Since the passage of Title VII of the US Civil Rights Act, the burden for rectifying discrimination increasingly falls on the party who experienced it (Nielsen et al., 2010; Nielson & Nelson, 2005). Rather than organizations being accountable and audited for discriminatory practices, individuals must file a claim of discrimination with either the EEOC or a state and local Fair Employment Practices Agency as a mandatory precursor to filing an employment discrimination claim in court. Reporting discrimination is a difficult decision, and it is powerfully shaped by psychological factors, including concerns with being targeted for retaliation and viewed as a complainer. Given these negative consequences, potential claimants’ decisions may be shaped by the likelihood that others will perceive their discrimination claims as legitimate, buffering them from character derogation. Indeed, even access to counsel depends upon a lawyer being willing to see the case as plausible and winnable. This paper investigates how ingroup discrimination, discrimination perpetuated by one’s own social group, can be an insidious form of discrimination that causes women to have their sexism claims dismissed as illegitimate, and leaves them vulnerable to derogation.

Theoretical Contributions

This work makes several contributions to the literature. First, we aim to replicate foundational work on ingroup discrimination that finds ingroup discrimination is underrecognized relative to outgroup discrimination (Inman & Baron, 1996). Second, we extend previous work by examining the phenomena in the context of formal claims of discrimination. We also test the interpersonal consequences of claiming ingroup discrimination to show how the delegitimization of ingroup discrimination can harm women who report it. Finally, we apply a prototype framework to examine the mechanisms underlying the delegitimization of ingroup discrimination and subsequent consequences for women who claim.

Ingroup Discrimination: An Insidious Manifestation of Bias

At first glance, ingroup discrimination (e.g., women discriminating against other women), may seem like a trivial issue because people tend to favor members of their ingroup over the outgroup (Hewstone et al., 2002), and women in particular display strong in-group biases, favoring women over men (Rudman & Goodwin, 2004). Indeed, some psychologists and sociologists have theorized that women leaders will not discriminate against the women subordinate to them, and they would favor other women by promoting their advancement (Cohen & Huffman, 2007; Nelson & Bridges, 1999). Empirical evidence, however, demonstrates that women leaders do not reliably favor other women, and sometimes discriminate against other women. For example, women (vs. men) leaders do not provide higher salaries or more promotion opportunities for the women subordinate to them (Maume, 2011; Penner et al., 2012; Srivastava & Sherman, 2015; Stainback et al., 2011) and sometimes actively discriminate against their female (vs. male) subordinates (Duguid, 2011; Duguid et al., 2012; Ellemers et al., 2004; Faniko et al., 2021; Kaiser & Spalding, 2015). Likewise, women laboratory leaders exhibit as much hiring bias as men leaders in favoring a man over a woman job candidate (Moss-Racusin et al., 2012). The prevalence of ingroup-directed sexism may be compounded as even in male-dominated industries, women are especially likely to have women rather than men supervisors (Moore, 2002; Reskin et al., 1999).

Structural factors may drive women to perpetuate discrimination against their ingroup. Organizational cultures often reproduce patriarchal structures by centering, normalizing, and valuing stereotypically masculine traits such as agency, confidence, and risk-taking (Cheryan & Markus, 2020; Forbes, 2002). In such environments, women must exhibit and internalize these values to succeed, which may lead them to discriminate against other women to protect their own status and wellbeing, inadvertently joining their male colleagues in perpetuating gender inequities (Derks et al., 2011; Ellemers et al., 2004; Napier et al., 2020). Although ingroup discrimination is ultimately harmful to women, powerful structural forces may encourage women to perpetuate this form of discrimination to protect themselves within patriarchal systems.

Ingroup Discrimination Violates Prototypes of Discrimination

Although women can experience discrimination at the hands of other women, we suspect that others might dismiss their sexism claims as illegitimate relative to when they make the same claims describing sexist treatment from men. Discrimination toward women by women may be underrecognized as discrimination because it violates prototypes, or expectations of what discrimination looks like and who is expected to perpetrate discrimination. People have prototypes of what discrimination looks like and who perpetuates it, and they compare observed actions to this expectancy when determining whether they are discriminatory (Inman & Baron, 1996; Rodin et al., 1990).

One component of the discrimination prototype is that people expect discrimination to be enacted by a member of a high-status group against a member of a lower status group (Bastart et al., 2021; Inman & Baron, 1996; O’Brien et al., 2008; Rodin et al., 1990). For example, the same negative behavior is seen as more sexist when it is committed by a man against a woman than by a woman against another woman (Inman & Baron, 1996; Krumm & Corning, 2008). People also hold prototypes of who perpetrates discrimination. Women may not fit people’s expectation of the perpetrator of bias regardless of whether the target of that bias is another woman or a member of another disadvantaged group. Women may be viewed as sensitive to all forms of bias given their membership in a group that experiences discrimination (Inman & Baron, 1996), and women are viewed as especially kind and moral (Eagly & Mladinic, 1994; Goodwin & Fiske, 2001), and may be viewed as unlikely to perpetrate discrimination. Because ingroup directed discrimination violates these prototypes of discrimination, perceivers might be less likely to recognize ingroup directed discrimination as legitimate. The current work builds upon this foundational scholarship on prototypes of discrimination and extends it to understanding the perceived legitimacy of and reactions to claims of ingroup discrimination.

Legitimacy of and Reaction to Ingroup Discrimination Claimants

Perceiving discrimination claimants as legitimate is central to whether plaintiffs gain access to the legal system and whether their cases can achieve the promise of civil rights laws. These judgements of legitimacy serve as a significant barrier to justice as discrimination claims rarely materialize into successful trial victories, as these claims experience steep attrition from the legal system (Kaiser & Quintanilla, 2014; Nielson & Nelson, 2005). If prototypes of sexist discrimination render ingroup discrimination to be viewed as implausible or unlikely, then women who make claims of ingroup discrimination risk having their claims perceived as insufficiently legitimate. Further, this delegitimization of ingroup discrimination claimants will be especially likely when discrimination is ambiguous as ambiguity creates opportunities for stereotype-based expectations to shape judgments (Major et al., 2003).

When discrimination claimants are viewed as illegitimate, perceivers may find fault in the character of discrimination claimants rather than the perpetrators. Illegitimate discrimination may be framed as unwillingness by the target to take personal accountability, leading to mistrust and character derogation (Schlenker et al., 2001). Specifically, claimants who are perceived as illegitimate may be judged as hypersensitive complainers and as interpersonally problematic (Bergman et al., 2002; Kaiser & Miller, 2001, 2003). If claims of ingroup discrimination violate expectations of prototypical discrimination and are thus delegitimized, these women claimants are vulnerable to heightened interpersonal consequences. These interpersonal consequences can, in turn, lead to worse personal and job-related outcomes for those who choose to report discrimination. Given the potential costs of reporting discrimination, women may not report ingroup discrimination at all (Shelton & Stewart, 2004; Stangor et al., 2002), leaving the mistreatment unchecked.

The Current Study

In the current work, we advance foundational literature on prototypes of discrimination by examining their downstream consequences, and the mechanisms that drive them, for women who formally claim ingroup vs. outgroup discrimination. Figure 1 provides a visual depiction of the associations that will be tested in this research. Across three experiments, we hypothesize that women making ingroup discrimination claims (claims of sexism from a woman manager) will have their claims viewed as less legitimate compared to those making outgroup discrimination claims (claims of sexism from a man manager (Hypothesis 1 [path a1]). In addition, women making ingroup discrimination claims (compared to outgroup discrimination claims) will be viewed as interpersonally problematic, specifically as complainers and unlikeable (Hypothesis 2 [path c1]). 

Fig. 1
figure 1

Summary of Hypotheses

Full size image

We also test the mechanisms driving the effect of supervisor gender on perceptions of the claimant. In the first two studies, we hypothesize that ingroup claimants will receive harsher interpersonal consequences than outgroup claimers because their claims are perceived as less legitimate (Hypothesis 3 [paths a1 and b1]). In the third study, we hypothesize that ingroup claims are perceived as less legitimate, and the claimants subsequently receive harsher interpersonal consequences, because their discrimination violates the prototype of discrimination and who perpetrates it (Hypothesis 4 [paths a2, b2, and b1]).

As preliminary analyses, in each study we test for gender differences to see if men and women similarly delegitimize ingroup discrimination and evaluate the claimants negatively. Finally, we test additional consequences of claiming ingroup discrimination that were not central to our main hypotheses in our supplemental analyses.

Analyses

To address outliers, measures with standardized residuals exceeding an absolute value of three were winsorized to three. All t-tests conducted are two-tailed. These studies were approved by an institutional review board prior to data collection. All data and code are available here: https://osf.io/tmz94/.

Study 1 Method

Participants

One hundred seventy-five participants were recruited via MTurk. Data from 167 participants were analyzed. Eight participants were eliminated for failing an attention check. Participant demographics are described in Table 1. In each study, participants had to select “female” or “male” for their gender or leave this question blank. We note the limitations of using a binary approach to assess gender and using sex category labels when asking for gender identity in the General Discussion. Using G*Power (Faul et al., 2009), a sensitivity analysis revealed the sample was powered to detect an effect size of d = 0.44 at 80% power, α = .05 for the main effect of perpetrator gender.

Table 1 Participant Demographics
Full size table

Procedure

In recruitment, participants were told that the purpose of this study was to learn about how people respond to incidents that take place in a work setting. If they chose to participate, they would view a report of an incident that took place within an organization, and then answer a series of questions about their response to the incident and evaluation of the parties involved. During the study, they viewed a woman employee’s discrimination complaint against her employer. In the complaint, the employee (Chelsea) described being passed over by a supervisor for promotion in favor of a man coworker with less experience and lower revenue generation. Participants were randomly assigned to view a complaint about either Rachel (woman supervisor) or Steven (man supervisor) (see Appendix). They had up to 30 min to complete the procedure. All included participants correctly identified the name of the supervisor, which served as the manipulation check.

Measures

All items were rated on a 5-point Likert scale (1 = strongly disagree; 5 = strongly agree). See Table 2 for means and standard deviations and Table 3 for correlations between measures. Seven additional exploratory measures were examined and reported in the online supplement.

Table 2 Means and Standard Deviations for all Dependent Variables by Condition
Full size table
Table 3 Means, Standard Deviations, and Correlations for Dependent Variables by Study
Full size table

Legitimacy of Discrimination Claim

Participants evaluated the legitimacy of the discrimination claim using a four-item scale of face valid items: “The supervisor’s decision was due to bias”; “The supervisor favored the male candidate due to his gender”; “Chelsea was denied the promotion because of sex discrimination”; and “I do not think this event was an instance of sex discrimination (reverse coded).” This scale was developed for use in the current research, and the internal reliability for this scale was excellent (α = .91).

Employee as Complainer

Participants indicated the extent to which they believed the employee was a complainer using a five-item scale: “The employee filing the report is irritating”; “The employee filing the report is a troublemaker”; “The employee filing the report is argumentative”; “The employee filing the report is hypersensitive”; “The employee filing the report is a complainer”. The internal reliability for this scale was excellent (α = .91). This measure was adapted from Kaiser and Miller’s (2001) 6-item measure.

Likeability of Employee

Participants indicated the extent to which the employee was likeable with two items: “The employee filing the report is likeable”; “The employee filing the report is friendly”. The internal reliability for this scale was adequate (α = .78). This measure was adapted from Kaiser and Miller’s (2001) 15-item measure.

Study 1 Results

For each outcome, we conducted independent sample t-tests to compare means between the woman (n = 74) and man (n = 93) manager conditions. Two participants did not respond to the Likeability of Employee measure, thus we analyzed 165 rather than 167 responses for that measure only. Shapiro–Wilk tests revealed the distribution for each measure’s data was significantly different from normal (Legitimacy: W = 0.96, p < .001; Complainer: W = 0.96, p < .001; Likeability: W = 0.82, p < .001); however, given the sample size, our analyses are robust to non-normality (Lumley et al., 2002).

Legitimacy of the Discrimination Claim

Consistent with Hypothesis 1, participants perceived the discrimination claim by the woman employee as less legitimate when the supervisor was a woman (M = 3.11, SD = 0.78) compared to a man (M = 3.51, SD = 0.59), t(165) = 3.84, p < .001, d = 0.59, 95% CI = [0.28, 0.90]. See Table 2 for means and standard deviations from each study.

Employee as Complainer

Consistent with Hypothesis 2, the employee was seen as more of a complainer when she made a claim of sexist discrimination by a woman supervisor (M = 2.49, SD = 0.96) compared to a man supervisor (M = 2.20, SD = 0.73), t(165) = -2.18, p = .031, d = -0.34, 95% CI = [-0.64, -0.03].

Likeability of Employee

Participants did not differ significantly on their impressions of the employee’s likeability regardless of whether her supervisor was a woman (M = 3.10, SD = 0.63) or a man (M = 3.20, SD = 0.48), t(163) = 1.12, p = .264, d = 0.17, 95% CI = [-0.13, 0.48].

Test of Mediation

Employee as Complainer

To test if women who claim discrimination from a woman are derogated more than those claiming discrimination from a man because their claims are considered less legitimate, we conducted a mediation analysis. We used the R package ‘mediation’ (Causal Mediation Analysis; Tingley et al., 2014) to estimate the indirect effect of supervisor gender on perceptions of the employee as a complainer through legitimacy of the claim with 10,000 bootstrapped samples. There was a significant indirect effect of supervisor gender on the employee as a complainer via legitimacy, b = 0.28, 95% CI = [0.12, 0.47], p < .001, suggesting that the effect of supervisor gender on perceptions of the employee as a complainer (i.e., employees with a woman vs. man supervisor were perceived as greater complainers) was positively related to the extent to which her claim was delegitimized. In this model, the direct effect of supervisor gender on perceptions of the employee as a complainer was no longer significant, b = 0.003, 95% CI = [-0.22, .23], p = .999.

Likeability of Employee

We also tested the mediation model described above with likeability as the outcome. There was a significant indirect effect of supervisor gender on perceptions of the employee’s likeability via legitimacy, b = -0.17, 95% CI = [-0.28, -0.07], p < .001. Although the direct effect of supervisor gender on likeability was not significant in the t-test or this model, b = 0.07, 95% CI = [-0.08, 0.23], p = .357, supervisor gender was related to perceptions of likeability when accounting for the extent to which the employee’s claim was perceived as legitimate.

Gender Moderation

We performed 2-way ANOVAs to test if supervisor gender, participant gender, or their 2-way interaction were associated with mean differences in any of the outcome measures. We only analyzed data from participants who responded to the gender identity demographic question. Neither participant gender nor the 2-way interaction was significant in any of the analyses (see Table 4).

Table 4 Interaction Results for 2 (Participant Gender: Man vs. Woman) × 2 (Supervisor Gender: Man vs. Woman) ANOVAs
Full size table

Study 1 Discussion

Study 1 provides initial support for Hypothesis 1 that claims of ingroup discrimination are delegitimized relative to claims of outgroup discrimination. Support for Hypothesis 2 is mixed, as women making ingroup discrimination claims (vs outgroup claims) were viewed more negatively as complainers but were not viewed as less likeable. Hypothesis 3 was supported in a mediation model showing perceptions of the claim’s legitimacy mediated the relationship between supervisor gender and perceptions of the employee as a complainer and the employee’s likeability. Finally, all results were consistent across participant gender. Study 2 further explores these effects with a replication of the first study within a university subject pool sample rather than an online sample.

Study 2 Method

Participants and Procedure

One hundred thirty-one undergraduate students participated in exchange for partial course credit. Data from 119 participants were analyzed after eliminating responses from 12 participants who failed an attention check. Cell sizes are unequal due to a procedural error in random assignment. Despite the unequal sample sizes, two-tailed t-tests were still used to analyze the data as equal sample sizes are not an assumption of t-tests. Importantly, using F-tests for equality of variances, we found that the variances were not significantly different for legitimacy of the claim (p = .635), employee as a complainer (p = .252), and likeability of employee (p = .757). Participant demographics are in Table 1. A sensitivity analysis revealed the sample was powered to detect an effect size of d = 0.53 at 80% power, α = .05 for the main effect of perpetrator gender (G*Power; Faul et al., 2009).

Participants received the same information during recruitment as in Study 1. During the study, they again reviewed the file of a woman employee who filed a discrimination complaint to her company against her supervisor who was either a man or a woman (see Appendix). The results for eight additional exploratory measures are reported in the online supplement.

Measures

After they reviewed the file, participants were asked to describe the incident that had taken place in their own words (“Please describe the incident described in the claim in your own words, including your impression of what happened between the employee and the supervisor”). They then completed the same measure of legitimacy of the discrimination claim (α = .93), perceptions of the employee as a complainer (α = .89), and likeability of employee (α = .67) on a 7-point Likert scale (1 = strongly disagree; 7 = strongly agree).

Study 2 Results

For each outcome, we conducted independent sample t-tests to compare means between the woman (n = 72) and man (n = 47) manager conditions. There were no missing data for the main outcomes in Study 2. Shapiro–Wilk tests revealed that the distribution of data in each measure differed significantly from a normal distribution (Legitimacy: W = 0.97, p = .010; Likeability: W = 0.87, p < .001; Complainer: W = 0.97, p = .006). However, given the sample size in this study, our analyses are robust to non-normality (Lumley et al., 2002).

Legitimacy of the Discrimination Claim

Participants perceived the discrimination claim as not significantly less legitimate when her supervisor was a woman (M = 4.43, SD = 1.53) compared to a man (M = 4.96, SD = 1.30), t(117) = 1.95, p = .054, d = 0.36, 95% CI = [-0.01, 0.72].

Employee as Complainer

The employee was seen as more of a complainer when she made a claim of sexism from a woman (M = 3.22, SD = 1.16) compared to a man (M = 2.70, SD = 1.24), t(117) = -2.31, p = .022, d = -0.42, 95% CI = [-0.79, -0.06].

Likeability of Employee

The employee was not seen as less likeable when her supervisor was a woman (M = 4.06, SD = 0.70) versus a man (M = 4.13, SD = 0.67), t(117) = 0.50, p = .615, d = 0.09, 95% CI = [-0.27, 0.45].

Test of Mediation

Employee as Complainer

To again test if the derogation of women who claim discrimination from a woman more than those claiming from a man is related to the perceived legitimacy of their claim, we conducted a mediation analysis using the same method described in Study 1. There was a significant indirect effect of supervisor gender on the employee as a complainer via legitimacy, b = 0.18, 95% CI = [0.01, 0.38], p = .045, and there was no longer a significant direct effect of supervisor gender on the employee as a complainer, b = 0.33, 95% CI = [-0.06, 0.73], p = .098. Similar to Study 1, this mediation suggests that ingroup discrimination claims are perceived as less legitimate relative to those outgroup discrimination claims and, in turn, ingroup claimants are perceived as greater complainers.

Likeability of Employee

In addition, we tested the mediation model described above with likeability as an outcome. The indirect effect of supervisor gender on the likeability of the employee was significant, b = -0.09, 95% CI = [-0.21, 0.00], p = .045. Despite the null direct effect in the t-test and in this model, b = 0.02, 95% CI = [-0.21, 0.27], p = .862, perpetrator gender was related to the employee’s likeability via the extent to which the employee’s claim was delegitimized.

Gender Moderation

We performed 2-way ANOVAs to test if supervisor gender, participant gender, or their 2-way interaction were associated with mean differences in any of the outcome measures. Neither participant gender nor the 2-way interaction was significant in any of the analyses (see Table 4).

Study 2 Discussion

In Study 2, we found no significant effect of ingroup vs. outgroup discrimination on claim legitimacy, although the means trended in the hypothesized direction (partial support for Hypothesis 1). As in Study 1, employees claiming ingroup (vs. outgroup) discrimination were perceived as more of complainers, but they were viewed as no less likeable (mixed support for Hypothesis 2). In addition, the relationship between supervisor gender and perceptions of the employee as a complainer and the employee’s likeability was related to the claim’s perceived legitimacy (Hypothesis 3). All results were consistent across participant gender. We conducted a third study to explore the robustness of the effect and to explicitly test whether prototypes drive the biases against ingroup claimants. This study again uses an undergraduate subject pool but uses a relatively larger sample size to increase power.

Study 3 Method

Participants and Procedure

Two hundred and two participants were recruited via an undergraduate participant pool in exchange for partial course credit. Data from 181 participants were analyzed; 21 participants were eliminated for failing an attention check. Participant demographics are in Table 1. A sensitivity analysis revealed the sample was powered to detect an effect size of d = 0.42 at 80% power, α = .05 for the main effect of supervisor gender (G*Power; Faul et al., 2009). Participants received the same study information during recruitment and reviewed the same employee claims as the previous two studies (see Appendix).

Measures

After reviewing the claim, participants completed the same measure of claim legitimacy (α = .89), perceptions of the extent to which the employee was a complainer (α = .89) and likeable (α = .74), and two additional measures, described below. Results from 12 additional exploratory measures are reported in the online supplement.

Violates Prototype of Discrimination

Perceptions that the claim violated the prototype of discrimination, i.e., the expectation that gender discrimination is perpetrated by a man against a woman, was assessed with two items on a 7-point Likert scale (1 = strongly disagree; 7 = strongly agree): “The supervisor wouldn’t discriminate against a woman”, and “The supervisor was unlikely to be biased against a woman”. This scale was developed for the current research, and the internal reliability of this scale was marginally acceptable (α = .65).

Violates Prototype of Perpetrator

Perceptions that the discrimination violated the prototype of a perpetrator, i.e., that a woman would not discriminate against anyone, was measured with three items on a 7-point Likert scale (1 = strongly disagree; 7 = strongly agree): “The supervisor wouldn’t discriminate against anyone”, “I believe this supervisor is capable of behaving in a discriminatory way” (reverse-coded), and “This supervisor is capable of being biased” (reverse-coded). This scale was developed for the current research, and the internal reliability was adequate (α = .77). The two prototype measures were significantly correlated (r = 0.50, p < .001).

Study 3 Results

For each outcome, we conducted independent sample t-tests to compare means between the woman (n = 92) and man (n = 89) manager conditions. There were no missing data for the main outcomes in Study 3. Shapiro–Wilk tests revealed that the distribution of data in the legitimacy measure was not significantly different from normal (W = 0.99, p = .073), but the data for the complainer (W = 0.98, p = .025), likeability (W = 0.81, p < .001), violated prototypes for discrimination (W = 0.93, p < .001) and perpetrator (W = 0.98, p = .006) were. Given the sample size, however, our analyses are robust to the non-normality in the data (Lumley et al., 2002).

Legitimacy of the Discrimination Claim

Participants perceived the discrimination claim as less legitimate when her supervisor was a woman (M = 4.17, SD = 1.03) compared to a man (M = 4.49, SD = 0.87), t(179) = 2.27, p = .024, d = 0.34, 95% CI = [0.04, 0.63].

Employee as Complainer

The employee was seen as more of a complainer when she made a claim of sexism from a woman (M = 3.57, SD = 0.95) compared to a man (M = 3.18, SD = 0.90), t(179) = -2.83, p = .005, d = -0.42, 95% CI = [-0.71, -0.13].

Likeability of Employee

Participants did not differ in likability ratings regardless of whether her supervisor was a woman (M = 4.02, SD = 0.63) or a man (M = 4.10, SD = 0.70), t(179) = 0.80, p = .424, d = 0.12, 95% CI = [-0.17, 0.41].

Violates Prototype of Discrimination

The claim of discrimination from a woman supervisor was viewed as violating the prototype of discrimination (M = 4.27, SD = 0.93) more than the claim from a man supervisor (M = 3.52, SD = 0.83), t(179) = -5.73, p < .001, d = -0.85, 95% CI = [-1.16, -0.55].

Violates Prototype of Perpetrator

Consistent with prototypes that women are less likely perpetrators of all types of discrimination, the woman supervisor was seen as more likely to violate the prototype of a perpetrator (M = 3.23, SD = 1.01) relative to the man supervisor (M = 2.92, SD = 0.82), t(179) = -2.24, p = .026, d = -0.33, 95% CI = [-0.63, -0.04].

Test of Serial Mediation

To examine whether perceptions that the ingroup claim violates the prototype of discrimination (Fig. 2) or who perpetrates it (Fig. 3) are related to the delegitimization and subsequent derogation of the claimant, we tested four serial mediation models. Following the method and R code described by Lemardelet and Caron (2022), we estimated the indirect effects with 10,000 bootstrapped samples.

Fig. 2
figure 2

Serial Mediation Model Linking Supervisor Gender to Employee as Complainer Through Violates Prototypes of Discrimination and then Perceived Legitimacy of the Claim. Note. * p < .05, ** p < .01, *** p < .001

Full size image
Fig. 3
figure 3

Serial Mediation Model Linking Supervisor Gender to Employee as Complainer Through Violates Prototypes of Perpetrator and then Perceived Legitimacy of the Claim. Note. * p < .05, ** p < .01, *** p < .001

Full size image

Employee as a Complainer

In the first model (see Fig. 2), we found a significant indirect effect of the supervisor’s gender on the extent to which the employee was perceived as a complainer through the extent to which the claim violated prototypes of discrimination and, in turn, the extent to which it was perceived as legitimate (b = 0.12, p = .010, 95% CI = [0.04, 0.22]). When accounting for the mediators, the direct effect of supervisor gender on perceptions of the employee as a complainer was no longer significant (b = 0.26, p = .067, 95% CI = [-0.02, 0.54]), suggesting full mediation. In the second model (see Fig. 3), the indirect effect of the supervisor’s gender on the extent to which the employee was perceived as a complainer via the extent to which the perpetrator violated prototypes and subsequent legitimacy judgments was not significant (b = 0.05, p = .106, 95% CI = [0.004, 0.11]). In sum, the effect of supervisor gender on perceptions of the employee as a complainer was related to the extent to which the claim violated the prototype of discrimination, and in turn, the extent to which it was perceived as legitimate. The effect was not related to the extent to which the supervisor violated prototypes of perpetrators of discrimination and the subsequent perceived legitimacy of the claim. These serial mediation models suggest that ingroup claimants may be considered complainers to a greater extent than outgroup claimants because their claims violate the prototype of discrimination (i.e., perpetrated by a man against a woman), which lowers their legitimacy.

Likeability of Employee

We tested the same models using likeability of the employee, instead of perceptions of the employee as a complainer, as an outcome and found neither indirect effect was significant (violates prototype of discrimination: b = -0.01, p = .464, 95% CI = [-0.06, 0.01]; violates prototype of perpetrator: b = -0.01, p = .471, 95% CI = [-0.06, 0.01]).

Gender Moderation

We performed 2-way ANOVAs to test if supervisor gender, participant gender, or their 2-way interaction were associated with mean differences in any of the outcome measures. Neither participant gender nor the 2-way interaction was significant in any of the analyses (see Table 4).

Study 3 Discussion

Findings from Study 3 supported Hypothesis 1, finding claims of discrimination toward a woman from a woman (vs. man) manager are perceived as less legitimate, partially supported Hypothesis 2, that the women filing the claims are perceived as bigger complainers, but as no less likeable, and found no differences in outcomes by participant gender. We also found cross-sectional evidence that ingroup claims are perceived to violate the prototype of discrimination more than outgroup claims, which is related to greater delegitimization of their claims and perceptions of them as complainers (but not perceptions of their likeability; partial support for Hypothesis 4).

Meta-Analyses

We conducted three separate meta-analyses of the measures we collected across all three studies (claim legitimacy, perceptions of the employee as a complainer: N = 467; likeability: N = 465) to test our research questions with greater power. We used a fixed effects approach based on Goh and colleagues’ (2016) process and macro.

Consistent with Hypothesis 1, there was a significant, small to medium effect of supervisor gender on perceptions of claim legitimacy such that ingroup discrimination claims were viewed as less legitimate than outgroup discrimination claims, d = 0.43, SE = 0.09, Z = 4.58, 95% CI = [0.25, 0.62], p < .001. Consistent with Hypothesis 2, women making ingroup discrimination claims were viewed as complainers to a greater extent than those making outgroup discrimination claims, d = -0.41, SE = 0.08, Z = -5.41, 95% CI = [-0.56, -0.26], p < .001. There was no overall effect on likeability, d = 0.13, SE = 0.09, Z = 1.37, 95% CI = [-0.06, 0.31], p = .172.

Moderation Analyses

For a more highly powered test of moderation than in the individual studies, we examined moderation meta-analytically, using the R package, ‘metafor’ (Viechtbauer, 2010). For each test, we report the Q statistic, which is a measure of heterogeneity. However, it is important to note that heterogeneity is difficult to assess with this small number of studies, and the measure has considerable uncertainty (Higgins et al., 2022).

Participant Gender

Consistent with the individual study results, participant gender (i.e., proportion of women who took the study) did not explain a significant amount of heterogeneity in perceptions of the claim’s legitimacy (QM = 1.38, df = 1, p = .241), the employee as a complainer (QM = 0.16, df = 1, p = .691), or the employee’s likeability (QM = 0.06, df = 1, p = .813), suggesting this factor did not moderate the effect of perpetrator gender on any outcomes.

Participant Race

Participant race (i.e., proportion of White people who took the study) also did not explain a significant amount of heterogeneity in perceptions of the claim’s legitimacy (QM = 1.43, df = 1, p = .231), the employee as a complainer (QM = 0.16, df = 1, p = .687), or the employee’s likeability (QM = 0.05, df = 1, p = .818)), suggesting this factor did not moderate the effect of perpetrator gender on any outcomes.

Sampling Method

Sampling method (i.e., recruiting participants via MTurk vs. an undergraduate subject pool) did not explain a significant amount of heterogeneity in perceptions of the claim’s legitimacy (QM = 1.50, df = 1, p = .220), the employee as a complainer (QM = 0.17, df = 1, p = .684), or the employee’s likeability (QM = 0.05, df = 1, p = .831), suggesting this factor did not moderate the effect of perpetrator gender on any outcomes.

General Discussion

Claiming discrimination is the primary mechanism for beginning to rectify unfair outcomes, yet these studies suggest women who experience and claim discrimination from a woman (vs. man) manager may face skepticism and character derogation. Across three experiments, women claiming discrimination from a woman supervisor had their claims perceived as less legitimate, and they were judged as complainers to a greater extent than those claiming the exact same discrimination from a man manager.

Confirming previous work, we found no evidence of moderation by participant gender, suggesting that both men and women hold prototypes of who perpetrates discrimination (Inman & Baron, 1996), which in turn, predicts delegitimizing a woman claiming ingroup discrimination and evaluating her negatively. Both genders may hold these prototypes because they are socialized in similar contexts and thus susceptible to various stereotypes and prototypes (Jost & Banaji, 1994; Nosek et al., 2007; Zaikman & Marks, 2014). Another possibility is that women endorse such beliefs protect their own wellbeing. Although gender discrimination is very real, endorsing the structures that uphold this disequilibrium can be protective for women such that greater system justification is positively related to women’s wellbeing (Napier et al., 2020).

The relative delegitimization of ingroup claimants is consistent with previous findings that under conditions of hostile sexism, gender-based discrimination perpetrated by a woman (vs. man) against a woman coworker (i.e., denying her request to join the company basketball team) is perceived as less discriminatory (Bastart et al., 2021). However, we add to this literature by demonstrating that ingroup discrimination is delegitimatized even when women make a formal claim pointing to evidence of gender discrimination. Our work also expands upon foundational literature on ingroup discrimination (Inman & Baron, 1996) by assessing the downstream interpersonal consequences of neglecting to recognize ingroup-directed discrimination (i.e., ingroup claimants are perceived as greater complainers). Finally, our research also provides correlational initial evidence that ingroup-directed discrimination claims are perceived as less legitimate than outgroup discrimination claims because the former violates prototypes of what discrimination looks.

Limitations and Future Research Directions

In all three studies, we found only partial support for Hypothesis 2; ingroup claimants were never seen as less likeable than outgroup claimants. Participants may have been reluctant to rate the woman claiming either type of discrimination as unlikeable because it would be socially unacceptable to express dislike of a victim of misfortune. Indeed, the average likeability ratings were above the midpoint in each condition across the three studies (see Table 2). Kaiser and Miller (2003) similarly found that while participants rated an African American person who claimed discrimination (vs. took personal responsibility for an event) as more of a complainer, they were not perceived as any less likeable. Interestingly, the indirect effect of perpetrator gender on likeability was significant in Study 1 and 2, suggesting that perpetrator gender may predict perceptions of a claimant’s likeability, but only via the claim’s perceived legitimacy.

In addition, the cross-sectional nature of the mediation analyses limits any conclusions of causality. We conducted mediation analyses to test why a woman claiming ingroup discrimination is perceived as a complainer to a greater extent than a woman claiming outgroup discrimination. However, we cannot claim causality with this specific ordering of variables without experimentally manipulating these mechanisms (i.e., perceived legitimacy and violation of prototypes of discrimination). Furthermore, these studies identified, but did not attempt to intervene on, the mechanisms driving the biased perceptions and outcomes of ingroup discrimination. Nonetheless, the generally consistent effects across all three studies, suggest that future studies could attempt to intervene by broadening the prototype of discrimination, or by bolstering the claim’s legitimacy (Danna et al., 2020).

In our studies, we did not specify the race of the claimant. Given that White women are prototypical victims of gender discrimination (Kaiser et al., 2022), it is likely participants assumed the claimant was White. However, women of color, in particular Black women, are perceived as less likely targets of gender discrimination and are more likely to have their claims discounted than White women (de Leon & Rosette, 2022). If we manipulated employee race so that there were White and Black claimants, participants may have found the Black woman’s claim illegitimate in both conditions, possibly washing out the effect of perpetrator gender. Future work should consider taking an intersectional approach to untangle which aspects of a woman’s identities and experiences make her particularly vulnerable to delegitimization and character derogation when claiming gender discrimination.

Finally, our sample is limited by both gender and race. Due to the framing of our gender identity demographic question, participants could only select male, female, or leave the question blank. By forcing participants to respond in a binary manner, we failed to capture responses from participants with different gender identities. Furthermore, the sample we recruited was majority White and Asian. Therefore, we note that the generalizability of our results may be limited to White and Asian people who identify along the gender binary.

Practice Implications

Although women who experience ingroup discrimination stand to benefit psychologically and economically from reporting it, awareness of the potential consequences may disincentivize filing a claim. Fears of retaliation and backlash that generally inhibit reports of discrimination by low status groups (Bergman et al., 2002; Stangor et al., 2002) may be heightened when women’s claims of ingroup (vs outgroup) discrimination are delegitimized. If women do not report their discrimination due to heightened fears, they avoid the potential costs, but also the potential benefits. Psychologically, confronting discrimination can lead to greater feelings of self-worth and empowerment (Gervais et al., 2010; Swim & Thomas, 2006), while not confronting can lead to guilt and rumination (Shelton et al., 2006; Swim & Thomas, 2006). Economically, women can experience financial gains when discrimination is reversed (e.g., being retroactively awarded a promotion or raise, or receiving compensatory damages; Burstein & Edwards, 1994). Women who experience discrimination at the hands of a woman superior find themselves in a difficult position in which they may experience especially harsh repercussions if they claim discrimination, yet by not claiming they miss out on the potential psychological and economic benefits.

Claims of discrimination can benefit not just the claimants, but also organizations more broadly. Following confrontation, perpetrators of discrimination are less likely to discriminate in the future (Czopp et al., 2006), which benefits both the current and future targets of mistreatment. Conversely, unconfronted discrimination creates norms that this behavior is acceptable (Mallett et al., 2021), which may embolden perpetrators to continue their abuse. Discrimination claims also hold organizations accountable and can motivate them to improve their equity practices. Indeed, following a discrimination lawsuit, companies became more diverse than they were before the lawsuit, indicating a switch to more egalitarian practices (Kalev et al., 2006). Given the legal importance of discrimination claims in changing organizational norms and preventing future discrimination, the costs of unreported ingroup discrimination are high.

Conclusion

Across three studies, women who claimed ingroup discrimination experienced more negative consequences than those who claimed discrimination by an outgroup member. The greater consequences they faced were related to their claim’s being perceived as less legitimate, and this delegitimization was related to their experiences violating the prototype of discrimination. Although both ingroup and outgroup discrimination are detrimental, ingroup discrimination is particularly insidious because it is delegitimized and thus overlooked. When women file claims of ingroup discrimination, these claims may be dismissed as unfounded more often than outgroup claims, and even if their claims are investigated, they face potentially lasting character derogation. By problematizing women claimants, the issue becomes individual women’s bad behavior, masking powerful structural factors (i.e., patriarchy and masculine norms) that may drive women to discriminate against other women (Sheppard & Aquino, 2017). The consequences of Ingroup discrimination are often overlooked, but recognizing the legitimacy of this discrimination may be an important step in the pursuit of workplace gender equity.