it appears that leaders

varying in thsir Ssteem for their Least Preferred Coworker are not differantlaXly

effective in a given situation because they perceive the daiiands of that situiition

in different ways.

^' frro^P-Task Situations - The influence of the eight group)-task situations

as determinants of jud^.nents about the way the HEX should behave is sunanarized en

line $ of Tables 3 anc1 U. There it can be seen that the three situatlonal variables

and all their interactions account for approximately 90 per -isnt of the controlled

variance in all four analyses. Thus, almost all of the variance that is accounted

for is directly attributable to variations in the group-task situations. Clearly

then, as Fiedlar has suggested, these different situations do seem to "demand1*

different types of leadership behaviors. However, in order to better underctand

the differences between these situations, a consideration of each group-task

situation variable and its interaction is necessary.

\. Leader-ilember Affective relations. In all four analyses, it can be seen

that subjects expect the inost effective leader to behave in a significantly dif-

ferent manner when he has "good" affective relations with his group members than

when he has "bad" affective relations, ilore specifically, the id. is expected to

ask more questions, ^ive more answers, and display more positive and less negative

socio-eiotional behavior when he has good than when he has bad affective relations

with his group members. The mean differences for each of the four behaviors way

be seen in Table 5»

Table $ about here

In addition to influencing all four types of behavior, in Tables 3 and ^ it

can be seen that consistent with Fiedler's hypothesis, the leader-member affective

11a

TABLE 5

The Influence of Affective Leader-iiember Relations on Expectations about the Most Effective Leader's Behaviors

Good Affective Bad Affective Type of Behavior Relations Relations F

Positive Socioemotional 18.66 15.69 120.68

Negative Socioemotional 9.8U 11.98 71.35

Giving Answers 18.90 17.30 38.20

Asking Questions 18. U5 16,2h I;9.12

All F^ significant beyond the .001 level of confidence. The higher the mean score, the more likely the MEL will engage in the particular type of behavior. Scores can range from 3 to 2k*

-12-

rslation dimension is the most important singie deterirdnant of expectations about

the HEL's behaviors. That is, in all cases, the affective relatic is dimension

accounts for more than $0 per cent of the controlled variance in expectations

about the IlX'a behavior.

B. Task Structure. Contrary to Fiedler's expectations, the degree of the

task structure appears to play a relatively minor role in determining expectations

about the MEL's behavior. Even though task structure does have a significant-

effect on expectations about the iCL's positive socio-emotional behavior (i.e., he

is expected to show significantly more positive socio-emotional behavior in strv.c-

tured than in unstructured situations - Z ■ 17.51 and 16.8U, respectively,

p v ,001), it accounts for less than 0.6 per cent of the total variance and less

than 3.3 per cent of the controlled variance in any of the four analyses. Here,

however, it must be noted that the present analysis is concerned with subjects'

expectations about a leader's behavior in hypothetical situations and not with

actual behavior in real situations. Fiedler's model is based on an analysis of

these latter situations and thus the finding does not necessarily mean that task

structure is an irrelevant dimension in analyzing ingoing groups, nor that in a

real situation, leaders do not take the task structure into account in determining

their course of action. Tne Unding does suggest, however, that the task atructure

diiaension may not be as important as Fiedler has indicated, and certainly deserves

2 a closer and .lore critical analysis.

C. Leader's Position Power. As in the case with affective leader member

relations, it can be seen that subjects expect the most effective leader to behave

in a significantly different iianner when he has high position power than when he

is in a position of low power. This effect is significant beyond the »001 level

in all four analyses. More specifically, the MEL is expected to ask fewor

-13-

questions, give ncre answers, and to show jaore negative and .less pcsitivs socio-

envotional behavior when he has a high power position than when he has a low power

position. Table 6 presents the mean differences for each of the four behaviors.

Table 6 about here

In addition, it should be noted that the leader's power position is the second

most important determinant of expectations in :ill four analyses. As might be

expected, position power appears to account for slightly more of the variance of

task behaviors than of socio-emotional behaviors, ilore specifically, while podition

power accounts for 23.9 per cent of the controlled variance in expectations about

'feiving answers" and 20.3 per cent of the controlled variance associated with

"asking questions," it only accounts for 16.1 per cent of ohe variance of negative

socio'emotional behavior and lU.8 per cent of the controlled variance of positive

socio-emotional behavior. In other words, it appears that unlike affective rela-

tions which appear to have a similar influence on all types of behavior, the

leader's power position is most important in influencing his task-related behaviors

and less importsnt with respect to intarpersonal behaviors.

Affective Relations X Task Structure Interaction - This interaction is not

significant in any of the four analyses. Further, in no case does this interaction

account for more than 0.2 per cent of the total variance or more than 2.1 per cent

of tbs controlled variance. Tius the relative unimportance of the task-structure

dimension as a determinant of expectations about the HEL'a behavior is again demon-

strated.

Affective relations X Power Position Interaction - In three of the four analy-

ses, this interaction is significant beyond the .CCi level of confidence. Similar

to the power variable, the affective relations by position power interaction

appears to have differential influence on different types of behaviors, i'lore

13a

TABIE 6

The Influence of the Leader's Position Power on Expectations About the Host Effective Leader's Behaviors

Hi Position Lo Position TjQ)© of Behavj or

Positive Socioemotional

Negative Socioemcticnal

Giving Answers

Asking Questions

All F's significant beyond the ,001 level of confidence. The higher the mepn scorej the more likely the HEL «ill ongage in the particular type of behavjor. Scores can range from 3 to 2li.

Power Power F

16. U6 17.73 23.55

11. UP 10.33 2U.5U

18.70 17. U9 32.5U

16.69 16.00 19.0li

-lii-

i^e fically, although this interaction is ths third largest determinant of vari-

anct . expectations in the thrae a;.alyseo wuere it is sigmi'icjnt, it is con-

sidprably more important v/ith respect to interpersonal bciiaviors tl.an with respect

to taflk behaviors. That is, this interaction accounts for ii4.2 per cent ol the

controlled vra'iance in öxpeetation^ about the ii^L's positive socio-emotional

behavior, and 16.1 per cent of the controlled variance associated with negative

socic-emotionax behavior. In contrast to this, it on?y accounts for 7.1 per cent

of the cortrolled variance associated with asking questions and none of tiia variance

associated w.'/.h giving anfvers.

Table 7 about here

The means for the three significant interactions may to seen in Table 7, In

all three cases, it appears that uhan the leader is in a task situation having

good effective relations, subjects do not expect tha HEL to bahave differently

when he lias high '■•ower than when he has low power. However, in tnose situations

where tie leader has bad affective relations with his r^oöe.J, the HEL is expected

to behave quite differently depending upon his power position. >iore specifically,

when there are poor affective relations and the leader has a high power position,

he Is expected to ask fewer questions, and show „lore negative ?..iti less positive

socio-c wtional behavior, than when he has a low power position.

Task Structure X Power Pusi^on Interaction. ir4 Tables 3 and It, it can be

seen that this interaction reaches the .01 level of significance in only one of the

four analyses. Interestingly en^ jh, this one significant interaction is with

respect to the only type cf behavior that was not influenced by the affective rela-

tions X position power interaction—na;jely, giving answers. Again, however, it

14a

TABLE 7

The Influence of the Affective Relations x Position Povir Interaction or Expectations about the ifast Effective Leader's Behavior

Affective Leader-Meinber Relations

Po!7-"tion Power Good Bad Good Bad Good Bad

High 10.61 Hi. 30 9.86 13.11 18.18 15.19

Lew 13.71 17.08 9.82 10.8U 18.72 17,28

A, Positive SocJ.emctionel Behavior

B. Negative Socioemotional Behavior

C, Asking Questions

V - $0,2( i F - 30. 2h F = 13.10

All F's significant beyond the .001 level of confidence* Scores range from 3 to 2U. The higher the socre, the more likely the MEL will «n^age in the behavior.

-15-

muat be noted that even though this one interaction is significant, it •>iust be

considered skeptically since it accounts for less than ono-.ialf of one per cent of

the total variance, and only $ par cent of the controlled variance of expectatio.is

about the iEL's'^answer giving" behavior. Further evidence cf the relative unimpor-

tance of this interaction can be seen in Table 8, which presents the Cleans for this

interacticn. There it can be seen t!;at although the power variable has a large

Table 8 about here

influence on expectations about question answering behavior, task structure has

relatively little The interaction indicates that whor a leader has a high power

position, he is expected to answer slightly roore questions when the task is

structured than when it is unstructured. In contrast, when he has a low power

position, he is expected to answer slightly more questions when the task is

unstructured than when it is structured. Ketther of these differences, howevei,

are significant at the .01 level of confidence.

Affective Relations X Task Structure X Power Position Interaction. In Tables

3 and U, it can be seen that the triple interaction does not approach significance

in any of the analyses. Further, and perhaps more importantly, it car. also Le seen

that in no case does this interaction account for tiore than 0.1 per cent of the

total variance or more than 1 1/2 per cent of the controlled variance.

To sunnari-Le briefly then, although more than 90 per cent of the controlled

variance in e;:pectatior.s about the way the »EL should behave is attributable to

variations in the group-task situation, al lost all of this variance can be account-

ed for by only two variables and t^.eir interaction. lore specifically, affective

relations, position power, and t' e affective relations X position power inter-

action accounts for 90.9 per cent nf the controlled variance of positive

15a

TABLE 8

The Influence of the Task Structure x Position Power Interaction on Expectations about the Host Effective Leader's "Answer Giving" Behavior

Task Structure

Posit ;.on Power Hl^h Low

High 18.98 18.42

Low 17.27 17.72

F - 9.52, p .01

Scores can range from 3 to 2U, The higher the score, the more likely the MEL engages in "answer giving" behavior.

-16-

socio-emotional acores, li$,9 per cent of the controlled variance of ne^ativf1 socio-

enotional scores, 6S,h per cent of the controlled variance jf expectations about

the IlEL's "asking questions" behavior, and 79.6 per cent of the controlled vari-

ance of expectations about the ICL's "giving answers" behavior.

Thus while consistent with Fiedler's notions—different groap-task situations

do appear to "derand" different kinds of leadership behavior for maxunum group

effectiveness—only two of the three dimensions isolated by Fiedler appear capable

of consistently explaining a significant per cent of the variation in these

"demands." Further, it must be recalled that although these two dimensions do

account for between 8ö and 90 per cent of the controlled variance, they are only

accounting for between 6 and 16 per cent of the total variance. Cltarly then, it

is quite obvi.as that other variables associated with the group-task situation

remain to be isolated. Indeed, it is worth noting that in his more recent articles,

Fiedler has been doing just this. For exanple, he has considered the homogeneity-

heterogeneity of the group and the amount of stress (either internal or external)

that the group is working under. There can be little doubt that considerably more

rese?rch is necessary before the various complexities of the group-task situation

are isolated and t; e imnortance of these variables as determinants of "de lands" are

assessed.

Before concluding, however, a note of caution ..mst be reintroduced. Earlier

in this pap r, it was pointed out that wi'ile Fiedler's contingency i.odel was based

on an analysis of "real" groups in "real" situations, the present paper has only

been concerned with subjects' beliefs about the way "the .aost effective leader"

would behave in eight hypothetical situations. Clearly, although a subject may

expect the lEL to behave in the same danner on structured and unstructured tasks,

this does not aean that the id will or does behave in the same msnner. SLoilarly,

-17-

although in Mi. hy[5otlif;ticol siUutions c^r-'SiciKrod in tais piper, Jia subjects

expect tliO MEL to b »have quite differently '.;hen hs has high pos tion power than

when ha hac low position power, this dojs aot mean the IffiL will or does behave

differently. To put this in & slightly Jifferent way, although the difference

between having high and low position power may appear quite large in these hypo-

Ihetical situatJins, in actual groopi: this dxüt notion may have little, if any,

practical -iirnificance.

Thus, in conclusion, although the present paper has presentedaridence sup-

porting some of Fif.-iler's assu-uptions underlying the continbency „»del (i.e., that

different situations demand different lerderahip bahaviors, and that subjects who

differ in LPC do not perceive those de.iands differently), and some evidence ques-

tioning other assu-iptions (i.e., that the demands covary with the three dimensions

of the consistency model in a specific wanner), it must be recalled that these

results c^n in no way be taken as a direct test of the model. Hather they can only

be viewed as supplementary to Fiedler's position, ./here they agree with his posi-

tion, they provide some convergent validity for his arguments; wh.,re they djiagreo,

uhey merely raiue questions about the validity of his assumptions and point to

directions where further research with "real groups" in "real sitaationa" must be

conducted.

Reft nances

Bales, S. ?. iBtsrsctlon proccas analysis. Combridga, Mass.; Addlson- WnsUy, 1951.

Boll«s, S. and Ksasick, S. Statistical utility in experimental inference, PsvcholQatcal Reports. 1958, &, 223-227.

Fl«dler, F. E. taadeir Attitudas and Jroup Effectiveness. Urbana, Illinois: Univ. of Illinois Press. 1958.

Fiedler, F. B. A contingency model of leadership effectiveness. In L. Berkowitz (Ed.)» Advances in experimental social psycholoRy. New York: Academic Press, 1964. pp. 149-190.

Fiedler, F. E. The contingency model: a theory of leadership effec- tiveness. In H. Proshansky and B. Seidenberg (Eds.), gaste studies in social psychology. Mew York: Bolt, 1965, pp. 538-551.

Fiedler, F. B. The effect of leadership and cultiral heterogeneity on group performance: a test of the contingency mu^al. Journal of Experimental Social Psychology. 1966a, 2, 237-264.

Fiedler, F. E. A review of research on AS and LPG scores as measures of leadership style. Technical Report #§3, Group Effectiveness Research Laboratory, Oniverslty of Illinois, 1966b.

Fisbbein, M. The perception of non-members: a test of Merton's reference group theory. Soclometrv. 1963, 2^, 271-286.

Millar, L. K. Explained variance. Paper presented at the 56th Annual Meeting of the American Sociological Aeaociation, August, 1961.

Shew, M. E. and Blum, J. M. Effects of leadership style upon group per- formance as a function o** task structure. Journal of Personality and Social Psychology. 1966, £, 238-242,

Triandie, B. C. and Triandi», L. M. Rac«, soclai class, religion, and nationality as determinants of social distance. Journal of Abnormal and Social Psychology. 1960, £1, 110-118.

Triandis, B. C. and Triandls, L. M. Some studies of social distance. In D. Steiner and M. Flshbein (Eds.), Current studies in social psychology. New York: Bolt. 1965, pp. 207-217.

19

Footnotes

1. For a more coinplete discussion of the use of the analysis of variance

twhnlque as an indicant of strength of relationship and the distinction

between total and controlled variance, see Fishbein (1963) and Triandls

and Triandis (1965).

2. In reviewing an earlier draft of this paper, Gordon O'Brien suggested

another reason why Task Structure may have had such a small effect in

ths pr«sent study. Specifically, he suggested that the use of Bales'

categories for describing leader behavior may have mitigated against

finding differences since Bales' categories were primarily designed

to handle behaviors in one type of task situation, namely, a situation

in which a group attempts to solve an unstructured verbal problem.

Although it is possible that significant differences might have been

found had other behavioral categories been used, it should be recalled

that "«Tos (19S1) considers his categories as being cap*1*-!? of

enveloping all the behavior that can occur in any small face-to-face

group. In keeping with this. Bales' Interaction Process Analysis

has been used in many different contexts. Similarly, Fiedler and his

associates have often used the BDQ in various group-task situations.

CD FORM 1473

DOCUMENT CONTROL DATA - RSD

1. ORIGiNATINC ACTIVITY (Corporate author)

Group Effecti.v..>_i Research Laboratory Department of Psychology University of Illinois, Urbana, lilinois

2a. REPORT SECURITY CLASSIFICATION

Unclassified

3. REPORT TITLE

A Consideration of Two Assumptions Underlying riedler's Contingency Model for the Prediction of Leadership Effectiveness

4. DESCRIPTIVE NOTES (Type of report and inclusive dates)

Technical Report

5. AUTriG., (S)

Martin Fislibein, Eva Landy, and Grace Hatch

6. REPORT DATE 7a. TOTAL NO. OF PAGES

May, 1967 28

7b. NO. OF REFS. ba. CONTRACT OR GRANT NO.

12 Nonr 1834(36)

8b. PROJECT NO. 9a. ORIGINATOR'S REPORT NUMBER

2870 Technical Report No. 52 (67-8) c. NR 177-472 d. ARPA Order #454

10. AVAILABILITY/LIMITATION NOTICES

DISTRIBUTION OF THIS DOCUMENT IS UNLIMITED

12. c,3iNSCRING MILIT*RV ACTI-ZITY

Departnent of Navy Office of Naval Research Group Psychology Branch

DL- FORM 1473 (continued;

13. ABSTRACT

The present paper can best be viewed as an attempt to explore two of the basic assumptions underlying Fiedler's (1964. 1965) Contingency Model: (I) the assumption that different group-task situations "xequire a different leader-group member interaction," i.e., "delnand,, different typer- of leadership behaviors; and (2) the assumption that these 'demands" will covary systematically with the three dimen-'ions of the group-task situation specified by the Contingency Mo4*l. One hundred forty-one male undergraduates rated the way they believed the "Most Effective Leader" (MEL) should perform in each of eight group-fask situations on a Behavioral Description Questionnaire. The resuii? indicated that although these ratings of the MEL's behavior did vary across the different group-task situations, the ratings were significantly influenced by only two of the three group-task dimansions isolated by Fiedler, namely, the leader-member relations and the position power dimensions. Additional hypotheses related to the Contingency Model were also investigated and discussed.

14. KEY WORDS

Contingency Model Leadership effectiveness Leadership behavior Leader-member relations Leader position power