Fiedler 1978

ADVANCES IN EXPERIMENTAL 8OcW PSYCHOL4XY, VOL.
11
THE CONTINGENCY MODEL AND THE

DYNAMICS OF THE LEADERSHIP
PROCESS'
Fred E. Fiedler
UNIVERSITY OF WASHINGTON
SEATTLE. WASHINGTON
1. Introduction ............................................................ 60
11. The Contingency Model 60
A. Motivational Structure . . . . . . . . . . . . . . . . . . . . 60
62
111. Validation Evidence. .. ............. 61
IV. Dynamics of the Leadership ...................... 13
75
80
C. Leadership Training ....................... 81
................................... 84
........... 86
86
................... 89
C. Leader Behavior and Organization 101
D. What does LPC "Really" Measur 102
103
........... 104
104
108
References .......... ......................... 108
'This chapter is based on research conducted largely under contracts with the Office of Naval
Research, the Advanced Research Projects Agency of the Department of Defense, the Office of the
Chief of Naval Education and Training, and the Army Research Institute for the Behavioral Sciences,
to all of whom I wish to express my sincere thanks for their support. I am also deeply appreciative of
my colleagues and friends who saw this chapter in various stages of completion. I am particularly
indebted to Alan R. Bass,Judith Fiedler, A. Koman, Gary Latham, Linda Mahar, Joseph McGrath,
Gary Yukl, and VaIann Valdeson, who gave me the benefit of their critical comments at various
stages of the manuscript.
59 .
Copyrighi @ 1978 by Academic Press. Inc.
All rights of reproductionin any form reserved.
ISBN 612-015211-8
60 FRED E. FIEDLER
I. Introduction
In the first volume of this series I described a contingency model of leadership

effectiveness which was based on research conducted between 1951 and 1963
(Fiedler, 1964).This model made specific predictions about the way in which cer-
tain personality attributes would correlate with leadership and organizational per-
formance under given conditions at a particular point in time. Since then, well over
300 studies have dealt with various aspects of this models earlier formulation.
The major task which now faces us is the development of a dynamic theory of
leadership which adequately accounts for the changes in a leaders behavior and
performance over time. This chapter proposes that the key for such a dynamic
leadership theory is an understanding of the factors which change the leaders
situational control and influence.
The first section of this chapter briefly reviews the model and major valida-
tion research; the second section presents a dynamic interpretation of the Con-
tingency Model; the third discusses important methodological issues related to
the situational control dimension and the main predictor, the Least Preferred
Co-worker score (LPC). The final section describes a new approach to leadership
training which is based on the Contingency Model as well as empirical studies
which validate the method and thus provide further support for the dynamic
interpretation which this chapter presents.
11. The Contingency Model
The Contingency Model states that effectiveness of interacting2 groups or

organizations depends, or is contingent, upon the appropriate match between
leader personality attributes, reflecting his or her motivational structure, and the
degree to which the leader has situational control and influence. Specifically,
task-motivated leaders perform best in situations in which their control is either
high or relatively low; relationship-motivated leaders perform best when they
have moderate control and influence.
A. MOTIVATIONAL STRUCTURE
The leaders motivational structure is measured by means of a simple bipo-

lar adjective scale which asks the individual to think of all the persons whom he
*The Contingency Model deals primarily with groups in which performance depends on the
interaction and collaborative contribution of the members (e.g., as in basketball teams), rather than
the relatively independent performance of coacting group members whose scores are. s u m m e d to yield
a team performance measure (e.g., bowling teams).
CONTINGENCY MODEL AND THE LEADERSHIP PROCESS 61
or she has ever known and then to describe the one person with whom it has been
most difficult to work (the least preferred co-worker or LPC). The description,
using a recently published scale (Fiedler, Chemers, & Mahar, 1976), is made on
18 bipolar items which follow the Semantic Differential format, for example,
Pleasant :_:-:-:-:-:-:_ :-: Unpleasant
87654321
Friendly :_ :- :- _: _: _: _: _: : Unfriendly3
87654321
An individual who describes the LPC in very negative, rejecting terms (low
LPC score, i.e., less than about 57) is considered task-motivated. In other words,
the completion of the task is of such ovemding importance that it completely
colors the perception of all other personality traits attributed to the LPC. In
effect, the individual says, If I cannot work with you, if you frustrate my need
to get the job done, then you cant be any good in other respects. You are.. .
unfriendly, unpleasant, tense, distant, etc.
The relationship-motivated individual who sees his or her LPC in relatively
more positive terms (high LPC score, i.e., about 63 and above) says, Getting a
job done is not everything. Therefore, even though I cant work with you, you
may still be friendly, relaxed, interesting, etc., in other words, someone with
whom I could get along quite well on a personal basis. Thus, the high LPC
person looks at the LPC in a more differentiated manner-more interested in the
personality of the individual than merely in whether this is or is not someone with
whom one can get a job done.
There is also a middle LPC group, consisting of perhaps 15-20% of the
population, clustered around the population mean, which appears to differ in
many respects from either the high or the low LPC. These individuals tend to be
socially independent, less concerned about the way others evaluate them, and
less eager to conform to the expectations of others or to take the leadership role
(Fiedler, Chemers, & Mahar, 1976; Mai-Dalton, 1975). Some individuals in this
middle category may also have mixed motivation or a combination of the two
motivational patterns. We currently know relatively little about this group, which
we have tentatively labeled socioindependent. For the sake of easier presenta-
tion, we will deal primarily with high- and low-LPC leaders.
It is important to stress that the LPC score does not generally predict leader
behavior. Thus, a high LPC score does not necessarily imply that a leader will be
considerate, nor does a low LPC score imply more structuring behavior. In fact,
3The other item pairs are rejecting-accepting, tense-relaxed, distant-close, cold-warm,

supportive-hostile, boring-interesting, quarrelsome-harmonious, gloomy-cheerful, open-guarded,
backbiting-loyal, untrustworthy-trustworthy, considerate-inconsiderate, nasty-nice, agreeable-
disagreeable, insincere-sincere, and kind-unkind.
62 FRED E. FIEDLER
as we shall discuss later in this chapter, leader behavior is itself the product of the
interaction between the LPC score and the degree to which the situation provides
the leader with control and influence.
B. SITUATIONAL CONTROL
The second major element in the Contingency Model is the degree to which
the situation provides the leader with control and influence. (In previous publica-
tions this dimension usually has been called situational favorableness, a term
which has since been discarded because it is frequently confused with task
difficulty.) A high degree of control and influence implies that the leader has
correspondingly high certainty that his decisions and actions will have predicta-
ble results, and that they will achieve the desired goals and gratify the leaders
needs in the situation. As we shall see, situational control and influence has
emerged as the key to the development of a dynamic theory of leadership. The
most common method of measuring situational control or favorableness has been
by means of three subscales.
1. Leader-Member Relations
The first subscale indicates the degree to which leader -member relations
are good, that is, the degree to which the leader enjoys the support and loyalty of
group members. This dimension has been measured by such methods as the
group members preference ratings or evaluations of the leader, or by a group
atmosphere scale which the leader fills out. The latter approach, which has
been used most frequently, is a 10-item semantic differential scale which asks the
leader himself to describe the atmosphere of the group in such terms as
friendly-unfriendly, pleasant-unpleasant, etc. A number of studies (e.g.,
McNamara, 1968) have shown that the scale correlates highly with the groups
loyalty and support of the leader but not with LPC (see Fiedler, 1967, p. 153;
McNamara, 1968; Stinson & Tracy, 1974). A new scale (shown in Fig. 1) allows
the leader to estimate leader-member relations with his own group as well as
between subordinate managers and their groups. The correlation between this
scale and the Group Atmosphere scale in a recent study was .88, n = 26.
The leader-member relations dimension is the most important of the three
subscales, since a leader who enjoys the support and loyalty of group members
can depend and rely on them. He can be certain that the group members will do
their best to comply with his wishes and directions. A leader who cannot count
on his group is, of course, in a very precarious position. He will need to be
considerably more circumspect in his dealings with subordinates and continu-
ously on guard to assure that his directions or policies are not subverted.
2 . Task Structure
The second most important dimension is task structure. We generally do not
LEADER-MEMBER RELATIONS SCALE
C i r c l e the number which best represents your response t o each item.
1. The people 1 supervise have trouble getting

along w i t h each other.
2. My subordtnates are r e l i a b l e and trustworthy.
3. There seems to be a f r i e n d l y atmosphere amng

the people I supervise.
4. My subordinates always cooperate w i t h me i n

g e t t i n g the job done.
5. There i s f r i c t i o n between my subordinates

and myself.
6. My subordinates give me a good deal of help

and support i n g e t t i n g the j o b done.
7. The people I supervise work well together i n

getting the j o b done.
8. I have good r e l a t i o n s w i t h the people

I supervise.
TOTAL SCORE
Fig. 1. Leader-member relations scale. From Fiedleref n l . , 1975. Reprinted by permission of

John Wiley & Sons, Inc.
think of the task as providing the leader with control and influence. However,
leaders who have a blueprint or detailed operating instructions are assured of the
support of their organization in directing the job. They.very rarely get any
arguments from subordinates as to the course the group should take. In contrast,
when the task is unstructured, as is the case with typical committee assignments
or research and development work, the control which leaders can exercise over
the task and the group is considerably diluted. Most committee chairpersons
would find it extremely difficult to dictate the committees decision or to predict
the acceptability of the outcome of these decisions. Such a prediction is easily
made when the leader has step-by-step instructions which guarantee an accepta-
ble completed task if they are faithfully followed.
The degree of task structure has been measured in a number of ways,
generally following four subcategories suggested by Shaw (1963). These are the
64 FRED E. FIEDLER
tasks decision verifiability, goal path multiplicity, goal clarity, and solution
specificity. Thurstone scales for each of thev four subscales have been de-
veloped by Hunt (1967) and have been described in previous publications (e.g.,
Fiedler, 1967; Fiedler & Chemers, 1974). Abbreviated scales also have been
extensively used and generally serve the purpose quite well. By and large, line
management, production supervisors, and military troop commanders have rela-
tively structured tasks. Leaders of research and development groups, commit-
tees, boards, or organizations which require creative effort tend to have unstruc-
tured tasks. A recent task structure scale (Fiedleret af., 1976) is shown in Fig. 2.
TASK STRUCTURE RATING SCALE C i r c l e one
7-/
Jsuall) Sometimes ieldom
-True True
IS THE GOAL CLEARLY STATE0 OR KNOWN?
1. Is there a blueprint, picture, &el, or

detailed description available of the
finished product or service? 2 1 0
2. Is there a person available t o advise and

give a description of the finished product
o r service, or how the job should be done? 2 1 0
IS THERE ONLY ONE NAY TO ACCOMPLISH THE TASK?
3. Is there a step-by-step procedura, or a standard

operating procedure which indicates i n d e t a i l
the process which i s t o be followed? 2 1 0
4. Is there a specific way t o subdivide the task

into separate parts or steps? 2 1 0
5. Are there some ways which are c l e a r l y recognized

as b e t t e r than others f a r performing t h i s task? 2 1 0
IS THERE ONLY ONE CORRECT ANSWER OR SOLUTION?
6. Is It obvious when the task i s finished and the

correct solution has been found? 2 1 0
7. Is there a book, manual, or job description
which indicates the best solutlon o r the best
outcome for the task? 2 1 0
IS I T EASY TO CHECK WHETHER THE JOB WAS DONE RIGHT?
8 . Is there a generally agreed understanding about

the standards the particular product or service
has t o meet ta be considered acceptable? 2 1 0
9. Is the evaluation o f t h i s task generally made
on $me quantitative basis? 2 1 0
10. Can the leader and the group f i n d out how well
the task has been accomplished in enough time
2 1 0
-
to improve future perfomnce?
-
Fig. 2. Task structure rating scale. From Fiedler et al., 1975. Reprinted by permission of
3 . Position Power
The third dimension which defines situational control is position power, that
is, the degree to which leaders are able to reward and punish, to recommend
sanctions, or otherwise to enforce compliance by subordinates. Here again, a
scale is available which provides norms for scoring (see Fig. 3) (Fiedler et al.,
1976).
Task structure and position power should be assessed by the leader's
superiors. Self assessments are subject to distortion. In one study, for example,
inexperienced high LPC leaders perceived their task structure and position power
as very low. However, self-ratings are useful for training purposes, as we will
see later.
POSITION POYER RATING SCALE
Circle the number which best represents your anwer.
1. Can the leader d i r e c t l y o r by reconmendation adninister rewards

and punishments t o his subordinates?
2 1 0
Can act d i r e c t l y o r wrecalmen'd-ht 7
can reconrncnd with with mixed results
high effectlveness
2. Can the leader d i r e c t l y or by recarmendation affect the p r m t i o n ,
demotion. h i r i n g or f i r l n q of h i s subordinates?
2 1 0
Can act d i r e c t l y o r Can recamend but 7
can reconmend with with mixed results
high effectiveness
3. Does the leader have tk knweldqe necessary t o assign tasks to
subordlnates and instruct them I n task ContpletfOn?
2 1 0
YES Soartimes or i n ~KJ
soar aspects
4, Is i t the leader's job t o evaluate the performance o f his subordlnatesl
2 1 0
YES Solnetimes o r i n NO
some aspects
5. Ha5 the leader been given some o f f i c i a l t i t l e of authority b the
organization (a.9.. foraun. departmnt herd. platoon leader{?
2 0
, YES wo
TOTAL. , ........... .
Fig. 3. Position power rating scale. From Fiedler et al., 1975. Reprinted by permission of
66 FRED E. FIEDLER
4 . Measuring Situational Control

The leaders situational control (or situational favorableness) is usually rep-
resented on an 8-point continuum. This is obtained first by dividing the groups
into those falling above and below the normative cutting score on leader-member
relations, then on task structure, and finally on position power. A continuous
scale of situational control can also be constructed by weighting the standardized
leader-member relations score by 4, the standardized task structure score by 2,
and adding these to the standardized position power score (Nebeker, 1975). The
formula thus reads:
Leader situational control = 4(leader-member relations) + 2(task structure) +
(position power).
This system of weighting has recently been supported by a number of empirical
studies which will be discussed in Section I11 of this chapter.
It must be pointed out, of course, that the three major subscales of leader-
member relations, task structure, and position power by no means represent the
only factors which determine the leaders situational control and influence. Other
studies (e.g., Ayer, 1968; Meuwese & Fiedler, 1965) have pointed to situational
stress as affecting the leaders control; cross-cultural studies (e.g., Fiedler, 1966;
Fiedler, Meuwese, & Oonk, 1961; Fiedler, OBrien, & Ilgen, 1969) have shown
that linguistic and cultural heterogeneity also play a major role in determining
leader control. And leader experience and training also increase control. In
unusual cases this formula may thus require appropriate modification, and spe-
cific rules governing these modifications still need to be developed.
5 . LeaderSituation Interaction
As already mentioned, task-motivated leaders tend to perform best in situa-
tions in which their control is either high or else relatively low, while the
relationship-motivated leaders perform best in situations in which they enjoy
moderate control. This relationship is seen in Fig. 4. The horizontal axis indi-
cates the eight cells or octants of the situational control dimension. The vertical
axis shows the correlations between the leaders LPC score and the group per-
formance scores. The median correlations of the original studies are connected
by the broken line on this graph.
Positive correlations, shown by median coefficients above the midline on
the graph, show that the relationship-motivated (high-LPC) leaders performed
better than did task-motivated (low-LPC) leaders. Negative median correlations
(i.e., below the midline) indicate that the task-motivated leaders were superior.
Thus, both the task- and the relationship-motivated leaders performed well under
some conditions and poorly under others. This implies that it is not appropriate to
speak of a good leader or a poor leader. Rather, a leader may perform well
in one situation and not another, depending on the proper match between leader
motivational structure and situational control.
- O r i g i n a l studies
V a l i d a t i o n studies
----- West Point study
1 .oo
+Rleader
elatlonship-mtivated
performs best
.---
-.20 +Task-motivated
leader perfoms best
-.40
,
-.60
-.80
-1 .oo J I I I I I 1 I I I
Favorable I I1 I11 IV V VI VII VIII Unfavorable
for Leader for Leader
Leader-Member
Relations Good Good Good Good Poor Poor Poor Poor
Task Structure Structured Unstructured Structured Unstructured
Leader
Power Position Strong Weak Strong Weak Strong Weak Strong Weak
Fig. 4. Median correlationsbetween leader performance and group performance for the origi-
nal studies, validation studies, and the West Point 1972 study. From Fiedler, 1977b.Reprinted with
permission.
III. Validation Evidence
The Contingency Model has been extensively tested in a wide variety of

laboratory and field conditions. Most of these studies have supported the model.
Since the Contingency Model makes different predictions for different situations,
for example, for each of the eight octants or cells of the situational control
dimension shown on Fig. 4,a sizable number of groups must be found which
perform comparable tasks under each of these conditions. That is, the groups
which are compared must have leaders who have (a) demonstrably good or poor
relations with group members, (b)structured versus unstructured tasks, and (c)
strong versus weak position power. Most validation studies have tested the
Contingency Model only in some of these conditions.
Extensive reviews of the validation evidence have been presented in a
number of publications (e.g., Fiedler, 1971b; Fiedler & Chemers, 1974; Mitch-
ell, Biglan, Oncken, & Fiedler, 1970), and the research will be presented here in
summary form.
Table I presents the correlation coefficients obtained in validation studies
between 1963 and 1973 which dealt with interacting groups and tested the appro-
priate octant. That is, the methodology and design of each study was rated by
four independentjudges who had been given training in classifying groups. They
TABLE I
SUMMARY OF FIELD AND LABORATORY STUDIES
TESTING THE CONTINGENCY MODELa
~ ~~
Octant Octant Octant Octant Octant Octant Octant Octant

Field studies I 2 3 4 5 6 7 8
Hunt (1967) -.a - .80 .21 .30

-.51 .60 - .30
Hill (1%9) -.I0 -.29 -.24 -62
Fiedler et al. (1969 - .21 .oo .67b -.51
O'Brien and Fiedler
(unpublished) - .46 .47 - .45 -.14
Tumes (1972) -. 4 7 .62e
Laboratory experimentsd
Belgian Navy
(Fiedler, 1966) -.72 -37 -.16 .08 .07 -26 -.37
-.77 SO -3 .13 .03 .I4 -.27 .60
Shima (1968) -.26 .7Ib
Mitchell (1970) -24 .43
.17 .38
Fiedler Exec. .34 .5 1
Chemers and Skrzypek
( 1972) -.43 -.32 .10 .35 .28 .I3 .08 -.33
Rice and Chemers (1973) 30 - .40
Sashkin (1972) -.2P
Schneier (1978) - .5P
Median, all studies -.59 -.I0 -.29 .40 .I9 .13 .17 -.35
Median,fieldstudies -51 -.21 -.29 .47 .21 -24 .30 -.33
Median, laboratory
experiments d.72 .21 -.23 .38 .I6 .14 .08 -.35
Medians in original
studies -.52 -3 -.33 -41 .42 .05 -.43
Number of correlations in the expected direction: 38f

Number of correlations opposite to expected direction: 9
p by biomial test: .01
"From Fiedler, 197la. Copyright 1971 by the American Psychological Association. Reprinted
with permission.
@pc .05.
' p < .01.
dAn additional study by Shiflett and Nealey (1972) should have been included in this table. This
laboratory experiment compared 3-man college groups with very high intellectual ability and with
moderate ability on performance in creative tasks in octants 3 and 4 (weak and strong position power).
The results of the moderate ability groups supported the prediction of the model (- .16, .34), while
those of very high ability gave contradictory findings (.a, - .49, all ns = 8. and nonsignificant).
These additional findings give rise to interesting speculations about high-ability college students but
do not change the basic interpretation of the table.
"PI c ;001.,
Qxclusive of octant 6, for which no prediction had been made.
CONTINGENCY MODEL. AND THE LEADERSHIP PROCESS 69
were to determine whether the groups were interacting or coacting, and into
which of the eight octants the various groups belonged. Groups were considered
properly classified if three of the four judges agreed in their placement of the
group. Most of the correlations are based on very small samples and were
individually not significant. Thirty-eight (81%) of the 47 separate correlation
coefficients were in the predicted direction. A binomial test of the proportion of
correctly and incorrectly predicted correlations is highly significant.
The similarity in the pattern of correlations also provides evidence of the
Contingency Model's predictive power. The median correlations obtained in the
validation studies under laboratory and field conditions are indicated on Table 11,
as are the median correlations obtained in the original studies. The medians of the
validation studies in Fig. 3 are connected by a solid line. The rank-order conela-
tion between the correspondingmedians is .75 (n=7) and statistically significant.
As Table I1 also shows, the only discrepancy between the predicted median
correlations and the validation study medians occurred in octant 2 of the labora-
tory studies, while the direction of the correlations in the field studies followed
the prediction of the model. These findings suggest either that the prediction for
octant 2 is incorrect or that the conditions of octant 2 are difficult to reproduce in
TABLE I1
COMPARISON OF MEDIAN CORRELATIONS FROM ORIGINAL STUDIES WITH
CORRESPONDING CORRELATIONS FROM VALIDATION STUDIES"
Median correlations between leader LPC scores and group performance
anginal All validation Laboratory

Octant studies studies Field studies studies West Point
-52 -59 -51 -12 -43

-58 - 10 -21 21 -32
-33 -29 -29 -23 10
47 40 47 38 35
42 19 22 16 28
(13) (-24) (14) (13)
05 17 30 08 05
-43 -35 -33 -35 -33
Rank correlations between medians of original studies and results of validation studies
~
All validation studies .79 p < .05 (one-tailed)

Field studies .75 p < .05 (one-tailed)
Laboratory studies .46 n.s.
West Point study .86 p < .01 (one-tailed)
~~
"From Fiedler, 1973. Reprinted with permission.

*No predictions were made for octant 1.
70 FRED E. FIEDLER
laboratory research. The latter seems more likely, since laboratory studies are by
their nature artificial. Moreover, situations in which the leader has low position
power while the task is highly structured are rather rare in real life. It is even
more unusual to find a leader who can manage a situation of this type unless he
has the full support of his group. This may account in large part for the difficulty
of finding such groups in real-life conditions.
Also reproduced are studies which permitted only an approximateclassifica-
tion of groups into those in which the leader had high, intermediate, and low
situational control (or favorableness) (Table 111). Here, 26 of the 35 (74%)
correlation coefficients are in the predicted direction, and the binomial test is
again highly significant.
This method of validation has been attacked on various grounds, e.g.,
because it interprets nonsignificant correlation coefficients and because some
correlation coefficients are based on subsamples in the same study where one
subsample might have had good and the other poor leader-member relations
(Ashour, 1973; Graen, Alvares, Onis, & Martella, 1970). While the latter argu-
ment has some validity, the former does not (Fisher, 1948). However, it has been
pointed out, quite correctly, that most validation studies were based on concur-
rent measurement of LPC, leader-member relations, and group performance
scores, and that leader-member relations measures and even LPC scores might
well be affected by the groups performance. We shall discuss this point shortly.
In addition, of course, a number of studies have failed to support the
Contingency Model. Most frequently cited have been two laboratory experiments
TABLE I11
SUMMARY OF LPC GROUP PERFORMANCE CORRELATIONS OF STUDIES EXTENDING
THE CONTINGENCY HYPOTHESISO
Study and date Favorable Intermediate Unfavorable
OBrien (1969) - .08 .77b -.I3

Anderson (1966) -.SO .21 - .22 .I2
Lawrence and Lorsch (1967y -.SO -.I0 -.13 .66
Nealey and Blood (1968) - .22 .79b
Nealey and Owen (1970) -.sob
Fiedler and Barron (1 967r
Task I -.42 -.56 -.32 .67 -.08 -.01 -.53 -.72 .18
Task II -.71 -.59 .69 .41 -.I5 -.20 -.47 -.61 -.I4
Median - .37 .20 - .30
Note: The location of the correlation coefficient in the table indicates degree of judged favorableness of the
leadership situation. The farther to the left, the more favorable the situation.
From Fiedler, 1971b. Copyright 1971 by the American Psychological Association. Reprinted with
permission.
bStudy not conducted by writer or his associates.
CInFiedler ( 1967).
by Graen, Orris, and Alvares (197 1). However, the Graen el al. experiments are,
themselves, methodologically faulty, as has been pointed out in the literature
(e.g., Fiedler, 197la; Chemers & Skrzypek, 1972). Specifically, all experimental
manipulations in the Graen et al. studies were weak. Some groups with sup-
posedly high position power had position power scores below the mean and vice
versa. Likewise, statistical tests showed that the classification of leaders as
belonging in the high or the low LPC groups was in a number of cases in~orrect.~
A methodologically sound validation study was conducted by Chemers and
Skrzypek (1972) at the U.S. Military Academy. This field experiment utilized
128 cadets from two cadet training companies. In order to assure that neither the
LPC score nor the leader-member relations measure would be influenced by the
groups performance, LPC as well as sociometric preference measures were
obtained 3 weeks prior to the study. High- and low-LPC leaders were selected
from those whose scores fell at least 1 standard deviation above or below the
mean of the LPC distribution. Groups with good or poor leader-member rela-
tions were assembled by bringing together men who chose one another as co-
workers or who expressed dislike for one another.
Each group performed one structured and one unstructured task in counter-
balanced order. The structured task consisted of drawing the plan for a barracks
building to scale from a set of specifications; the unstructured task required the
group to design a program to facilitate an interest in world politics among
enlisted men assigned overseas. Furthermore, groups were randomly assigned to
a high or to a low position power condition. In the former, the leader had the
responsibility to evaluate the performance of each group member and assign a
score which would become a part of the cadets permanent service record. In the
low position power condition, the leader was instructed to act as chairman, and
the group was told that the leader had no real power to reward or punish group
members.
In summary, leader LPC and leader-member relations were experimentally
determined in udvunce, and task structure and leader position power were
strongly manipulated, yielding the full eight-celled situational favorableness di-
mension of the Contingency Model.
Leader LPC and group effectiveness scores were correlated for each octant
of the situational control dimension. The dash-dot line on Fig. 4 plots the
A laboratory experiment by Vecchio (1977) also failed to confirm the Contingency Model.
Although the author claims that he conducted a very rigorous test of the model, groups with sup-
posedly good leader-member relations were assembled so that leaders were assigned to supervise
two favorable ranked classmates and one unfavorably ranked classmate (Favorable Group Atmo-
sphere) or two unfavorably ranked classmates (Unfavorable Group Atmosphere). In other words,
each of the groups was deliberately designed to have at least one group member with whom the leader
did not wish to work. Unless this study was designed for some other purpose, I am at a complete loss
to understand the rationale for this curious procedure which deliberately weakens the manipulation of
the single most important component of the situational favorableness dimension.
72 FRED E. FIEDLER
resultant correlations against the predicted curve. As Table I1 shows, the rank-
order correlation of the two sets of correlation coefficients was .86 @ < .02). The
interaction of LPC, leader acceptance, and task structure was also highly signifi-
cant (F = 6.19, p < .025) and accounted for about 28% of the group perfor-
mance variability (Shiflett, 1973), and thus strongly supports the model.
In addition, a number of subsequent laboratory and field studies (e.g., Rice
& Chemers, 1975; Sashkin, 1972) provide further evidence supporting the pre-
dictive ability of the Contingency Model. Three carefully designed experiments
were conducted by Hardy (1971, 1975) and Hardy, Sack, and Harpine (1973) in
which LPC scores also were obtained 1 or 2 weeks prior to the study. In two of
these studies (Hardy, 1975; Hardy et al., 1973) leader-member relations were
experimentally manipulated by assigning subjects to groups on the basis of
preassessed sociometric scores. The 1971 Hardy study used 56 4-person groups
of college students to test cells 1 4 , Hardy et al. used 56 groups of high school
students to test cells with weak position power, namely, 2 , 4 , 6, and 8, and the
1975 Hardy study tested 39 groups of elementary school students from the fourth
grade in cells 5-8 of the model.
As in most laboratory tests, the results obtained for octant 2 did not support
the prediction. The results for all other seven octants were significant in the
predicted direction. Table IV summarizes the findings.
On the whole, there is now little question about the basic validity of the
Contingency Model. The evidence is very clear that low-LPC leaders perform
better than high-LPC leaders in situations in which they have very high or
relatively low control and influence, and high-LPC leaders perform better in
TABLE IV
SUMMARY OF VALIDATION EXPERIMENTS CONDUCTED BY HARDY AND HIS ASSOCIATES
Contingency model &ant
Group I 2 3 4 -5 6 7 8
56 3-man group of college

students (Hardy, 1971) L > Ha H< L L > HU H > La - - - -
56 Groups of high school
students (Hardy et al.,
1973) - H<L - H>Lb - H>Lb - L>H
39 Elementary school
groups (Hardy, 1975) - - - - H>L" H>L" H>La L > H
Abbreviations: H = High LFC leader; L = Low LPC leader.

'F-test significant in predicted direction p < .05.
*F-test significant in predicted directionp < .01.
cF-testsignificant in predicted directionp < 10.
situations in which their control is moderate: These findings are quite strong for
octants 1 , 4, 5, and 8, in which correlations account for substantial portions of
the variance, that is, about 36%, 16%, 5%, and 12% respectively. Octant 3
appears to account for about 8% of the variance, while the variance accounted for
by octants 2, 6, and 7 is negligible. Whether further refinements in scaling the
situational control dimension and improvements in the LPC score will result in
substantial increases in the accuracy of these predictions remains to be seen. The
unreliability of the typical group performance criteria and of the various personal-
ity and situational measures in leadership are likely to remain serious limitations
in our efforts to add to the precision of predicting leadership performance. It must
also be remembered, however, that the Contingency Model does not take into
account a number of other important factors. Among these are the leaders and
the groups abilities, motivation, the leaders and members sex, and the relation
of the leader and the group to key persons at higher levels of the organization.
Further research will be needed to integrate these other factors into a more
general theory of leadership.
IV. Dynamics of the Leadership Process
A viable leadership theory must be able to explain leadership as an integral

part of a continuously changing organizational process. Frequent shifts in man-
agement, personnel turnover, new assignments, and turbulence in the political
and economic environment are part of organizational life. But in addition, the
leader himself changes. While personality attributes remain reasonably constant
over the space of a few months or even a few years, the way in which a leader
approaches the job may undergo considerable change as he or she learns by
experience or by training how to cope more confidently with job related prob-
lems. Since these organizational factors as well as the leaders own development
affect situational control, the effectiveness with which the leader performs the
job will also vary.
The dynamic nature of this process is most easily visualized in the schematic
drawing of the Contingency Model, shown in Fig. 5. Situational control is again
indicated on the horizontal axis while the vertical axis of the figure indicates
leadership effectiveness. The solid and the broken lines trace the predicted per-
formance of the relationship- and of the task-motivated leader, respectively.
As this schematic drawing suggests, the effectiveness of the leader will
change as his motivational structure changes or as situational control and influ-
ence change. As we will show in a later section, the LPC score remains fairly
constant over time. Our major concern is, therefore, with changes in situational
control. These principally take place as leaders acquire experience and training or
as they are given new assignments, and as personnel turnover among their
74 FRED E. FlEDLER
\I
\
1
1 2 ,3 4 5 6 7 8
Leader-Member
Relations + t +
Task S t r u c t u r e t t +
P o s i t i o n Power + - + + +
Favorable S i t u a t i o n s Moderately Unfavorable S i t u a t i o n s
(High C o n t r o l ) Favorable (Low C o n t r o l )
Fig. 5 . Schematic representation of the Contingency Model relationship between leader LPC
scores and task performance in different conditions of leader situational control. From Fiedler,
1977a. Reprinted by permission of W. H. Winston & Sons.
subordinates or superiors require them to develop new patterns of interpersonal

relations with these important others in the work environment.
The effects which changes in situational control will have on the leaders
performance can readily be seen on Fig. 5. For example, let us assume a situation
in which the leaders control is initially in the very low zone (i.e., at the far right
of Fig. 5 ) . Here, the task-motivated leaders performance should be better than
that of the relationship-motivated leader. As the leader gains in ability to cope
with the situation, or as the situation becomes more predictable, the leader will
increase his situational control. (This can be visualized as a movement toward the
moderate-control zone in the middle of the graph.) We should then find that the
performance of the task-motivated leader decreases, while that of the
relationship-motivated leader increases, as situational control shifts from low to
moderate. With a still further increase in situational control to the high-control
zone, the performance of the relationship-motivated leader will again decrease,
while that of the task-motivated leader will increase.
The changes in the leaders situation which mainly affect his or her control
and influence are operationally defined by the Contingency Model as changes in
the leader-member relationship, task structure, and position power. However,
situational control can also be interpreted in terms of the leaders certainty about
the outcomes and consequences of his or her behavior and decisions, as has been
shown by Nebeker (1975), Mai-Dalton (1975), and Beach, Mitchell, and Beach
(1978). Having group support, knowing exactly what is to be done and how it is
to be done, and having the power to reward and punish subordinates as a means
of assuring compliance provides the leader with a high degree of confidence that
his directions will be carried out and that they will have predictable conse-
quences. There is, therefore, little need to be insecure and anxious about the-out-
come. Any environmental condition which introduces instability and uncertainty
will reduce situational control and increase the leaders anxiety about the out-
come. This is likely to cause the leader to take some compensatory action in
order to regain control and to reduce the uncertainty about the outcomes and the
attendant anxiety. This change in leader behavior will bring about an increase or
decrease in leadership performance depending on how well the particular leader
behavior happens to match the requirements of the situation.
A . LEADERSHIP EXPERIENCE
One factor which almost inevitably changes situational control comes with
the experience which leaders gain on the job (e.g., Bons & Fiedler, 1976). Tasks
tend to become more routine, leaders get to know their subordinates, and in most
instances they are able to handle them better as time goes on. Likewise, with time
on the job, the leader will learn the bosss standards and expectations.
As the Contingency Model predicts, leadership experience may increase
situational control but it does not generally increase leadership performance.
While some managers or supervisors get better with time on the job, others tend
to get worse. Everyone knows of some young and inexperienced leaders who
have outperformed every veteran in sight. History provides such examples as
Joan of Arc who was only 17 years old when she defeated her much more
experienced opponents at the Battle of Orleans, and William Pitt who was only
24 at the time he became one of Englands most distinguished prime ministers.
A review of various empirical studies on leader experience and performance
(Fiedler, 1970) summarized the findings on 13 different organizations. These
included military units and ad hoc teams, industrial and business organizations,
meat markets, and voluntary organizations. The median correlation between the
leaders years of experience and rated performance was -. 12. One of the studies
compared the leadership performance of 48 Belgian Navy petty officers with an
average of 9.8 years of experience and the performance of 48 new recruits
without any formal leadership experience. The groups of leaders were matched
on leader and group member intelligence and leadership style. The teams worked
on four simulated military tasks developed in cooperation with responsible Navy
officers. Yet, the performance of experienced and inexperienced leaders was
almost identical on all four of these tasks.
A replication compared leadership effectiveness on experimental tasks for
one group of captains and majors with 5-17 years of experience and 4 years of
military college and for a group of new recruits who had just completed their
76 FRED E. IWDLER
basic training course. Here again, the average performance of groups led by
inexperienced trainees was about as high as the average performance of groups
led by the expenenced and trained officers. Moreover, the officers with relatively
more training and experience performed no better than did the officers with
relatively little experience. And studies of post office supervisors, business man-
agers, foremen, and police sergeants also showed no consistent relationship
between years of leadership experience and leadership effectiveness (Fiedler,
1972a).
As Fig. 5 shows, the Contingency Model predicts that leadership experience
will have different effects on the performance of relationship- and task-motivated
leaders. A number of investigations have supported this hypothesis. One exam-
ple comes from a study of infantry squads (Fiedler, Bons, & Hastings, 1975a).
These basic infantry units consist of 8-1 1 enlisted men. A group of 28 sergeants,
who served as squad leaders, was evaluated at the time the units were formed and
when the organization was, therefore, still in a state of flux. These same
sergeants were again evaluated by the same superiors after they had had 5 months
of experience in working with their units. While these sergeants judged the
situation as providing moderate situational control at the first time of evaluation,
they judged their situational control to be relatively high at the time of the second
evaluation.
Figure 6 shows that the prediction based on the Contingency Model was
supported. The relationship-motivated leaders performed relatively well while
the situation was still uncertain and relatively unstructured. However, the task-
motivated leaders performed relatively better 5 months later when the situation
had become more stabilized and their control had presumably moved into the
high zone.
A similar finding was obtained in a study of general managers of 32 con-
sumer cooperatives (Godfrey , Fiedler, & Hall, 1959). This federation consisted
of small companies, each typically serving one county of the State of Illinois.
The federation maintained excellent records of company performance which
were used as criteria for bonuses as well as for subsequent promotion of mana-
gers to more desirable jobs. The major criteria were the percent of net income to
total sales and the percent of overhead costs to total sales. An analysis of various
economic indicators showed that the economy of the county had relatively little
influence on the performance of the organization, since a company in a booming
economy also tended to have more competitors. The leadership situation was
judged as high in control for experienced managers but only of moderate control
for the less experienced general managers.
The group was first divided into task- and relationship-motivated managers
and then further subdivided into an experienced and a relatively inexperienced
group. As Fig. 7 shows, the relationship-motivated managers who were inex-
perienced performed better than those who were experienced. In contrast, task-
I
I
I
Time 1 Time 2
LOW Experience Hlgh Experience
T k L e a d e r s S i t u a t i o n a l Control
Moderate High
Fig. 6. Change in performance of high- and low-LPC leaders as a function of increased

experience. Interaction significant. Adapted from Fiedler et al. (1975a).
motivated managers who were experienced performed better than those with
relatively low experience.
If the situation is i n the low-control zone for the inexperienced leader, we
would, of course, expect that the increasing experience will move the leader into
the moderate-control zone. Hence, the inexperienced leader should perform well
if he or she is task-motivated, while the experienced leader should perform
relatively well if he or she is relationship-motivated.
This can be inferred from a study which McNamara (1 968) conducted with
Canadian elementary and secondary school principals. The performance evalua-
tion of elementary principals was based on superintendents ratings, while that of
secondary principals was based on the average achievement test scores of ninth-
and eleventh-grade students on a province-wide test.
In general, experienced elementary school principals enjoy relatively high
control in their leadership situation. Their schools are small and they can easily
78 FRED E. FIEDLER
3
a
E
Mean
*
8
Fig. 7. Effect of experience on the performance of high- and low-LPC managers. From
Fiedler, 1972a. reprinted by permission.
get to know their staff members. Their tasks tend to be structured since policies
and cumcula are determined by the school district. Moreover, elementary school
students present relatively few disciplinary problems. Thus, the situation gives
the experienced principal a great deal of control; in contrast, the inexperienced
principal, who still has to learn the ropes, will have only a moderate degree of
situational control.
The secondary schools were complex organizations, employing 30-40
teachers who, in turn, reported to various department heads. Moreover, the task
of high school principals is relatively unstructured since it requires them to
develop policies, interact with various civic and governmental agencies, and
maintain discipline over teenage students who are notoriously difficult to deal
with. We would estimate, therefore, that the inexperienced principals have low
situational control. The experienced principals would have moderate situational
control. We thus have the following conditions:
High control: Experienced elementary school principals
Moderate control: Inexperienced elementary and experienced
secondary school principals
Low control: Inexperienced secondary school principals
McNamara divided the principals into those with little experienced on the job
(less than 2 years) and those who were established (more than 3 years). The
resulting average performance scores for elementary and for secondary school
principals with high and low LPC scores are shown on Figs. 8A and B. As the
Contingency Model predicts, secondary school principals who were task-
motivated performed best if they were inexperienced, that is, in low-control
situations; relationship-motivated principals performed best if they were experi-
enced, placing them in moderate-control situations. The performance of experi-
enced and inexperienced elementary school principals followed the same pattern
as that seen for general managers in the study of cooperatives or the study of
infantry sergeants, in which experienced leaders were rated as having high situa-
tional control. Similar findings, also on college teachers have been reported by
Hardy and Bohren (1975). In brief, the data indicate that a dynamic interpretation
of the Contingency Model permits us to explain as well as to predict the effects of
leadership experience on task performance.
Relationshlp-
M o t i v a t e d (High LPC)
U l a r k - M o t { vated
(Low LPC)
'.'Ot
1.50 A
E r t a b l ished
2.20 (Two Years (Three Years
o r Less) or mre)
40.00
35.00 \
32.50
30.00
27.50
:;a' Established
25.00
(Two Years (Three Years
22.50 or Less) or More)
Fig. 8 . Average performance of relationship- and taskmotivated school principals with rela-
tively low and high levels of experience. A. Elementary school principals. B . Secondary school
principals. Interaction significant. From Fiedler. I972a. reprinted by permission.
80 FRED E. FIEDLER
B . SELECTION AND ROTATION

Figures 6,7,and 8 also have important implications for leadership selection
and rotation. At the time of selection most leaders are, of course, inexperienced.
As is evident from our data, the performance of task-and relationship-motivated
leaders changes in different directions as they gain in experience and training.
Selection is, therefore, not the one-time problem of putting square pegs into
square holes and round pegs into round holes. Rather, we have to make our
selection in the knowledge that the shape of the pegs and the holes changes over
time. In selecting leaders we must obviously decide whether we want optimum
performance immediately or whether we can afford to wait until the leader's
experience and training have made the situation optimal for the particular type of
personality the individual brings to the job.
The problem is illustrated schematically in Fig. 9, which shows the ex-
pected performance curves of relationship- and task-motivated (high- and low-
LPC) leaders who begin their management jobs in situations in which they have
very low control (points A and A'), or in moderate-control situations (points E
and E'). That is, the performance of the relationship-motivated leader will
gradually increase as he moves from the low- to the moderate-controlzone (point
A to point B ) . The organization must then either wait until the relationship-
motivated leader's control and influence increase with experience, or else it must
provide such aids to increasing the leader's control as giving him a harmonious
-High LPC
_-_ Lon LK
Effect o f r o t a t i n g the leader

to a new job
c'
with similar characteristics
/- -y
c 8' II
\ -
E f f e c t of Experience
High Moderate Lon
Control
Fig. 9. Schematic illustration indicating the points in time at which task- and relationship-
motivated leaders should be rotated to different jobs. Curved arrows indicate one possible rotation
strategy.
group or intensive leadership and task training. This same strategy will also apply
to the task-motivated leader whose situational control at the time of his selection
is moderate but expected to increase to a high-control situation with time and
experience (that is, from point B to point C on the graph).
If the organization opts for immediate performance, it will, of course;
choose the task-motivated leader for the situation which is, at the time of selec-
tion, low in control (point A ); it will select the relationship-motivated leader for
the moderate-control situation. We must then expect that the performance of
these leaders will gradually decrease as their experience provides them with
increasing control and influence. The ideal strategy will, therefore call for rotat-
ing or transferring these leaders to a different situation at pointsX and Y. In other
words, the task-motivated leader who reaches point X should be reassigned to a
position which provides him either high or relatively low situational control,
while the relationship-motivatedleader who reaches point Y should be reassigned
to another position corresponding to point B. How much time will be required
before points X and Y are reached and require reassignment will depend, of
course, on the complexity of the task, the nature of the group, and the intellectual
resources of the leader, as well as the availability of training. An organization
may well consider giving a task-motivated leader in situation A , or a
relationship-motivatedleader in situation B., a difficult group of subordinates,
or a more challenging task. There may even be reason for discouraging
leaders in these situations from obtaining task training. As we shall see, this
counterintuitiverecommendation is in part supported by several studies described
below.
C. LEADERSHIP TRAINING
The effects of leadership training programs are often difficult to identify

since the specific procedures and approaches differ widely from one program to
another. The results of leadership training have been generally disappointing,
however. For example, Campbell, Dunnette, Lawler, and Weick (1970, p. 325)
summarize the literature to that time by saying,
Thus, with regard to the bulk of the literature on training effects, it remains to be demon-
strated whether the changes in the criteria used to measure training effects have any impor-
.
tance for the organizations goals. . . it musf be demonstrated that what is learned in a
training program contributes to making an individual a better manager.
Stogdills survey of the literature reported in his massive Handbook of Leader-

ship (Stogdill, 1974, p. 199) led him to a similar conclusion.
It seems possible, however, that many leadership training programs may
well be effective in changing behavior or increasing the leaders situational
control without affecting average performance. That is, the overall effects of
82 FRED E. FIEDLER
increasing the situational control of high- and of low-LPC leaders may simply
cancel each other out by making some leaders more and others less effective.
This is well-illustrated by an ingenious experiment conducted by Chemers, Rice,
Sundstrom, and Butler (1975) at the University of Utah. In this study, ROTC
cadets were given the LPC scales, and those whose scores fell into the upper and
the lower third of the distribution were assigned to leadership positions of 4-man
teams. The other ROTC cadets, along with students from a psychology class,
served as group members. The groups task consisted of deciphering simple
coded messages, that is, cryptograms. The number of letters, correctly identified
within a given time, constituted the criterion of performance. The groups were
randomly divided so that half of the high- and half of the low-LFC leaders
received training in decoding. This training consisted of teaching the leaders a
number of simple rules by which these tasks can be accomplished (e.g., the most
frequent letters and words in the English language is e). This training proce-
dure had the effect of structuring the task since the leaders now were given a
specific best procedure to follow, as well as guidelines which indicated
whether they were on the right track.
As it happened, the group atmosphere was rated as quite low by leaders (in
large part because there was considerable ill feeling between the ROTC and
psychology students, and because the ROTC cadets felt that they were on the
spot). None of the leaders had high position power. Thus, given poor leader-
member relations, low position power, and an unstructured task, the situation for
untrained leaders fell into the low situational control zone (octant 8), while the
trained leaders also had poor leader-member relations and low position power
but a structured task. Their leadership situation, therefore, provided corre-
spondingly greater control (i.e., approximately octant 6).
Figure 10 shows the results of this study. In the untrained condition the
task-motivated (low-LPC) leaders performed better than did the relationship-
motivated leaders. However, in the trained condition the relationship-motivated
leaders were more effective. Most importantly, however, the task-motivated
leaders with training had less effective groups than did the task-motivated leaders
without training. In other words, the interaction was highly significant. Thus, we
can infer that the training given to task-motivated leaders actually decreased their
effectiveness, as the Contingency Model would predict. In a personal communi-
cation, Chemers pointed out that the low-LPC leaders with training had irritated
their group members by their expertise and their disdain for group member
contributions, and that the resulting arguments seemingly led to poor perfor-
mance.
Various contaminating factors make it difficult to replicate these results in
field studies. For example, managers with many years of experience in an or-
ganization are likely to have had more training than managers who are new, and
many organizations send their managers to training courses as a reward for good
- High LPC
__-- LOW LPC
:If
1
\
/
TRAINED
I
UNTRAINED
Fig. 10. Comparison of task performance by high- and low-LFC leaders with and without task
training. From Chemers et al. (1975).
performance or as a sign of esteem rather than as a means for remedying poor

performance.
It should also be noted that some training programs deliberately or unwit-
tingly decrease leader control and influence. Training in participative manage-
ment obviously requires the leader to relinquish some of his power over the
decision-making process. Some programs strongly impress leaders with the
complexity and difficulty of the job and thus decrease the leaders confidence
that they have much control over their subordinates or the task. Thus, in one of
our unpublished analyses of Army personnel, noncomissioned officers with rela-
tively more training tended to perceive their tasks as somewhat more complex
and their position power as lower than was the case for those with relatively little
training.
Data have been collected in several field studies which measured the amount
of training i n terms of weeks of attendance in various leadership courses and
performance in terms of superiors evaluations. While these predictor and crite-
rion measures leave much to be desired, the results have generally been in the
expected direction. For example, in one study of police patrol sergeants, situa-
tional control was judged to be moderate for experienced and trained sergeants
and low for inexperienced sergeants. Here the correlation between amount of
training and performance was significantly positive (.84, n = 7, p < .01) for
relationship-motivated sergeants, but slightly negative (- .14, n = 8) for task-
motivated sergeants (Fiedler, 1972a).
84 FRED E. FIEDLER
The effects of training are obviously complex. We need to understand how

different approaches to task training and human relations training affect the
leaders actual and perceived situational control, and we need to consider
whether training might also affect the leaders personality or motivational struc-
ture. As the final section of this chapter will show, the Contingency Model has
led to the development of a training program which teaches the leader to change
the situation to fit his or her personality rather than attempting to change leader
personality or behavior to fit the situation.
D. ORGANIZATIONALTURBULENCE
Turbulent and unstable conditions in an organization result in uncertainty

and thus lower the leaders situational control (Bons & Fiedler, 1976). These
decreases in situational control might arise from any number of reasons. The
most common of these are shake-ups in the organizational hierarchy, the
reassignment of leaders to different units or jobs, or shifts of the organizations
position in response to new economic or political demand (e.g., new product
lines, new competitors, new legislative constraints).
The effects of certain organizational changes on the leaders performance
were recently investigated in a longitudinal study of an infantry division (Bons &
Fiedler, 1976) in which 135 Army squad leaders were tested and evaluated
shortly after the units were formed and again 5-8 months later.
The organizational change was defined as occurring if a squad leader was
given new subordinates, new bosses, or a new job. (Note that this was a different
gmup from that described in Fig. 5 . ) A military setting for a study of this type is
particularly appropriate, since personnel and job changes in military units are
quite frequent during the first year of an organizations existence and, therefore,
permit the investigator to follow personnel through an entire assignment cycle.
The typical squad leader is a regular Army sergeant with 5-6 years of
service. He commands a group of 8-10 enlisted men and he reports to two
superiors, namely a second lieutenant who is the platoon leader, and a senior
noncommissioned officer who is the platoon sergeant. In the present study, the
performance of the squad leaders was evaluated by both of these superiors. The
performance scale consisted of eight factored items relevant to task performance
and eight items relevant to personnel performance, specifically morale, disci-
pline, and job satisfaction. In addition, as we shall discuss in more detail later on,
the leaders behavior was also rated by his subordinates on a variety of scales
including the Consideration and Structuring items of Stogdills (1963) Leader
Behavior Description Questionnaire XI1 and rewarding and punitive behavior.
Inexperienced squad leaders, as a group, generally described their situation
at the first testing session as providing low situational control. Situational control
increased to the moderate level for those squad leaders whose organizational
environment remained stable, that is, who remained with their squad throughout
the entire 6- to 8-month period and served under the same platoon leader and
platoon sergeant throughout that time. The experienced squad leaders, with more
than 5 or 6 years of Army service, rated themselves as having moderate situa-
tional control at first and relatively high control at the termination of the training
period, provided their situation remained stable. Since the changes in subordinate
units, or in assignment to a different superior, were an old story for experienced
leaders but were much more unsettling for inexperienced leaders, we shall here
deal with the latter group.
The Contingency Model enables us to make the following predictions: (a)in
the stable condition, the inexperienced leaders will move from the low into the
moderate-control zone and relationship motivated leaders should perform better
than task-motivated leaders; (b) in the condition of organizational turbulence,
which decreases the situational control of the leader from moderate to low control
(or confines him to the low-control situation), the task-motivated leader should
perform better than the relationship-motivatedleader. In other words, a decrease
in situational control should increase the effectiveness of the low-LPC leader but
decrease that of the high-LPC leader.
The results supported these hypotheses. The data points in Fig. 11 show
performance at time 2, when performance for time 1 and leader intelligence are
statistically controlled by covariance analysis. In effect, the graph permits us to
infer that the performance of the low-LPC leader increased somewhat in the
turbulent condition between time 1 and time 2, while that of the high-LPC leader
markedly decreased. In the stable condition, the performance of the high LPC
400
380
-
---
High LPC leaders
Low LPC leaders
360
340
320
.
304
- -$--- --- -
280
260
240
220
200
NO Change Change
Fig. I 1 . Task performance at time 2 by inexperienced high- and low-LPC leaders in stable and
unstable organizational conditions with time 1 performance and leader intelligence controlled by
covariance analysis. Midline indicates estimated performance at time 1.
86 FRED E. FIEDL.ER
leader increased while that of the low-LPC leader remained relatively un-
changed.
To summarize this discussion, the importance of these findings is threefold.
First and foremost, our work suggests that the Contingency Model provides a
framework for explaining the dynamic nature of the leadership process. Second,
the data contradict theories of leadership which implicitly or explicitly hold that
one particular managerial style determines good leadership performance.
Likewise, our results contradict the assumption underlying many leadership
training programs which seek to increase leadership performance by increasing
the leaders control and influence. Rather, our data show that an increase in the
leaders situational control and influence will have a differential effect on the
performance of relationship- and task-motivated leaders. Likewise, a decrease in
situational control may improve performance by making the task seem more
challenging and exciting. In large part, the Contingency Model thus helps to
explain why leadership experience and leadership training have usually failed to
increase leadership performance across the board, and why managerial selection
and rotation generally have yielded mixed results.
V. Meaning of Situational Control and LPC
As the preceding sections have shown, the performance of a group or

organization depends to a considerable part on the proper match of the leaders
personality attributes (measured by LPC) and the degree to which the situation
provides him or her with control and influence. We shall here consider these two
main elements of the theory in greater detail.
A. SITUATIONAL CONTROL
An operational definition of situational control or favorableness has been

presented in the earlier part of this chapter in terms of leader-member relations,
task structure, and position power, The situational control dimension can be
conceptually defined as the degree to which the situation enables leaders to
determine the outcome of their decisions and actions. A high degree of control
will therefore make the leader feel less anxious and more certain that the task will
be successfully completed.
A number of studies support this interpretation. First, the reported tension
and anxiety of leaders is significantly higher in situations in which leader-
member relations are poor, the task is relatively unstructured, and position power
is low (Fiedler, 1966, 1967). Second, Nebeker (1975) showed that high situa-
tional control can be interpreted as low decision uncertainty. He constructed a
scale which asked subjects to indicate the probability that such behaviors as
criticizing poor work, supervising, instructing, and dealing with subordinates

will have the intended results. These managers were also asked to indicate how
much control and influence they thought they had over the productivity and
morale of their work group, using a continuous measure of situational control (see
page 67).
In a study of 43 naval aviation maintenance shop supervisors, Nebeker
found a correlation of -.56 (v < .01) between situational control and decision
uncertainty, and a correlation of - .30 (D < .05) between perceived control and
decision uncertainty. In a second study of 49 public works managers, the correla-
tion between situational control and decision uncertainty was - .25 (D < .05) and
between situational control and self-perceived control and influence .55 (v
< .01).
The scaled situational control or favorableness dimension is a theoretically
weighted combination of three scales. The question now arises whether the
theoretical weights (giving four times as much weight to leader-member rela-
tions and twice as much to task structure as to position power) reflect an optimal
combination of these three factors. Nebeker compared
a multiple regression solution, where the components of situational favorability are used to
predict decision uncertainty, with the zero-order coefficients obtained by the theoreiical
~ombination.~ In both studies, the actual coefficients obtained were equivalent for either
method [r = .56 vs. R = .58 in one study, and I = .40vs. R = .39 in the other].
On the basis of these findings Nebeker concluded,
It is impressive that the zero-order coefficients account for just as much of the variance as do
the multiple regression coefficients. Such a comparison actually favors the theoretical
combination, since the multiple regression solution is a best fit and can, therefore,
capitalize on error variance. It appears that Fiedlers theoretical combination of the compo-
nent variables of situational favorability is about as close to optimal in predicting uncer-
tainty as we could possibly expect. [Nebeker, 1975, p. 2921
A subsequent study by Mai-Dalton (1 975) found an almost identical relationship

(R = .54) between rated decision uncertainty and the situational control sub-
scales.
Still another experiment by Beach et al. (1978, in press) asked students to
estimate the probability of success and the situational favorableness of a series of
hypothetical leadership situations. The experiment presented these leadership
situations as having either good or poor leader-member relations, high or low
task structure, and high or low position power, using a 2 X 2 X 2 factorial design.
The two measures estimated probability of success and situational favorableness,
correlated .89 (p < .Owl), suggesting that these are essentially identical con-
cepts. A multiple correlation was then computed with situational favorableness
or control as the criterion and the three situational control subscales as predictors.
88 FRED E. FIEDLER
The Beta weights obtained by Hoffmans (1960) method were quite similar to
those predicted by the Contingency Model, that is, .45 for leader-member rela-
tions, .33 for task structure, and . 1 1 for position power, roughly in the propor-
tion of 4:3:1.
These findings have important implications for integrating the Contingency
Model with the work of other decision theorists (e.g., MacCrimmon & Taylor,
1976; Schroder, Driver, & Streufert, 1967). This can be seen by examining the
information searching behavior of high- and low-LPC persons under condi-
tions of high and of low uncertainty. For example, a study by Mai-Dalton (1975)
asked subjects to complete an In-Basket Test which requires responses to letters,
memoranda, or messages addressed to the leader within a given period of time.
In moderate-control situations in which high-LPC leaders tend to be most effec-
tive, these individuals were also most likely to ask for additional information,
while low-LPC persons manifested most information searching behavior in
high-control situations.
A strong argument also can be made for considering such variables as
mechanistic versus organic climates (Bums & Stalker, 1961), environmental
uncertainty (Lawrence & Lorsch, 1967), and organizational climate (House &
Rizzo, 1971) as special cases of situational control. A mechanistic climate is
defined as one in which the organization is stable and therefore predictable, while
an organic climate is more ambiguous and fluid and thus less predictable. Csoka
(1975) conducted a study of Army dining hall stewards in 52 Army mess halls.
He predicted that the high-LPC leaders would perform better in an organic
climate while low-LPC leaders would excel in a mechanistic climate. Organiza-
tional performance was rated on standardized scales by company commanders
and by food service officers who evaluated the quality of the meals as well as the
attractiveness and sanitation of the service.
As predicted by the Contingency Model, task-motivated (low-LPC) leaders
performed better in dining halls rated as having a relatively mechanistic climate
(-.Sl, n = 17, p < .OS), while relationship-motivated(high-LPC) leaders per-
formed better in the dining halls rated as having an organic climate (+S 7 , n =
17, p < .OS). (Leader LPC and group performance for all of 52 dining halls
correlated - .08.)In an effort to approximate the Contingency Model situational
control dimensions still further, Csoka categorized the groups as high or low on
leader-member relations, as having a mechanistic or organic climate (to substi-
tute for high or low task structure), and as having a high position power in all
groups. The resulting correlations shown in Table V again support the predic-
tions. The Csoka study thus provides an important bridge between the Con-
tingency Model and other theories by demonstrating that a mechanistic or organic
organizational climate is neither conducive nor detrimental to performance.
Rather, it affects the leaders control and influence, and thus determines in part
whether high- or low-LPC leaders will be more effective under the given condi-
tions.
TABLE V
CORRELATIONS BETWEEN LPC AND PERFORMANCE
RATINGS IN EACH OCTANT REORDERED ON A
MECHANISTIC-ORGANIC CONTINUM
Octant I Octant 6 Octant 3 Octant 5
- .66' - ,806 - .69' .80a

n = 14 n = 20 n = II n = 7
B. LPCSCORE
LPC has generally been seen as the weakest part of the Contingency Model.
The measure has been difficult to interpret, and a number of questions remain,
even after more than 25 years of continuous research. This section briefly dis-
cusses some methodological issues related to LPC, and the interpretation of the
score.
I . Reliability and Stability of LPC
The internal consistency of the LPC score has never been in question. A
recent review by Rice ( 1 978) reports a mean split-half reliability from various
reported studies of .88. Some investigatorshave found separate interpersonal and
task factors in previous scales (e.g., Fox, Hill, & Guertin, 1973; Yukl, 1970) but
the task factor has been relatively unimportant in most of these analyses. A new
18-item scale shown on page 61 of this chapter has been designed to mini-
mize task factor items and has somewhat higher internal reliability than did
the previously reported scales.
The more critical problem is the retest reliability or stability of the LPC
score, since it determines whether the measure can serve as a predictor of leader-
ship. The stability of the score also has bearing on whether LPC measures a
transitory attitude, as has been claimed by Fishbein, Landy, and Hatch (1969b);
and Stinson and Tracy (1974), or a more stable personality attribute.
The inference that the LPC score must have some stability over time is
supported by such investigations as Chemers and Skrzypek (1972), Fiedler et
al. (1969), and Hardy and his associates (1973; Hardy, 1971, 1975; Hardy &
Bohren, 1975). since all of these obtained the leader's LPC scores several weeks
prior to the study. If LPC measured a transitory attitude, it is unlikely that
significant results in the hypothesized direction could have been obtained.
The question of stability has now been extensively investigated by Rice
(1978) whose survey of the literature on LPC uncovered 23 retest correlations.
The retest reliabilities in these studies ranged from .01 to .91 with a median of .67
and a mean of .64(SD= .36). A correlation between the length of the test-retest
90 FRED E. FIEDLFR
interval and the magnitude of the stability coefficient was -.30 (n = 23, n.s.).
Since the interval between testing ranged from a few days to 2% years, these
findings indicate that the magnitude of the stability of LPC is not primarily a
function of time. Bons (1974) obtained a retest reliability for a group of higher
level Army leaders over a 5-month period of .72 (n = 45, p C .001) and Prothero
and Fiedler (1974) obtained a retest correlation of .67 (n = 18, p < .01) for
faculty members of a school of nursing over a 16- to 24-month period. There is
evidence that a very unsatisfactory work experience in a laboratory experiment
can change the LPC score temporarily, and that a major life change, e.g., a
stressful combat assignment may affect LPC (Bons, Bass, & Komorita, 1970).
On the other hand, to place the median retest correlation for LPC of .67 in some
perspective, it should be noted by way of comparison that the test-retest correla-
tion of the Minnesota Multiphasic Inventory (MMPI) over a period of 1 week
was .60, and the median stability coefficient of the Hartshorne and May honesty
scales over 6 months was.50 (Sax, 1974). Mehrens and Lehmann (1969) re-
ported the stability of the California Psychological Inventory for 13,OOO subjects
over a 1-year period as .65 for males and .68 for females. Thus the stability of
LPC clearly falls within the acceptable range for established personality attribute
measures. The variance of retest reliabilities appears to be high. It should be
pointed out, however, that the lowest correlations of .01 came from participants
in a civil service executive development workshop, and the next lowest correla-
tions of .23 came from a group of undergraduates participating in an 8-week
business game. In both of these cases, the training program may well have
created demand characteristics or even implicit instructions as to the proper
attitude one should adapt toward poor co-workers (Rice, 1978). It is, in any
event, difficult to compare the variance of retest correlations of LPC with the
variance of other social psychological measures since it is rare to find even one or
two stability estimates reported (Robinson & Shaver, 1973).
2. Meaning of LPC
There has been considerable controversy about the precise interpretation of
LPC. It has been described as a measure of an attitude (Fishbein, Landy, &
Hatch, 1969b), of cognitive complexity (Mitchell, 1970), of work orientation
(Rice, 1975), and of a motivational hierarchy or goal structure (Fiedler, 1972b).
There is very little question, however, that individuals who perceive their LPCs
in very negative and perjorative terms (low LPC) are basically more concerned
with the task than with their interpersonal relations; those who perceive their
LPCs in a more positive and differentiated manner (high LPC) are basically more
concerned with interpersonal relations than with task accomplishment.
The main problem of defining the exact meaning of the LPC score stems
from two sources. First, LPC does not consistently correlate with any of the usual
personality tests. While this is good insofar as it tells us that we are measuring a
unique trait or attribute, it also precludes the usual construct validation procedure
of defining one test score in terms of other tests or personality measures. A
second source of difficulty in defining LPC derives from the fact that high- and
low-LPC persons tend to behave differently in situations in which their control is
high and in which their control is low. We shall return to this second problem
after reviewing the main effects attributable to high and low LPC scores.
At the simplest level, the LPC score can be seen as measuring an attitude of
the leader toward the person with whom he or she finds it most difficult to work,
that is, someone who is extremely frustrating when a leader tries to get a job done
(Fishbein et al., 1969b). Different people get frustrated for different reasons.
Someone who cares passionately about getting a job done will react with a great
deal of acrimony and resentment when someone poses a threat to job accom-
plishment. On the other hand, someone to whom the task is of lesser importance
will react in a more moderate and detached manner toward a poor co-worker. He
will be less emphatic in his condemnation and perhaps less infuriated by the
person who does not do his share. The extremely negative ratings by the low-LPC
leader suggest, therefore, that the poor work relationship is of such ovemding
importance to the leader that it completely distorts and colors his perception of
the poor co-workers other personality attributes. A high-LPC leader seems to
look at the poor co-worker first as an individual, as someone who could still be
valued for other characteristics rather than just for his contribution to the
In addition, there are also qualitative differences in the way in which people
with high and with low LPC scores perceive those with whom they cannot work.
Fishbein et al. found that high-LPC persons tend to describe their LPCs as
bull-headed, bossy, dogmatic people who avoid work, are know-it-alls, and thus
threaten smooth interpersonal relations in a work group. Low-LPC persons tend
to see their LPCs as unintelligent, unreliable, self-centered, and lazy-in other
words as poor team players who are detrimental to getting the job done.
The relatively greater concern for interpersonal relations by high-LPC lead-
ers, and for task performance by low-LPC leaders, is also shown by Bordens
( I 976) recent study which correlated the degree to which leaders perceived their
job as valuable and the way they perceived their interpersonal relations with
superiors and with subordinates. Borden noted that the leaders boss as a rule not
only defines the task, but that his evaluation of the leader also defines the
effectiveness with which the task has been accomplished. A good relationship
sThere is some evidence of a middle-LPC group which differs in behavior and attitudes from
both high- and low-LPC persons. These individuals, who comprise about 7-101 above and below
the population mean of 60,either have a mixed motivation pattern or else a distinctly different
leadership approach. They seem less concerned about the opinions and attitudes of others and less
dependent on superiors and subordinates evaluations, as well as less desirous of accepting a
leadership position (Bass,Fiedler, & Krueger, 1964, Mai-Dalton, 1975). Further research is needed
to describe this group in a more adequate manner.
92 FRED E.FIEDLER
with the boss has a number of advantages for a leader who is concerned about his
job. It enables him to communicate more freely with the person who assigns and
evaluates task performance and, consequently, also to obtain better guidance on
how the task is to be done. According to Borden's hypothesis, the low-LPC
leader's evaluation of the job will correlate with the degree to which he has a
good relationship with his boss. In contrast, a good relationship with subordi-
nates implies that the leader can rely on his group for emotional support and
satisfaction of interpersonal needs. The value which the job has for the high-LPC
leader should correlate with the leader's perceptions of leader-subordinate rela-
tions. This relationship will not hold for low-LPC leaders. As Table VI shows,
the results of analysis for a sample of platoon sergeants and a sample of platoon
leaders generally support these hypotheses.
In a similar vein, Hansson and Fiedler (1973) found evidence that high-
L K persons tend to be more concerned with being in harmony with their interper-
sonal environment than are low-LPC persons. High-LPC employees of a park
department expressed greater interest in continued employment the more they
agreed with management on policies, goals, and departmental purposes (.30,
.47,* and .39,* ns = 24;p* = < .05). In contrast, the corresponding correlations
obtained for low-LPC employees were low and insignificant (-. 12, .01, and .09;
11s = 19), indicating that the feeling of solidarity and identification with the or-
ganization is more important to high- than to low-LPC persons.
The same conclusion emerged from an unpublished analysis of men in 32
TABLE VI
CORRELATIONS BETWEEN LEADER-SUBORDINATE RELATIONS
AND THE DEGREE TO WHICH THE LEADER
CONSIDERS HIS JOB TO BE VALUABLE
Leader-superior relations Leader-subordinate relations
Leader High LPC Low LPC High LPC Low LPC
Squad leaders
(sample 1)
Squad leaders
(sample 2)
Platoon sergeants
(sample 3)
Platoon leaders
(sample 4)
Number of cases is indicated in parentheses.

ap < .05.
bp < .01.
=p < .001.
CONTINGENCY MODEL AND THE LEADERSHlP PROCESS 93
infantry squads. In each of these squads, one high- and one low-LPC man was
asked to describe himself and his squad leader on identical scale items. (As
previous research has shown, individuals feel closer to a person they perceive as
similar to themselves than to someone they perceive to be different, e.g., Fiedler,
Warrington, & Blaisdell, 1952; Secord, Backman, & Eachus, 1964.) Each of
these men also indicated how much he (a) enjoyed being a member of his squad;
(b) was interested in making a career of the Army, and ( c ) was satisfied and
contented with military life. Feeling close to the squad leader (perceiving him as
similar) was significantly correlated with satisfaction with military life for high-
LPC squad members (S4, .31, .45, respectively), but not for low-LPC squad
members (.W, - .I 1, - .05, respectively). Thus, the high LPC person associates
a good relationship with his boss (in this case, his squad leader) with a feeling of
wanting to belong to the organization; the low LPC person associates a good
relationship with his boss with a feeling that his job is satisfying and valuable.
Chemers and Skrzypek (1972) found that high-LPC leaders generally made
significantly more social emotional comments, while low-LPC leaders generally
displayed considerably more task-relevant behavior in various task situations in
their West Point study. Nebeker and Hansson (1972) found high-LPC persons to
be significantly more optimistic about human nature and positive in their at-
titudes toward others than were low-LPC persons, and hence more likely to seek
close interpersonal relations with others. The evidence from these and similar
studies is, therefore, quite consistent in showing that high-LPC persons are more
concerned and affected by interpersonal relations while low-LPC persons are
more concerned with task performance.
3 . LPC and the PersonSituation Interaction
The interpretation of the LPC score is complicated by the frequent finding
that the behavior of high- and low-LPC leaders is strongly affected by the control
and influence which the situation provides. In low-control situations and under
stress, high-LPC leaders generally attend to interpersonal relations while low-
LPC leaders attend to the task. However, in situations in which the leader enjoys
high control and influence, the high-LPC leaders behavior is seen as more task
relevant and less considerate while the low-LPC leader is seen as less concerned
with the task and as more considerate. In other words, not only the individuals
leadership performance but also his leadership behavior is affected by the interac-
tion of personality attributes reflected by LPC and situational control.
This is shown in a number of empirical studies. For example, Meuwese and
Fiedler (1965) compared the behavior of Reserve Officer Training Corps cadets
under relatively stress-free and relatively stressful conditions. Each group, con-
sisting of 3 men, performed two creative tasks (i.e., developing a new pay
system for the ROTC and inventing a fable about the Army for school children).
In both of these conditions the highest ranking cadet was appointed to be the
94 FRED E. FIEDLER
leader. In the low-stress condition the sessions were quite informal, the students
were told to come in civilian clothes, and were assured that their performance
would not affect their Army career. In the high-stress condition, the men came in
uniform and they addressed each other by rank. An Army officer with the rank of
major, lieutenant colonel, or colonel, who was unknown to the cadets, sat across
the table from the group and continually evaluated their performance. Sub-
sequent postsession ratings indicated that this last condition had been quite stress-
ful.
As Fig. 12 shows, in the low-stress condition, low-LPC leaders were rated
as democratic and promoting group participation while high-LPC leaders were
seen as showing more concern with the task, that is, as frequently proposing new
ideas and as integrating the ideas of others. In the high-stress condition, how-
ever, the low-LPC leader tended to behave in a task-relevant manner while the
high-LPC leaders behaved more frequently in a democratic, participative manner
which showed concern with interpersonal relations (Fiedler, 1967, pp. 191-192).
Similar results have been obtained in a study of group creativity by Fiedler
t o w stress External stress
&-
n
0 " L
55
LOW LPC +/'
/
Fig. 12. The effect of leader motivation (LF'C) and situational stress on interpersonal, social-
emotional, and job-relevant leader behavior. A. Relationship-oriented categories. B . Task-oriented
categories. From Meuwese and Fiedler (1965).
et af. (1961), an experiment on Air Force instructor behaviors in stressful and

nonstressful situations (Ayer, 19681, and in a study of school principals by
Hawley (1969).
These complex findings can be explained by considering the high and the
low LPC score as indicating a different motivational or goal hierarchy. High-
LPC persons are seen as giving highest priority in work situations to having close
and emotionally supportive interpersonal relations with subordinates. Low-LPC
persons, on the other hand, give highest priority to task accomplishment which
provides them with tangible evidence of their competence. (Intelligent and ex-
perienced persons may, of course, seek to attain these goals in a somewhat
different manner than will dull and inexperienced persons.)
Maslow (1954) has pointed out that satisfied needs cease to motivate. In
other words, assured that they will attain their primary goal, individuals will then
seek to achieve the goals which are next in priority. In groups which provide high
situational control the leaders have this assurance: High-LPC leaders have the
support of their subordinates; low-LPC leaders know that the task will be com-
pleted (Beach er al., 1975). Under these high-control conditions, the relationship-
motivated leader can then turn his attention to other aspects of his situation,
for example, seeking the approbation of such important others as the boss or
faculty member who serves as experimenter, as well as gaining more complete
control of the group. As one astute observer noted, in a high-control situation,
the relationship-motivated leader acts like an underemployed professional
mother, that is, in an overly controlling, managing, and demanding manner. The
low-LPC leader who knows the task will get done can sit back, relax, and behave
in a generally considerate and pleasant manner.
Several studies can be interpreted as providing direct support for the
motivational hierarchy hypothesis. Larson and Rowland (1973) worked with a
group of highway department managers in an executive development seminar. In
the course of this program, high- and low-LPC managers were asked to complete
an In-Basket Test which requires the individual to respond to various letters,
memoranda, requests, and messages which he supposedly finds on his desk.
Larson and Rowland randomly assigned the managers in their study to one of two
conditions. In a nonthreatening, low-stress condition, the managers were told
that the procedure was experimental and was administered so that their opinion
about its usefulness could be obtained. In the high-stress condition, the managers
were urged to do their best since the test was a well-known predictor of
administrative ability. The effectiveness of the stress manipulation was verified

by means of palmar sweat measures.

The responses to In-Basket items were rated as primarily showing concern
with the task or concern with interpersonal relations. As predicted, the high-LPC
managers gave more task-relevant responses in the low-stress condition and more
relationship-oriented responses in the high-stress condition. The reverse was the
96 FRED 8. FIEDLER
case for low-LPC managers who gave more relationship-oriented statements in

the low-stress condition and more task-relevant statements in the high-stress
condition.
Similar results were obtained in a laboratory experiment by Green,
Nebeker, and Boni (1976) in which 26 high-LPC and 26 low-LPC persons played
the part of a committee chairperson in a role-played and prerecorded committee
meeting. Each of the subjects listened to the tape-recorded committee session
which was interrupted at seven different times to permit the committee chair-
man to say just what you would say if you were actually there. He was to
respond to disagreements among committee members, requests for more infor-
mation, the need to deal with difficult committee members, etc. The leaders were
given either relatively high- or low-position power, and the prerecorded meetings
were designed to be either harmonious and cooperative or tom by conflict and
disharmony.
Ratings of tape-recorded responses showed that the chairmen with high
LPC were somewhat less considerate than were those with low LPC in the
high-control condition. In the condition which involved low situational control,
high-LPC chairmen were significantly more considerate and concerned with
interpersonal relations while low-LPC chairmen were significantly more task
oriented.
4. Changes in Situational Control and Their Efects on the Behavior of
High- and Low-LPC Persons
The dynamic nature of the relationship between situational control and
leader behavior becomes apparent in a number of studies. In one particularly
important experiment of this type by Sample and Wilson (1965), members of a
large introductory psychology class were divided into small groups under high-
and low-LPC leaders with the instruction to plan, run, and report an experiment
with a laboratory rat. Unknown to the students, an observer was assigned to each
group to record the same leaders behavior for each of these three phases, using
Bales (195 1) interaction process analysis system. The three subtasks could be
scaled in terms of the leaders situational control. Planning, the least structured
task, gave the leader low control. Reporting ranked next, then running the
experiment which gave the leader relatively high situationalcontrol since the task
at that point was very highly structured.
Figure 13 presents the results-of this study. The vertical axes of the graphs
show the leaders task-relevant behavior and socioemotional behavior, the hori-
zontal axes show the three phases in order of high, medium, and low task
structure, respectively, that is, running, reporting, and planning. As can be seen,
the low-LPC leaders made more positive social-emotional comments and fewer
task-relevant statements in the running phase than did high-LPC leaders. How-
ever, in the planning phase in which the leaders situational control was low,
-
\
RUNNING WRITING PLANNING
:,
2,
0
w
t-
%
c
4
RUNNING WRITING PLANNING
Fig. 13A and B . Changes in social-emotional and task-relevant behavior by relationship- and
task-motivated leaders in different phases of the same group task. Data from Sample and Wilson
(1965). Reprinted from Fiedler, 1971c with permission.
low-LPC leaders showed concern with the task and high-LPC leaders made more
socioemotional comments.
The dynamic interaction between leader LPC and situational control be-
comes especially clear in an experiment involving human relations training for
leaders. If the motivational hierarchy interpretation of LPC is correct, training
which increases leader-group relations will increase the leader's situational con-
trol. With greater situational control (a shift to the left on Fig. 13) high-LPC
leaders should become less considerate and less concerned with their interper-
sonal relations while task-motivated leaders would become more concerned with
interpersonal relations and, hence, more considerate. An experiment was con-
ducted in Iran by Chemers (1969) which used 48 Americans living in Iran who
served as leaders and 96 Iranians who volunteered to serve as group members.
The Americans were split into high- and low-LPC groups. Half of each group
was given a control training on the physical geography of Iran and the other half
received culture assimilator training designed to increase the ability of Americans
to get along with Iranians and to interact with them in a more effective manner.
[The effectiveness of this culture training program has been reported in a number
of studies (see Mitchell, Dossett, Fiedler, & Triandis, 1972).]
At the conclusion of three tasks, the Iranian group members rated the
98 FRED E. FIEDLER
American leader, the leaders consideration behavior, and the group climate. The
mean ratings, converted toe-scores, are shown on Table VII. As can be seen, all
interactions were significant in the predicted direction: The low-LPC leaders who
received training were seen as more considerate (Fig. 14), the group climate was
rated as better, and leaders were rated highly. The high-LPC leaders who had
received the identical training were less esteemed, were rated as less considerate,
and the group climate was seen as less pleasant. No differences were found
between high- and low-LPC leaders in the control condition. We may thus infer
that the training increased the leaders situational control and that this increased
situational control had a beneficial effect on the interpersonal behavior of low-
LPC leaders but a detrimental effect on that of high-LPC leaders.
Similar changes in leader behavior also occur in real-life conditions when
factors in the organizational environment increase or decrease the leaders situa-
tional control. This is shown in the previously mentioned study of infantry squad
leaders (Bons & Fiedler, 1976). As will be recalled, leaders were assigned to
new squads shortly after formation of the units. They were tested and evaluated
at that time and again some 5-8 months later. In the intervening period, some of
the squad leaders remained in a stable environment in which they worked with
the same superiors and the same group of subordinates and, therefore, increased
their situational control as they gained experience on the job. Other squad leaders
had to adjust to new bosses and to new subordinates. We expect, of course, that a
turbulent environment decreases the leaders situational control. In the latter
situation, the high-LPC leaders should seek closer interpersonal ties, while the
low-LPC leaders should seek greater control over the group in order to accom-
plish the task.
Control in a military organization is primarily exercised by offering rewards
or by implicit or explicit threat of punishment to assure compliance in the per-
TABLE VII
INTERACTION OF TRAINING AND LEADERSHIP STYLE ON
SEVERAL MEASURES OF INTERPERSONAL RELATIONSa
Culture Geography
Scale
Member ratings of: HighLPC Low LPC High LPC Low LPC F p
Group atmosphere - .354 .I18 .091 -.029 5.00 .05

Leaders consideration behavior - ,366 .342 .033 -.216 11.00 .01
Evaluation of leader - .403 .205 .037 -.001 6.73 .025
Climate: liking for situation -.421 .231 .002 .017 6.27 ,025
aStandardized (2) scores.

*From Chemers (1969).
.30
.O 1 n
Human Relations Control

Training Training
Fig. 14. Subordinates behavior ratings of high- and low-LPC leaders with human relations
training and with control training. From Chemers (1969).
formance of the task. Rewarding behavior clearly is more conducive to the

development of good interpersonal relations than is punitive behavior. We would
expect, therefore, that these two types of behavior will be used differently by
high- and by low-LPC leaders as their situational control changes. Figure 15
shows the subordinates ratings of their squad leaders rewarding behavior. The
points on the graph indicate the rated behavior at time 2, with behavior ratings
and leader intelligence statistically controlled by covariance analysis. The dashed
midline on the graph indicates the mean behavior score at time 1 , that is, the
presumed ratings leaders obtained at the first rating period.
At the left side of the graph we find that the leaders rewarding behavior in a
stable leadership environment did not change noticeably between time 1 and time
2. However, in the unstable condition, in which the leader changed jobs, high-
LPC squad leaders became dramatically more rewarding in their interactions with
subordinates, while low-LPC leaders became drastically less rewarding, that is,
less concerned with good interpersonal relations.
Figure 16 shows the results of a similar analysis for threatened or actual
punitive behaviors (official punishment or court-martial). The analysis compared
leaders who were or who were not given a new squad in the time intervening
between the first and the second evaluation and testing periods.
As can be seen, all leaders tended to become more punitive by time 2, and
there was no difference in the behavior of relationship-motivated leaders in the
100 FRED E. FIEDLER
700 I Hfgh LPC
Fig. 15. Rewarding behavior of high- and low-LPC leaders at time 2 in stable and unstable
organizational conditions, with leader intelligence and behavior scores at time 1 controlled by
covariance and analysis. Midline indicates mean scores at time 1.
600
580
560
41
460 T
Fig. 16. Administrative-punitive behavior of high- and low-LPC leaders at time 2 in stable and
unstable organizational conditions, with leader intelligence and behavior scores at time 1 controlled
by covariance analysis. Midline indicates mean scores at time I .
stable and in the unstable leadership condition. However, task-motivated leaders

became significantly more punitive in the unstable condition in which their
situational control decreased, presumably because they felt a need for more
control over group members in order to accomplish the task.
C. LEADER BEHAVIOR AND ORGANIZATIONAL PERFORMANCE

In light of the foregoing discussion, let us now attempt to explain effective
leadership performance in behavioral terms for the typical case.
In high-control situations, high-LPC leaders are seen as inconsiderate and
more concerned with the task than with their interpersonal relations. Low-LPC
leaders are rated as relatively unconcerned with the task, but considerate, pleas-
ant, and relaxed in their interactions with group members. Recall here that a
high-control situation implies that the leader has the groups support as well as a
highly structured task and high position power. Under these conditions, the
group knows what the task involves and wants to perform it. A group of
motivated and knowledgeable subordinates obviously requires no more than
minimal guidance from the leader.
The low-LPC leader in a high-control situation provides minimal guidance.
If the motivational hierarchy hypothesis of LPC is correct, the low-LPC leader
behaves in this fashion because his primary goal of task accomplishment is being
satisfied, he trusts his group, and he can now focus on his secondary goal of good
interpersonal relations. The high-LPC leader, whose primary goal of securing
group support is satisfied, now turns to the task. He is generally described in such
situations as directive, managing, autocratic, and structuring. He provides strong
leadership where none is needed. In fact, his behavior may very well interfere
with task accomplishment: His overcontrolling behavior may be resented by
group members who are already doing their best on the job. In short, the high-
LPC leaders behavior may be maladaptive in high-control situations.
In moderate-control situations, in which high-LPC leaders generally per-
form best, low-LPC leaders become increasingly concerned with the task. They
are described as relatively more task-oriented, structuring, and controlling, and
as less considerate and less concerned with interpersonal relations. High-LPC
leaders, on the other hand, are generally rated as considerate and interpersonally
oriented.
To understand the full significanceof these behavior patterns on subordinate
performance, we must recall the nature of the two moderate-control situations.
One of these (octant 4) is usually characterized by tasks which involve free
expression of members opinions, attitudes, and information. These are tasks
which require tact and sensitivity on the part of the leader. Examples are policy
and decision-making committees, research and development teams, and organi-
zations in which the leader plays a coordinating rather than a controlling role.
102 FRED E. FIEDLFR
These situations frequently require the leader to reconcile conflicting viewpoints,

to negotiate between subgroups, and to encourage and support new ideas. The
second moderate-control situation involves work units in which the leader has
high task structure and position power, but relatively poor relations with the
group (octant 5 ) . Again, the leader must be tactful and sensitive under these
conditions. While he may know exactly what to do, and while he may have the
legitimate power to enforce his decisions, prudence and concern for productivity
demand that he behave with circumspection lest the group turn against him or
engage ih subtle sabotage. The high-LPC leader, who attends to the interpersonal
problems in his group, is likely to succeed here. The low-LPC leader, who
withdraws from interpersonal relations and focuses on the task to the neglect of
group members, is likely to find an unwilling and uncooperative work team.
In the low-control situation, which is characterized by poor leader-member
relations, low task structure, low position power, and/or high stress, the low-
LPC leaders tend to perform well.
The high-LPC leader tends to become anxious in this situation and seeks
emotional support. He is relatively nondirective, nonstructuring, and he pays less
attention to the task. In conditions of very high stress, high-LPC leaders may
withdraw almost completely from the leadership role, as indicated by the way
their subordinates tend to describe them, namely as nondirective, nonstructuring,
nonsupportive, and inconsiderate. It is then not surprising that the task perfor-
mance of high-LPC leaders is poor in low-control situations.
Low-LPC leaders, on the other hand, are described as directive, structuring,
and concerned with the task, but inconsiderate and unconcerned with their rela-
tions with subordinates. In other words, the low-LPC leader is likely to be
effective because he is firmly committed to the task regardless of the conse-
quences which this might have on his relations with subordinates. If they do not
perform well, he will not hesitate to reprimand or punish them or to discard them.
This pattern of behavior is well-illustrated by company troubleshooters and
start-up men who are sent to new or ailing offices and plants, or to organiza-
tions which are in a state of crisis. These individuals enjoy the challenge of a new
job and they perform well under stress, but they rarely remain highly effective
after the organization has again settled into a comfortable routine. Under these
circumstances, which are likely to provide moderate situational control, the
organization is likely to call on a manager who has the ability to keep people
working together in harmony and who has the knack for keeping the organization
running smoothly, namely, a high-LPC leader.
D. WHAT DOES LFC REALLY MEASURE?

Let us now briefly recapitulate what we know about LPC. First, and perhaps
foremost, LPC is an internally consistent measure which is relatively stable over
time. It thus appears to be a personality attribute which is transsituational in

nature.
It is, no doubt, discouraging that there should still be questions about the
interpretation of this score after more than 25 years of continuous research.
However, this is a problem which LPC shares with many other social-
psychological measures including, for example, cognitive complexity, locus of
control, or the K Scale of the MMPI. In fact, as recent controversies show, there
is even doubt that we know, after all these years, what intelligence tests really
measure. Admittedly, LPC poses a more difficult problem.
As has been said earlier, it may be argued whether LPC is an orientation
toward work, an attitude, a cognitive complexity measure, or an index of a
motivational hierarchy. However, there is general consensus among those who
have worked with this measure that a high LPC score identifies a person whose
major concern and first priority is the development and maintenance of good
interpersonal relations; a low LPC score identifies a person whose major concern
and priority is the accomplishment of the task or, perhaps in a more general
sense, to obtain tangible evidence of his or her own competence. These concerns
are especially apparent in conditions of low situational control, high stress, and
high uncertainty. There is consistent but less compelling evidence that the oppo-
site types of behaviors occur in situations which provide the leader with a high
degree of control and certainty. These findings favor a motivational hierarchy
interpretation, although other interpretations, consistent with these findings, are
also tenable. Whatever the precise and final interpretation of the LPC score might
turn out to be, there is very little question that it measures a personality attribute
which has very important consequences for organizational behavior.
VI. A New Approach to Leadership Training6
One acid test of a dynamic leadership theory is its ability to increase the
leaders effectiveness and the organizations performance. Our ability to control
a process is, of course, a strong indication that we understand it. The Con-
tingency Model suggests that we can improve leadership performance by chang-
ing the match between the leaders personality and his or her situational control
EThiseffect may have broader implications for relating leadership theory and personality theory.
A possible bridge between personality and leadership theories becomes apparent when we recall (a)
that leader uncertainty and anxiety tend to be high when situational control and influence are low, and
(b) that high- and low-LPC persons attempt to cope with anxiety and uncertainty in a different
manner. There is also growing evidence that the intelligence of the leader complexly interacts with
the LPC score and the situation in affecting behavior. Since intelligent people generally have more
resources for dealing with anxiety and uncertainty, this finding is not surprising. The exact role which
intelligence plays in these interactions is now under investigation.
104 FRED E. FTEDLER
and influence. Increasing leadership perfonnance implies then either changing

the leaders personality attributes which are reflected by LPC or else changing
the leaders situational control. The latter obviously presents the easier task,
particularly since organizational changes as well as leadership experience and
training imply almost constant changes in the leaders situational control. One
method of dealing with the problem of maintaining the right match between the
leaders personality and the control and influence the situation provides is to
teach the leader to modify the situation to match his personality. A training
program to accomplish this purpose has recently been devised.
A . GENERAL DESCRIPTION OF THE PROGRAM
The training program, called Leader Match (Fiedler et al., 1976), is pre-
sented in the form of a self-administeredworkbook, which can also be augmented
by appropriate lecture-discussions and films. Completing the manual takes from
4 to 8 hours.
Each chapter of the workbook consists of a short explanation of the various
basic concepts of the Contingency Model. The explanations are followed by
several short problems, case studies, or leadership episodes (probes) for which
the trainee selects the best answer. The trainee is then given immediate feedback
on the response. An incorrect response requires review of the chapter and trying
the probe again in order to assure that the material is completely understood. The
manual also contains two review sections and sever$ short tests to identify
problem areas which need additional review.
The first part of the manual requires trainees to complete the LPC scale
which identifies and interprets their motivational structure. The next four chap-
ters are devoted to teaching trainees how to measure their situational control, that
is, leader-member relations, task structure, and position power, and to determine
the degree to which the trainees current leadership job ,and personality are
appropriately matched.
In the final chapters trainees learn to modify factors in their situation in
order to obtain the appropriate match between their personality and situational
control and to apply the program to the leadership situations of their subordinates
in leadership positions.
B. VALIDATION STUDIES
Nine validation studies of Leader Match have now been completed. Four of
these were conducted in civilian organizations and five in military settings. The
The trainees ability to measure situational control accurately, in an absolute sense, is, of
course, less important for this purpose than teaching them to focus on the relevant factors in the
leadership situation and how to modify these as needed.
latter are of particular importance, since a large number of leaders perform

identical tasks and, therefore, can be compared under more controlled condi-
tions.
I. Civilian Organizations
The subjects for the four civilian studies consisted of ( a ) volunteers of a
public health agency, (b)middle managers of a county government, (c) police
sergeants, and (a) public works supervisors and managers (Fiedler, Mahar, &
Schmidt, 1975). I n each study those eligible for training were randomly as-
signed to either the trained or the control condition; 2 4 months after training a
performance evaluation was obtained from at least two of the subjects super-
visors.
In all four studies, trained leaders were rated as performing better than
untrained leaders, with 80% of the performance items in the predicted direction
and 33% significantly so. However, because of serious attrition problems, and
the possibility that the attrition may have reflected poor motivation on the part of
some trainees, the results of these studies had to be interpreted with considerable
caution, and studies under more controlled conditions were essential.
2 . Military Organizations
Two tests were made with naval personnel, namely 52 junior officers and
chief petty officers from eight naval air squadrons, and 20 officers and petty
officers from a naval destroyer (Leister, Borden, Fiedler, 1977). One half of the
officers and petty officers were randomly assigned to a training group which
received the Leader Match program, while the other half were assigned to the
control group. The subjects in both groups completed several questionnaires. The
trainees then participated in two 4-hour supervised training sessions during which
they read the manual, viewed a 28-minute film, Leadership: Style or Circum-
stance (CRM Productions, 1974), and participated in a short discussion.
Performance ratings were completed prior to training for each subject by
three of his or her immediate supervisors. Ratings were again obtained 6 months
later from at least two of the same supervisors. It was thus possible to determine
the degree to which a particular leaders performance had increased or decreased
subsequent to training. The specific method involved conversion of pre- and post-
training performance scores to standardized scores and comparing the mean
change in performance scores of trained and control groups. This method of using
change scores is considered appropriate for a study of this nature (Overall &
Woodward, 1975).
The results indicated that the performance of the trained group had increased
significantly over the 6-month period when compared with the control group
on task performance items as well as on persona! performance items (p <
.001 and .002,respectively) (See Fig. 17). This means that the increase in
106 FRED E. FIEDLER
32
31
r
G
30
f2
r
5
.o
29
28
Untrained Trained untrained Trained

Prior t o Leader Match Trafning 5 i x Months After Leader Match Training
Fig. 17. Comparison of performance ratings of Navy officers and petty officers in the training
and control conditions before and after Leader Match training.
performance occurred in the interpersonal relations area, morale, and discipline,

as well as in assigned tasks. Moreover, a separate analysis of the naval air station
and destroyer samples indicated almost identical findings.
Two studies were conducted independently by investigators at West Point
using Army personnel (Csoka & Bons, 1978). The first of these dealt with
trainees who were scheduled to serve as acting platoon leaders in various battal-
ions throughout the United States during the summer months. One-third of 114
trainees were randomly selected for Leader Match training, while the others
served as controls. At the end of the assignment, the acting platoon leaders were
rated by superior officers who had no information about the study. Leaders
trained with Leader Match received significantly higher ratings in their unit than
did those in the control group (see Fig. 18). The second study involved cadets who
served as the leaders of 53 training platoons, with one half of these young men
randomly chosen to receive Leader Match training. At the end of the semester, all
platoon leaders were evaluated by their superior officers. As shown in Table VIII,
the platoon leaders with training were significantlymore often consideredto be the
best in their company than those without training. Interviews indicated that trained
leaders had been able to modify their situational control as recommended by the
training manual.
A ninth study has just been completed on 18 university and college ROTC
programs in which Leader Match was used in 9 randomly selected schools while
the other 9 schools served as controls. At the end of the training, cadets from
60
u 50
i
e
&
40
30
20
Leader Match Informed Control Uninformed Control
Trained Group Group
Fig. 18. Performance ratings given to acting platoon leaders who had received Leader Match
training and who had been assigned to the control conditions. From Csoka and Bons (1978).
these 18 schools along with cadets from other schools in the ROTC regional
command, were assigned at random to platoons at an advanced ROTC summer
training camp, with no more than two cadets from each school in the same
platoon. Cadets were randomly assigned to four or five leadership positions
while at camp, and their leadership performance was evaluated by an officer and
a noncommissioned officer as well as by means of peer group ratings. The
average performance ratings of cadets from the schools with Leader Match
training was significantly higher than performance of control school cadets on all
measures, with training accounting for 28% of the variance in officer and NCO
TABLE VllI
PERFORMANCE RANKS IN COMPANY FOR PLATOON
LEADERS WITH OR WITHOUT LEADER MATCH TRAINING"
Rank order in company

~ ~ ~~~
Second
Best best Poorest
Trained 14b 4 8
Untrained 7 9 11
"From Csoka and Bons (1978).

bChi Square 5 . 3 5 , ~< .05.
108 FRED E. FIEDLER
ratings and 15% and 23% in peer group ratings on combat and administrative
leadership, respectively (Fiedler & Mahar, 1977, unpublished).
VII. Conclusion
The Contingency Model provides a conceptual framework which enables us

to explain the effects of such change-inducing conditions as organizational turbu-
lence, leadership experience, training, and job rotation. The integrating concept
in the dynamic interpretation of this theory is the leaders situational control and
influence which, as it changes, brings about a corresponding change in the
leaders behavior and performance. Further support of the theory is provided by
studies which show that the concepts of this theory can be applied to a leadership
training program which teaches the leader to match situational control to his or
her personality.
The Contingency Model has a number of shortcomings which need to be
remedied as new research data become available. Specifically, we need a concep-
tually cleaner definition and a better metric of the situational control dimension,
and we need to continue research on personality and behavioral correlates of
LPC, as well as on the effect of this variable on interpersonal encounters outside
the leadership context. Further research also is called for to determine the role
which such other leader attributes as intelligence, sex cognitive complexity, and
task ability play in determining organizational performance and the role which
subordinates play in the leadership process (Fiedler & Leister, 1977).
Above all, we need to seek concepts which integrate different leadership
theories of proven worth. The interpretation of the situational control dimension
as a correlate of uncertainty and anxiety is a promising development in this
direction. Leadership, and the authority relationship of which it is a part, is a
central and important phenomenon in our everyday life. It plays a powerful role
in the governance of our institutions and our society, and it makes obvious the
critical need for understanding, developing, and improving the leadership re-
sources at our disposal.
REFERENCES
Anderson, L. R. Leader behavior, member attitudes, and task performance of intercultural discussion
groups. Journal of Social Psychology, 1966.69, 305-319.
Ashour, A. S. The Contingency Model of Leadership Effectiveness: An evaluation. Organizarional
Behavior and Human Pe~ormance.1913.9, 339-355.
Ayer, J . G.Effects of success and failure of interpersonal and task performance upon leader percep-
tion and behavior. Unpublished masters thesis, University of Illinois, 1968.
Bales, R. F. Interaction process analysis. Cambridge, Mass.: Addison-Wesley, 195 I .
Bass, A . R., Fiedler, F. E., & Kreuger, S. Personality correlates of assumed similarity (ASo) and
related scores. Urbana, Illinois: Group Effectiveness Research Laboratory, University of
Illinois. 1964.
Beach, B. H., Mitchell, T. R.. & Beach, L. R. Components of situational favorableness and
probability of success. Tech. Rep. 75-66. Seattle: University of Washington, Organizational
Research Group, 1975.
Bons, P. M. The effect of changes in leadership environment on the behavior of relationship and
task-motivated leaders. Unpublished doctoral dissertation, University of Washington, 1974.
Bons, P. M., Bass, A. R., & Komorita, S . S. Changes in leadership style as a function of military
experience and type of command. Personnel Psychology, 1970,23, 551-568.
Bons, P. M., & Fiedler, F. E. Changes in organizational leadership and the behavior of relationship-
and task-motivated leaders. Administrative Science Quarterly, 1976. 21, 4 3 3 4 7 2 .
Borden. D. F. The effects of leadership style on leader-supervisor and leader-subordinate interper-
sonal relations. Unpublished master's thesis, University of Washington, 1976.
Bums, T., & Stalker, G. M. The managemeni of innovation. London: Tavistock, 1961.
Campbell, J . P.. Dunnette, M. D., Lawler, E. E., & Weick, K . E. Managerial behavior, perfor-
mance, and effectiveness. New York: McGraw-Hill, 1970.
Chemers, M. M. Cultural training as a means for improving situational favorableness. Human
Rehatiom. 1%9,22,53 1-546.
Chemers, M. M., Rice, R. W.,Sundstrom, E., & Butler, W. Leader esteem for the least preferred
co-worker score, training, and effectiveness: An experimental examination. Journal of Personal-
iry and Social Psychology. 1975, 31, 401409.
Chemers, M. M., & Skrzypek, G. J. An experimental test of the Contingency Model of leadership
effectiveness. Journal of Personality and Social Psychology, I972,24, 172- 177.
CRM Productions. L e d r s h i p : Style or circumstance. Los Angeles, 1974. (Film)
Csoka, L. S. Organic and mechanistic organizational climates and the Contingency Model. Journal
of Applied Psychology. 1975,60, 273-277.
Csoka, L. S . , & Bons, P. M. Manipulating the situation to fit the leader's style-Two validation
studies of Leader Match. Journal of Applied Psychology, 1978, in press.
Fiedler, F. E. A contingency model of leadership effectiveness. In L. Berkowitz (Ed.), Advances in
experimental social psychology. Vol. I . New York: Academic Press, 1964. Pp. 149-190.
Fiedler, F. E. The effect of leadership and cultural heterogeneity on group performance: A test of the
contingency model. Journal of Experimental Social Psychology, 1966.2, 237-264.
Fiedler, F. E. A theory of leadership effectiveness. New York: McGraw-Hill, 1967.
Fiedler, F. E. Leadership experience and leader performance-Another hypothesis shot to hell.
Organizational Behavior and Human Performance, 1970,5, 1-14.
Fiedler, F. E. Note on the methodology of the Graen, Oms, and Alvarez studies testing the Con-
tingency Model. Journal of Applied Psychology, 1971, 55, 202-204. (a)
Fiedler, F. E. Validation and extension of the Contingency Model of Leadership Effectiveness: A
review of empirical findings. Psychological Bulletin, 1971, 76, 128-148. (b)
Fiedler, F. E. Personality. motivational systems, and behavior of high and low LPC persons. Human
Relations, 1971, 25, 391-412. (c)
Fiedler, F. E. Leadership experience and leadership trainingSome new answers to an old problem.
Administrative Science Quarterly. 1972, 17, 453470. (a)
Fiedler, F. E. Personality, motivational systems, and behavior of high and low LPC persons. Human
Relations, 1972, 25, 391-412. (b)
Fiedler, F. E. The contingency model-A reply to Ashour. Organizaiional Behavior and Human
Performance. 1973, 9, 356-368.
Fiedler, F. E. Situational Control and a dynamic theory of leadership. In B. King, F. E. Fiedler, & S.
Streufert (Eds.), Managerial Control and Organizational Democracy. Washington, D.C.: W. H.
Winston & Sons, 1977. (a)
Fiedler, F. E. What triggers the person-situation interaction in leadership. In D. Magnusson & N. S.
Endler (Eds.), Personality at the Crossroads: Current issues in Interaciional Psychology. Hills-
dale, N.J.: Lawrence Erlbaum Assoc., Inc., 1977. Pp. 154-163. (b)
Fiedler, F. E., Bons, P. M., & Hastings. L. The utilization of leadership resources. In W. T.
110 FRED E. FIEDLER
Singleton & P. Spurgeon (Eds,), Measurement of human resources. London: Taylor & Francis,
1975. Pp. 233-244. (a)
Fiedler, F. E., & Chemers, M. M. Leadership and effective management. Glenview, Ill.: Scott,
Foreman, 1974.
Fiedler, F. E., Chemers, M. M., & Mahar, L. Improving leadership effectiveness: The LeaderMarch
Concept. New Yo& Wiley, 1976.
Fiedler, F. E., & Leister, A. F. Leader intelligence and task performance: A test of a multiple screen
model, Organizational Behavior and H u e Performance, 1977, 20, 1-14.
Fiedler, F. E., Mahar, L., & Schmidt, D. Four validation studies of contingency model training.
Tech. Rep. 75-70. Seattle: University of Washington, Organizational Research, 1975. (b)
Fiedler, F. E.. Meuwese, W., & Oonk, S. An exploratory study of group creativity in laboratory
tasks. Acra Psychologica, 1961.18, 100-119.
Fiedler, F. E., OBrien, G. E., & Ilgen, D. R. The effect of leadership style upon the performance
and adjustment of volunteer teams operating in a stressful foreign environment. Human Rela-
tions, 1969,22, 503-514.
Fiedler, F. E., Wamington, W. B., & Blaisdell, F. J. Unconscious attitudesascorrelatesof sociomet-
ric choice in a social group. Journal of Abnormal and Social Psychology, 1952,47, 790-796.
Fishbein, M., Landy, E., & Hatch, G. A consideration of two assumptions underlying Fiedlers
Contingency Model for prediction of leadership effectiveness. American Journal of Psychology,
1969,82, 457473. (a)
Fishbein, M., Landy, E., & Hatch, G.Some determinants of an individuals esteem for the least
preferred co-worker: An attitudinal analysis. Human Relations, 1969, 22, 173-188. (b)
Fox, W. M., Hill, W. A , , & Guertin, W. H. Dimensional analysis of the least preferred co-worker
scales. Journal of Applied Psychology, 1973.57, 192-194.
Godfrey, E., Fiedler, F. E., & Hall, D. M. Boards, management, andcompany success. Danville,
Ill.: Interstate Press, 1959.
Graen, G. B., Alvares, D., Oms,I. B., & Martella. J. A. The contingency model of leadership
effectiveness: Antecedent and evidential results. Psychological Bulletin, 1970, 74, 285-296.
Graen. G. B., Oms, J. B., & Alvares, K.M. Contingency Model of Leadership Effectiveness: Some
experimental results. Journal of Applied Psychology, 1971, 55, 196-201.
Green, S., Nebeker, D., & Boni, A. Personality and situational effects on leader behavior. Academy
of Management Journal, 1976,19, 184-194.
Hansson, R. O., & Fiedler, F. E. Perceived similarity, personality, and attraction to large organiza-
tions. Journal of Applied Social Psychology, 1973,3, 258-266.
Hardy, R. C. Effect of leadership style on the performance of small classroom groups: A test of the
Contingency Model. Journal of Personality and Social Psychology, 1971, 19, 367-374.
Hardy, R. C. A test of poor leader-member relations cells of the contingency model on elementary
school children. Child Development, 1975.45, 958-964.
Hardy, R. C., & Bohren, J. F. The effect of experience on teacher effectiveness: A test of the
contingency model. Journal of Psychology. 1975.89, 159-163.
Hardy, R. C., Sack, S., & Harpine, F. An experimental test of the contingency model on small
classroom groups. Journal of Psychology, 1973,85, 3-16.
Hawley, D. E. A study of the relationship between the leader behavior and attitudes of elementary
school principals. Unpublished masters (M.Ed.) thesis, University of Saskatchewan, 1969.
Hill, W.The Validation and extension of Fiedlers theory of leadership effectiveness. Academy of
Management Journal, March 1969, 33-47,
Hoffman, P. A. A paramorphic representation of clinical judgment. Psychological Bulletin. 1960,
57, 116-131.
House, R. J., & Rizzo, J. R. Toward the measurement of organizational practices: A scale develop-
ment and validation. Experimenral Publication System (Washingron, D.C.), 1971,12, Ms. No.
481-1.
Hunt, J. G. Fiedlers leadership contingency model: An empirical test in three organizations. Orguni-
zarional Behavior and Human Performance, 1967, 2, 290-308.
Larson, L. L., & Rowland, K. Leadership style, stress, and behavior in task performance. Organiza-
tional Behavior and Human Pedormance. 1973.9, 407421.
Lawrence, P. R., & Lorsch, J . W. Organization and environment: Managing dlfferentiarion and
integration. Cambridge, Mass.: Harvard University Press, 1967.
Leister, A. F . , Borden, D., & Fiedler, F. E. The effect of contingency model leadership training on
the performance of Navy leaders. Academy of Management Journal, 1977, 20,464-470.
MacCrimmon, K . R., & Taylor, R. N. Decision making and problem solving. In M. D. Dunnette
(Ed.), Handbook of industrial and organizational psychology. Chicago: Rand McNally, 1976.
Mai-Dalton, R. The influence of training andposition power on leader behavior. Tech. Rep. 75-72.
Seattle: University of Washington, Organizational Research, 1975.
Maslow, A. H. Morivafion and personality. Chicago: Harper, 1954.
McNamara, V. D. Leadership, staff, and school effectiveness. Unpublished doctoral dissertation, Uni-
versity of Alberta, 1%8.
Mehrens, W. A., & Lehmann. I. J. Standardized rests in education. New York: Holt, 1969.
Meuwese. W., & Fiedler, F. E. Leadership and group creativity under varying conditions of stress.
Urbana: University of Illinois, Group Effectiveness Research Laboratory, 1965.
Mitchell, T. R. Leader complexity and leadership style. Journal of Personality and Social Psychol-
OD. 1970, 16, 166-174.
Mitchell, T. R . , Biglan, A,, Oncken, G., & Fiedler, F. E. The Contingency Model: Criticisms and
suggestions. Academy of Management Journal, September 1970. 13, 253-268.
Mitchell, T. R . , Dossett, D. L.. Fiedler, F. E., & Triandis. H. C. Culture training: Validation
evidence for the culture assimilator. International Journal of Psychology. 1972,7, 97-104.
Nealey, S . M . , & Blood, M. R. Leadership performance of nursing supervisors at two organizational
levels. Journal of Applied Psychology. 1968, 52, 414-422.
Nealey, S. M., & Owen, T. W. A multitrait-multimethod analysis of predictors and criteria of
nursing performance. Organizational Behavior and Human Performance, 1970.5, 348-365.
Nebeker, D. M. Situational favorability and environmental uncertainty: An integrative study. Admin-
istrative Science Quarterly, 1975, 20, 28 1-294.
Nebeker, D. M . , & Hansson, R. 0.Confidence in human nature and leader style. Tech. Rep. 72-37.
Seattle: University of Washington, Organizational Research, May 1972.
OBrien, G. E. Group structure and the measurement of potential leader influence. Australian
Journal of Psychology, 1969. 21, 277-289.
Overall, J . E., & Woodwad, 1. A. Unreliability of difference scores: A paradox for measurement of
change. Psychological Bulletin, 1975.82, 85-86.
Prothero. J., & Fiedler, F. E. The effect of situational change on individual behavior andperfor-
mance: A n extension of the contingency model. Tech. Rep. 74-59. Seattle: University of
Washington, Organizational Research, 1974.
Rice, R. W. The esteem for least preferred co-worker (LPC) scores: What do& it measure? Unpub-
lished doctoral dissertation, University of Utah, 1975.
Rice, R. W . Psychometric propenies of the esteem for least preferred co-worker (LPC) scale.
Academy of Management Review. 1978, 3, No. I , 106-1 18.
Rice, R. W., & Chemers, M. M. Predicting the emergence of leaders using Fiedlers Contingency
Model of Leadership Effectiveness. Journal of Applied Psychology, 1973, 57, 28 1-287.
Rice, R. W . , & Chemers, M . M. Personality and situational determinantsof leader behavior. Journal
of Applied Psychology, 1975.60, 20-27.
Robinson, J. P., & Shaver, P. R. Measures of social psychological attitudes. Ann Arbor, Mich.:
University of Michigan, Institute for Social Research, 1973.
Sample, J. A , , & Wilson, T. R. Leader behavior, group productivity, and rating of least preferred
co-worker. Journal of Personalify and Social Psychology, 1965, 1, 266-270.
112 FRED E. FIEDLER
Sashkin. M.Leadership style and group decision effectiveness: Correlational and behavioral tests of
Fiedler's Contingency Model. Organizational Behavior and Human P erfonnance. 1972, 8,
347-362.
Sax, G. Principles of education measurement and evaluation. Belmont, Calif.: Wadsworth, 1974.
Schneier, C. E. The contingency model of leadership: An extension to emergent leadership and leader's
sex. Organizational Behavior and Human Performance. 1978, 21, 220-239.
Schroder, H. M., Driver, M. J., & Streufert, S. Human information processing. New York: Holt,
1967,
Secord, P. F., Backman, C. W.,& Eachus, H.T. Effects of imbalance in the self-concept on the
perception of persons. Journal of Abnormal and Social Psychology, 1964,68, 442446.
Shaw, M. E. Scaling group tasks: A method of dimensional analysis. Tech. Rep. No. 1. GainesviUe:
University of Florida, 1963.
Shiflett, S. C. The contingency model of leadership effectiveness: Some implications of its statistical
and methodological properties. Behavioral Science, 1973, 18, 4 2 9 4 .
Shiflett, S. C . , & Nealey, S . M. The effects of changing leader power; A test of situational
engineering. Organizational Behavior and Human Pe@ormance. 1972,7, 37 1-382.
Shima, H . The relationship between the leader's modes of interpersonal cognition and the perfor-
mance of the gmup. Jupanese Psychological Research, 1968. 10, 13-30.
Stinson, J . E., & Tracy, L. Some disturbing characteristics of the LPC score. PersonnelPsychoIogy,
1974,24, 477485.
Stogdill, R. M. Manual for the Leader Behavior Description Quesn'onnaire-Fonn XI]: An eqeri-
mental revision. Columbus: Ohio State University, 1963.
Stogdill. R. M. Handbook of leadership. New York Free Press, 1974.
Taylor, D. An investigation of the relative stabilities of some prominent measures of leadership.
Unpublished master's thesis, University of Washington, 1975.
Tumes, J . The Contingency Theory of Leadership: A behavioral investigation. Paper presented at the
meeting of the Eastern Academy of Mangement, Boston, May 1972.
Vecchio, R. P. An empirical examination of the validity of Fiedler's model of leadership effective-
ness. Organizational Behavior and Human Performance, 1977, 19, 180-206.
Yukl. G. Leader LPC scores: Attitude dimensions and behavioral correlates. Journal of Social
P~chology,1970,80, 207-212.

Fiedler 1978

Uploaded by

Copyright:

Available Formats

Fiedler 1978

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fiedler 1978

Uploaded by

Copyright:

Available Formats

What is the contingency model and how does it relate leadership effectiveness to situational factors?

What is the contingency model and how does it relate leadership effectiveness to situational factors?

What are the main components and predictions of the contingency model?

What are the main components and predictions of the contingency model?

ADVANCES IN EXPERIMENTAL 8OcW PSYCHOL4XY, VOL.

THE CONTINGENCY MODEL AND THE

In the first volume of this series I described a contingency model of leadership

11. The Contingency Model

The Contingency Model states that effectiveness of interacting2 groups or

The leaders motivational structure is measured by means of a simple bipo-

3The other item pairs are rejecting-accepting, tense-relaxed, distant-close, cold-warm,

LEADER-MEMBER RELATIONS SCALE

C i r c l e the number which best represents your response t o each item.

1. The people 1 supervise have trouble getting

2. My subordtnates are r e l i a b l e and trustworthy.

3. There seems to be a f r i e n d l y atmosphere amng

4. My subordinates always cooperate w i t h me i n

5. There i s f r i c t i o n between my subordinates

6. My subordinates give me a good deal of help

7. The people I supervise work well together i n

8. I have good r e l a t i o n s w i t h the people

Fig. 1. Leader-member relations scale. From Fiedleref n l . , 1975. Reprinted by permission of

TASK STRUCTURE RATING SCALE C i r c l e one

1. Is there a blueprint, picture, &el, or

2. Is there a person available t o advise and

IS THERE ONLY ONE NAY TO ACCOMPLISH THE TASK?

3. Is there a step-by-step procedura, or a standard

4. Is there a specific way t o subdivide the task

5. Are there some ways which are c l e a r l y recognized

6. Is It obvious when the task i s finished and the

8 . Is there a generally agreed understanding about

POSITION POYER RATING SCALE

Circle the number which best represents your anwer.

1. Can the leader d i r e c t l y o r by reconmendation adninister rewards

4, Is i t the leader's job t o evaluate the performance o f his subordlnatesl

4 . Measuring Situational Control

Task Structure Structured Unstructured Structured Unstructured

III. Validation Evidence

The Contingency Model has been extensively tested in a wide variety of

Octant Octant Octant Octant Octant Octant Octant Octant

Hunt (1967) -.a - .80 .21 .30

Number of correlations in the expected direction: 38f

Median correlations between leader LPC scores and group performance

anginal All validation Laboratory

-52 -59 -51 -12 -43

All validation studies .79 p < .05 (one-tailed)

"From Fiedler, 1973. Reprinted with permission.

Study and date Favorable Intermediate Unfavorable

OBrien (1969) - .08 .77b -.I3

Contingency model &ant

56 3-man group of college

Abbreviations: H = High LFC leader; L = Low LPC leader.

IV. Dynamics of the Leadership Process

A viable leadership theory must be able to explain leadership as an integral

subordinates or superiors require them to develop new patterns of interpersonal

Fig. 6. Change in performance of high- and low-LPC leaders as a function of increased

B . SELECTION AND ROTATION

Effect o f r o t a t i n g the leader

High Moderate Lon

The effects of leadership training programs are often difficult to identify