Leader group prototypicality and reactions to leader transgressions

Abstract

Leader group prototypicality—the extent to which a leader embodies and represents the group—is a characteristic that typically engenders favorable attitudes toward the leader. The present research examines how leader group prototypicality shapes group member reactions to leaders who commit wrongful acts, that is, transgressions. In three studies, I propose and demonstrate that outsiders’ knowledge of leader transgressions moderates how leader group prototypicality influences group member reactions to transgressing leaders. Specifically, when leaders commit transgressions that are known to outsiders, leader group prototypicality leads to more punitive reactions by eliciting an acute threat to the group image. Furthermore, punitive responses against transgressing leaders driven by group image threat are more pronounced among highly identified group members. The current research shows when and why the oft-observed advantage of leader group prototypicality might be reversed.

Keywords

group image threat group prototypicality leader transgression

Being representative of a group is one of the fundamental qualities that determine group members’ support for their leaders (Hogg, 2001; Hogg & van Knippenberg, 2003; D. van Knippenberg, 2011). By embodying the group’s norms and identity, representative—or group prototypical—leaders are seen as more effective and are endorsed more strongly than leaders who are less so (Hogg et al., 1998; Platow & van Knippenberg, 2001). Evidence also indicates that group prototypical leaders receive support even in the wake of events that would typically lead to a loss of endorsement, for example, when they fail to achieve successful group outcomes (Giessner & van Knippenberg, 2008).

What happens if leaders fail to live up to their groups’ expectations by violating accepted norms, rules, or laws? When leaders engage in actions whose appropriateness is questionable as judged by norms, policies, or professional codes of conduct—that is, when they commit transgressions—group members typically respond negatively and punitively by displaying outrage or condemning their leaders (M. Brown & Treviño, 2006; Joshi et al., 2007). However, not all leaders are evaluated equally punitively for committing wrongful acts. Research on leader transgressions suggests that group members are more forgiving of a transgressing leader when they have positive regard for him/her, such as when they trust him/her and view him/her as being capable of ensuring the group’s success (Bies et al., 2016; Epitropaki et al., 2020; Shapiro et al., 2011). Considering the body of evidence showing that positive perceptions of a leader are more likely when the leader is group prototypical than when the leader is less or non-group-prototypical (for reviews, see Hogg, 2001; Hogg & van Knippenberg, 2003; D. van Knippenberg, 2011), when leaders transgress, their group prototypicality might help them fend off adverse reactions from group members.

In the present research, I suggest that leader group prototypicality might not protect leaders from losing followers’ support in the wake of a transgression. Instead, I propose that it can invite more punitive reactions from the group, especially when leaders commit transgressions of which outsiders are aware. Leaders in general stand as a lens through which outsiders perceive the group, and their behaviors often reflect upon the entire group and shape the group’s image in the eyes of outsiders (Haslam et al., 2011; Hogg, 2001; D. van Knippenberg, 2011). Thus, leaders’ wrongful acts, when seen or known by outsiders, can elicit image threat—the concern that the leader’s wrongdoings can damage the image that outsiders hold of the group (Abrams et al., 2013, 2014). I predict that group prototypical leaders’ transgressions will pose a more acute image threat than those of less or non-group-prototypical leaders, especially when their transgressions are known to outsiders. As a result, when committing transgressions that are known to outsiders, group prototypical leaders will be evaluated more punitively (i.e., receive less support) for having caused damage to the group image than less group prototypical leaders. Put differently, I expect that the oft-observed advantage for leaders that are group prototypical will be reversed when they have committed publicly known wrongdoings.

Leader Transgression and Group Image Threat

A central idea behind the social identity perspective is that people desire a positive social identity: individuals want to feel good about their membership in a particular group (Branscombe & Miron, 2004; Ellemers et al., 2002; Hogg & Abrams, 1988; Tajfel & Turner, 1979). Crucial to achieving and sustaining a positive social identity is to secure validation of the group’s positive characteristics from outside by ensuring that outsiders perceive the ingroup favorably (Ellemers et al., 2002; Luhtanen & Crocker, 1992). Groups indeed seek external validation of their positive characteristics—such as status, ranking, and reputation, all of which reflect and shape the groups’ image in the eyes of outsiders—and are keen on avoiding damage to the groups’ image (Ellemers et al., 2002; Elsbach & Kramer, 1996; Rabinovich & Morton, 2015).

Group leaders play a central role in forging the group’s external image: characteristics or behaviors of a leader often reflect on the group as a whole and influence the image outsiders have of the group (Bass & Bass, 2008; D. van Knippenberg, 2011). Thus, when leaders display inappropriate or questionable behaviors, the overall image of the groups that these leaders oversee can be negatively affected. For instance, in the corporate world, firms with directors and executives charged with having committed misconduct are evaluated less favorably by outsiders, such as security analysts (Cowen & Marcel, 2011). In politics, too, the external image of a country is compromised when the leader of the country commits inappropriate or potentially questionable deeds (Wike et al., 2017).

Importantly, research suggests that group members do not perceive all ingroup transgressions as equally damaging to the group’s image; instead, the wrongdoings of a member who is perceived to embody the group’s identity and its essential characteristics are more threatening to the group’s image (R. Brown et al., 2008; Iyer et al., 2007; Lickel et al., 2004, 2011). If the strength of the image threat following ingroup transgression depends on the transgressor’s representativeness of the group, then it follows that group members are more likely to experience a more acute image threat the more prototypical the transgressing leader is—a leader who better embodies the group identity and represents “who we are.”

Outsider Visibility of Leader Transgression

The preceding analysis suggests that when group leaders commit wrongful acts, group members might experience a stronger image threat when their leader is more, rather than less, group prototypical. I further propose that this effect of leader group prototypicality on group image threat following leader transgression will be more likely when transgressions are visible or known to outsiders than when they are not.

Outsider visibility of leader transgression refers to awareness of the leader’s wrongdoing by people who are not part of the group (Reicher & Levine, 1994). Outsiders might become aware of leaders’ wrongful acts by being direct victims of such wrongdoings (Abrams et al., 2013; Branscombe et al., 1999), or by accidentally witnessing the misbehaviors (O’Reilly & Aquino, 2011; Skarlicki & Kulik, 2004). On the other hand, outsiders might not be aware of leaders’ transgressions at all if they have not witnessed the wrongdoings or if they are not affected by such actions. Examples of leader transgressions low in outsider visibility include a leader of a marketing team insulting a subordinate during an individual meeting, or a politician making racist remarks during their party’s private caucus. By contrast, examples of leader transgressions high in outsider visibility are a leader of a marketing department insulting a subordinate in the office corridor thus rendering their behavior noticeable to people from different departments, or a politician making racist remarks in public.

Group image threat refers to group members’ perceptions of (potential) damage to how outsiders regard the ingroup (Lickel et al., 2005; Shuman et al., 2018). When transgressions are committed solely in an intragroup context, not noticed by outsiders, such transgressions might influence the value that group members accord to their ingroup and group membership, but not necessarily the group’s image in the eyes of outsiders (Bies et al., 2016; Epitropaki et al., 2020). That is, the effect of leader group prototypicality on group image threat will be weaker, if not nonexistent, when outsiders are not aware of leader’s transgressions. By contrast, leader transgressions that are known to outsiders naturally involve the possibility that outsiders’ perceptions of the group as a whole be negatively affected (Abrams et al., 2013, 2014; Branscombe et al., 1999). Thus, transgressions of a leader who better represents the group—a group prototypical leader—will present a stronger group image threat than those of a less group prototypical leader.

Hypothesis 1: A transgressing leader’s group prototypicality and the outsider visibility of his/her transgression will interact to predict group members’ experience of group image threat, such that the effect of leader group prototypicality on group image threat will be positive and stronger when the leader’s transgressions are visible to outsiders than when they are not.

A threat to their group’s image impedes members’ ability to maintain a positive identity; thus, group members will react negatively—punitively—to other ingroup members who pose a such threat. Groups often punish ingroup transgressors by terminating their association with them or by distancing themselves from them (Iyer et al., 2007; Knowles et al., 2014; Marques et al., 2001). However, there are many instances where such disconnection or distancing is not possible (e.g., when transgressors share group membership based on immutable characteristics such as race; Jackson et al., 1996). More importantly, in the case of leader transgressions, group members typically do not have the outright authority to dismiss their leader (Abrams et al., 2013, 2014).

Although group members might not be able to punish transgressing leaders by rejecting them, they can sanction them by withdrawing their support for them. Withdrawing support for a leader is a means for members who do not possess formal authority to sanction the leader, to express their discontent with them for failing to present the group in a positive light to outsiders (Bass & Bass, 2008; Brown & Trevino, 2006). Furthermore, withdrawing support for an existing leader can potentially spur a change in leadership, which can signal to outsiders that the group will not tolerate the transgressing leader and that he/she is not reflective of the group (Chemers, 2001; Knowles et al., 2014).

From this, I hypothesize that group members’ experience of image threat—which I predict to be especially likely when a group prototypical leader commits wrongful acts that are known to outsiders—will negatively impact their support for the leader. Stated differently, I expect that the interactive effect of a transgressing leader’s group prototypicality and outsider visibility of the leader’s transgression on group image threat will affect members’ support for the leader.

Hypothesis 2: Group image threat will mediate the interactive effect of a transgressing leader’s group prototypicality and outsider visibility of the leader’s transgression on members’ support for the leader, such that the negative indirect effect of leader group prototypicality on leader support via group image threat will be stronger when the leader’s transgressions are visible to outsiders than when they are not.

The Role of Member Ingroup Identification

I have argued that transgressions by group prototypical leaders can invite more unfavorable reactions from the group than transgressions by less group prototypical leaders because of group image threat, which is more likely when leaders commit transgressions known to outsiders. Importantly, concerns about maintaining a positive group image, and desire to prevent the tainting of the group’s image, should be stronger for members who consider their membership to be an important part of their self-concept, that is, those who identify strongly with their group (Ellemers et al., 2002; Lickel et al., 2005). In support of the role of ingroup identification, evidence shows that high identifiers are more likely to worry about the group’s image following ingroup transgressions and are more likely to take action to protect and restore it (Johns et al., 2005; Shuman et al., 2018).

Similarly, when leaders transgress, it is high identifiers who might be more likely to become concerned about the group’s image, and thus more likely to withdraw their support for the transgressing leaders. I propose that the experience of group image threat and subsequent withdrawal of leader support in response to a group prototypical leader’s transgressions known to outsiders will be more pronounced among highly identified group members than among their less identified peers.

Hypothesis 3: Member ingroup identification will moderate the interactive effect of a transgressing leader’s group prototypicality and the outsider visibility of the leader’s transgression on members’ experience of group image threat, such that the interactive effect will be stronger among high identifiers than among low identifiers.

Hypothesis 4: Member ingroup identification will moderate the conditional indirect effect of leader group prototypicality on member support for the transgressing leader via group image threat when outsider visibility of the leader’s transgression is high, such that the conditional indirect effect will be stronger among high identifiers than among low identifiers.

Overview of Studies

I report three studies (one preregistered) that tested the aforementioned predictions. Study 1 tested Hypotheses 1 and 2 using a hypothetical workplace scenario in which leader group prototypicality and outsider visibility of the leader’s transgressions were manipulated. Study 2 examined the role of ingroup identification in the experience of group image threat and leader support in the workplace context by having participants rate the group prototypicality of their current leader and recall their experience of the leader’s transgressions. Study 3 examined the predictions in a political context, using a sample of U.S. participants from the Democratic and Republican parties in a hypothetical scenario where a politician from their party had been charged with committing wrongful acts. In all studies, I report all manipulations and measures, explain how sample size was determined, and report all data exclusions.

Study 1

Method

Participants and design

Study 1 employed a 2 (leader group prototypicality: high vs. low) × 2 (outsider visibility of leader transgression: high vs. low) between-participant design. I planned a priori to collect data from approximately 200 participants (50 participants per condition). A total of 218 U.S. citizens (74 women; M_age = 34, SD_age = 8.34) were recruited via Amazon’s Mechanical Turk. Five participants failed to pass an attention check item (Oppenheimer et al., 2009),¹ resulting in a final sample of 213 participants. A sensitivity power analysis using G*Power (Faul et al., 2009) revealed that a sample size of 213 had 80% power to detect a minimum effect size of f = .19 for the primary test of my hypotheses in Study 1 (i.e., a two-way interaction within an analysis of variance [ANOVA]).

Procedure

Upon accessing the website containing the study materials, participants were asked to imagine themselves in a specific workplace scenario and were randomly assigned to one of the four experimental conditions. Participants then read descriptions of the department of which they were a member, the norms and values of the department, and the department’s leader. Leader group prototypicality was manipulated with this description. Next, participants received information indicating the leader’s transgression, which varied with regard to its visibility to people outside the department. Finally, participants rated the extent to which they would be concerned about the department’s image and their punitiveness toward (i.e., support for) the transgressing leader.

Manipulations and measures

Leader group prototypicality manipulation

All participants first read that they were a member of the claims department at an insurance company, and that the claims department had long prioritized financial efficiency over customer orientation. Next, the head of the claims department, named David Rogers, was introduced; I manipulated leader group prototypicality through this introduction. Specifically, in the high (low) leader group prototypicality condition, the head was described as someone who had shown that he favored the ideals of financial efficiency (the ideals of “customer first”); who believed that once the company gained cost leadership, customer issues will be resolved naturally (once the company gained support from clients, financial issues will be resolved naturally); and who had made remarks that financial efficiency should be the primary consideration (customers should always be the primary consideration).

Outsider visibility of leader transgression manipulation

Next, the outsider visibility of leader transgression was manipulated. Specifically, participants in the high (low) outsider visibility condition read:

Within the company [Within the claims department], David has been rumored to have taken credit for his subordinates’ work. In particular, he worked with some of his direct reports on an unusually complex case with one of the company’s largest corporate clients, the case that he successfully managed and that ended up as a great success. It has recently been uncovered that David was given an extraordinary bonus of $5,000 by the top management, but none of his subordinates who worked with him on the project were rewarded in any way. Although it is not clear what the terms of his bonus payment were exactly, this story has been circulated both within and outside the claims department [Because it is not clear what the terms of his bonus payment were exactly, this story has been known only to employees within the claims department].

Group image threat

After reading the description of the leader transgression, participants indicated the extent to which they would be concerned about the department’s image by answering four items adapted from Shuman et al. (2018). Sample items were: “I would be concerned that the claims department is not well-respected in this company” and “People in this company might view the entire claims department in a negative light” (α = .93; 1 = strongly disagree, 7 = strongly agree).

Support for the leader

Support for the leader was measured with two items (α = .94): “How likely would you be to support David Rogers as head of the claims department?” (1 = not at all likely, 7 = very likely) and “I would like to see David Rogers continue to serve as the head of the claims department” (1 = strongly disagree, 7 = strongly agree).

Manipulation checks

I assessed the effectiveness of the leader group prototypicality manipulation with the five-item Leader Group Prototypicality Scale (α = .96; B. van Knippenberg & van Knippenberg, 2005). Sample items are: “To what extent do you think David Rogers represents what is characteristic about the claims department?” (1 = not at all, 7 = to a great extent) and “How typical is David Rogers of the claims department?” (1 = not at all typical, 7 = very typical). Because judgments of leader group prototypicality might be affected by the knowledge of the leader’s transgression (Hogg, 2001; D. van Knippenberg, 2011), I measured leader group prototypicality right after the manipulation but before the leader transgression information. As a check of the outsider visibility manipulation, I used a single item: “A rumor that David Rogers took credit for subordinates’ work is widely known within the company” (1 = strongly agree, 7 = strongly disagree).

Additional measures

In addition to measuring perceived group image threat and support for the leader, in Study 1, I measured two variables that the social identity analysis of leadership would predict to be affected by leader group prototypicality: (a) general evaluations of the leader and (b) leader’s group identity affirmation—the extent to which group members perceive that their group norms, values, and identity are affirmed by their leader (Hogg, 2001; D. van Knippenberg, 2011). General leader evaluation was measured with five items: “nice,” “loyal,” “honest,” “generous,” and “competent” (α = .95; 1 = not at all, 7 = very much). Leader group identity affirmation was measured with three items adapted from Scheepers et al. (2002): “David Rogers helps the company correctly understand the true worth of the claims department,” “David Rogers exemplifies what the claims department is about,” and “David Rogers demonstrates what the claims department stands for” (α = .93; 1 = strongly disagree, 7 = strongly agree). Both of these variables were measured after participants had read the information about the leader’s transgression.

Results

Table 1 presents descriptive statistics and correlations among variables in Study 1.

Table 1.

Descriptive statistics and correlations among study variables: Study 1.

Variable	M	SD	1	2	3
1. Leader group prototypicality	0.00	1.00
2. Outsider visibility of leader transgression	−0.01	1.00	−.09
3. Group image threat	4.96	1.44	.05	.08
4. Leader support	4.25	1.88	−.08	−.10	−.22**

Note. N = 213. For leader group prototypicality, the high leader group prototypicality condition was coded 1, and the low leader group prototypicality condition was coded −1. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1.

p < .05.

Manipulation checks

Participants in the high leader group prototypicality condition perceived that the leader was significantly more group prototypical (M = 5.82, SD = 1.08) than participants in the low leader group prototypicality condition did (M = 3.82, SD = 1.93), t(165.00)² = 9.36, p < .001, f = .64. Regarding the outsider visibility manipulation check, participants in the high outsider visibility condition perceived that the leader’s wrongdoing was significantly more widely known within the company (M = 4.17, SD = 1.12) than participants in the low outsider visibility condition did (M = 3.07, SD = 1.22), t(211) = 6.83, p < .001, f = .47.

Hypothesis test

In Hypothesis 1, I predicted that a transgressing leader’s group prototypicality and the outsider visibility of the leader’s transgression would interact to predict members’ concern about the group’s image, such that the effect of leader group prototypicality would be positive and stronger when the leader’s transgressions were known to outsiders than when they were not. To test this hypothesis, I conducted a 2 (leader group prototypicality: high vs. low) × 2 (outsider visibility: high vs. low) ANOVA. In this analysis, neither of the main effects was significant, ps > .209, fs < .09. Yet, results revealed a significant Leader Group Prototypicality × Outsider Visibility interaction, F(1, 209) = 8.87, p = .003, f = .21 (see Figure 1, and Table 2, Model 1). Specifically, in the high outsider visibility condition, group image threat rated by participants in the high leader group prototypicality condition (M = 5.49, SD = 1.49) was significantly higher than group image threat rated by participants in the low leader group prototypicality condition (M = 4.73, SD = 1.30), F(1, 209) = 7.48, p = .007, f = .19. Furthermore, in the low outsider visibility condition, group image threat rated by participants in both leader group prototypicality conditions did not differ significantly (M_{high leader group prototypicality} = 4.67, SD_{high leader group prototypicality} = 1.56; M_{low leader group prototypicality} = 5.07, SD_{low leader group prototypicality} = 1.27), F(1, 209) = 2.16, p = .144, f = .10. Hypothesis 1 was thus supported.

Figure 1.

Interactive effect of leader group prototypicality and outgroup visibility of leader transgression on group image threat: Study 1.

Table 2.

Results of regression analyses: Study 1.

Variable	Group image threat	Leader support
	Model 1	Model 2	Model 3
Leader group prototypicality	.09	−.17	−.15
Outsider visibility of leader transgression	.12	−.20	−.18
Leader Group Prototypicality × Outsider Visibility	.29**	−.49***	−.43***
Group image threat			−.21*
R ²	.05*	.09***	.11***
ΔR²			.02*
Conditional direct effect of leader group prototypicality on leader support
Moderator	Bootstrap b	Bootstrap SE	95% CI
Low outsider visibility	0.28	0.17	[−0.06, 0.63]
High outsider visibility	−0.58	0.18	[−0.94, −0.23]
Conditional indirect effect of leader group prototypicality on leader support via group image threat
Moderator	Bootstrap b	Bootstrap SE	95% CI
Low outsider visibility	0.04	0.04	[−0.01, 0.13]
High outsider visibility	−0.08	0.04	[−0.17, −0.01]

Note. N = 213. Entries are unstandardized coefficients. For leader group prototypicality, the high leader group prototypicality condition was coded 1, and the low leader group prototypicality condition was coded −1. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1. Bootstrap sample size = 10,000.

p < .05. **p < .01. ***p < .001.

Hypothesis 2 predicted that group image threat would mediate the interactive effect of leader group prototypicality and outsider visibility of leader transgression on member support for the transgressing leader. To test this prediction, I examined conditional indirect effects of leader group prototypicality on leader support through group image threat when outsider visibility of leader transgression is high and when it is low, using leader group prototypicality as the predictor (1 = high leader group prototypicality, −1 = low leader group prototypicality), group image threat as the mediator, leader support as the dependent variable, and outsider visibility as the moderator (1 = high outsider visibility, −1 = low outsider visibility). Analyses revealed that, in the high outsider visibility condition, a 95% percentile bootstrap CI with 10,000 bootstrap samples for the indirect effect of leader group prototypicality on leader support via group image threat was negative and did not include zero, b = −0.08, SE = 0.04, 95% CI [−0.17, −0.01], whereas a 95% percentile bootstrap CI for the indirect effect in the low outsider visibility condition included zero, b = 0.04, SE = 0.04, 95% CI [−0.01, 0.13], see Table 2. These results support Hypothesis 2.

Supplementary analyses

I also explored how information about leader group prototypicality and leader transgressions affected general evaluations of the leader and the extent to which group members perceived him/her as serving to affirm their group identity. A two-way ANOVA using leader evaluation as the dependent variable revealed only a significant Leader Group Prototypicality × Outsider Visibility interaction, F(1, 209) = 4.04, p = .046, f = .14. Specifically, in the high outsider visibility condition, a group prototypical leader was rated significantly less positively (M = 3.89, SD = 1.90) than a non-group-prototypical leader (M = 4.58, SD = 1.59), F(1, 209) = 4.08, p = .045, f = .14. However, in the low outsider visibility condition, leader evaluations across group prototypical and non-group-prototypical leader conditions were not significantly different, F(1, 209) = 0.67, p = .415, f = .05. The same ANOVA using the leader’s group identity affirmation also revealed only a significant Leader Group Prototypicality × Outsider Visibility interaction, F(1, 209) = 5.43, p = .021, f = .16, such that the group prototypical leader was seen as significantly more affirming of the group identity (M = 4.92, SD = 1.44) than the non-group-prototypical leader (M = 4.10, SD = 1.81), F(1, 209) = 7.13, p = .008, f = .18, in the low outsider visibility condition; however, the difference was nonsignificant across the leader group prototypicality conditions in the high outsider visibility condition, F(1, 209) = 0.41, p = .525, f = .04.

Discussion

The results of Study 1 provided empirical support for the core argument of this paper: when leaders commit transgressions that are known to outsiders, their group prototypicality might invite more punitive reactions (lower support) from the group because group prototypical leaders’ public transgressions engender a more acute image threat. However, two features of Study 1 warrant caution in interpreting these results. First, the use of a hypothetical workplace scenario, with which participants could not identify strongly, likely constrained the operation of the social identity-based processes proposed in this paper. Second, the use of the departmental norm (i.e., an emphasis on financial efficiency over customer orientation) to operationalize leader group prototypicality might have affected how participants perceived the leader and the leader’s questionable behavior. Specifically, it might be that participants were reacting to a leader who prioritizes financial performance over customer service rather than, or along with, reacting to the leader’s group prototypicality. Therefore, I conducted a second study to address these limitations and to test the full theoretical model of the paper using a context with which participants could identify more or less strongly (i.e., their actual work group). In addition, in Study 2, instead of manipulating leader group prototypicality, I measured participants’ perceptions of their own leader’s group prototypicality.

Study 2

Method

Participants, design, and procedure

For Study 2, I aimed to collect data from 200 full-time working adults. A total of 191 working adults (86 women; M_age = 29, SD_age = 10.48) were recruited from Prolific Academic and participated in the study. Participants were employed in various industries, including finance, education, technology, and manufacturing. Four participants failed to pass an attention check item and thus were not included in the analyses reported next. Study 2 employed a one-factor (outsider visibility of leader transgression: high vs. low) between-participant design.

The study was introduced as being about people’s experience in the workplace. Participants were first asked to think of a group (or the smallest unit) to which they belonged at work, and to indicate the extent to which they identified with that group. Next, participants provided the initials of the leader of their group and rated how group prototypical he/she was. After indicating their leader’s group prototypicality, participants were asked to write about the leader’s transgressions, either known or unknown to outsiders depending on the condition to which they were assigned. Thus, in Study 2, leader group prototypicality and ingroup identification were measured, while outsider visibility of the leader’s transgression was manipulated.

Participants’ recall of transgressions by their leaders was crucial to Study 2 (for a similar approach, see Shapiro et al., 2011). However, 22 participants indicated that they could not think of any instances of their leader committing transgressions that fit the prompt; of these 22 participants, 16 were in the high outsider visibility condition, χ²(df = 1) = 4.13, p = .042. Consequently, these 22 participants were also excluded from the analyses, resulting in a final sample of 165 participants. A sensitivity power analysis using G*Power indicated that a sample of 165 participants could detect a minimum effect size of f = .26 with 80% power for the analysis of the main hypothesis in Study 2 (i.e., a three-way interaction within a regression framework).

Measures and manipulation

Ingroup identification

After thinking about their work group, participants indicated the extent to which they identified with it, using five items from Mael and Ashforth’s (1992) Ingroup Identification Scale. Sample items included: “When I talk about my group, I usually say ‘we’ rather than ‘they’” and “When someone praises my group, it feels like a personal compliment” (α = .70; 1 = strongly disagree, 7 = strongly agree).

Leader group prototypicality

Next, I asked participants to think about their group leader and to rate him/her on three items from the Leader Group Prototypicality Scale used in Study 1 (α = .90; B. van Knippenberg & van Knippenberg, 2005).

Outsider visibility of leader transgression manipulation

Participants were then asked to think and write briefly about their leader’s specific action(s) whose appropriateness was questionable in light of moral, professional, or legal standards. Participants in the high outsider visibility condition were asked to think about specific actions publicly committed by their leader and whose appropriateness was questionable, that is, transgressions committed in the presence of, toward, or being known to anyone outside their group. By contrast, participants in the low outsider visibility condition were asked to think and write about questionable actions of their leader known only to them or to members of their own group, but not to anyone outside the group.

To ensure that the transgressions participants recalled in the high and low outsider visibility conditions did not differ substantively, I had two independent coders, blind to the study’s conditions and hypotheses, categorize participants’ open-ended responses. To do so, I used the eight types of leader transgressions identified by Shapiro et al. (2011; i.e., negligence of duty, verbal/physical abuse, discrimination, favoritism, dishonesty, incompetence, interpersonal sabotage, and miscellaneous). The two coders showed substantial initial agreement (Cohen’s κ = .76), and disagreements were resolved through discussion. The 165 open-ended responses were sorted into the eight transgression types: negligence of duty (23.0%), verbal/physical abuse (12.7%), discrimination (0.6%), favoritism (7.9%), dishonesty (16.4%), incompetence (28.5%), interpersonal sabotage (0%), and miscellaneous (10.9%). There was no significant difference between the high and low outsider visibility conditions in the distribution of types of transgressions recalled by participants, χ²(df = 6) = 3.85, p = .699.

The two coders also rated the severity of transgressions on two items: “How much harm did this person cause?” (1 = not at all, 5 = very much) and “I think this is a very severe transgression for leaders to commit” (1 = strongly disagree, 5 = strongly agree; α_{Rater 1} = .97, α_{Rater 2} = .92, r = .88, p < .001). They did this prior to discussing the transgressions to resolve disagreements. I averaged their ratings to create a perceived severity variable, which was found not to differ between the high and low outsider visibility conditions (M_{high outsider visibility} = 2.27, SD_{high outsider visibility} = 0.95; M_{low outsider visibility} = 2.33, SD_{low outsider visibility} = 1.24), t(146.06) = 0.32, p = .751, f = .03.

Group image threat

The same four-item scale used in Study 1 was used as a measure of group image threat (α = .87).

Support for the leader

Support for the leader was measured with the same two items used in Study 1 (α = .93).

Additional measure

In Study 2, I also measured perceived leader effectiveness with three items: “[Leader initials] is an excellent leader,” “[Leader initials] is enthusing,” and “[Leader initials] awakens my feelings of commitment to the team” (α = .92; 1 = strongly disagree, 5 = strongly agree).

Results

Descriptive statistics and correlations among Study 2 variables are presented in Table 3.

Table 3.

Descriptive statistics and correlations among study variables: Study 2.

Variable	M	SD	1	2	3	4
1. Leader group prototypicality	4.66	1.61
2. Outsider visibility of leader transgression	−0.01	1.00	.03
3. Ingroup identification	5.44	0.81	.20**	.00
4. Group image threat	4.29	1.41	.07	.21**	.16*
5. Leader support	4.85	1.87	.49***	−.05	.04	−.24**

Note. N = 165. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1.

p < .05. **p < .01. ***p < .001.

Hypothesis 1 predicted that when leaders transgressed, leader group prototypicality would increase members’ experience of group image threat more so when the transgressions were known to outsiders than when they were not. To test this prediction, I regressed group image threat on leader group prototypicality (mean-centered), outsider visibility of leader transgression (contrast coded: 1 = high outsider visibility, −1 = low outsider visibility), and the interaction between these two variables. This analysis revealed a significant main effect of outsider visibility, b = 0.32, SE = 0.11, p = .003, 95% CI [0.11, 0.53]. Importantly, this main effect was qualified by a significant Leader Group Prototypicality × Outsider Visibility interaction, b = 0.19, SE = 0.07, p = .005, 95% CI [0.06, 0.32]. Simple slopes analysis further showed that the effect of leader group prototypicality on group image threat was positive and significant when participants recalled leader transgressions known to outsiders, b = 0.24, SE = 0.09, p = .010, 95% CI [0.06, 0.42], whereas the effect was not significant when participants recalled leader transgressions not known to outsiders, b = −0.14, SE = 0.10, 95% CI [−0.33, 0.05].

To test Hypothesis 2—which concerned the mediating effect of group image threat on the relationship between the Leader Group Prototypicality × Outsider Visibility interaction and leader support—I conducted conditional indirect effects analyses of leader group prototypicality on leader support through group image threat as a function of outsider visibility of leader transgression (1 = high outsider visibility, −1 = low outsider visibility). Results revealed that in the high outsider visibility condition, a 95% percentile bootstrap CI with 10,000 bootstrap samples for the indirect effect did not include zero, b = −0.09, SE = 0.06, 95% CI [−0.22, −0.01], whereas in the low outsider visibility conditions, a 95% percentile bootstrap CI for the indirect effect included zero, b = 0.05, SE = 0.04, 95% CI [−0.02, 0.15]. These results provided support for Hypotheses 1 and 2.

In Hypothesis 3, I predicted that ingroup identification would further moderate the Leader Prototypicality × Outsider Visibility interaction on group image threat, such that the interactive effect would be more pronounced among high identifiers than among low identifiers. To test this prediction, I regressed group image threat on leader group prototypicality (mean-centered), outsider visibility of leader transgression (contrast coded), ingroup identification (mean-centered), and the two- and three-way interaction terms associated with these variables. This analysis revealed significant main effects for outsider visibility (b = 0.26, SE = 0.10, p = .013, 95% CI [0.05, 0.47]), ingroup identification (b = 0.34, SE = 0.14, p = .015, 95% CI [0.07, 0.62]), the Leader Group Prototypicality × Outsider Visibility interaction (b = 0.20, SE = 0.07, p = .002, 95% CI [0.08, 0.33]), and the Leader Group Prototypicality × Ingroup Identification interaction (b = 0.23, SE = 0.08, p = .005, 95% CI [0.07, 0.39]). These significant effects were qualified by a significant Leader Group Prototypicality × Outsider Visibility × Ingroup Identification interaction (b = 0.17, SE = 0.08, p = .034, 95% CI [0.01, 0.33]; see Figure 2, and Table 4, Model 2). Consistent with Hypothesis 3, the strength of the interactive effect of Leader Group Prototypicality × Outsider Visibility on group image threat varied as a function of ingroup identification; for high identifiers (1 SD above the mean): b = 0.34, SE = 0.09, p < .001, 95% CI [0.16, 0.52]; for low identifiers (1 SD below the mean): b = 0.06, SE = 0.09, p = .487, 95% CI [−0.12, 0.25]. Simple slopes analyses further showed that the effect of leader group prototypicality on group image threat was positive and significant when participants recalled leader transgressions known to outsiders and when their identification with the group was high (1 SD above the mean), b = 0.56, SE = 0.14, p < .001, 95% CI [0.29, 0.83], whereas all other simple slopes were not significant at the conventional p = .05 level.

Figure 2.

Interactive effect of leader group prototypicality and outsider visibility of leader transgression on group image threat as a function of ingroup identification: Study 2.

Table 4.

Results of regression analyses: Study 2.

Variable	Group image threat		Leader support
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Leader group prototypicality	.04	.03	.58***	.60***	.60***	.61***
Outsider visibility of leader transgression	.32**	.26*	−.16	−.07	−.02	.03
Leader Group Prototypicality × Outsider Visibility	.19**	.20**	−.31***	−.26**	−.28***	−.24**
Ingroup identification		.34*			−.36*	−.29
Group Prototypicality × Identification		.23**			−.38***	−.34***
Outsider Visibility × Identification		−.08			−.54***	−.56***
Group Prototypicality × Outsider Visibility × Identification		.17*			−.42***	−.39***
Group image threat				−.30**		−.20*
R ²	.10**	.18***	.31***	.36***	.46***	.48***
ΔR²		.08***		.05**	.15***	.02*
Conditional direct effect of leader group prototypicality on leader support
Moderator		Bootstrap b	Bootstrap SE		95% CI
Low outsider visibility–low ingroup identification		0.80	0.14		[0.53, 1.08]
Low outsider visibility–high ingroup identification		0.88	0.13		[0.63, 1.14]
High outsider visibility–low ingroup identification		0.95	0.14		[0.67, 1.24]
High outsider visibility–high ingroup identification		−0.22	0.15		[−0.52, 0.09]
Conditional indirect effect of leader group prototypicality on leader support via group image threat
Moderator		Bootstrap b	Bootstrap SE		95% CI
Low outsider visibility–low ingroup identification		0.04	0.03		[−0.01, 0.12]
Low outsider visibility–high ingroup identification		0.03	0.04		[−0.04, 0.10]
High outsider visibility–low ingroup identification		0.02	0.03		[−0.05, 0.09]
High outsider visibility–high ingroup identification		−0.11	0.06		[−0.25, −0.01]

Note. N = 165. Entries are unstandardized coefficients. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1. Bootstrap sample size = 10,000.

p < .05. **p < .01. ***p < .001.

Lastly, I examined the role of ingroup identification in moderating the conditional indirect effect of leader group prototypicality on leader support via group image threat at different levels of transgression visibility (Hypothesis 4). Tests of conditional indirect effects using 10,000 bootstrap samples revealed that when ingroup identification was high (1 SD above the mean) and when participants recalled leader transgressions known to outsiders, a percentile bootstrap estimate of the indirect effect of leader group prototypicality on leader support via group image threat was −.11, SE = 0.06, 95% CI [−0.25, −0.01]. However, 95% percentile bootstrap CIs for the conditional indirect effects included zero when ingroup identification was high and outsider visibility was low (b = 0.03, SE = 0.04, 95% CI [−0.04, 0.10]), when ingroup identification was low and outsider visibility was high (b = 0.02, SE = 0.03, 95% CI [−0.05, 0.09]), and when ingroup identification was low and outsider visibility was low (b = 0.04, SE = 0.03, 95% CI [−0.01, 0.12]); see Table 4.

Supplementary analyses

I also examined whether perceived leader group prototypicality, recalling of leader transgressions, and ingroup identification had any independent or joint effects on general perceptions of leadership effectiveness, using the perceived leader effectiveness variable measured for exploratory purposes in Study 2. I regressed perceived leadership effectiveness on leader group prototypicality (mean-centered), outsider visibility (contrast coded), ingroup identification (mean-centered), and the two- and three-way interaction terms associated with these variables. This analysis revealed only a significant main effect of perceived leader group prototypicality, b = 0.77, SE = 0.05, p < .001, 95% CI [0.68, 0.86]. No other effects were significant, ps > .146.

Discussion

Study 2 constructively replicated the findings of Study 1: when facing leader transgressions known to outsiders, group members perceived a greater group image threat when the leader was group prototypical than when the leader was less group prototypical, which in turn led them to be less supportive of their transgressing leader. Study 2 further showed that such harsher reactions to group prototypical transgressing leaders were more likely among highly identified members than among less identified members, offering further credence to my argument that the reversal of leader group prototypicality advantage might occur because of group members’ desire to maintain a positive social identity. Yet, the sample size of Study 2 was small, and the manipulation of outsider visibility of leader transgressions was not as clean as it could be. I conducted Study 3 to address these limitations.

Study 3

Study 3 had two primary objectives in addition to addressing the limitations of Study 2. First, I sought to replicate the findings in a different context to ensure that the current theoretical arguments are generalizable across broader group situations and not limited to the context of small work groups examined in Studies 1 and 2. To this end, I examined my predictions in a political context, investigating how the transgressions of an elected politician from a particular political party might influence the experience of image threat of people affiliated with that party, and their subsequent support for the transgressing politician. Elected politicians are typically considered leaders of those whom they represent, serving important leadership functions such as articulating the visions and protecting the interests of the population (e.g., Reicher et al., 2005). Of particular importance among these leadership functions are expressing and upholding values, norms, and identity of the group they represent (Hogg et al., 2012; Krosnick et al., 2010). Thus, when elected political leaders commit transgressions, they might elicit a threat to the group image among their constituents—especially those who share the same political affiliation—as strong as, if not stronger than, the threat to the group image elicited by leaders of task groups when they commit transgressions.

The second objective of Study 3 was to demonstrate that the decrease in support for group prototypical transgressing leaders is driven by a threat to the image that outsiders hold of the group, not to the image that members have of their own group. Specifically, considering that leader behavior can influence how group members perceive and feel about their group (e.g., feeling proud or shameful of being part of the group, or emotionally attached to or detached from the group; Tyler & Blader, 2003), there is a possibility that a transgressing leader might influence members’ own evaluation of their group, independent of the group’s external image, which in turn would affect support for the leader rather than, or in addition to, perceived group image threat. Study 3 examined this possibility by measuring group members’ private regard for their group after learning about their leader’s transgressions.

Method

Participants and design

I recruited a total of 516 U.S. citizens from Prolific Academic (259 Republican, 257 Democrat; M_age = 37.36, SD_age = 12.93). I used the effect size observed in Study 2 for the test of Hypothesis 3 (an increase in R² as a result of including the Leader Group Prototypicality × Outsider Visibility × Ingroup Identification interaction; f = .17) to determine the minimum required sample size (N = 256). I then sought to recruit twice as many participants for the minimum sample size, as members of different political parties (i.e., the Democratic and Republican parties) might exhibit different responses.

Per preregistration (https://osf.io/2bfcj/?view_only=82961fb4af9e48219852176cc10ed91c), I excluded 11 participants who failed to pass an attention check item. This resulted in a final sample of 505 participants, who were randomly assigned to one of the four conditions created by a 2 (leader group prototypicality: high vs. low) × 2 (outsider visibility: high vs. low) between-participant design. A sample size of 505 had 80% power to detect a minimum effect size of f = .15 for the primary analysis in Study 3 (i.e., a three-way interaction within a regression framework).

Procedure

To ensure that I recruited participants from both the Democratic and Republican parties, I set up two online experiments: one for Democrats and the other for Republicans. The studies were advertised as concerning U.S. citizens’ political attitudes. Participants first indicated their citizenship and party affiliation. If participants were affiliated with neither the Democratic nor the Republican party, they were given an explanation as to why they were not eligible to participate.

After providing consent, participants indicated how strongly they identified with their political party and read a scenario about a fictional district representative from their party who was charged with attempted tax evasion. I manipulated leader group prototypicality and outsider visibility of the leader’s transgression through different versions of the scenario. After the scenario, participants reported image threat and indicated their support for the congressperson.

Manipulations and measures

Leader group prototypicality manipulation

Participants were asked to imagine that they lived in a district currently represented by the party they were affiliated with. They then read one of two descriptions of a fictional congressperson in which I manipulated leader group prototypicality. Specifically, participants in the high (low) leader group prototypicality condition read:

According to VoteMatch, a nonpartisan organization that rates politicians on their adherence to party objectives, your district’s representative is rated 93% [57%]. This means that he would be expected to vote along party lines 93% [57%] of the time. The times that he would not vote with the party, he would be expected to split evenly between voting in a more liberal or more conservative direction.

Outsider visibility of leader transgression manipulation

Participants in the high (low) outsider visibility condition read:

Recently, your district representative has been facing charges from the public [within the party] that he had filed tax forms that were materially misleading, incomplete, and in error. These allegations have been widely circulated in your district, although [These allegations have not been made public yet, as] it is not clear at this point if your district representative deliberately understated the amount of tax that he owed to the federal government by “faking ignorance.”

Group image threat

I used the same four items to measure group image threat used in Studies 1 and 2 (α = .90).

Support for the leader

The same two items used in Studies 1 and 2 were used in Study 3 to measure participants’ support for their leader (α = .86).

Ingroup identification

Before reading the scenario, participants indicated their identification with their political party using the four-item scale (α = .88) from Luhtanen and Crocker (1992). A sample item was: “In general, belonging to the Democratic [Republican] Party is an important part of my self-image” (1 = strongly disagree, 7 = strongly agree).

Manipulation checks

To assess the effectiveness of the group prototypicality manipulation, I used the same five-item scale (α = .93) used in Study 1. This was measured before the transgression information. As an outsider visibility manipulation check, I asked two questions (e.g., “People in this district are aware of the allegations against the representative”; α = .94; 1 = strongly disagree, 7 = strongly agree).

Additional measures

In Study 3, I also measured participants’ negative regard for their political party, to examine how group leaders’ transgressions might affect the evaluations and perceptions members have of their group. A sample item was: “If I were in this district, I would have regretted that I belong to the Democratic [Republican] Party” (α = .68; 1 = strongly disagree, 7 = strongly agree). In addition, participants rated the leader’s group prototypicality once again after reading about the transgression, to examine whether knowing about it caused them to change their evaluations. The same five-item scale was used (α = .95).

Results

Descriptive statistics and bivariate correlations are presented in Table 5.

Table 5.

Descriptive statistics and correlations among study variables: Study 3.

Variable	M	SD	1	2	3	4
1. Leader group prototypicality	−0.01	1.00
2. Outsider visibility of leader transgression	0.01	1.00	−.02
3. Ingroup identification	3.98	1.45	.02	−.02
4. Group image threat	4.51	1.31	.06	.06	.06
5. Leader support	2.98	1.54	−.01	−.01	.12**	−.26***

Note. N = 505. For leader group prototypicality, the high leader group prototypicality condition was coded 1, and the low leader group prototypicality condition was coded −1. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1.

p < .01. ***p < .001.

Manipulation checks

I first conducted a 2 (party affiliation: Democratic vs. Republican) × 2 (leader group prototypicality: high vs. low) × 2 (outsider visibility: high vs. low) ANOVA using the perceived leader group prototypicality measure administered before the transgression information was read by participants. This analysis revealed only a significant main effect of leader group prototypicality, F(1, 497) = 133.47, p < .001, f = .89. Specifically, participants in the high leader group prototypicality condition perceived the congressperson to be significantly more group prototypical (M = 5.24, SD = 0.87) than participants in the low leader group prototypicality condition (M = 4.12, SD = 1.28), t(449.88) = 11.55, p < .001. The same analysis of the outsider visibility manipulation check revealed only a significant main effect of outsider visibility, F(1, 497) = 590.83, p < .001, f = 1.09; participants in the high outsider visibility condition indicated that the congressperson’s charges were significantly more widely known within the district (M = 5.54, SD = 1.01) than participants in the low outsider visibility condition did (M = 2.59, SD = 1.64), t(415.86) = 24.34, p < .001. No other effects were significant in either analysis.

Hypothesis tests

Party affiliation had neither a main effect nor interacted with other variables to predict group image threat or leader support, and thus is not discussed further.

Hypothesis 1 predicted that a leader’s group prototypicality and outsider visibility of that leader’s transgressions would interact to predict members’ concern about group image. I tested this prediction by conducting a 2 (leader group prototypicality: high vs. low) × 2 (outsider visibility: high vs. low) ANOVA using group image threat as the dependent variable. This analysis revealed only a significant Leader Group Prototypicality × Outsider Visibility interaction, F(1, 501) = 4.03, p = .045, f = .09. Planned contrasts further revealed that, consistent with Hypothesis 1, in the high outsider visibility condition, participants in the high leader group prototypicality condition perceived significantly higher group image threat (M = 4.79, SD = 1.22) than participants in the low leader group prototypicality condition did (M = 4.40, SD = 1.37), F(1, 501) = 5.58, p = .019, f = .11. However, in the low outsider visibility condition, perceived group image threat scores by participants in the low (M = 4.46, SD = 1.36) and high leader group prototypicality (M = 4.38, SD = 1.26) conditions were not significantly different from each other, F(1, 501) = 0.24, p = .627, f = .00.

Hypothesis 2 predicted that group image threat would mediate the interactive effect of leader group prototypicality and outsider visibility on leader support. Tests of the conditional indirect effect of leader group prototypicality on leader support through group image threat at different levels of outsider visibility using 10,000 bootstrap samples revealed that, in the high outsider visibility condition, the 95% percentile bootstrap CI for the indirect effect of leader group prototypicality on leader support via group image threat did not include zero, b = −0.06, SE = 0.03, 95% CI [−0.12, −0.01], whereas in the low outsider visibility conditions, the 95% percentile bootstrap CI for the indirect effect did include zero, b = 0.01, SE = 0.03, 95% CI [−0.04, 0.07].

To test Hypothesis 3, which predicted that the interactive effect of leader group prototypicality and outsider visibility on group image threat would be more pronounced among high identifiers than among low identifiers, I regressed perceived group image threat on leader group prototypicality (contrast coded: 1 = high leader group prototypicality, −1 = low leader group prototypicality), outsider visibility of leader transgression (contrast coded: 1 = high outsider visibility, −1 = low outsider visibility), ingroup identification (mean-centered), and all two- and three-way interactions among these variables. This analysis revealed a significant Leader Group Prototypicality × Outsider Visibility interaction, b = 0.12, SE = 0.06, p = .041, 95% CI [0.01, 0.23]. In addition, the Leader Group Prototypicality × Ingroup Identification interaction was also significant, b = 0.11, SE = 0.04, p = .009, 95% CI [0.03, 0.18]. However, these significant two-way interactions were qualified by a significant Leader Group Prototypicality × Outsider Visibility × Ingroup Identification interaction, b = 0.08, SE = 0.04, p = .039, 95% CI [0.01, 0.16]. Consistent with the prediction, the Leader Group Prototypicality × Outsider Visibility interaction was significant when participants’ identification with their party was high (1 SD above the mean), b = 0.24, F(1, 497) = 8.52, p = .004, whereas the same interaction was not significant when ingroup identification was low (1 SD below the mean), b = −0.01, F(1, 497) = 0.01, p = .990. Simple slopes analyses further showed that the effect of leader group prototypicality on perceived group image threat was positive and significant when transgressions were known to outsiders and participants’ identification was high, b = 0.47, SE = 0.11, p < .001, 95% CI [0.24, 0.69], whereas no other simple effects were significant (all ps > .502). Thus, Hypothesis 3 was supported (see Figure 3, and Table 6, Model 2).

Figure 3.

Interactive effect of leader group prototypicality and outsider visibility of leader transgression on group image threat as a function of ingroup identification: Study 3.

Table 6.

Results of regression analyses: Study 3.

Variable	Group image threat		Leader support
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Leader group prototypicality	.08	.08	−.03	−.01	−.04	−.01
Outsider visibility of leader transgression	.09	.08	−.02	.01	−.02	.01
Leader Group Prototypicality × Outsider Visibility	.12*	.12*	−.13	−.09	−.13	−.09
Ingroup identification		.04			.13**	.14**
Group Prototypicality × Identification		.11**			.00	.03
Outsider Visibility × Identification		.02			.01	.01
Group Prototypicality × Outsider Visibility × Identification		.08*			−.09	−.06
Group image threat				−.30***		−.31***
R ²	.02*	.04*	.01	.07***	.03*	.09***
ΔR²		.02**		.06***	.02*	.06***
Conditional direct effect of leader group prototypicality on leader support
Moderator		Bootstrap b	Bootstrap SE		95% CI
Low outsider visibility–low ingroup identification		−0.05	0.14		[−0.32, 0.22]
Low outsider visibility–high ingroup identification		0.21	0.13		[−0.05, 0.47]
High outsider visibility–low ingroup identification		−0.07	0.13		[−0.32, 0.18]
High outsider visibility–high ingroup identification		−0.15	0.13		[−0.41, 0.11]
Conditional indirect effect of leader group prototypicality on leader support via group image threat
Moderator		Bootstrap b	Bootstrap SE		95% CI
Low outsider visibility–low ingroup identification		0.02	0.04		[−0.05, 0.10]
Low outsider visibility–high ingroup identification		0.00	0.04		[−0.08, 0.09]
High outsider visibility–low ingroup identification		0.02	0.03		[−0.04, 0.09]
High outsider visibility–high ingroup identification		−0.15	0.05		[−0.26, −0.05]

Note. N = 505. Entries are unstandardized coefficients. For leader group prototypicality, the high leader group prototypicality condition was coded 1, and the low leader group prototypicality condition was coded −1. For outsider visibility of leader transgression, the high outsider visibility condition was coded 1, and the low outsider visibility was coded −1. Bootstrap sample size = 10,000.

p < .05. **p < .01. ***p < .001.

Hypothesis 4, which predicted that the conditional indirect effect of leader group prototypicality on leader support via group image threat as a function of outsider visibility would be more pronounced when participants’ group identification was high than when it was low, was examined using 10,000 bootstrap samples. The 95% percentile bootstrap CI for the indirect effect of leader group prototypicality on leader support via group image threat for high identifiers in the high outsider visibility condition did not include zero, b = −0.15, SE = 0.05, 95% CI [−0.26, −0.05]. However, 95% percentile bootstrap CIs for all of the other indirect effects did include zero: for high identifiers in the low outsider visibility condition (b = 0.00, SE = 0.04, 95% CI [−0.08, 0.09]); for low identifiers in the high outsider visibility condition (b = 0.02, SE = 0.03, 95% CI [−0.04, 0.09]); and for low identifiers in the low outsider visibility condition (b = 0.02, SE = 0.04, 95% CI [−0.05, 0.10]); see Table 6.

Supplementary analyses

Finally, I explored the implications of the transgressing leader’s group prototypicality, outsider visibility of leader transgression, and ingroup identification for participants’ own perception of the group and potential changes in perceived leader group prototypicality. First, I examined participants’ negative regard for the group after learning about the transgression, by regressing the measured negative group regard on the manipulated leader group prototypicality, outsider visibility, ingroup identification, and all the two- and three-way interactions among these variables. This analysis showed only a significant main effect of ingroup identification, b = −0.20, SE = 0.04, p < .001, 95% CI [−0.27, −0.13], indicating that high identifiers felt less negatively about their group even after learning about their leader’s transgressions, regardless of the leader’s group prototypicality, and whether or not the leader’s transgressions were known to outsiders. The measured negative private ingroup regard was positively associated with group image threat (r = .17, p < .001), and negatively associated with leader support (r = −.16, p < .001). None of the results reported before changed when this negative private ingroup regard was entered as a covariate.

Next, I conducted a repeated measures ANOVA using leader group prototypicality ratings measured before and after reading about the transgression as a within-participant factor, and leader group prototypicality and outsider visibility as between-participant factors. This analysis revealed a significant main effect of measurement timing, F(1, 501) = 503.33, p < .001, f = 1.00, such that participants perceived the congressperson to be significantly less group prototypical (M = 3.24, SD = 1.43) after learning about the transgression than beforehand (M = 4.67, SD = 1.23). Importantly, a Measurement Timing × Leader Group Prototypicality interaction was also significant, F(1, 501) = 28.57, p < .001, f = .24. Specifically, the decrease in leader group prototypicality rating was significantly greater for participants in the high leader group prototypicality condition than for participants in the low leader group prototypicality condition, suggesting that transgressions had a greater negative impact for leaders who were initially perceived as group prototypical.

Discussion

Study 3 replicated the findings of Studies 1 and 2, and demonstrated that the withdrawal of leader support, driven by group image threat, can occur not only in small work groups but also in a broader social context, such as a political one. I also found that high identifiers felt less negatively about their group when their leaders were charged with committing wrongful acts, regardless of how group prototypical the transgressing leader was, or whether the transgressions were known to outsiders.

General Discussion

Ensuring endorsement from their group is crucial for a leader’s success (Bass & Bass, 2008; Chemers, 2001; Haslam et al., 2011). Research on the social identity of leadership has amassed evidence indicating that being group prototypical helps leaders garner support from followers (Hogg, 2001; Hogg & van Knippenberg, 2003; D. van Knippenberg, 2011). The present research investigated the possibility that being group prototypical might cause group leaders to lose support from the group more than being less or non-group-prototypical when they do wrong. In three studies, I found that after committing publicly known transgressions, group prototypical leaders were supported less than leaders who were less group prototypical, especially by highly identified members. These punitive reactions are driven by a more acute threat to the group image experienced by group members due to group prototypical leaders’ transgressions.

Implications

The theory and findings of this paper make several contributions. First, the current findings contribute to the social identity analysis of leadership by identifying a specific condition under which the widely documented advantages of being group prototypical might be reversed: when leaders commit transgressions that are visible to outsiders. Researchers have recently begun to identify contingencies associated with the beneficial impact of leader group prototypicality. Specifically, group prototypical leaders do not necessarily enjoy advantages over less group prototypical leaders when group members experience high levels of uncertainty (Rast et al., 2012, 2015), when members have a strong desire to avoid uncertainty (Pierro et al., 2005), or when members have to work cooperatively with outsiders (Teixeira et al., 2011). The present findings extend this line of research showing that when group prototypical leaders violate accepted rules, norms, or laws, and such violations are known to outsiders, group members become concerned about the image outsiders might hold of them, and consequently withdraw their support for those leaders to a greater degree than they would for less group prototypical leaders.

Second, this paper contributes to research on ingroup transgression and the subsequent experience of group-based emotions. Research on collective shame has highlighted that group members experience concern about the group image in the eyes of outsiders when other ingroup members commit wrongful acts, which in turn elicits behaviors aimed to restore or maintain a positive group image (Branscombe et al., 1999; Lickel et al., 2004, 2005; Shuman et al., 2018). Although existing research has documented various strategies that group members employ to manage the group image (e.g., Knowles et al., 2014; Lickel et al., 2011), relatively less is known about when group members are more or less likely to experience group image threat. I distinguished ingroup transgressions that occur privately within the group from those that are visible to outsiders, and showed that ingroup transgressions are likely to elicit image threat when they are visible to outsiders. Furthermore, the present research highlights that, although any publicly committed transgressions can elicit worries about the group’s external image, transgressions committed by more representative—group prototypical—leaders are especially concerning.

The present research thus tells a cautionary tale about the use of the “one of us” strategy as a means to pursue leadership success. Emphasizing the leader’s similarity to other group members and embodiment of the ingroup’s identity is a popular leadership and identity-constructing rhetoric observed in a wide range of contexts (Reicher & Haslam, 2017), and evidence attests to its effectiveness (Hogg & van Knippenberg, 2003; D. van Knippenberg, 2011). Yet, the present results suggest that group prototypical leaders might be hit harder when being accused of or actually having committed questionable deeds, because their transgression can have greater negative implications for the image outsiders have of the group. Leaders who have gained support with the help of, or by promoting, their group prototypicality will therefore have to be mindful that their leadership might be more vulnerable to the effects of accusations of wrongful acts than would be the case for less group prototypical leaders.

Strengths, Limitations, and Future Research Directions

Consistent findings across three studies using different operationalizations of core constructs and different group contexts lend credence to the validity of the present theoretical framework. In addition, showing that group identification further moderates the interactive effect of leader group prototypicality and outsider visibility of leader transgression corroborates the argument that the reversal of the leader group prototypicality advantage is—like the emergence of this advantage—an outcome of identity-based processes.

Yet, there are several limitations of these studies, which also suggest promising avenues for future research. First, although the present research distinguished leader transgressions along the dimension of visibility to outsiders, there are other theoretically and practically important differentiations that can be made with respect to leaders’ wrongful acts. For example, some transgressions are clearly self-focused, whereas others are more group-oriented (Abrams et al., 2013). Transgressions might also vary with respect to whether they are made against laws (and are thus subject to legal prosecution) or against more general ethical standards. Transgressions also differ with respect to who the victims are—ingroup versus outgroup—or whether or not there are any identifiable victims. Future research will likely benefit from examining the implications of different types of leader transgressions, and how leader group prototypicality might affect reactions to them.

The group contexts examined in the present research—task groups and political parties—were limited in the sense that membership in such groups is inherently more changeable than membership in certain social groups is, such as racial or nationality-based groups. Investigations of social groups with different levels of membership changeability or group boundary permeability (Jackson et al., 1996; Tajfel & Turner, 1979) might offer additional insight into how group members react to ingroup transgressions (e.g., leaving the group feeling ashamed because of an ingroup transgression vs. protesting against that transgression).

It is also worth noting that high identifiers felt less negatively about their group after knowing their leader’s transgressions, regardless of the leader’s group prototypicality and outsider visibility of the transgressions (Study 3). This finding highlights the distinction between a group’s public image and members’ private perceptions of their group (Luhtanen & Crocker, 1992). More importantly, it suggests that high identifiers might be more tolerant of transgressions committed by ingroup members and yet still respond more strongly to events that might damage their external image (Allpress et al., 2014; Knowles et al., 2014; Shuman et al., 2018). Additional research is needed to clarify the role of group identification in how members perceive and react to ingroup transgressions.

Worthy of an additional note is the main effect of leader group prototypicality and its conditional effect when leader transgressions are not known to outsiders. In all three studies reported in the present paper, neither the main effects of leader group prototypicality on image threat nor its simple effects in the low outsider visibility conditions were significant. One possible interpretation of these nonsignificant findings might be that, in light of the definition of group image threat (concern about the group image being undermined in the eyes of outsiders), when leaders’ transgressions are not known to outsiders, their group prototypicality might matter less, or might not even matter at all, to the group’s experience of image threat. Yet, the finding from the supplementary analysis in Study 3—that groups react more strongly to transgressing group prototypical leaders than to their non-group-prototypical counterparts (as per group prototypicality ratings before and after knowing about the transgressions)—suggests that group prototypical leaders’ transgressions, even when they are not known to outsiders, might elicit a different kind of concern. While the present studies cannot offer any conclusive interpretations, further examinations of potential concerns experienced by groups as a result of knowing about their leaders’ transgressions, and the groups’ subsequent reactions in relation to their leaders’ group prototypicality, will be a fruitful avenue for future research.

Finally, future research will also likely benefit from exploring outsiders’ perspectives on ingroup transgressions. For one, even though group members are more punitive toward an ingroup prototypical transgressor, outsiders might not differentiate transgressors according to their group prototypicality. Furthermore, additional work is needed to understand how outsiders react when groups punish ingroup transgressors. If outsider perceptions of ingroup image are not affected by whether the transgressor is punished, and instead are affected only by the transgression itself, censuring a transgressor does not constitute an effective means of dealing with ingroup transgression, at least as an image management tactic. Incorporating an outsider perspective should provide additional insight into how ingroup transgression shapes intra- and intergroup identity dynamics.

Footnotes

Acknowledgements

The author would like to thank Juhyang Baek and Yejin Yoo for coding the data in Study 2.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partially supported by a Korea University Business School research grant.

ORCID iD

Jin Wook Chang

Notes

References

Abrams

Randsley

Moura

Travaglino

G. A.

(2013). A double standard when group members behave badly: Transgression credit to ingroup leaders. Journal of Personality and Social Psychology, 105, 799–815. https://doi.org/10.1037/a0033600

Abrams

Travaglino

G. A.

Randsley

Moura

May

P. J.

(2014). A step too far? Leader racism inhibits transgression credit. European Journal of Social Psychology, 44, 730–735. https://doi.org/10.1002/ejsp.2063

Allpress

J. A.

Brown

Giner-Sorolla

Deonna

J. A.

Teroni

(2014). Two faces of group-based shame: Moral shame and image shame differentially predict positive and negative orientations to ingroup wrongdoing. Personality and Social Psychology Bulletin, 40, 1270–1284. https://doi.org/10.1177/0146167214540724

Bass

B. M.

Bass

(2008). Leadership and performance beyond expectations. Free Press.

Bies

R. J.

Tripp

T. M.

Shapiro

D. L.

(2016). Abusive leaders or master motivators? “Abusive” is in the eye of the beholder. In Ashkanasy

N. M.

Bennett

R. J.

Martinko

M. J.

(Eds.), Understanding the high-performance workplace: The line between motivation and abuse (pp. 252–276). Routledge.

Branscombe

N. R.

Ellemers

Spears

Doosje

(1999). The context and content of social identity threat. In Ellemers

Spears

Doosje

(Eds.), Social identity: Context, commitment, content (pp. 35–58). Blackwell.

Branscombe

N. R.

Miron

(2004). Interpreting the ingroup negative actions towards another group: Emotional reactions to appraised harm. In Tiedens

L. Z.

Leach

C. W.

(Eds.), The social life of emotions (pp. 314–345). Cambridge University Press.

Brown

Treviño

L. K.

(2006). Ethical leadership: A review and future directions. The Leadership Quarterly, 17, 595–616. https://doi.org/10.1016/j.leaqua.2006.10.004

Brown

González

Zagefka

Manzi

Čehajić

(2008). Nuestra culpa: Collective guilt and shame as predictors of reparation for historical wrongdoing. Journal of Personality and Social Psychology, 94, 75–90. https://doi.org/10.1037/0022-3514.94.1.75

10.

Chemers

M. M.

(2001). Leadership effectiveness: An integrative review. In Hogg

M. A.

Tindale

R. S.

(Eds.), Blackwell handbook of social psychology: Group processes (pp. 376–399). Blackwell.

11.

Cowen

A. P.

Marcel

J. J

(2011). Damaged goods: Board decisions to dismiss reputationally compromised directors. Academy of Management Journal, 54, 509–527. https://doi.org/10.5465/amj.2011.61967992

12.

Ellemers

Spears

Doosje

(2002). Self and social identity. Annual Review of Psychology, 53, 161–186. https://doi.org/10.1146/annurev.psych.53.100901.135228

13.

Elsbach

K. D.

Kramer

(1996). Members’ responses to organizational identity threats: Encountering and countering the Business Week ratings. Administrative Science Quarterly, 41, 442–476. https://doi.org/10.2307/2393938

14.

Epitropaki

Radulovic

A. B.

Ete

Thomas

Martin

(2020). Leader–follower transgressions, relationship repair strategies and outcomes: A state-of-the-science review and a way forward. The Leadership Quarterly, 31, 1–19. https://doi.org/10.1016/j.leaqua.2019.101376

15.

Faul

Erdfelder

Buchner

Lang

A.-G.

(2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149–1160. https://doi.org/10.3758/BRM.41.4.1149

16.

Giessner

S. R.

van Knippenberg

(2008). “License to fail”: Goal definition, leader group prototypicality, and perceptions of leadership effectiveness after leader failure. Organizational Behavior and Human Decision Processes, 105, 14–35. https://doi.org/10.1016/j.obhdp.2007.04.002

17.

Haslam

S. A.

Reicher

S. D.

Platow

(2011). The new psychology of leadership. Psychology Press.

18.

Hogg

M. A.

(2001). A social identity theory of leadership. Personality and Social Psychology Review, 5, 184–200. https://doi.org/10.1207/S15327957PSPR0503_1

19.

Hogg

M. A.

Abrams

(1988). Social identifications: A social psychology of intergroup relations and group processes. Routledge.

20.

Hogg

M. A.

Hains

S. C.

Mason

(1998). Identification and leadership in small groups: Salience, frame of reference, and leader stereotypicality effects on leader evaluations. Journal of Personality and Social Psychology, 75, 1248–1263. https://doi.org/10.1037/0022-3514.75.5.1248

21.

Hogg

M. A.

van Knippenberg

(2003). Social identity and leadership processes in groups. Advances in Experimental Social Psychology, 35, 1–52. https://doi.org/10.1016/S0065-2601(03)01001-3

22.

Hogg

M. A.

van Knippenberg

Rast

D. E.

III . (2012). The social identity theory of leadership: Theoretical origins, research findings, and conceptual developments. European Review of Social Psychology, 23, 258–304. https://doi.org/10.1080/10463283.2012.741134

23.

Iyer

Schmader

Lickel

(2007). Why individuals protest the perceived transgressions of their country: The role of anger, shame and guilt. Personality and Social Psychology Bulletin, 33, 572–587. https://doi.org/10.1177/0146167206297402

24.

Jackson

L. A.

Sullivan

L. A.

Harnish

Hodge

C. N.

(1996). Achieving positive social identity: Social mobility, social creativity, and permeability of group boundaries. Journal of Personality and Social Psychology, 70, 241–254. https://doi.org/10.1037/0022-3514.70.2.241

25.

Johns

Schmader

Lickel

(2005). Ashamed to be an American? The role of identification in predicting vicarious shame for anti-Arab prejudice after 9–11. Self and Identity, 4, 331–348. https://doi.org/10.1080/15298860500145822

26.

Joshi

Anand

Henderson

(2007). The role of organizational practices and routines in facilitating normalized corruption. In Langan-Fox

Cooper

C. L.

Klimoski

R. J.

(Eds.), Research companion to the dysfunctional workplace: Management challenges and symptoms (pp. 235–251). Edward Elgar Publishing.

27.

Knowles

E. D.

Lowery

B. S.

Chow

R. M.

Unzueta

M. M.

(2014). Deny, distance, or dismantle? How White Americans manage a privileged identity. Perspectives on Psychological Science, 9, 594–609. https://doi.org/10.1177/1745691614554658

28.

Krosnick

J. A.

Visser

P. S.

Harder

(2010). The psychological underpinnings of political behavior. In Fiske

S. T.

Gilbert

D. T.

Lindzey

(Eds.), Handbook of social psychology (5th ed., Vol. 2, pp. 1288–1342). Wiley.

29.

Lickel

Schmader

Barquissau

(2004). The evocation of moral emotions in intergroup contexts: The distinction between collective guilt and collective shame. In Branscombe

N. R.

Doosje

(Eds.), Collective guilt: International perspectives (pp. 35–55). Cambridge University Press.

30.

Lickel

Schmader

Curtis

Scarnier

Ames

D. R.

(2005). Vicarious shame and guilt. Group Processes & Intergroup Relations, 8, 145–157. https://doi.org/10.1177/1368430205051064

31.

Lickel

Steele

R. R.

Schmader

(2011). Group-based shame and guilt: Emerging directions in research. Social and Personality Psychology Compass, 5, 153–163. https://doi.org/10.1111/j.1751-9004.2010.00340.x

32.

Luhtanen

Crocker

(1992). A collective self-esteem scale: Self-evaluation of one’s social identity. Personality and Social Psychology Bulletin, 18, 317–334. https://doi.org/10.1177/0146167292183006

33.

Mael

Ashforth

(1992). Alumni and their alma maters: A partial test of the reformulated model of organizational identification. Journal of Organizational Behavior, 13, 103–123. https://doi.org/10.1002/job.4030130202

34.

Marques

J. M.

Abrams

Serôdio

(2001). Being better by being right: Subjective group dynamics and derogation of in-group deviants when generic norms are undermined. Journal of Personality and Social Psychology, 81, 436–447. https://doi.org/10.1037/0022-3514.81.3.436

35.

Oppenheimer

D. M.

Meyvis

Davidenko

(2009). Instructional manipulation checks: Detecting satisficing to increase statistical power. Journal of Experimental Social Psychology, 45, 867–872. https://doi.org/10.1016/j.jesp.2009.03.009

36.

O’Reilly

Aquino

(2011). A model of third parties’ morally motivated responses to mistreatment in organizations. Academy of Management Review, 36, 526–543. https://doi.org/10.5465/amr.2009.0311

37.

Pierro

Cicero

Bonaiuto

van Knippenberg

Kruglanski

A. W.

(2005). Leader group prototypicality and leadership effectiveness: The moderating role of need for cognitive closure. The Leadership Quarterly, 16, 503–516. https://doi.org/10.1016/j.leaqua.2005.06.002

38.

Platow

M. J.

van Knippenberg

(2001). A social identity analysis of leadership endorsement: The effects of leader ingroup prototypicality and distributive intergroup fairness. Personality and Social Psychology Bulletin, 27, 1508–1519. https://doi.org/10.1177/01461672012711011

39.

Rabinovich

Morton

T. A.

(2015). Things we (don’t) want to hear: Exploring responses to group-based feedback. European Review of Social Psychology, 26, 126–161. https://doi.org/10.1080/10463283.2015.1115214

40.

Rast

D. E.

III Gaffney

A. M.

Hogg

M. A.

Crisp

R. J.

(2012). Leadership under uncertainty: When leaders who are non-prototypical group members can gain support. Journal of Experimental Social Psychology, 48, 646–653. https://doi.org/10.1016/j.jesp.2011.12.013

41.

Rast

D. E.

III Hogg

M. A.

Tomory

J. J.

(2015). Prototypical leaders do not always get our support: Impact of self-uncertainty and need for cognition. Self and Identity, 14, 135–146. https://doi.org/10.1080/15298868.2014.964755

42.

Reicher

Haslam

S. A.

(2017). The politics of hope: Donald Trump as an entrepreneur of identity. In Fitzduff

(Ed.), Why irrational politics appeals: Understanding the allure of Trump (pp. 22–35). Praege.

43.

Reicher

Haslam

S. A.

Hopkins

(2005). Social identity and the dynamics of leadership: Leaders and followers as collaborative agents in the transformation of social reality. The Leadership Quarterly, 16, 547–568. https://doi.org/10.1016/j.leaqua.2005.06.007

44.

Reicher

Levine

(1994). Deindividuation, power relations between groups and the expression of social identity: The effects of visibility to the outgroup. British Journal of Social Psychology, 33, 145–163. https://doi.org/10.1111/j.2044-8309.1994.tb01015.x

45.

Scheepers

Spears

Doosje

Manstead

A. S. R.

(2002). Integrating identity and instrumental approaches to intergroup differentiation: Different contexts, different motives. Personality and Social Psychology Bulletin, 28, 1455–1467. https://doi.org/10.1177/014616702237574

46.

Shapiro

D. L.

Boss

A. D.

Salas

Tangirala

von Glinow

M. A.

(2011). When are transgressing leaders punitively judged? An empirical test. Journal of Applied Psychology, 96, 412– 422. https://doi.org/10.1037/a0021442

47.

Shuman

Johnson

Saguy

Halperin

(2018). Threat to the group’s image can motivate high identifiers to take action against in-group transgressions. Personality and Social Psychology Bulletin, 44, 1523–1544. https://doi.org/10.1177/0146167218768800

48.

Skarlicki

D. P.

Kulik

C. T.

(2004). Third-party reactions to employee (mis)treatment: A justice perspective. Research in Organizational Behavior, 26, 183–229. https://doi.org/10.1016/S0191-3085(04)26005-1

49.

Tajfel

Turner

J. C.

(1979). An integrative theory of intergroup conflict. In Austin

W. G.

Worchel

(Eds.), The social psychology of intergroup relations (pp. 33–47). Brooks/Cole.

50.

Teixeira

C. P.

Demoulin

Yzerbyt

V. Y.

(2011). Choosing the best means to an end: The influence of ingroup goals on the selection of representatives in intergroup negotiations. Journal of Experimental Social Psychology, 47, 228–234. https://doi.org/10.1016/j.jesp.2010.10.003

51.

Tyler

T. R.

Blader

S. L.

(2003). The group engagement model: Procedural justice, social identity, and cooperative behavior. Personality and Social Psychology Review, 7, 349–361. https://doi.org/10.1207/S15327957PSPR0704_07

52.

Van Knippenberg

. (2011). Embodying who we are: Leader group prototypicality and leadership effectiveness. The Leadership Quarterly, 22, 1078–1091. https://doi.org/10.1016/j.leaqua.2011.09.004

53.

Van Knippenberg

van Knippenberg

. (2005). Leader self-sacrifice and leadership effectiveness: The moderating role of leader prototypicality. Journal of Applied Psychology, 90, 25–37. https://doi.org/10.1037/0021-9010.90.1.25

54.

Wike

Stokes

Poushter

Fetterolf

(2017). U.S. image suffers as publics around world question Trump’s leadership. Pew Research Center. https://www.pewresearch.org/global/2017/06/26/u-s-image-suffers-as-publics-around-world-question-trumps-leadership/