The impact of leader power and behavior on leadership perceptions: a LISREL test of an expanded categorization theory of leadership model. (includes appendix)

How individuals become perceived as leaders has long captivated the interest of management scholars and practitioners alike. No doubt, this attention reflects the widespread belief that managers who are reputed to be effective leaders enjoy greater subordinate acceptance of their decisions and policies as well as higher subordinate commitment and morale (Kanter, 1977; Lord,1985). Although some authors argue that leadership is phenomenological in nature (Meindl, Ehrlich, & Dukerich, 1985; Pfeffer, 1977), this contention cannot discount the body of evidence affirming that superior and subordinate feelings and actions toward a manager greatly depend on their perceptions of his or her leadership (see Gould & Penley, 1984; Hinkin & Schriesheim, 1989). Given such significance, research on the formation of leadership reputation has steadily grown.

Lord and his colleagues conceived the prevailing theory for leadership perceptions (e.g., Cronshaw & Lord, 1987; Lord, 1985; Lord, Foti, & deVader, 1984; Lord & Maher, 1990; Lord, Phillips, & Rush,1980; Phillips, 1984; Rush, Phillips, & Lord, 1981). Human observers cannot process and remember all information stimuli that arise from leader-follower interactions. Thus they employ cognitive heuristics to manage complex social cues. One common heuristic involves matching observed stimuli with a mental "prototype," the most representative member of a cognitive category (Phillips, 1984; Shaw, 1990; Walsh, 1988). Applied to leadership, Lord and his colleagues proposed that if observers discern resemblance between salient actions or qualities of an ostensible leader with those of a leader prototype, they would classify that person as a leader and store that impression in long-term memory (Cronshaw & Lord, 1987). Over time, the person may become more closely identified with the category and be ascribed other category properties that did not appear during encoding (Fiske & Taylor, 1991). If observed features or actions do not fit the leader prototype, then the person is encoded into memory as a nonleader. When making later judgments about the encoded individual, perceivers then rely on the corresponding prototype to outline their impressions (Cronshaw & Lord, 1987).

Empirical research upheld Lord et al.'s contention that observers rely on categorization when they form leadership perceptions (Cronshaw & Lord, 1987; Lord et al., 1984; Phillips, 1984). This research primarily examined prototypical leader behavior and group performance as bases for leadership impressions. In a typical paradigm, subjects read or viewed a description about a target person, who displayed either many or few leader behaviors toward other work group members (e.g., planning, coordinating work). Invariably, observers reported stronger leadership impressions of the stimulus person exhibiting frequent leader actions. Likewise, other studies provided observers with information about the leader's group outcomes (e.g., Lord, Binning, Rush, & Thomas, 1978; Mitchell, Larson, & Green, 1977; Rush et al., 1981; Rush, Thomas, & Lord, 1977). Feedback reporting effective group performance increased leadership attributions compared with ineffective group feedback (Larson, 1982).

Although expanding our understanding of leader impressions, leader schema studies overlooked a quintessential feature of leadership: the bases of the leader's power (French & Raven, 1959; Kanter, 1977). By most conceptualizations, leadership is the use of influence to direct and coordinate subordinates toward organizational goals (Dahl, 1957; Jago, 1982; Mulder, deJong, Koppelaar, & Verhage, 1986; Yukl, 1989). That influence, however, depends on one or more power sources, such as those identified by French and Raven (1959)--namely, legitimate, expert, referent, coercive, or reward power. In the absence of any implied or ascribed power, even leaders displaying leader behaviors cannot effectively mobilize their subordinates. Supporting the necessity of power, Hinkin and Schriesheim (1989) showed that subordinates of powerless supervisors express more dissatisfaction and disloyalty, and Graen, Cashman, Ginsburg, and Schiemann (1977) found greater disaffection among employees who report to managers lacking upward influence.

Besides real empowerment, the ascribed or implied basis of a leader's power may shape leadership impressions. Just as formal leadership conceptualizations often involve power or the capacity for influence (Yukl, 1989), so may everyday leader schemas. This is evident from a comparison of behavioral components of leader schemas (Foti, Fraser, & Lord, 1982; Lord et al., 1984; Phillips & Lord, 1982) with layman (behavioral) definitions of French and Raven's (1959) power bases (Hinkin & Schriesheim, 1989). For example, Lord and his colleagues identified proposing solutions, making suggestions, and showing sympathy to followers as prototypical of leadership. These same actions, however, also convey expert and referent power for people (see Schriesheim, Hinkin, & Podsakoff, 1991). Leader schemas also comprise personality traits (e.g., charismatic, trustworthy, intelligent, and informed; Foti et al., 1982; Lord et al., 1984), some of which connote power bases (e.g., referent and expert power;

Hinkin & Schriesheim, 1989). Given that power is implicit in leadership schemas, prototypical attributes may enhance leadership impressions partly through allusions to power. Apart from implied power, direct ascriptions of power may induce leadership perceptions. To illustrate, a manager holding an Ivy League MBA (high expert power) is more likely to project leadership than will a high school graduate (low expert power).

Moreover, implied leader power may account for demonstrated effects of leader behavior and group performance in categorization studies. That is, descriptions of influential acts directed at others (e.g., assign tasks, set goals, maintain performance standards; Phillips & Lord, 1982) may automatically suggest that the stimulus person has some implicit authority or legitimacy for exercising such influence. Otherwise, why would group members (portrayed in these vignettes) so readily comply with the target person's request? Similarly, communications about the effectiveness of a leader's work group may provide cues about his or her power. Through attributional reasoning, people may regard group accomplishment as a leadership outcome and thus decide that leaders of …