Personality and Situatto^al Determinants of Leader Behavior

This paper examines the relationship between such behaviors as the leader's consideration and structurine and two key variables of the Continpencv Model. The variables, the Least Preferred Coworker (LPC) score and the "favorabieness" of the leadership situation, interact in a highly consistent manner in affecting these leader behaviors. The evidence that leader behavior depends, in large part, on the interaction of LPC and situational favorabieness indicates (a) that we should avoid defining leadership style on the basis of leader behaviors, and (b) that the prediction of leader behavior must be based on the joint effects of personality and situational factors. PERSONALITY AND SITUATIONAL DETERMINANTS OF LEADER BEHAVIOR Fred E. Fiedler University of Washington The behavior of leaders has been an important and early focus of social psychological research. Carter and his associates (Cavter, et al., 1951), Bales (1965) and Cattell (1951) are but a few of the more prominent workers in this area. Many others have contributed to the measurement of leader behavior, and have developed methods for investigating how the leader's behavior influences the performance as well as the satisfaction and morale of group members. Of particular importance has been the work of Shartle and his associates at Ohio State» University, which eventuated in the now well-known leader behavior dimensions of Consideration and Initiation of Structure in Interaction (see Stogdill & Coons, 1957). These studies as well as most others in the area have led to a convergence of findings which point to two major classes of leader behaviors on which supervisors and managers as well as others in leadership positions can be meaningfully described. These are (a) the concern and effort directed toward establishing and maintaining good Interpersonal relations, high morale, job satisfaction, etc., and (b) the concern with performance which expresses itself in directing and structuring the group effort toward the common goal. The hope of earlier studies was to identify specific leader behaviors which would be related, if not causal, to effective group performance as well as member satisfaction so that leaders might then be trained to engage in these behaviors. 1 Fiedler These hopes were disappointed. A review of the available literature by Korraan (1966)showed that leader behaviors did not correlate consistently with group effectiveness. No consistent relations emerged between the structuring behaviors of the leader and either effectiveness or member satisfaction. And while a moderate and consistent relationship appears to exist between considerate behavior and member satisfaction, it must be kept in mind that satisfied employees are more likely than unsatisfied employees to describe their supervisor as considerate. A study by Halpin (1955) had shown that the effective leaders of air crews were both structuring and considerate, while this was not true of leaders of educational institutions. Fleishman and Harris (1962) had found an Interaction between consideration and structuring leader behaviors on group satisfaction. However, these findings were not replicated in other investigations. It is, therefore, clear that we are dealing with a complex problem. In brief, there can be no doubt that the consideration and the initiation of structure dimensions and similar behavior categories describe important leader behaviors. At the same time there is disappointingly little empirical evidence that these dimensions affect performance (Korman, 1966; Campbell, et al., 1970). Why these behaviors do not predict, or correlate with, group performance represents a major theoretical problem. One clue which might assist us toward a satisfactory explanation is the finding that situatlonal factors and certain personality attributes interact in determining leadership effectiveness (Fiedler, 1967, 1971). Could similar interactions determine leader behaviors? Fiedler 3 Consistency of Leader Behavior Before we can address these problems we must first ask whether there is consistency In leader behavior over different situations. Writers in the field have generally spoken of autocratic, structuring, task-centered leadership styles, or of such styles as democratic, employee-centered, considerate. By personality style—or leadership style—we generally mean a trans-situational mode of relating and interacting with others. If there is such a thing as, for example, a "considerate leadership style" or an "employee-centered leadership style," then we should expect that the Individual with this leadership style should behave in a relatively more employee-centered or cor.slderate manner than leaders who are not so rated, regardless of the situation. But if the same leaders are autocratic or Job-centered in one situation, but democratic or employee-centered and considerate In another situation, they clearly do not have a particular leadership style which manifests itself in these behavior categories. To speak of considerate or democratic leaders, or of joband employee-centered leaders, may then be unwarranted and misleading. It is, of course, obvious that the behavior of leaders, or for that matter, the behavior of anyone else, will to some extent depend upon the situation. People behave differently at funerals than at cocktail parties. The problem at issue is whether (a) the tendency to behave in a considerate, employee-centered manner is an attribute of the leader's personality, and therefore properly considered to be his leadership style; or (b) whether the leader's personality and the situation Interact, and the person who is considerate under one condition tends to be 4 Fiedler relatively leas considerate under other conditions. If the latter Is the caae, It will have major Implications for current leadership theory as well as our currently widespread attempts to train people to become more considerate or human relations oriented In their Interactions with subordinates. Everyday observation suggests that some supervisors are quite considerate and concerned for the feelings and welfare of their men when all Is going well, but they become tense and Inconsiderate toward subordinates when they are under pressure. In contrast, we frequently see supervisors who are aloof and distant. If not truly Inconsiderate, under routine conditions, but quite considerate and employee-centered under conditions of stress and emergencies. If these are not Isolated Instances but represent a lawful relationship, then it seems likely that we are dealing with an interaction between the situation and relevant personality attributes. The problen, then, is to identify the relevant personality attributes as well as the situational factors which determine how individuals in leadership positions will behave. Personality and leader behavior. What, first of all, is a "relevant personality attribute?" As far as this inquiry is concerned, it is a reliable personality variable of the leader which affects such other Important types of group behaviors as performance. On« variable of this type is the Least Preferred Coworker (LPC) score, This score has been extensively used in a large number of leadership studies and it is a key variable of the writer's Contingency Theory of leadership effectiveness (Fiedler, 1967). It is obtained by asking an individual (a) to think of all the people with whom he has ever worked. Fiedler 5 and (b) to describe the one person in hia life with whom he found it most difficult to work on a common task. This description is made by marking fi set of 16to 20-items of an eight point bi-polar scale of the Semantic Differential type (^ >e Osgood, 1952). The LP^: score is simply the sum of item scores, with the "good" pole of each item scale giving a high score. A high LPC score, indicating that the individual described his LPC in relatively favorable terms, is interpreted as reflecting a basic motivation to be related to others. A low LPC score is Interpreted as reflecting a motivation for task accomplishment. Note that high LPC has been interpreted by various writers as conceptually similar to consideration, 'vhile low LPC has been interpreted as similar to initiation of structure. As we shall show, such an interpretation is clearly not warranted by the empirical evidence and it is quite inconsistent with the theoretical position underlying the current interpretation of LPC. The relevance of LPC for the present inquiry lies in the wellsubstantiated findings which show an interactxrm between LPC and situational favorableness in affecting group performance. The majority of studies have found that low LPC leaders tend to perform better in very favorable as well as in unfavorable leadership situations, while high LPC leaders tend to perform better in situations which are intermediate in favorableness (Fiedler, 1971). It is not unreasonable to expect, therefore, that we should find a similar interaction between LPC and situational favorableness in determining leader behavior. Situational Favorableness The favorableness dimension is conceptually defined as the degree to which the situation itself gives the eader power and influence in 6 Fiedler his group. This dimension has been operational!zed In a number of ways. It Is most frequently defined on the basis of three component dimensions: (a) the degree to which the leader feels accepted by his group (the "group atmosphere" score); (b) the degree to which the task is structured; and (c) the degree to which the leader position has power and Influence. The importance of the situational favorableness dimension is not too surprising in retrospect. It is one thing to tackle a leadership task when one has complete control over the situation e.g., the captain of a navy ship. It is quite a different story when the leader's influence is "ninimal; e.g., the disliked chairman of a volunteer group which has been asked to devise a new school buss fig policy. The remainder of this paper will deal with the interaction of the leader's LPC score and various situational factors in determining the behavior of the leader. These results have been not only in the studies which are described but a number of others as well (e.g., Fiedler, O'Brien, and IIgen, 1966; Nayar, Touzard and Rummers, 1968; and Ninane and Fiedler, 1970; etc.). Interaction between LPC and Situational Favorableness The Dutch Study (Fiedler. Meuwese and Oonk, 1961). The first study which suggested an interaction between LPC snd situational factors was conducted in Holland. In this experiment, four-man groups were given creative tasks Ce.g., writing three different stories about the same TAT card). The groups were assembled either with appointed leaders or without appointed leaders, and either homogeneous In terms of religion and regional sub-culture or heterogeneous. Homogeneous groups with appointed leaders are considered a more favorable leadership situation Fiedler 7 than heterogeneous groups with emergent leaders. The sessions were tape recorded and content analyzed. However, It was not possible to differentiate the leader from group members on the basis of the recordings, and the analyses were based on total interactions, irrespective of who made various comments. It is nevertheless reasonable to expect that the contribution by leaders would be greater than that bv members, and that he content analysis would, therefore, reflect the leader's behavior to a greater extent than the behavior of members. These analyses suggested that the high LPC groups (i.e., groups led by high LPC leaders) made more comments related to the task in the relatively favorable situation (appointed leader, homogeneous group), but more comments related to the maintenance of interpersonal relations in the unfavorable situation (emergent leader, heterogeneous group). Groups led by lew LPC leiders made relatively more comments related to group maintenance in the favorable situation but more comments related to the task in the unfavorable situation. ROTC (Meuwese and Fiedler. 1965). A second study utilized KOTC cadets who were assigned to 54 three-man groups and given creative tasks under three conditions of stress. These were (a) "low stress," where the leader's and the members' anxiety was minimized; (b) "interpersonal stress," where conflict among group members was experimentally generated; and (c) "external stress," which was created by having a very formal task situation in which a high ranking officer faced and continually evaluated the group during the entire period of interaccion. The high LPC leaders under the less stressful conditions (low stress, high group atmosphere) tended to be concerned with the task while the low LPC leaders tended 8 Fiedler to concern themselves with relation-maintaining statements. In the least favoruole situation (external stress, low group atmosphere), the high LPC leaders were concerned with the relationship while the low LPC leaders were concerned with the task. This, of course, supported the findings of the Dutch study (see Figure 1). Each point on the Figure is based on nine groups. Insert Figure 1 about here Belgian navy teams (Fiedler, 1966). A third study was conducted with Belgian naval force personnel. This Involved a large field experiraant in which three-man teams performed structured as well as unstructured tasks under different conditions of situational favorableness. Specifically, half the groups were under the leadership of a petty officer, half under the leadership of a recruit; half the groups were culturally homogeneous while half were heterogeneous in culture and language; and half the groups began with a structured task while the other half began with an unstructured task. Moreover, half the groups were of relatively high intelligence while the other half were of relatively lower Intelligence. After each task session, the proup members described their leader on a number of rating scales. These scales were factor analyzed and yielded clusters which could be interpreted as "motivating and involved leader," "considerate," "noncritlcal and not aggressive," "directive and structuring." Since all clusters were positively Intercorrelated and interpreted as indicating "intensity of interaction," the cluster score was then correlated with LPC. The present analysis was confined to homogeneous groups since the stereotypes held by each language group toward the other tend to obscure other ratings. In these homogeneous