نقش اسناد در چگونگی واکنش دنبال کنندگان به ابراز هیجانی رهبران زن و مرد

Abstract We investigated how leaders' gender interacts with anger and sadness expression and followers' attributions for their emotional expression on evaluations made by followers. In a laboratory study concerning hypothetical leaders, people evaluated the competence of male and female leaders differently depending on their emotional displays (anger vs. sadness). These differences were observed only when observers made high dispositional attributions for leaders' emotional expression. Perceived leader role-based emotional expression mediated these relationships. In a field study concerning actual leaders, dispositional attributions for sadness and anger expression were negatively related to follower perceptions of the leader, and these attributions moderated the relationships between leader gender and various follower perceptions of the leader. Leader prototype fit mediated these relationships. We discuss the implications of these findings for leadership research and practice.

Introduction Research on the emotional expression of leaders in organizations focuses mainly on positive emotions which have been suggested to inspire and motivate employees (Bass and Avolio, 1994 and Conger and Kanungo, 1998). For example, the positivity or negativity of leaders' emotional expressions has been linked to followers' experienced emotions (Johnson, 2009 and Sy et al., 2005) and to follower ratings of leader competence and attraction to the leader (Bono & Ilies, 2006). A leader's negative emotions can also have important implications for his or her followers' perceptions. For example, often followers accept or even prefer that their leaders express negative emotions, such as during times of stress and crisis (see Humphrey et al., 2008 and Madera and Smith, 2009). Effective leaders are seen to “lead with emotional labor,” meaning that they not only display appropriate emotions but they also utilize emotional expression as a way to motivate and guide followers who take their cues from them (Humphrey, 2008 and Humphrey et al., 2008). Brotheridge and Lee (2008) noted that leader roles are very high in emotional labor (Hochschild, 1983), meaning that for effective performance of the role the leader must not only seek to dampen undesirable emotional displays, but also display certain emotions that derive from social norms. They also noted that in particular, leaders need to be especially cognizant of gender role expectations for emotional expression. Studies of how leaders' negative emotional expressions affect followers and other observers suggest that the expression of specific negative emotions influences observers' perceptions of the leader, such as satisfaction with the leader and leader competence, and that these effects vary depending on the gender of the leader (Glomb and Hulin, 1997 and Lewis, 2000). For example, Lewis (2000) found that observers rated female leaders who displayed anger as being less effective than female leaders who displayed sadness following the same adverse event (i.e., poor financial performance). On the other hand, male leaders who displayed anger were rated as more effective than male leaders who displayed sadness or no emotion following the same event. We sought to extend the literature of leader emotional expression in two ways. First, we examined how the attributions people make for the emotional expression of leaders affects their evaluations of them. Studies that have examined particular forms of leader emotional expression (e.g., Glomb and Hulin, 1997 and Lewis, 2000) are limited to short-term hypothetical situations and raise questions about whether a particular type of emotional expression will remain influential to followers after repeated exposure to the leader's various emotional expressions under different circumstances. Thus, our studies drew from attribution theory (Kelley, 1967) to investigate the role of attributions on observers' reactions to these displays. Secondly, we sought to better understand the cognitive processes linking leader gender, emotional display, and attribution to perceptions when (1) observers examine leaders responding to specific situations; and (2) followers rate their leaders' general tendencies in terms of expressing emotion. We anticipated that distinct states would mediate the influences of attributions, depending on whether one examines specific instances of behavior engaged by hypothetical leaders versus the general behavior patterns of actual leaders. Role congruence and leadership perceptions Leadership prototype theory Leadership prototype theory suggests that people match a leader's behavior or character with a prototype of ideal leaders they hold in memory. This prototype is essentially a benchmark, and mismatches of actual leaders and these benchmarks are thought to affect general impressions that followers have about their leaders. Among the empirical evidence supporting this proposition, Lord, Foti, and De Vader (1984) demonstrated that followers' ratings of a target leader's behavior are strongly influenced by their recognition of prototypic traits. Six leadership prototype dimensions have been found to be held across most individuals: sensitivity, dedication, charisma, attractiveness, intelligence, and strength (Offermann, Kennedy, & Wirtz, 1994). Lord and Maher (1993) proposed that leaders will be evaluated unfavorably by their followers if their behaviors are inconsistent with the followers' leadership prototype. Social role theory Gender stereotypes have been found to influence people's perceptions of women's and men's emotional displays (e.g., Brody and Hall, 2000 and Shields, 1987). In addition, social role theory proposes that when an individual behaves in a manner that is inconsistent with his or her sex role, that person is judged negatively (Eagly, 1987 and Heilman, 2001). Thus, just as leader prototype theory suggests that individuals have expectations for how leaders should behave, social role theory highlights the importance of others' expectations for how persons of different genders should behave. Role congruency theory is an extension of social role theory. It is based on the assumption that followers evaluate female leaders more stringently than male leaders. This theory argues that the stereotypical female sex role and the leadership role are incongruent (Eagly & Karau, 2002). For example, several stereotypically male characteristics are found to overlap with leader prototypes across various societal cultures (e.g., Schein, 2001). Within the six commonly held leadership prototypes, only sensitivity is related to stereotypically female characteristics, (Johnson, Murphy, Zewdie, & Reichard, 2008), although arguably only “strength” is consistent with the male prototype. Female leaders are seen to be evaluated unfavorably either when they adopt more masculine leader characteristics and violate their sex role stereotypes, or when they behave in a manner that is consistent with feminine role expectations, thus departing from leader role expectations (Eagly and Karau, 2002 and Heilman, 2001). Gender, emotional expression and perceived leader competence The manner in which leaders express emotion has a substantial influence on how observers evaluate their competence. Studies have found that verbal and paraverbal cues, many of which signal emotional content, differentiate the competence of leaders (Frese et al., 2003 and Towler, 2003). As noted above, studies of gender stereotypes and leadership conducted in both laboratory and field settings have found that female leaders are evaluated as being less effective when they exhibit a more masculine style (see Eagly, Karau, & Makhijani, 1995). Along the same lines, Fischer (1993) suggested that women and men are evaluated more favorably when they express emotions that are congruent with their gender identity. Research on gender-specific emotional expression stereotypes indicates that anger is a male gender-congruent but female gender-incongruent emotion (e.g., Briton and Hall, 1995 and Plant et al., 2000). Anger conveys images of strength, competence, intelligence, and higher social status (Keating, 1985, Plant et al., 2000 and Tiedens, 2001). These have been suggested to be agentic traits that are prototypical of effective leaders. As such, anger display may result in higher ratings of leader competence for male than for female leaders because it is inconsistent with gender role expectations for women (Eagly et al., 1995). It should be noted, however, that anger conveys little warmth, friendliness, or kindness, and therefore frequent and/or inappropriately expressed anger may be antithetical to effective leadership regardless of the leader's gender. Conversely, the same studies found that sadness is female gender-congruent and male gender-incongruent, and it is associated with likability, sensitivity, warmth, and kindness (Briton and Hall, 1995, Eagly et al., 1995, Keating, 1985 and Plant et al., 2000). These are desirable traits that may be conducive to the formation of trustful bonds between leaders and followers. However, sadness also conveys weakness, neediness, submissiveness, and low social status, all of which are seen to deviate from effective leadership (Keating, 1985, Plant et al., 2000 and Tiedens, 2001). One may therefore anticipate that, ceteris paribus, sadness display will result in lower ratings of leader competence for male than for female leaders. Attribution theory and perceived leader competence Attribution theory suggests that people look for clues to explain others' behavior, especially when that behavior is unexpected (Bradfield and Aquino, 1999 and Geddes and Callister, 2007). In the workplace, leaders' displays of emotions are often perceived as outside of leader role norms and in such cases they are salient to observers (Rafaeli & Sutton, 1987). Based on the information they gather, observers may attribute the leaders' emotional expression either to dispositional causes (e.g., anger is reflective of trait of being an angry person or to the individual being in a bad mood) or to external causes (e.g., anger is attributed to a situational context for which anger is a suitable or typical human response). When causality is assigned to an external factor, people infer that the person has no control over the event and hence is not responsible for his/her behavior. When causality is assigned to an internal factor, people infer that the person is responsible for his/her behavior (Kelley, 1967). Previous research has examined the role of attribution in how followers evaluate leader's intentionality and sincerity (Dasborough and Ashkanasy, 2002 and Dasborough and Ashkanasy, 2004). Prior studies on leader emotional display have assessed observers' ratings of hypothetical leaders based on one-shot simulation materials (Glomb and Hulin, 1997, Lewis, 2000 and Tiedens, 2001). These studies provide insights into how people evaluate leaders' displays of specific emotions based on single, isolated observations. However, we argue that in such scenarios it is difficult for the observer to understand the context surrounding the event and the target's routine behavioral patterns, and therefore we expect that the normal attributional judgment process is bypassed. Because information about the leader is very incomplete, the observer is left to project the missing information needed to make an inference about the individual. This encourages the retrieval and use of salient stereotype information as a basis of judgment. Observers are more likely to rely on stereotypes to judge actors' behavior when little other information is available (see Fiske & Taylor, 1991). Stereotypes are less important when observers focus closely on actors' individuating information or when observers are highly motivated to form accurate impressions (Fiske and Neuberg, 1991 and Madon et al., 2006). In actual interactions between leaders and followers, followers' perceptions of their leaders are influenced by both current and past experiences observing the leader. This historical information provides a context for making attributions about the leaders' emotional expression. To some extent this historical information can be incorporated into laboratory studies and allow for attributions to be based on information other than stereotypes held in memory. We suggest that when people make weaker attributions to the traits and/or other dispositions of the target for his or her emotional expressions because he/she has not consistently exhibited the same behavior in past similar situations, the leader's gender will less strongly predict evaluations. For example, historical information about a female leader's behavior may indicate that she does not always react angrily to situations of a general nature, and thus observers' evaluations of her anger episode may incline to situational attributions. Conversely, when there is no historical information (as in the standard scenario laboratory study) or if historical information indicates that she has repeatedly acted angrily, then prejudicial judgments about anger and women are more likely.