Pergamon Journal of Management 27 (2001) 537–561
Gender and discipline in the workplace: Wait until your father gets home Leanne E. Atwater*, James A. Carey, David A. Waldman Arizona State University West, School of Management, P.O. Box 37100, Phoenix, AZ 85069, USA Received 9 April 2000; received in revised form 28 September 2000; accepted 29 November 2000
Abstract This study examined recipients’ perceptions of workplace discipline. Females delivering discipline were perceived to be less effective and less fair than males. Both recipients’ biases and behavior differences by male and female supervisors appear to contribute to reduced effectiveness. These results suggest the need to raise the awareness of managers and subordinates regarding potential negative reactions to females administering discipline. Special training in discipline delivery for female managers may also be warranted. © 2001 Elsevier Science Inc. All rights reserved.
1. Introduction Prompted by the influx of women into supervisory positions in the workforce, the issue of women and leadership has been the subject of much research. Despite the fact that most of the recent literature suggests that female managers differ little from males in their leadership style or overall effectiveness (cf. Powell, 1988), that does not necessarily mean that males and females are equally effective at all aspects of supervision. For example, studies have demonstrated that females display more transformational leadership behavior (e.g., individual consideration) than males, which according to transformational leadership theory promotes higher performance in followers (Bass, Avolio & Atwater, 1996). However, an important aspect of supervision that has received little attention, and to our knowledge, no attention with regard to gender differences, is discipline behavior. Discipline has been described as “an important albeit distasteful function in almost every manager’s job” (Greer
* Corresponding author. Tel.: ⫹1-602-543-6114; fax: ⫹1-602-543-6221. E-mail address:
[email protected] (L.E. Atwater). 0149-2063/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved. PII: S 0 1 4 9 - 2 0 6 3 ( 0 1 ) 0 0 1 0 7 - 6
538
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
& Labig, 1987, p. 507). Sex-role stereotypes and socialization differences between males and females would suggest that differences exist that may put women at a distinct disadvantage when it comes to discipline. However, whether females are perceived as less effective at administering discipline than their male counterparts has not been previously investigated. Two factors potentially contribute to disadvantage female supervisors when delivering discipline. Males and females may be perceived differently or they may behave differently when delivering discipline. Specifically, person-centered theory (Morrison & Von Glinow, 1990) suggests that female personality traits and behavior patterns may make females less suited for leadership roles, particularly those involving dominance and assertiveness. This theory proposes that socialization differences promote nurturing rather than assertive behavior patterns, thus suggesting that males and females may behave differently. Alternatively, bias-centered theory (Morrison & Von Glinow, 1990) asserts that individuals have beliefs about what behaviors women should exhibit, and behaviors that violate those beliefs are negatively regarded. Females are expected to be warm, sensitive, passive, and supportive. Behavior contrary to that stereotype is viewed negatively. Both person-centered theory and bias centered theory would suggest that females may suffer when it comes to delivering discipline as compared to their male counterparts. The primary purpose of the current study was to assess whether females are seen by their subordinates as less effective at administering discipline. Additionally, based on socialization differences, males and females may react differently to discipline when they are the recipients. The present study also investigated differences between male and female recipients’ perceptions of discipline events.
2. Person-centered theory 2.1. Socialization In line with person-centered theory, research has accumulated showing that girls receive strong and consistent indications of their powerlessness in two areas: influence over people and mastery over tasks. Differences in power and influence among males and females begin as early as the age of three. Jones (1983) reports that between the ages of three and five, boys become increasingly likely to use direct influence attempts such as “give me that.” Girls learn that they are less effective influencers than boys and retreat to indirect, polite styles. Girls who do not retreat to the accepted styles by age seven and continue to demonstrate dominance tend to be disliked and rejected (Jones, 1983). Through socialization, males learn to be task-oriented, confident, and independent while females learn to be supportive, submissive, and dependent. Socialization may also cause women to be less self-confident about their abilities than men. This stems in part from the type of feedback girls receive in school. Feedback given to boys in school is often directed toward nonintellectual qualities such as misconduct, while feedback for girls is more often specifically relevant to ability. As such, girls have a more difficult time attributing negative feedback to external causes and more often view criticism as an indication of their lack of ability (Dweck & Bush, 1976; Dweck, Davidson, Nelson & Enna, 1978).
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
539
Socialization also contributes to differences in language patterns for men and women. For example, traditional female conversational styles may put females at a disadvantage in discipline interactions by sending messages to recipients that they (the female supervisors) are less confident or less sure about their decisions to discipline. “From childhood, most girls learn that sounding too sure of themselves will make them unpopular” (Tannen, 1995, p. 140). Tannen (1995) highlights other ways in which women’s linguistic styles differ from those of men. Women are more likely to downplay their certainty whereas men minimize their doubts. Apologizing and mitigating criticism with praise are other female linguistic styles that may interfere with effective discipline delivery. Men also tend to be more sensitive to the power dynamics of interaction, speaking in ways that position them as “one-up.” Women, on the other hand, tend to use speaking patterns that avoid putting others in a one-down position. Males’ tendencies to dominate conversation and interrupt may also lead others to perceive them as more competent and confident (Tannen, 1995). Socialization differences also operate through the expectations others have of males and females. Social role theory (Eagly, 1987) purports that people are expected to behave in a manner that is consistent with societal gender roles. Expectations associated with gender roles can foster behavior in men and women consistent with these roles. In social interactions, people communicate their gender stereotypic expectations about how others should behave and thereby directly induce people to behave in ways consistent with those expectations (Wood & Karsten, 1986). As such, self-regulatory and expectancy confirmation processes (similar to the self-fulfilling prophecy, Rosenthal, 1995) can induce gender differences in behavior (Olson, Roese & Zanna, 1996). Also of interest regarding others’ expectations and the self-fulfilling prophecy are two studies that suggest that people still harbor negative attitudes toward women in leadership or authority roles. Butler and Geis (1990) studied nonverbal indicators of affect toward males and females exerting leadership behavior. They observed subjects’ nonverbal reactions to male and female confederates in a group discussion from behind one-way mirrors. All subjects’ nonverbal behaviors were recorded and coded. Female leaders received more negative and fewer positive nonverbal responses to displays of leadership than males. Yet when subjects were asked on a paper and pencil measure to rate the leaders’ competencies, there were no differences in ratings of male and female leaders. Rudman and Kilianski (1999) showed student participants drawings of male and female authority figures (e.g., doctors or bosses) and assessed the latency of response to positive and negative adjectives. They concluded from their results that both males and females had negative attitudes toward female authority figures and neutral attitudes toward male authority figures. These studies suggest that females may be receiving negative cues about their leadership behavior that may even be unconsciously delivered by observers or subordinates. These negative cues could be perceived as disapproval and thereby reduce female leaders’ confidence in their leadership skills. In support of the confidence argument, Benedict and Levine (1988) found that females were more likely to positively distort performance ratings of male subordinates than were males. They suggested that this may have been because of a lack of confidence or feelings of intimidation on the part of the females.
540
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
2.2. Status and power differences Later in life, socialization differences may become perpetuated because of status and power differences. Social role theory (Eagly, 1983; 1987; Eagly & Steffen, 1984) asserts that women and men are distributed differently into social roles; men tend to occupy higher status roles, and women occupy lower status roles. Female-typed occupations generally have had less prestige, less power and less pay than male-typed occupations (Dexter, 1985). In addition, self-selection into gender-typed specialities has contributed to decreased status for women. For example, women in management may gravitate toward specialties such as office management (staff positions) whereas men enter production or operations (line positions). As such, female managers may, in general, achieve lower status than male managers. With respect to power, like status, the reality remains that in almost every domain outside the family, women still lag behind men (Rhode, 1990). Women suffer from diminished position power as well as from perceived powerlessness on the part of others. Schlueter, Barge and Blankenship (1990) found that male managers believed they had more power over subordinates than did female managers, and they were also perceived by others to have more power. Even conversational styles can put females at a disadvantage when it comes to influence. Female managers tend to use more disguised requests and find subtle rather than direct ways to tell subordinates what to do (Schlueter, Barge & Blankenship, 1990). While power and influence differences between the sexes begin early in childhood, they persist into adulthood. Instone, Major and Bunker (1983) found that women used fewer influence attempts, used a smaller range of strategies than did men, and had lower selfconfidence in their abilities to influence others. Consistent with this finding, Falbo and Peplau (1980) found that women and individuals with less power tended to use indirect, unilateral influence strategies (e.g., negative affect, withdrawal) whereas men used bilateral strategies such as bargaining, persuasion or reasoning. They also found that men in positions of power expect compliance with their influence attempts while women frequently do not. They reasoned that because women anticipate noncompliance they are more likely to use strategies that do not require another’s cooperation. Other researchers (e.g., Instone, Major & Bunker, 1983) have found that females are more likely than males to use coercive power which also may stem from fears of noncompliance. The criteria women and men use to decide when power is needed also differ. With regard to discipline, women punish based on a norm of equality whereas men punish based on a norm of equity (Dobbins, 1985). For example, females are less sensitive to causes of poor performance (e.g., internal vs. external; stable vs. unstable) than males and tend to punish poor performers equally, regardless of type of cause. Males, on the other hand, are more sensitive to the cause of poor performance. For example, males supported training rather than punishment for poor performance when the cause was stable and external (e.g., ability) (Dobbins, 1985). The combination of reduced power, reduced confidence in compliance, a preference for indirect, unilateral, or coercive influence tactics, and preference for a norm of equality all work to a woman’s disadvantage in successfully administering discipline.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
541
2.3. Leadership Although males and females may show only modest differences in their actual leadership behaviors (e.g., female managers tend to be more participative and male managers more autocratic [Eagly & Johnson, 1990]), some stylistic differences may make it more difficult for female managers to effectively discipline. For example, Statham (1987) found that females in supervisory roles focused more on the task to be done and the people working for them, interacting with others a great deal. Males on the other hand, were seen as focused on themselves, and emphasized the power they had. Males felt that the ideal way to manage was to keep their distance from subordinates (Statham, 1987). While a female manager’s tendency to develop closer relationships with subordinates may help motivate them, it may also make it more difficult for females to discipline them when necessary and may influence how discipline delivered by a woman is perceived. In sum, with regard to socialization, status and power differences, it appears that women may be at a disadvantage regarding discipline due to their behavioral styles (person-centered theory).
3. Bias-centered theory 3.1. Socialization Males and females are socialized to expect different behaviors from men and women. Differences are expected in terms of language, dress, body language, and even the jobs males and females are expected to hold. Social role theory (Eagly, 1987) suggests that when individuals behave in ways that are inconsistent with those expectations, they are regarded negatively. If females in a supervisory role delivering discipline are seen by others as “out-of-role,” their behavior is likely to be regarded more negatively than males delivering discipline because that behavior is more likely “in-role” for males. 3.2. Status and power differences Everyday we are surrounded by images and messages that males are powerful and strong and females are powerless and weak. Consistent with bias-centered theory, research has consistently shown that status attributed to females is lower than that of males even if their achieved level or position is the same (Wagner, Ford & Ford, 1986; Wagner & Berger, 1997). When a woman is portrayed as powerful, she is often also evil like Glenn Close in Fatal Attraction or Cruella DeVille in 101 Dalmations. But regardless of level or position, females often suffer from lower ascribed status merely because they are female. Expectations State Theory indicates that because gender is associated with prestige and status, people have different expectations about the status of men and women that are generalized across situations and affect their behavior toward those individuals (Berger, Rosenboltz & Zelditch, 1980). Lucas and Lovaglia (1998) have described
542
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
gender as a diffuse status characteristic, with males being more highly valued and respected than females (Ridgeway & Diekema, 1992; Carli, 1991; Wagner & Berger, 1997). Heilman and Kram (1978) found that both males and females took less personal responsibility for failure when working with a female rather than a male, which may have stemmed in part from the female’s perceived lower status. Because females are seen as having lower status, they may have more difficulty engendering the respect necessary to successfully discipline. If women are seen as less competent and engender less respect, recipients of discipline administered by a woman may be less willing to accept responsibility for their unacceptable behavior. Because males have higher ascribed status in our society, they are more likely to be listened to and taken seriously. Alternatively, a female holding a management position may find herself in a position of status incongruity where her achieved level and ascribed status conflict. That is, the high achieved status of her role as manager conflicts with her lower attributed status of her sex. This may cause discomfort for those who must interact with female managers and may cause resentment by those who see a female in a management position as a threat to their own status or the status of their position. For example, in a study by Harlan and Weiss (1982) male managers said they preferred male supervisors and reported that it was an “insult to their intelligence” to be supervised by a woman. Sex-role stereotypes of power and status could reduce a woman’s credibility and impede her effectiveness in administering discipline. 3.3. Leadership In line with bias-centered theory, leadership is another domain in which stereotypes and perceptual biases put females at a disadvantage. The stereotypic woman is thought to be kind, nurturing, sensitive, warm, and expressive; the stereotypic man is thought to be dynamic, assertive, competitive, task-oriented, and competent. Although stereotypic female qualities are actually associated with positive aspects of considerate leadership, they may not be especially effective in the realm of applying discipline and may influence the ways males and females are evaluated “far beyond what the actual facts may dictate” (Powell, 1988, p. 65). Empirical studies examining the relationship between an individual’s gender role orientation (e.g., masculine or feminine) and leadership have generally demonstrated evidence of a strong association between leadership and a masculine gender identity (Kent & Moss, 1994). Stereotypical masculine characteristics also are beneficial to leader emergence (Fagenson, 1990). When Heilman, Block, Martell and Simon (1989) asked managers to describe successful managers, descriptions of women in general were far less congruent with descriptions of successful managers than were descriptions of men in general. Other studies have supported the notion that sex-role stereotypes currently exist and are barriers to women who desire leadership positions (Dobbins, Cardy & Truxillo, 1988; Sharp & Post, 1980; Walsh, Weinberg & Fairfield, 1987). For example, bias against female executives by males abated consistently between 1965 and 1985 but by 1985 it had not disappeared (Bowman, Worthy & Greyser, 1965; Sutton & Moore, 1985). Gallup Poll results as late as 1995 concluded that the preference for male bosses over female bosses was still present for both sexes.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
543
A powerful woman seems to be a contradiction in terms (Dubno, 1985; Kahn, 1984). This incompatibility between the female gender role and a leadership role may result in negative attitudes toward female managers. If she behaves like a woman she is rejected as an unacceptable manager. If she acts in a leader role, she is condemned as unfeminine (Koonce, 1997). This conflict could be especially troubling when a female manager is required to discipline a subordinate. Clearly, power, status and sex-role stereotypes could impact a female supervisor’s success in delivering discipline to a subordinate. 3.4. Reactions from recipients For discipline to be effective, it also ultimately requires behavior change on the part of the recipient with minimal negative side effects. Ball, Trevino and Sims (1994) proposed that recipients’ perceptions of the discipline event were important factors influencing the discipline recipient’s subsequent behavior and attitudes. Ridgeway and Diekema (1992) found that status asserting behavior was less acceptable when it came from a woman. Kunda and Sinclair (1999) found that people were more likely to activate negative stereotypes when their egos were threatened by the person. Thus, a context in which a female must discipline a subordinate may elicit strong negative stereotypes about female managers that would not be elicited by male managers. Female managers could thus be less successful at administering discipline simply because recipients are less willing to accept and constructively react to discipline when it comes from a woman rather than a man.
4. Hypotheses Based on the above discussion, we expect females to face more difficulties administering discipline and thus, to be less effective in this role than men. Women’s lowered ascribed and achieved power and status, and individuals’ reduced willingness to accept assertiveness, dominance, and influence attempts from women could create less than a supportive environment for women attempting to discipline a subordinate. Because subordinates may see the behavior as inappropriate and because females may lack confidence and respect from subordinates, negative repercussions would be expected to be greater when a woman delivers discipline than when a man does. For example, subordinates may feel more negatively toward a female manager or be less likely to modify subsequent behavior. Additionally, if discipline is coming from a person seen to lack power and authority, subordinates may be less willing to accept responsibility for their behavior. Because discipline behavior is contrary to expected female gender roles, (e.g., females are supposed to be kind and nurturing), discipline delivered by a female may be more likely to be seen as mishandled and less fair. Given the above discussion, we believe that both person-centered theory and bias-centered theory will negatively impact female supervisors in the discipline process. Specifically, we hypothesize: H1: Female supervisors will be less effective at delivering discipline than male supervisors.
544
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
H1a. Overall effectiveness of the discipline incident will be lower when the supervisor is female. H1b. Discipline will be less likely to result in behavior change by the recipient when it is delivered by a female supervisor. H1c. More negative outcomes will result when discipline is delivered by female supervisors. H1d. Female supervisors will be perceived by recipients as handling the incident more poorly than male supervisors. H1e. Discipline delivered by female supervisors will be perceived by recipients as less fair. H1f. The recipient will be less likely to accept responsibility for the event when discipline is administered by a female supervisor. H1g. Recipients will have more negative emotional reactions (e.g., they will be more angry and upset) when discipline is administered by a female supervisor. 4.1. Expected differences between male and female discipline recipients Research by Roberts (1991) and Roberts and Nolen–Hoeksema (1994) suggests that adult females are more sensitive and responsive to evaluative, negative feedback than males which may stem from their differing exposure to and interpretation of negative feedback (Roberts, 1991). For example, as mentioned earlier, boys receive more criticism than girls in school, but that criticism is directed toward disruptive behavior rather than incompetence. Girls receive less criticism, but what they do receive is directed at academic mistakes (Dweck, Davidson, Nelson & Enna, 1978). The academic teaching pattern suggests that girls will learn that acceptance and praise come when they demonstrate their ability to perform well. Criticism comes from failure that then threatens their self-esteem. For girls, criticism is taken personally, that is, they feel like a “bad person.” Failure for boys is a sign that they are not trying hard enough and is less likely to threaten their self-esteem (Dweck, 1986). For boys, criticism is not taken personally. Rather, they believe what they have done is “bad.” And, because boys receive more negative feedback, they tend to become desensitized to it. Given their greater sensitivity to negative feedback developed early in life, females may be less willing to accept responsibility for their mistakes. They may also react more negatively emotionally because they are more likely to take criticism personally, that is, as a threat to their self-esteem. In sum, these differences may contribute to differences between males and females in their reactions to and constructive use of discipline. We hypothesize the following: H2: Female recipients will react more negatively to discipline than male recipients. H2a. Female recipients will be more emotionally upset by discipline than male recipients. H2b. Female recipients will be more likely to perceive the discipline as unfair.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
545
H2c. Female recipients will be less willing to accept responsibility for their behavior than male recipients. The ways in which the sexes interact with one another is also at issue in a discipline context when males are disciplining females or females are disciplining males. Gutek and Cohen (1987) discuss “sex-role spillover” and contend that both men and women are usually comfortable interacting with the opposite sex in ways that are congruent with their gender roles, even if this conflicts with optimally carrying out one’s work role. Sex-role spillover would suggest that women supervisors may have particular difficulty when faced with the need to discipline a male subordinate. How does a female supervisor integrate the submissiveness society expects of women when interacting with men and the dominance required of women when disciplining men? Because male recipients are likely to be sensitive to the ascribed and achieved differences in power and status between themselves and their female supervisors, we anticipate that they will have more difficulty accepting discipline when it comes from a female rather than a male. Specifically, we hypothesize the following: H3. Male recipients will react more negatively to discipline from a female supervisor than will female recipients. H3a. There will be more negative outcomes when discipline is administered by females to males than by females to females. H3b. Males will be more likely than females to believe a female supervisor handled the discipline poorly. H4. Male recipients will react more negatively to discipline from a female supervisor than from a male supervisor. H4a. More negative outcomes will result when discipline is delivered by female supervisors to males than when discipline is delivered by males to males. H4b. Discipline delivered by females to males will be less effective than discipline delivered by males to males.
5. Method 5.1. Sample and data collection Data collected for this study were part of a larger study designed to assess recipients’ observers’ and managers’ perceptions of discipline events. We employed the qualitative interview approach used by Butterfield, Trevino and Ball (1996) to collect retrospective accounts of discipline events. While managers (or those delivering discipline) were interviewed by Butterfield et al. (1996), in our study, recipients of discipline were interviewed. As in the Butterfield et al. (1996) study, individuals interviewed for our study were contacted by undergraduate students studying Human Resources Management at a university in the Southwestern region of the U.S. Students were asked to identify individuals they knew who
546
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
had been disciplined on the job. Discipline could take the form of oral reprimand, written reprimand, suspension, or termination. The discipline had to be delivered by a boss, supervisor, or superior, rather than a co-worker. Students received training in how to conduct interviews as part of the course. Students were provided with a set of specific questions and were instructed to work in pairs and to follow the protocol exactly. All interviews were tape-recorded. Each interview lasted approximately 20 min. Students’ grades on the interview project depended in part upon the extent to which the protocol was followed. Each interviewee also completed a demographic sheet that asked a number of questions including the races, ages, and genders of the discipliner and recipient as well as job type. A total of 170 interviews were conducted, but seven interviews were dropped because the protocol was not followed or the tapes were inaudible. Our usable sample included 163 interviews. Sixty-eight of these interviews were conducted with female recipients; 95 with male recipients. Recipients were asked to indicate the gender of the supervisor who administered the discipline, as well as the approximate age of the supervisor and the type of work that they were doing when disciplined. The gender breakdown for the supervisors administering the discipline was 68 males and 95 females. (Note: In our original set of interviews, there were too few cases of males disciplined by females. We added to the database by specifying that students had to interview males disciplined by females. This accounts for the higher number of male as opposed to female recipients and the higher number of female supervisors.) Male and female supervisors did not differ in age nor in the types of work supervised. For example, banking, construction, medical/dental, restaurant, and retail were represented in each of the four possible gender combinations. The interview protocol was adapted from the protocol used by Butterfield et al. (1996). Questions were modified slightly where necessary to reflect the target of the interview (in our case the discipline recipient rather than the discipliner). 5.2. Content analyses The tape-recorded interviews were transcribed verbatim. Data were then analyzed in the following manner. First, each interview was broken down into “thought units” (Butterfield et al., 1996; Gioia & Sims, 1986) by a member of the research team. Each thought unit was preceded by numbers representing the interview number and numerically coded demographic information provided by the interviewee. Thought units ranged from a phrase to several sentences. The goal was to capture every complete and unique thought or idea. Every idea that pertained to the questions or to punishment/discipline (2900 total thought units) was coded into the categories adapted from Butterfield et al. (1996). Seven categories were specifically relevant to the hypotheses tested in this study (i.e., other categories such as co-worker reactions and manager reactions were not relevant to the purpose of this study or had too few entries to make meaningful gender comparisons). The categories used in this study were: Punishment didn’t work; general negative outcomes; manager made a mistake or didn’t know how to punish; recipient thought punishment was unfair; recipient did not accept responsibility for his or her behavior; recipient felt sad, disappointed, upset; recipient felt angry. An example from each category is provided in Table 1. In total 577 thought units
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
547
Table 1 Punishment categories and sample thought units Category
Sample thought unit
Punishment didn’t work General negative outcomes a) Bad feelings toward job or organization
“To be honest I don’t think it did any good at all.”
b) Bad feelings toward supervisor or co-workers c) Negative impact on co-workers or supervisor d) Negative impact on recipient Manager made a mistake or didn’t know how Recipient thought punishment was unfair Recipient did not accept responsibility
Recipient angry Recipient sad, disappointed, upset
“Afterwards I had a negative attitude and really didn’t want to go above and beyond to help with the facility in working overtime or pitching in.” “Whenever I see her (the supervisor) now, I have no respect.” “It created more fears among employees.” “Built up a lot of stress, anxiety . . . irritable with the family, my wife, silly disputes over things that used to be senseless.” “I believe the person doing the discipline was a robot following by rote and (she) had no idea what was happening.” “I felt that I was singled out and not given a chance to respond to the incident at the time.” “I didn’t do anything to deserve any disciplinary action. I was not delinquent on any of my qualifications. They just wanted me do more than I was doing and I was unwilling to do more.” “I was outraged that they were doing this.” “I shed a couple tears and after I got over that initial phase, I became somewhat bitter.”
were coded into these seven categories. Because a recipient could have endorsed a category more than once in an interview, we coded categories in terms of their presence or absence in an interview/incident. While breaking the interviews into thought units, the researchers also numerically coded characteristics of the discipline event. The following were coded: (1) the type of discipline (oral reprimand ⫽ 1; written reprimand ⫽ 2; suspension, probation or termination ⫽ 3), and (2) whether the incident could be classified as effective or ineffective. The first issue was straightforward and came from questions in the interview protocol. The issue of effectiveness required a judgment made by the coder based on the interview questions about outcomes of the discipline and other remarks made by the interviewee. Interviews were coded as 1 ⫽ effective, and 2 ⫽ ineffective. Effective incidents included cases where the outcome was clearly positive, or was positive but was accompanied by a more minor negative outcome (e.g., the individual improved, but was embarrassed). Alternatively, ineffective incidents included those where the outcome was clearly negative or primarily negative, but a minor positive outcome occurred (e.g., the recipient had a very poor relationship with the supervisor and wanted to quit his job, but was rarely late to work). The proportion of interviews coded as effective was 47%, while those coded ineffective was 53%. Agreement rates for effectiveness coding into two categories, prior to discussion were 95% and after discussion were 100%. The thought units were randomly separated into four sets. Five coders (two males and three females) worked in pairs to categorize the sets of thought units into the categories. The
548
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
Table 2 Frequencies and proportion of interviews where category was endorsed Male supervisor Female supervisor 2 Effectivenessa
Ineffective 29/68 .43 Effective 39/68 .57 Punishment didn’t work (did not change behavior) 26/68 .38 General negative outcomes (e.g., bad feelings toward 43/68 .63 organization or supervisor) Manager made a mistake or didn’t know how to 12/68 .18 discipline Recipient thought punishment was unfair 30/68 .44 Recipient did not accept responsibility for his or her 21/68 .31 behavior Recipient felt sad, disappointed, upset 21/68 .31 Recipient felt angry 33/68 .49
57/95 .60 38/95 .40 40/95 .42 69/95 .73
4.79* .25 1.63
34/95 .36
6.44**
55/95 .58 46/95 .48
3.02* 5.04*
24/95 .25 54/95 .57
.63 1.10
a This analysis was a 2 ⫻ 2 2 (effectiveness by gender). For the remaining categories the proportions indicate the proportion of interviews for which the category was present versus the number of categories where it was not present. * p ⬍ .05; ** p ⬍ .01.
pairs were formed randomly so that a different pair coded each set to minimize the possibility that the pairs would begin to “think alike.” After a sample of thought units was coded, pre-discussion agreement rates were determined to be 71%. Thought units for which there was disagreement were discussed by the pair of coders, and agreement rates after discussion were 100%. Thought units that did not fit into existing categories were coded as miscellaneous. A complete listing of categories, subcategories and their relative frequencies is available from the first author.
6. Results 6.1. Effectiveness of male versus female supervisors Hypotheses 1a through 1g proposed that female supervisors would be less effective at administering discipline than male supervisors. Hypotheses were tested with chi square analyses comparing the proportion of interviews in which each category related to effectiveness was mentioned. As can be seen from Table 2, there were four areas in which females were seen as less effective. In support of H1a, the discipline was judged less effective overall when administered by a female supervisor. Supporting H1d, females were more likely to be seen as making a mistake in administering the discipline or not knowing how to administer discipline. For example, recipients made comments such as, “She didn’t have to be so frank and nasty, and the way she came across was not in a professional manner,” or “I was completely dumbfounded. I had no idea what she was talking about,” Another, after correcting her supervisor’s facts, stated, “She just said, ‘Oh.” She didn’t even apologize.” These are contrasted with a comment regarding a male discipliner, “He stopped me from
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
549
doing it again, but he didn’t crush my spirit.” Recipients were more likely to see the discipline as unfair (supporting H1e) and were less likely to accept responsibility for their behavior when the supervisor was female (supporting H1f). Not infrequently, recipients disciplined by females cited favoritism. For example a recipient commented, “She had her little clique of friends and they abused many, many things. I was singled out and that was my biggest problem.” Regarding Hypotheses 1b, 1c, and 1g, the discipline was reported to work equally well in changing future behavior, and recipients were no more, yet no less, likely to be angry or upset when the discipline was delivered by a female, rather than a male, supervisor. In sum, hypotheses 1a, 1d, 1e, and 1f were supported, but 1b, 1c, and 1g were not. Overall, partial support is provided for Hypothesis 1. 6.2. Reactions of male versus female recipients Hypotheses 2a and 2b were tested with chi-square tests for differences between proportions. Females were no more likely to be upset, in general than males when disciplined (2 ⫽ 1.02), thus Hypothesis 2a was not supported. Female recipients were, however, more likely to perceive the discipline as unfair than were males supporting Hypothesis 2b. In 62% of the interviews, the female recipient found the incident to be unfair whereas in 45% of the interviews with males the recipient found the incident to be unfair (2 ⫽ 4.96, p ⬍ .05). Some female recipients appeared to be adamant about this subject stating for instance, “In no way was it fair. In no way was it appropriate,” “It was an injustice,” and “I think it was completely uncalled for.” Although males also found some incidents to be unfair, they more often accepted the equity of the situation. Females were not significantly less likely to accept responsibility than males (41% for females; 42% for males). Hypothesis 2c was not supported. 6.3. Reactions of male and female recipients to male and female discipliners Hypothesis 3a was not supported (see Table 3). Negative outcomes were not reported significantly more often when a female supervisor disciplined a male than when a female disciplined a female. Accordingly, female and male recipient comments tended to be similar, focusing largely on the strained relationship with the supervisor that was created by the discipline. A male commented, “The negative outcomes were that a lot of anger built up toward the manager and that is something I will never forget,” while a female likewise noted, “It made me dislike her more.” Additionally, both males and females cited reduced motivation as a negative outcome, for example, “I have become selfish about what I will and will not do for the company” from a female and, “I mean, I still do my job, but to go the extra mile for that person was very difficult after that time,” from a male. Hypothesis 3b was supported, however. Males disciplined by females were significantly more likely to believe that the manager made a mistake or did not know how to discipline, than were females disciplined by females (43% versus 28%). In fact, perceptions that the discipline was mishandled were more likely when males were disciplined by females, than any in any other gender combination.
550
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
Table 3 Frequencies and proportions of interviews where category was endorsed Male supervisor
Incident was effective (vs. ineffective) Punishment didn’t work (did not change behavior) General negative outcomes (e.g., bad feelings toward organization of supervisor) Manager made a mistake or didn’t know how to discipline Recipient thought punishment was unfair Recipient did not accept responsibility for his or her behavior Recipient felt sad, disappointed, upset Recipient felt angry
Female supervisor
Significant differencesa*
male subordinate (A)
female subordinate (B)
male subordinate (C)
female subordinate (D)
25/46 .54
14/22 .64
22/49 .45
16/46 .35
17/46 .37
9/22 .41
23/49 .47
17/46 .37
27/46 .59
16/22 .73
37/49 .76
32/46 .70
7/46 .15
5/22 .23
21/49 .43
13/46 .28
AC, AD BC, CD
17/46 .37
13/22 .59
25/49 .51
30/46 .65
AB, AD, CD
17/46 .37
4/22 .18
22/49 .45
24/46 .52
BC, BD
18/46 .39
15/22 .70
29/49 .59
25/46 .54
AB, AC
13/46 .28
8/22 .36
13/49 .27
11/46 .24
AD, BD
AC
a Z-tests for differences between proportions were significantly different (e.g., (A) .15 and (C) .43 differ significantly). * p ⬍ .05.
Hypothesis 4a also was supported. Males disciplined by females reported significantly more negative outcomes than males disciplined by males (76% vs. 59%). With regard to effectiveness (Hypothesis 4b), discipline events in which males were disciplined by females were not significantly less effective than when males were disciplined by males. However, while not hypothesized, it is interesting to note that events in which both males and females were disciplined by males were more effective than those in which females were disciplined by females (54% and 64% vs. 35%) (see Table 3). Male comments tended to be more resolute when disciplined by males, stating, “It was very effective” and “When it comes to being disciplined, I accept it.” But males appeared more resentful when disciplined by females, stating, “I was getting tired of hearing about it, and she just never really got over it.” The difference in acceptance of responsibility by females when she was disciplined by a male (only 18% did not accept) versus a female (52% did not accept) is also noteworthy.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
551
7. Discussion Results from this study suggest that, in general, situations involving females delivering discipline are perceived by recipients to be less effective overall, and females delivering discipline to male recipients are particularly troublesome. Recipients were less likely to believe that the punishment was fair, were less likely to accept responsibility for their behavior, and were more likely to think the manager did not know how to deliver the discipline when the discipliner was female. However, discipline delivered by males and females had similar success in terms of changing subsequent behavior. In roughly 40% of the cases for both male and female managers, the discipline reportedly did not change the recipient’s subsequent behavior or fix a problem. (These were self-reports of recipients. Managers may have had a different view.) It appears that while recipients are not less likely to change behavior in response to discipline by a female, their perceptions of the incident in terms of fairness, etc., are more negative. Male recipients had a more negative reaction to discipline when the discipliner was female than when the discipliner was male. Negative outcomes resulting from the discipline such as bad feelings toward the organization or supervisor were more likely when males received discipline from females than when they received discipline from males. Males who were disciplined by females also were much more likely to believe the discipline had been mishandled. In addition, male recipients also were more likely to react negatively in terms of emotions when the discipliner was female rather than male. In short, females delivering discipline to males appear to be problematic. This could result from discomfort on both the female supervisor’s and male recipient’s parts because of perceived status and stereotype differences. It would be interesting in future research to determine whether females actually handle discipline to males less well, or whether this is merely a result of less acceptance on the part of male recipients. Females were less likely in general to believe punishment was fair than were males. Given the differences in communication styles between men and women mentioned earlier, some of this difference may be attributable to females’ greater willingness to discuss feelings of fairness. There were also differences for females in terms of their reactions to male and female discipliners. In contrast to our expectations, discipline incidents for female recipients were far more likely to be effective when the discipline was delivered by a male, rather than by a female. Females were also much less likely to accept responsibility for their behavior when the discipline came from a woman, rather than a man. Again, this could be the result of sex role stereotypes and perceived power and status differences. Females may be more willing to accept discipline from a male because males are not expected to be compassionate and considerate. They also may be more accepting of their subordinate position to a male than a female. As an anecdote, a secretary once told one of the authors that it was more difficult for her to work for a woman because it was difficult to accept the status differences between herself and another female. It was easy to accept status differences when the supervisor was male. Acceptance of status differences may make discipline easier to accept. Because our data all came from the recipient’s perspective we could not discern whether females actually behaved differently than males or whether they were merely perceived differently. However, we do know that recipients were not reacting more negatively because
552
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
discipline delivered by females was more severe. In fact, the opposite was true. Males were more likely to terminate while females were more likely to orally reprimand. We also could not attribute differences to age of supervisor or type of job. Nevertheless, because the effectiveness of a discipline event depends upon the recipient’s reaction to or acceptance of the discipline, we must conclude that women are somewhat less effective in discipline delivery. While recipients were no less likely to change their behavior when discipline came from a woman, they were less likely to accept responsibility, more likely to think the discipline was unfair, and more likely to think the event was mishandled. Based on our results, we conducted two follow-up studies in an attempt to address the probable cause of the differences in effectiveness for male and female supervisors in administering discipline. These follow-up studies are described below.
8. Follow-up studies 1 and 2 Whether females were perceived as less effective at administering discipline could be best explained by either person-centered theory (e.g., socialization and behavior differences) versus bias-centered theory (recipient’s negative reactions to females administering discipline) was the question addressed in two follow-up studies. Based on bias-centered theory, one explanation is that recipients are less accepting of discipline when it is delivered by a female due to stereotypes and biases. We address this question in our first follow-up study by standardizing the discipline episode and manipulating the gender of the supervisor delivering it. In this way, we hoped to isolate biases toward females if they existed. That is, if in a standardized hypothetical discipline scenario, recipients reacted more negatively to females, we could conclude that negative reactions were based on gender bias alone since the behaviors of the male and female supervisor would be the same. 8.1. Follow-up study 1 Surveys were designed to present an appropriate discipline scenario and an inappropriate discipline scenario as delivered by a male or a female supervisor. Both scenarios involved a violation of a dress code by an employee. In the appropriate scenario, the recipient was called into the supervisor’s office and the violation was calmly explained. The recipient was respectfully asked to go home and change into acceptable clothing. In the inappropriate scenario, the supervisor made a scene in front of others and told the employee to leave work, and never come into work in such an outfit again. Male subjects received a survey in which “Steve” was the recipient, and either Mary or Bob was the supervisor. Female subjects received a survey in which Linda was the recipient, and either Mary or Bob was the supervisor. (We felt that if we were asking the subject to put him or herself into the position of the recipient, the recipient and subject gender had to be the same.) One hundred seventy business students (89 males and 81 females) read the scenarios and completed surveys. Forty-one males and 45 females were presented with the inappropriate scenario; 48 males and 36 females read the appropriate scenario. Attached to the scenario was a survey that asked recipients to put themselves in the place of the person being
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
553
disciplined. They were also asked the following questions: (1) would you follow the dress code in the future? (2) how would you feel about your supervisor following the interaction? (3) how would you feel about working for this store in the future? These questions were responded to on a five-point scale ranging from 1 ⫽ very positive to 5 ⫽ very negative. Subjects were also asked: (4) how did you feel that the supervisor handled the situation? This question was also answered on a five-point scale ranging from 1 ⫽ very well to 5 ⫽ very poorly. Questions 5 and 6 (do you think the punishment was fair, and would you accept responsibility for making a mistake) were answered on a three-point scale (1 ⫽ yes, 2 ⫽ maybe, 3 ⫽ no). Subjects were also asked whether they would or would not feel sad, upset, angry, disappointed, and embarrassed. When these surveys were turned in, subjects completed a second survey that assessed demographic information and served as a manipulation check. Specifically, students provided their age, number of years of work experience and were asked about the genders of the supervisor and recipient in the scenario. The average age of subjects was 27 years. The average number of years of work experience was 10 years. The manipulation check indicated that in all cases the subject accurately recalled the gender of the supervisor and recipient. 8.2. Results and discussion Two by two analyses of variance (male/female supervisor by male/female recipient) were performed on the six outcomes, as well as the five emotional reactions. These analyses were performed separately for the appropriate and inappropriate incidents. Results are presented in Table 4. It should be noted that subjects were significantly more likely to indicate that if they had been the disciplinee they would observe the dress code in the future in the appropriate incident (89%) as compared to the inappropriate incident (74%) (F(1,166) ⫽ 7.57, p ⬍ .01). In the appropriate incident, there were supervisor gender by subordinate gender interactions for feelings about the organization and supervisor’s handling of the situation. Tukey HSD tests revealed that both males and females reading scenarios where the supervisor was female indicated that they would feel more positively toward the organization following the incident than males reading scenarios where the supervisor was male. Males reading scenarios where they reported to females were also more positive than females reading scenarios where they reported to females. Regarding supervisors’ handling of the situation, Tukey HSD tests indicated that females reading scenarios involving female supervisors were significantly more negative than were males reading scenarios where the supervisor was female. This is in contrast to the findings from the interview data where males reporting to females were most likely to report that the situation had been poorly handled. Supervisor gender effects indicated that both males and females reported they would feel more angry when the supervisor was female than when the supervisor was male. Subordinate gender effects indicated that female recipients were more likely than male recipients to report embarrassment. In the inappropriate scenario, there was a significant interaction effect for fairness. Tukey HSD tests indicated that males reading scenarios where the supervisor was male saw the situation as more fair than females with male supervisors or males with female supervisors. Males reading scenarios where the supervisor was female saw the situation as
554
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
Table 4 Means and F-ratios for survey results Appropriate scenario Means
F-ratios
Male supervisor
Female supervisor
Outcomes
Male recip A
Female recip B
Male recip C
Female recip D
Follow dress code Feel about supva Feel about orga Supv handled situationa Fair/unfair Recip accept responsibility Emotions Sad Upset Angry Disappointed Embarrassed
1.08 3.43 3.30 3.00 1.75 1.29
1.05 3.38 3.00 2.71 1.71 1.42
1.17 3.04 2.50 2.58 1.71 1.29
1.17 3.25 2.92 3.25 1.67 1.08
.11 .04 .01 .13 .35 .09
1.97 1.04 3.12 .06 .00 1.39
.05 .51* 4.09b* 4.09c* .04 1.39
.00 .67 .29 .50 .46
.00 .50 .29 .25 .75
.00 .71 .50 .25 .54
.00 .67 .58 .42 .92
.18 .73 .13 .15 9.85**
1.62 .73 4.61* .15 1.39
.18 .26 .13 3.72 .15
Supv gender
Recip ⫻ supv gender and cell differences
Recip gender
Supv gender
Recip ⫻ supv gender and cell differences
Inappropriate scenario Means
F-ratios
Male supervisor
Female supervisor
Outcomes
Male recip A
Female recip B
Male recip C
Female recip D
Recip gender
Follow dress code Feel about supv Feel about org Supv handled situation Fair/unfair Recip accept responsibility Emotions Sad Upset Angry Disappointed Embarrassed
1.19 3.90 3.62 4.24
1.45 4.36 3.86 4.64
1.25 4.20 3.95 4.45
1.30 4.48 4.00 4.65
2.22 3.76 .68 3.22
.29 .84 .94 .32
.50 .12 .11 .23
1.95 1.29
2.32 1.68
2.53 1.40
2.17 1.35
.00 1.57
1.66 .64
4.62d* 2.67
.19 .62 .57 .57 .57
.18 .77 .86 .41 .91
.15 .70 .75 .40 .80
.13 .74 .74 .26 .96
.03 .94 2.19 2.03 9.66**
.32 .06 .08 2.28 3.01
.00 .33 2.54 .01 1.30
a
Low score is more positive. Significant difference for cells AC, AD, BC, BD. c Significant difference for cells BC, BD. d Significant difference for cells AB, AC, AD. * p ⬍ .05; ** p ⬍ .01. b
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
555
less fair than females with female supervisors. In this situation, females again reported more embarrassment than males, but there was no supervisor gender effect for anger. In summary, the survey results suggest that when presented with the same behavior from a male and female supervisor, when the incident is handled appropriately, male and female recipients generally do not perceive more negative outcomes when the supervisor is female. However, they do report significantly more anger when disciplined by a female. When the incident is handled inappropriately, female supervisors are perceived somewhat more negatively in terms of fairness. Particularly noticeable was the fact that males were more likely to see the discipline event as more fair when the supervisor was male than when she was female, though not all outcomes were more negative when the supervisor was female. Notably males reading scenarios where the supervisor was female were least negative in terms of subsequent feelings about the supervisor or organization. These data suggest that negative stereotypes toward females are not pronounced, but recipients’ tendency to feel more angry toward females provides some support for bias-centered theory. Combined with our earlier results, results from the follow-up study suggest that perhaps female supervisors in real work settings are less effective in delivering discipline than their male counterparts, in part because recipients are less accepting and more angry. However, the possibility that they are behaving differently cannot be ruled out. Females were not seen as mishandling the situation more often than males when the behavior by each was identical. Perhaps females are behaving differently than males. This question was addressed in the second follow-up study. 8.3. Follow-up study 2 In the second follow-up study we addressed person-centered theory by asking males and females to indicate how likely they would be to engage in a variety of behaviors in response to a male or female employee who had been reprimanded for inappropriate use of the phone and was caught again inappropriately using the phone. One hundred ten male subjects and 101 female subjects rated 10 different behaviors indicating how likely they would be to engage in each behavior. The scale ranged from 1-very unlikely to 4-very likely. Subjects were on average 29 years old, 79% were currently employed and 75% had worked as a supervisor at some time. Behavior alternatives rated by subjects included: ignore the behavior, talk to the person about what necessitated phone calls, politely orally reprimand, angrily orally reprimand, politely issue written reprimand, angrily issue written reprimand, politely suspend, angrily suspend, politely terminate, angrily terminate. We also combined the two types of oral reprimand into one measure of oral reprimand and combined the two types of termination into one measure of termination. 8.4. Results and discussion Data were analyzed with two-way analyses of variance (subject gender by recipient gender). Results are presented in Table 5. As can be seen, there were two behaviors in which a subject (supervisor) gender effect was found. Females were more likely to report that they
556
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
Table 5 Means and F-values for likelihood of responses to rule violation Gender of hypothetical subordinate
Ignore Talk about what’s going on Polite oral rep Angry oral rep Polite written Angry written Polite suspension Angry suspension Polite termination Angry termination Oral reprimand (polite & angry) Termination (polite & angry) a
Male
Female
Subject (SUPV) Gender
Subject (SUPV) Gender
Female
Male
Female
Male
1.48 2.62 2.78 1.90 2.68 1.36 1.44 1.30 1.32 1.18 2.34 1.25
1.64 2.92 3.08 2.09 2.64 1.25 1.47 1.34 1.51 1.30 2.58 1.41
1.55 2.83 2.78 2.08 2.53 1.55 1.80 1.62 1.78 1.62 2.43 1.70
1.72 2.72 2.91 2.02 2.77 1.47 1.79 1.34 1.60 1.34 2.48 1.46
F-values Subj gender
Subord gender
Subj ⫻ subord
3.17a .63 3.77* .36 .62 .98 .00 1.53 .00 .67 3.13a .19
.58 .01 .50 .26 .00 3.73* 7.98** 2.48 5.67* 5.64* .01 6.42*
.00 2.61 .54 .86 1.14 .01 .07 2.69 2.98a 4.22* 1.46 4.01*
p ⬍ .10; * p ⬍ .05; ** p ⬍ .01.
would ignore the behavior or issue polite oral reprimand than were males. There were also two significant interaction effects. Male subjects were more likely to report that they would terminate female recipients either politely or impolitely than they were to terminate male recipients. When oral reprimands were compared for male and female subjects, females were more likely to report that they would orally reprimand than were males. These results concerning harshness of discipline support our interview findings where males were more likely to terminate while females are more likely to orally reprimand. There were also effects for recipient gender that indicated that subjects reported more willingness to issue tougher penalties when the recipient was female than when he was male. While these effects were not hypothesized, given that the offense committed by the male and female recipient was identical, it is another indication that females are reacted to differently in the workplace than males. Overall, the results of this follow-up study tentatively suggest that males and females behave somewhat differently when faced with a discipline event. Females were more likely to report that they would deliver oral reprimands or ignore the event than were males. This supports person-centered theory. Also, the males were particularly likely to report that they would terminate the recipient if she was female, and generally both males and females were more likely to endorse tougher penalties for females than for males, suggesting some presence of bias-centered theory as well. 9. General discussion and conclusions Results from this research suggest that in terms of recipient perceptions, both outcomes and recipient emotions were more negative when the discipline was delivered by a female.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
557
Given that recipient reactions are a critical component of the success of a discipline event, we can conclude that females are less effective in this domain. The results from the follow-up studies suggest that this decrease in effectiveness may be a result of bias, supporting bias-centered theory, as well as behavior, supporting person centered theory. While the results from neither follow-up study were strong, indications were that there was some bias operating, and some behavior differences likely. We could also speculate that in the real world, biases may actually influence behaviors and vice versa, combining to create an environment that is less conducive for females in this aspect of their managerial roles. For example, females may get cues from subordinates that discipline from them is unwelcome. This may serve to reduce their confidence at discipline delivery, which in turn impacts their behavior and the way discipline from females is then perceived. The result is a downward spiral. With regard to male recipients’ negative reactions to female supervision in the interview data (i.e., their perception that females mishandled the discipline), and their responses that the situation was less fair when delivered by a female in the survey data, Hale (1996) suggests that males may perceive women in positions of power as a threat. “Men may see women’s equality in organizations as a resource depletion problem; that is, women in the workforce deplete job resources historically available to men. As the primary beneficiaries of hierarchical reward systems, men may also believe that sharing power really means yielding (their) privilege and doing so is both dangerous and costly” (p. 9). Certainly being disciplined by a female boss is a clear sign of female power which may be perceived as threatening to males in a variety of ways. 9.1. Limitations Several limitations should be mentioned. One of the limitations of the interview-based study is that we do not have data on supervisors’ actual behavior, but rather have only perceptions from discipline recipients. Nevertheless, as suggested before, with regard to the effectiveness of discipline, it is the recipient’s subsequent attitudes and behaviors that determine discipline’s success. A second limitation is that we do not know the make-up of the workforce in the jobs reported. For example, males working in traditionally “female” jobs could have been sensitized to their minority status as one of a few males in a predominantly female organization. This minority status could have caused overly negative reactions to discipline by a female. Our data suggest, however, that female recipients responded better to male supervisors than to female supervisors which would indicate that their reactions were based more on status and power differences between males and females than upon the gender type of the job. It should also be noted that accounts of discipline interviews from female recipients could be more emotion laden because females are more likely to express emotions. This could have contributed to their increased reports of embarrassment relative to males. An additional limitation is the convenience sampling that was done by having students select interviewees. Future research should attempt to identify males and females working in similar job contexts with supervisors at the same level in the organization in order to more systematically test male and female differences. Regarding the survey studies, reactions to “paper people” have received criticism in
558
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
previous research for not being representative of real work decisions. However, most of the criticism has suggested that stronger effects are produced when paper people are used than occur in direct observation of real life, or videotaped vignettes (Murphy, Herr, Lockhart & Maguire, 1986). As such, we would have expected more gender bias in the first survey study if bias in the workplace was likely. Alternatively, we may have seen more behavior differences in the second study than would actually occur. Yet our interviews suggest that there may be stronger effects in “real life.” Perhaps in “paper people” studies subjects are more influenced by what they feel they should say or do rather than what they would actually feel or do. In conclusion, while females have been shown in numerous studies to be equally effective as males at managing, they may not be equally effective at all aspects of management. In fact, our results would suggest that females may be compensating for their decreased success in the discipline arena with increased success in other arenas such as team-building and motivation. Overall, because our findings suggest that discipline is more acceptable to recipients when it is delivered by a male, female managers may need special training in how to deliver discipline in ways that are both effective and minimally threatening to recipients. Alternatively, perhaps employees need to be made aware of stylistic differences so they can become more accepting of a wider range of behaviors from female supervisors. The survey results regarding anger are also provocative. If both males and females express more anger when disciplined by a female even when the situation is handled appropriately, this may contribute to other negative outcomes and reactions that undermine a female supervisor’s effectiveness. Given that generally it appeared that females were likely to deliver less harsh discipline than males, and the follow-up study suggested that they were more likely to ignore a rule violation, perhaps females are sensitive to how they are perceived as discipline deliverers and tend to avoid it or deliver “softer” discipline. Perhaps this contributes to decreased credibility of females as discipline administrators in the eyes of subordinates. While our results are somewhat disappointing to those hoping that females were acquiring equality in the workplace, we can be optimistic that as more women enter the workforce and become role models for their children that behavioral expectations for boys and girls will adapt, and socialization patterns will become more similar. Additionally, as more women enter supervisory positions, recipients’ comfort with women as disciplinarians should increase. It is our hope that the results of this study will raise awareness on the part of managers and subordinates regarding stereotypes and expectations, including gender, and their impact on supervisory effectiveness. Perhaps if female supervisors become aware of how they may be perceived, they can approach discipline events more constructively with more sensitivity to recipients’ potential reactions. To be sure, our recommendation is not for women to “wait until father comes home.”
Acknowledgments We wish to thank Joan Brett, Priscilla Cartier, Barry Nathan, and Cheri Ostroff for their assistance and comments on earlier drafts of this work.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
559
Leanne Atwater is a Professor of Management at Arizona State University West. Her research areas include leadership, 360 degree feedback and gender and discipline. She is also the President of Atwater Management Consulting in Phoenix. James Carey is a Lecturer in the School of Management at Arizona State University West. He received his Ph.D. from Arizona State University and is certified as a Senior Professional in Human Resources (SPHR). His research interests include gender issues in the workplace, performance appraisal, and human resource issues affecting competitiveness. David A. Waldman is a professor of management at Arizona State University West and an affiliated faculty member of the Department of Management at Arizona State University Main. He has also taught at Concordia University and Binghamton University. His research interests include the effects of leadership across levels of analysis, multisource feedback processes, and the application of discipline in work settings. Professor Waldman is a fellow of the Society for Industrial and Organizational Psychology and the American Psychological Association. References Ball, G., Trevino, L., & Sims, H. (1994). Just and unjust punishment: influences on subordinate performance and citizenship. Academy of Management Journal, 37, 299 –322. Bass, B., Avolio, B., & Atwater, L. (1996). The transformational and transactional leadership of men and women: an extension of some old comparisons. Applied Psychology: An International Review, 45 (1), 5–34. Benedict, M., & Levine, E. (1988). Delay and distortion: tacit influences on performance appraisal effectiveness. Journal of Applied Psychology, 73, 507–514. Berger, J., Rosenholtz, S. J., & Zelditch, M., Jr. (1980). Status organizational process. Annual Review of Sociology, 6, 479 –508. Bowman, G., Worthy, N., & Greyser, S. (1965). Are women executives people? Harvard Business Review, 43 (4), 14 –28 and 164 –178. Butler, D., & Geis, F. (1990). Nonverbal affect responses to male and female leaders: implications for leadership evaluations. Journal of Personality and Social Psychology, 58, 48 –59. Butterfield, K., Trevino, L., & Ball, G. (1996). Punishment from the manager’s perspective: a grounded investigation and inductive model. Academy of Management Journal, 39, 1479 –1512. Carli, L. (1991). Gender, status and influence. In E. J. Lawler, G. Markovsky, C. Ridgeway, & H. Walker (Eds.), Advances in group processes (pp. 89 –113). Greenwhich, CT: JAI. Dexter, C. R. (1985). Women and the exercise of power in organizations: from ascribed to achieved status. In L. Larwood, A. H. Stromberg, & B. A. Gutek (Eds.), Women and work (pp. 239 –258). Beverly Hills, CA: Sage. Dobbins, G. (1985). Effects of gender on leaders’ responses to poor performers: an attributional interpretation. Academy of Management Journal, 28, 587–598. Dobbins, G. H., Cardy, R. L., & Truxillo, D. M. (1988). The effects of purpose of appraisal and individual differences in stereotypes of women on sex differences in performance ratings: a laboratory and field study. Journal of Applied Psychology, 73, 551–558. Dubno, P. (1985). Attitudes towards women executives: a longitudinal approach. Academy of Management Journal, 28, 235–239. Dweck, C. S. (1986). Motivational processes affecting learning. American Psychologist, 41, 1040 –1048. Dweck, C. S., & Bush, E. S. (1976). Sex differences in learned helplessness: I. differential debilitation with peer and adult evaluators. Developmental Psychology, 12, 147–156. Dweck, C., Davidson, W., Nelson, S., & Enna, B. (1978). Sex differences in learned helplessness: II. The
560
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
contingencies of evaluative feedback in the classroom and III. An experimental analyses. Developmental Psychology, 14, 268 –276. Eagly, A. H. (1983). Gender and social influence: a social psychological analysis. American Psychologist, 38, 971–981. Eagly, A. H. (1987). Sex differences in social behavior: a social role interpretation. Hillsdale, NJ: Erlbaum. Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: a meta-analysis. Psychological Bulletin, 108, 233–256. Eagly, A. H., & Steffen, V. J. (1984). Gender stereotypes stem from the distribution of women and men into social roles. Journal of Personality and Social Psychology, 46, 735–754. Fagenson, E. A. (1990). Perceived masculine and feminine attributes examined as a function of individuals’ sex and level in the organizational power hierarchy: a test of four theoretical perspectives. Journal of Applied Psychology, 75, 204 –211. Falbo, T., & Peplau, L. A. (1980). Power strategies in intimate relationships. Journal of Personality and Social Psychology, 38, 618 – 628. Gioia, D. A., & Sims, H. P. Jr. (1986). Social cognition in organizations. In H. P. Sims, Jr. and D. A. Gioia (Eds.), The thinking organization (pp. 20 – 48). San Francisco, CA: Jossey–Bass Inc. Greer, C., & Labig, C. (1987). Employee reactions to discipline. Human Relations, 40, 507–522. Gutek, B., & Cohen, A. (1987). Sex ratios, sex role spillover and sex at work: a comparison of men’s and women’s experiences. Human Relations, 40, 97–115. Hale, M. (1996). Gender equality in organizations. Review of Public Personnel Administration, Winter, 1996, 7–18. Harlan, A., & Weiss, C. (1982). Sex differences in factors affecting managerial career advancement. In P. A. Wallace (Ed.), Women in the workplace (pp. 59 –96). Boston, MA: Auburn House. Heilman, M. E., Block, C. J., Martell, R. F., & Simon, M. C. (1989). Has anything changed? Current characterizations of men, women and managers. Journal of Applied Psychology, 74, 935–942. Heilman, M., & Kram, K. (1978). Self-derogating behavior in women—fixed or flexible: the effects of co-worker sex. Organizational Behavior and Human Performance, 22, 497–507. Instone, D., Major, B., & Bunker, B. (1983). Gender self confidence and social influence strategies: an organizational simulation. Journal of Personality and Social Psychology, 44, 322–333. Jones, D. (1983). Power structures and perceptions of power holders in same-sex groups of young children. Women and Politics, 3, 147–164. Kahn, A. (1984). The power war: Male response to power loss under equality. Psychology of Women Quarterly, 8, 234 –247. Kent, R. L., & Moss, S. E. (1994). Effects of sex and gender role leader emergence. Academy of Management Journal, 37, 1335–1346. Koonce, R. (1997). Language, sex and power: women and men in the workplace. Training and Development Journal, September, pp. 34 –39. Kunda, Z., & Sinclair, L. (1999). Motivated reasoning with stereotypes: activation, application and inhibition. Psychological Inquiry, 10, 12–22. Lucas, J., & Lovaglia, M. (1998). Leadership status, gender, group size and emotion in face-to-face groups. Sociological Perspectives, 41, 617– 637. Morrison, A., & Von Glinow, M. (1990). Women and minorities in management. American Psychologist, 45, 200 –208. Murphy, K. R., Herr, B. M., Lockhart, M. C., & Maguire, E. (1986). Evaluating the performance of paper people. Journal of Applied Psychology, 71, 654 – 661. Olson, J., Roese, N., & Zanna, M. (1996). Expectancies. In E. T. Higgins, & A. Kruglanski (Eds.), Social psychology: handbook of basic principles (pp. 211–238). New York: Guilford. Powell, G. (1988). Women and men in management. Sage: California. Rhode, D. (1990). Gender equality and employment policy. In S. E. Rix (Ed.), The american woman 1990 –1991: a status report (pp. 170 –200). New York: W. W. Norton & Co.
L.E. Atwater et al. / Journal of Management 27 (2001) 537–561
561
Ridgeway, C. L., & Diatema, D. (1992). Are gender differences status differences? In C. L. Ridgeway (Ed.), Gender, interaction, and inequality (pp. 157–180). New York: Springer–Verlag. Roberts, T. (1991). Gender and the influence of evaluations on self-assessments in achievement settings. Psychological Bulletin, 109, 297–308. Roberts, T., & Nolen Hocksema, S. (1994). Gender comparisons in responsiveness to others’ evaluations in achievement settings. Psychology of Women Quarterly, 18, 221–240. Rosenthal, R. (1995). Interpersonal expectancy effects: a 30-year perspective. Current Directions in Psychological Science, 3, 176 –179. Rudman, L., & Kilianski, S. (2000). Implicit and explicit attitudes toward female authority. Personality and Social Psychology Bulletin, 26, 11,1315–1328. Schlueter, D. W., Barge, J. K., & Blankenship, D. (1990). A comparative analysis of influence strategies used by upper and lower level male and female managers. Western Journal of Speech Communication, 54, 42– 65. Statham, A. (1987). The gender model revisited: differences in the management styles of men and women. Sex Roles, 16, 409 – 429. Sutton, C., & Moore, K. (1985). Executive women—20 years later. Harvard Business Review, June, pp. 42– 66. Tannen, D. (1995). The power of talk: who gets heard and why? Harvard Business Review, September, pp. 138 –148. Wagner, D. G., Ford, R. S., & Ford, T. W. (1986). Can gender inequalities be reduced? American Sociological Review, 51, 47– 61. Wagner, D. G., & Berger, J. (1997). Gender and interpersonal task behaviors: status expectation counts. Sociological Perspectives, 40, 1–32. Walsh, J. P., Weinberg, R. M., & Fairfiled, J. L. (1987). The effects of gender on assessment center evaluations. Journal of Occupational Psychology, 60, 305–309. Wood, W., & Karsten, S. (1986). Sex differences in interaction style as a product of perceived sex differences in competence. Journal of Personality and Social Psychology, 50, 341–347.