Sex Roles

, Volume 62, Issue 9, pp 623–634

Bystander Sexism in the Intergroup Context: The Impact of Cat-calls on Women’s Reactions Towards Men

Authors

    • Department of PsychologyUniversity of Connecticut, 406 Babbidge Rd.
    • Bradley University
  • Diane M. Quinn
    • Department of PsychologyUniversity of Connecticut, 406 Babbidge Rd.
Original Article

DOI: 10.1007/s11199-009-9735-0

Cite this article as:
Chaudoir, S.R. & Quinn, D.M. Sex Roles (2010) 62: 623. doi:10.1007/s11199-009-9735-0

Abstract

Despite the fact that sexism is an inherently intergroup phenomenon, women’s group-level responses to sexism have received relatively little empirical attention. We examine the intergroup reactions experienced by 114 female students at a U.S. university in New England who imagined being a bystander to a sexist cat-call remark or control greeting. Results indicate that women experienced greater negative intergroup emotions and motivations towards the outgroup of men after overhearing the cat-call remark. Further, the experience of group-based anger mediated the relationship between the effect of study condition on the motivation to move against, or oppose, men. Results indicate that bystanders can be affected by sexism and highlights how the collective groups of men and women can be implicated in individual instances of sexism.

Keywords

SexismBystanderIntergroup emotionsCat-callGender identity

Introduction

In nearly all cultures, patriarchal social systems ensure that women will occupy a lower-power status than men. Sexist behaviors, such as sexual harassment, job discrimination, and cat-calls, are just a few of the many types of social phenomena that maintain this group-based hierarchy (e.g., Sidanius and Pratto 1999). While this group-based conflict lies at the heart of most sexist behavior, little research has examined how women’s psychological responses may take the form of group-level reactions. That is, while the preponderance of research to date has demonstrated that women who are targets of sexism experience a host of deleterious intraindividual outcomes such as increased negative affect and lowered self-esteem (e.g., Crocker et al. 1991; Fitzgerald 1993), little research has examined how experiences of sexism may shape intergroup outcomes such as women’s group-level emotions and behavioral intentions towards the outgroup of men, in general (for an exception, see Pennekamp et al. 2007). In order to address this gap, we apply insights from social identity perspectives (Tajfel and Turner 1986; Turner et al. 1987) and intergroup emotions (Mackie et al. 2000; Smith 1993, 1999) to examine the possibility that exposure to an instance of bystander sexism will elicit group-based responses from U.S. undergraduate women. Specifically, we examine how exposure to bystander sexism—imagining oneself as a bystander to a “cat-call” towards another woman—may elicit group-based emotions (i.e., anger and fear) and behavioral intentions (i.e., desire to move against or away from) towards men, in general.

Sexism and the Intergroup Context

Several decades of research demonstrates that sexism is a frequent occurrence in American women’s personal and professional lives and can be detrimental to their psychological well-being, health, and job satisfaction (e.g., Crocker et al. 1991; Fitzgerald 1993; Fitzgerald et al. 1997; Major et al. 2003; Schneider et al. 1997, 2001; Swim et al. 1998; Swim and Hyers 1999; Swim et al. 2001; see Lee et al. 2007 for a discussion of cross-cultural incidence and origin of sexism). While a great deal of research has focused on understanding these intraindividual outcomes, very little work has examined the possibility that sexism may also have ramifications for how women view the outgroup of men, more generally (for an exception, see Pennekamp et al. 2007). Because sexism necessarily implicates group-based identities—prejudiced acts towards women based on their group membership, where the perpetrator is typically male—specific experiences of sexism may also have important implications for how women perceive and react to men, in general. Put differently, the actions of one sexist man may serve to taint women’s perceptions of all men.

Under what conditions might we expect that the sexist actions of one man will affect women’s perceptions of all men? Here, we suggest that women’s group-based emotions and behavioral intentions towards the outgroup of men may become more negative when their gender group identity is salient. Drawing on insights from social identity perspectives (Tajfel and Turner 1986; Turner et al. 1987), individuals may vary to the extent that they view themselves as individuals vs. interchangeable members of a social group. In the context of gender, for example, environmental stimuli can prompt women to shift from thinking about themselves as unique individuals to thinking about themselves as interchangeable members of the larger social group of women, as a whole. Intergroup emotions theory (Mackie et al. 2000; Smith 1993, 1999), an extension of social identity perspectives, suggests that when women’s group identities become salient, their emotions and motivations shift to reflect their group, rather than individual, concerns. Thus, when a woman views an instance of sexism, her group identity as a woman may become salient and she may subsequently experience emotions and motivations on behalf of her gender group (i.e., intergroup response) rather than as an individual person (i.e., intraindividual response). Because she is now thinking about herself as an ingroup member, she also necessarily becomes concerned with the outgroup (i.e., men). Thus, when gender group identity is made salient, women may experience emotions and motivations towards the outgroup based on whether the current situation may help or harm women as a whole (Mackie et al. 2000; for a review see Mackie and Smith 2002).

Research in the domain of intergroup emotions suggests that when individuals detect harm or threat to their group, these appraisals lead them to experience predictable patterns of emotional and behavioral responses (Cottrell and Neuberg 2005; Frijda 1986; Mackie et al. 2000). Two of the most common emotions elicited by group-based threats are anger and fear. When individuals perceive that an outgroup threatens their ingroup but believe they also possess the strength and resources to counteract this threat, they are likely to feel anger towards the outgroup (e.g., Cottrell and Neuberg 2005; Frijda 1986; Mackie et al. 2000). This anger, in turn, should increase the ingroup members’ desire to move against, or approach, the outgroup (e.g., Crisp et al. 2007; Mackie et al. 2000; Yzerbyt et al. 2003). However, when individuals perceive the same threat but believe they do not have sufficient resources to counteract the threat, they are likely to feel fear towards the outgroup (e.g., Cottrell and Neuberg 2005; Frijda 1986; Mackie et al. 2000). This fear, in turn, should increase the ingroup members’ desire to move away from, or avoid, the outgroup (e.g., Mackie et al. 2000; Crisp et al. 2007).

While the intergroup emotions model (Mackie et al. 2000; Smith 1993, 1999) has been applied to many types of intergroup relations (e.g., differences in beliefs in controversial issues, race/ethnicity, and nationality; Butz and Plant 2006; Mackie et al. 2000; Maitner et al. 2006), its insights have rarely been applied to gender. Few studies have examined whether sexism leads women to experience group-based emotions and behavioral intentions towards men. Pennekamp and colleagues (2007) examined whether evidence of pervasive sexism would lead Dutch female undergraduates to experience feelings of anger towards men and increased behavioral intentions to improve the position of women in their society. Their results demonstrate that women who more strongly identified with their gender group reported more intergroup anger towards men which, in turn, was related to stronger motivations to demand reparations from men. While the Pennekamp and colleagues’ (2007) study provides some initial evidence that intergroup emotions theory may offer a useful framework to consider women’s group-level responses to sexism, it only examines one, approach-focused emotional response to sexism (i.e., anger). In the current work, we examine women’s feelings of anger and fear towards the outgroup of men in order to capture both approach-and avoidance-related emotions.

In the current study, we also extend previous research by examining whether the actions of an individual outgroup member, as opposed to the actions of the entire outgroup, can elicit group-based reactions. The majority of previous research in the domain of intergroup emotions has elicited group-based reactions by prompting participants to think about their group memberships, per se, or by presenting group-level threats (e.g., Mackie et al. 2000; Maitner et al. 2006; Pennekamp et al. 2007). For example, as we noted above, Pennekamp and colleagues (2007) prompted Dutch women to think about the existence of pervasive sexism—where the outgroup of men currently threatens the equality of their ingroup. The fact that this study procedure elicited anger towards all men may not be particularly surprising in light of the fact that the procedure itself prompted participants to think and respond at a group level.

Do daily, individual acts of sexism also render group-based responses from women? To our knowledge, this possibility has not been empirically examined. In the current study, we considered whether participants will experience group-based emotions and their concomitant motivations in a situation where they are not asked to think about their group membership explicitly and they are presented with an individual-, rather than group-, level threat. By definition, the motivation behind all forms of sexism is, in and of itself, group-based. Sexism represents prejudiced acts towards women based on their group membership. However, the manner in which this group-based prejudice or threat is expressed can vary widely. In some situations, such as those represented in Pennekamp and colleagues’ (2007) work, women are presented with evidence of group inequality (e.g., gender pay gap). In these situations, women are likely to perceive that men, as a group, present a threat to women, as a group. Because these situations explicitly call attention to gender groups, it follows that women will also respond to the situation on behalf of their group and express emotions and motivations directed at men, in general.

In other situations, however, a sexist threat can be expressed through an individual group member’s behavior, and women may not readily attribute the offense to gender group status. That is, when women make attributions about an individual man’s sexist behavior, they may be just as likely to attribute the behavior to his individual self (e.g., his rude personality) as they are to attribute the behavior to his group membership (e.g., a sexist man whose behavior represents the group-based threat of sexism). When sexism is expressed via an individual man’s behavior, American women frequently fail to attribute the behavior to the group-level threat of sexism (e.g., Crosby 1984; Inman and Baron 1996; Sechrist and Delmar 2009; for a review, see Barrett and Swim 1996). Unlike threats that are expressed in group-based terms, threats expressed in individual behaviors may fail to elicit group-level attributions and, therefore, group-based responses.

Bystander Sexism and Cat-calls

While individual instances of sexism can be expressed in any number of ways, bystander sexism is one expression that has received minimal empirical attention. Bystander sexism is an instance of sexism wherein a woman is not directly involved in the immediate social context of the sexist event targeted at another woman, but is exposed to the event nonetheless. To our knowledge, the only work that has directly examined the impact of bystander sexism (or a related construct) on women’s well-being is that of Hitlan and colleagues (Hitlan et al. 2006; Walsh and Hitlan 2007). These researchers have examined the impact of bystander sexual harassment—“experiences where one observes or knows about the sexual harassment of others but is not directly the target of the harassment.” In their sample of U.S. female employed undergraduate students, 69% of participants reported being a bystander to sexual harassment and these researchers find that the experience of bystander sexual harassment exacerbated the negative emotional responses women had in their own personal experiences with sexual harassment (Hitlan et al. 2006).

This work provides some preliminary evidence to suggest that being a bystander can elicit negative, intraindividual psychological consequences for women. However, it does not address the possibility that bystander sexism may elicit group-based responses. Further, compared to the work by Hitlan and colleagues, our research adopts a more restrictive definition of what it means to be a bystander. That is, we examine how observing a specific sexist incident, rather than observing or knowing about chronic sexist behavior in one’s workplace (i.e., sexual harassment), impacts women’s outcomes.

One situation in which women are likely to be bystanders to sexist situations is when other women are targets of cat-calls. Cat-calls are directed at women as a way to highlight a sexualized part of her body (e.g., breasts, hips, butt). As Gardner (1980) points out, women in America are frequently targets of evaluative and objectifying cat-calls about their bodies when they are in public. Although it is possible that men may intend to make cat-call remarks in order to compliment or attract women, researchers have consistently emphasized the derogatory and sexist nature of these comments (Bowman 1993).

Cat-calls are a frequent way in which women are the targets of sexism in their daily lives (Swim et al. 2001), with 42% of U.S. female college students reporting that they are the direct targets of cat-calls at least once a month and an additional 31% reporting these experiences every few days (Fairchild and Rudman 2008). Recent work reports that the experience of street harassment is directly related to greater preoccupation with physical appearance and body shame, and is indirectly related to heightened fears of rape for U.S. undergraduate women (Fairchild and Rudman 2008).

The negative effects of cat-calls may not be confined solely to women who are targets. An important feature of cat-call remarks is that they are given in public contexts, such as on city streets. Because of the public nature of these comments, they are likely to be overheard by other female bystanders. Thus, overhearing and attending to cat-calls directed at other women may also affect female bystanders.

Overview of Present Research

In the current work, we examine female undergraduates’ psychological responses to bystander sexism. To do so, we asked women to watch a video and imagine themselves as a bystander to an interaction where a man made either a sexist cat-call remark or a control greeting directed at another woman. We examined how exposure to these two different types of comments (i.e., sexist cat-call vs. control greeting) would impact the salience of their gender identity, their individual and intergroup emotional reactions, and their intergroup motivations towards men, in general.

We expected that participants would find the cat-call scenario to be more prejudiced than the neutral scenario, a difference that would serve as a test of the validity of our manipulation. Based on our theorizing noted above, we tested four main hypotheses in this study. In accordance with social identity perspectives (Tajfel and Turner 1986; Turner et al. 1987), we expected that women who were bystanders to the cat-call comment would be more likely to experience gender identity salience—thoughts about their gender group membership.
  • Hypothesis 1. Women in the bystander sexism condition will be more likely than women in the control condition to list thoughts about their gender group membership (e.g., girl, woman) on a measure of working self-concept.

Because women will be more likely to be thinking of themselves in terms of their gender identity, we expected that women exposed to the cat-call will also experience greater group-level emotions (i.e., anger and fear towards men) than women exposed to the control comment. However, to what degree will women experience anger compared to fear? Previous work in the intergroup emotions domain has often manipulated the relative power of the ingroup in order to elicit one emotion over the other (e.g., anger vs. fear; Mackie et al. 2000). For purposes of the present work, however, we chose a stimulus that could conceivably elicit either emotion: intergroup anger or fear. A cat-call is a particularly ambiguous sexist situation because it could be perceived as having either a harmful or complimentary intent (Bowman 1993). That is, some women may perceive the comment to be demeaning and overtly offensive while others may perceive the comment to be harmless and flattering. Thus, appraisals of the intent of the cat-call remark could vary widely. Further, we did not give women information about the relative power of the gender groups, making it possible that women’s appraisals of relative group strength would also vary widely. In addition, prior work has demonstrated that participants experience increases of both intergroup anger and fear in response to group-based threats (e.g., Maitner et al. 2006). Thus, in our study, it is possible that participants could feel intergroup anger or fear in response to the cat-call remark.
  • Hypothesis 2a. Women in the bystander sexism condition will report greater intergroup anger and greater intergroup fear than women in the control condition.

Although we predicted that women would report more of both negative emotions (i.e., anger and fear) when they overheard the cat-call compared to the neutral comment, we expected feelings of anger to be stronger than feelings of fear. Although fear is often a relevant emotional reaction in response to sexism (e.g., Woodzicka and LaFrance 2001), we expected that women would experience stronger feelings of anger because they were not the direct targets of the sexist commentary. Previous work indicates that women often anticipate that they will react with more anger than fear (Woodzicka and LaFrance 2001) and that they will be more likely to directly confront a sexist perpetrator (Shelton and Stewart 2004) when they are asked to indicate their expected responses to a hypothetical sexist situation. That is, women tend to overestimate the extent to which they will feel anger and exhibit confrontational behavior when they are asked to indicate how they think they will react. Although these results do not directly parallel the procedure used in the current work, these results do suggest that women who are bystanders watching a sexist scenario, but removed from the direct situation, will experience more intergroup anger than intergroup fear.
  • Hypothesis 2b. Among women in the bystander sexism condition, intergroup anger will be greater than intergroup fear.

Because intergroup emotions are hypothesized to elicit concomitant motivations towards the outgroup of men (Mackie et al. 2000; Smith 1993, 1999)—feelings of anger give rise to motivations to move against men while feelings of fear give rise to motivations to move away from men—our predictions for intergroup motivations were parallel to those noted above for intergroup emotions.
  • Hypothesis 3a. Women in the bystander sexism condition will report greater intergroup motivations to move against men and greater intergroup motivations to move away from men compared to women in the control condition.

  • Hypothesis 3b. Among women in the bystander sexism condition, women will report greater motivation to move against men than to move away from men.

According to intergroup emotions theory (Mackie et al. 2000; Smith 1993, 1999), intergroup emotions should mediate the effect of condition on their respective intergroup motivations. If, however, intergroup anger is the predominant response of female bystanders as we have theorized, intergroup anger should mediate the effect of condition on motivations to move against men, but intergroup fear should not mediate the effect of condition on motivations to move away from men.
  • Hypothesis 4. Intergroup anger will mediate the effect of condition on motivation to move against men.

Additionally, intergroup emotions should not mediate the effect of condition on their opposite motivational orientation. Anger—an approach-oriented emotion—should not mediate the effect of condition on motivation to move away from men. Fear—an avoidance-oriented emotion—should not mediate the effect of condition on motivation to move against men.

Ultimately, we expect that women who are bystanders to a cat-call remark will be more likely to respond to this situation on the basis of their group membership, rather than individual identity. If it is the case that women in the bystander sexism condition are more likely to exhibit gender identity salience and, therefore, exhibit greater group-based emotions, it also follows that these women should not necessarily experience greater individual-based emotions compared to women in the control condition. Because group-and individual-based emotions are distinct affective experiences (Seger et al. 2009; Smith et al. 2007), we expected that individual-based emotions would not be affected by our manipulation. In order to check this assumption, we also included a measure of individual level negative affect.

Method

Participants

One-hundred fourteen female students from a large public New England university in the U.S. participated in this study for partial course credit during the spring semester of 2005. Participants were predominantly Caucasian (79.8%), and the mean age of this sample was 18.6 (SD = 1.26) years.

Measure

State Negative Affect

The Multiple Affect Adjective Checklist (MAACL; Zuckerman and Lubin 1965) served as a measure of respondents’ experience of overall state negative affect. Respondents indicate on a 5-point Likert scale the extent to which each emotion adjective describes their current emotional state (1 = not at all, 5 = very much). The MAACL is a 20-item measure composed of three subscales measuring anxiety (e.g., nervous), depression (e.g., discouraged), and hostility (e.g., angry), and items are averaged to create a composite measure of each subscale (αs = .79, .78, and .81, respectively). The MAACL has been used to assess changes in state negative affect in response to sexism in several prior studies (e.g., Samoluk and Pretty 1994; Schmitt et al. 2003).

Intergroup Emotions

A measure of intergroup emotions (Mackie et al. 2000) was included in order to assess other-directed emotions (i.e., emotions directed towards men). The intergroup emotions measure is comprised of two, 4-item subscales measuring anger (e.g., irritated, furious) and fear (e.g., anxious, afraid). Respondents indicate on a 7-point Likert scale ranging from 1 (not at all) to 7 (extremely) the extent to which the out-group (i.e., men) makes them feel each emotion (e.g., “Men, in general, make me feel irritated”). Subscale items are averaged to create a composite measure of each emotion αs = .93 and .85, respectively.

Prejudice Appraisal Manipulation Check

In order to verify that women perceived the cat-call remark to be more prejudiced than the greeting, they were asked to make ratings about the extent to which they perceived the comment as prejudiced. We also included several filler rating items (e.g., intelligent, humorous) in order to reduce participant demand characteristics (adapted from Swim and Hyers 1999). Ratings of this one-item measure of prejudice were made on a 7-point Likert scale (1 = not at all, 7 = very).

Intergroup Motivations

In order to assess motivations toward the out-group (i.e., men) after overhearing the male confederate’s comment, the intergroup behavioral tendencies scale (Mackie et al. 2000) was utilized. This measure is comprised of two, 3-item subscales measuring the desire to move against (e.g., “Men, in general, make me want to oppose them”) or away (e.g., “Men, in general, make me want to avoid them”) from the out-group. Respondents indicate on a 7-point Likert scale ranging from 1 (not at all) to 7 (extremely) the extent to which the out-group (men) makes them want to engage in each behavior. Subscale items were averaged to create a composite score for each behavioral tendency (αs = .86 and .93, respectively).

Gender Identity Salience

Participants completed the Twenty Statements Test (TST; Kuhn and McPartland 1954)—a measure of working self-concept—in order to assess whether women were thinking about themselves in terms of their gender group. Participants were asked to fill in a series of twenty statements that complete the sentence “I am _____.” We created a dichotomous coding scheme in order to code for the presence or absence of gender identity (e.g., “I am a girl”) in each of participant’s 20 responses to this measure. Two trained raters coded each of these 20 statements for the presence of 3 target words representing gender identity: “girl,” “woman,” and “female.” These two raters demonstrated 99% agreement. Of the participants who wrote a gender identity response to the TST, only 1 participant wrote more than one gender identity response (i.e., 2 gender identity responses); the rest wrote only 1 gender identity response. Therefore, given this lack of variability in our sample, we created a final dichotomous measure that assessed whether women mentioned their gender identity (yes vs. no) in any of their 20 responses to the TST.

Procedure

Participants were tested in individual sessions and told that the purpose of the study was to examine how people form first impressions of others. After completing a consent form, participants were asked to view a videotape of an experimental session that was conducted during the prior semester and imagine that they were the participant in that session. Participants were told that they would be asked to make ratings about their impressions of a person in the video after they had finished watching it and that we were interested in seeing how their impressions compared to those of the participants we examined in the previous semester. They were told that we would be asking them to recall information about what they saw in the video and that they should make sure to pay attention to the video.

Participants were randomly assigned to hear one of two procedural manipulations in the video in which they believed they would later be interacting with the male participant as part of the study, or they believed that he was at the session due to a scheduling error. This manipulation did not affect the results described below, so we do not discuss the effect of this manipulation further.

The video was a recording of the following scenario involving a male participant and a female experimenter. The video was recorded from the perspective of the participant, where the participant would be seated facing the open doorway of the experiment room, and a female experimenter stood in front of the participant. Approximately one minute after the experimenter finished delivering verbal instructions regarding the nature of the study, a male confederate arrived at the experiment, paused in the open doorway leading into the room and made a brief statement to a fictitious female friend in the hallway outside. From their seated position, “participants” viewed the profile of the male confederate as he made the comment in the hallway. Participants were randomly assigned to hear either a sexist cat-call remark (N = 58) or a control greeting (N = 56) in the video. In the sexist condition, the male confederate made a “cat-call” remark into the hallway, directed at the fictitious female target, saying, “Hey Kelly, your boobs look great in that shirt!” The “cat-call” was intended to serve as a sexual objectification of the target female by drawing attention to a sexualized part of her body (i.e., breasts). In the control condition, the male confederate directed a greeting to the fictitious female target, saying, “Hey Kelly, what’s up?”

After the male confederate made his comment, the experimenter directed him to wait in a separate room. After the male confederate left the room, the experimenter closed the door to begin the experimental session, and the video stopped. After viewing the videotape, participants completed ratings of the male participant and the comment he made. Measures were completed in the order listed in the section above.

In order to ensure that participants paid attention to the main study manipulation, we asked participants the following question after completing the materials noted above: “In the video, which of the following do you recall about what happened prior to the start of the experiment?” Participants chose from 1 of 4 options: (1) A man greeted his friend in the hallway, (2) A man made a comment about a girl’s appearance in the hallway, (3) Don’t remember what the man said, and (4) Don’t remember ever seeing a man. Participants in the cat-call condition who chose options 1, 3 or 4 and participants in the control condition who chose options 2, 3, or 4 were excluded. Based on these criteria, 14 of the participants incorrectly recalled what they heard the male participant say in the video, so these participants’ data were dropped from all analyses. Therefore, our final sample included 100 women (54 in the sexism condition; 46 in the control condition).

Results

Manipulation Check and Descriptive Statistics

Table 1 presents the descriptive statistics for our main study variables. We first used our manipulation check to determine whether the bystander sexism condition was perceived to be more prejudiced than the control condition and, indeed, it was, t(98) = 13.00, p < .001, d = 2.52. However, although participants viewed the cat-call remark as more prejudiced than the greeting, their mean ratings of prejudice were just above the midpoint of the scale. Thus, the cat-call remark was rated as moderately prejudiced.
Table 1

Descriptive statistics based on condition.

 

Bystander sexism (n = 54)

Control (n = 46)

Between-group p-value

Appraisal of prejudice

4.67 (1.93)

1.15 (0.42)

<.001

Gender identity salience

15 (28%)

6 (13%)

.06

State negative affect

   

  Anxiety

2.02 (0.54)

2.14 (0.67)

n.s.

 Depression

1.95 (0.53)

2.11 (0.61)

n.s.

 Hostility

1.75 (0.62)

1.71 (0.67)

n.s.

Intergroup emotions

   

 Anger

4.08 (1.30)

1.58 (0.83)

<.001

 Fear

2.11 (0.98)

1.40 (0.77)

<.001

Intergroup motivations

   

 Move against

3.67 (1.48)

1.59 (0.83)

<.001

 Move away

4.65 (1.30)

2.07 (1.47)

<.001

Gender identity salience is dichotomous measure. Appraisal of prejudice, intergroup emotions, and intergroup behavioral intentions were measured on a 1 (not at all) to 7 (very/extremely) scale. State negative affect was measured on a 1 (not at all) to 5 (very much) scale. Comparison is based on t-test for appraisal of prejudice, Chi-square for gender identity salience, and univariate ANOVAs for remaining variables.

Table 2 presents the bivariate correlations among the main study variables. Intercorrelations among the state negative affect subscales (rs = .57 to .74), intergroup emotions (r = .62), and intergroup motivation (r = 57) were all significant and in the expected direction (all ps < .05). Additionally, the intergroup emotions of anger and fear were each correlated with their respective intergroup motivation in the expected direction (r = .82 and .28, respectively; all ps < .05).
Table 2

Correlations among study variables.

 

1.

2.

3.

4.

5.

6.

7.

8.

9.

1. Gender identity saliencea

         

2. Prejudice appraisal

.18

        

State negative affect

         

 3. Anxiety

−.04

−.04

       

 4. Hostility

.04

.02

.57**

      

 5. Depression

.07

−.19

.67**

.74**

     

Intergroup Emotions

         

 6. Anger

.14

.71**

.08

.26**

.09

    

 7. Fear

.02

.38**

.20*

.26**

.17

.62**

   

Intergroup Motivations

         

 8. Move against

.10

.63**

−.04

.18

−.05

.82**

.43**

  

 9. Move away

.26*

.58**

−.08

.05

−.04

.56**

.28**

.57**

 

 10. Conditionb

.20

.77**

−.11

.03

−.15

.77**

.42*

.67**

.71**

a0 = no, 1 = yes. b0 = control, 1 = sexist

**p < .01. *p < .05

  • Hypothesis 1: Gender identity salience

We conducted a chi-square analysis in order to determine whether participants who overheard the cat-call remark would be more likely to think of themselves in terms of their gender group identity. Results indicate that women in the bystander sexism condition were marginally more likely to exhibit gender identity salience compared to women in the control condition (28% vs. 13%), χ2(1) = 3.4, p = .06. Thus, there is a marginally significant trend for women to be more likely to think of themselves in terms of their gender identity (i.e., gender identity salience) after overhearing the cat-call remark compared to a neutral greeting.
  • Hypothesis 2a: Intergroup emotions across condition

We hypothesized that women in the bystander sexism condition would report greater intergroup emotions of anger and fear compared to those in the control condition. In order to examine this hypothesis, we conducted a one-way multivariate analysis of variance (MANOVA) of the intergroup anger and intergroup fear scales with condition as a between-subject factor. This multivariate analysis supported our hypothesis, Wilks’ λ = .40, F(2, 97) = 72.31, p < .01. Follow-up univariate tests indicated that women in the bystander sexism condition reported both more intergroup anger, F(1, 98) = 142.95, p < .001, and intergroup fear, F(1, 98) = 20.91, p < .01, compared to women in the control condition.
  • Hypothesis 2b: Intergroup anger vs. intergroup fear in bystander sexism condition

We hypothesized that women in the bystander sexism condition would report greater intergroup anger compared to intergroup fear. In order to examine this hypothesis, we conducted a one-way ANOVA among women in the bystander sexism condition using intergroup emotion (anger vs. fear) as a within-subject factor. Consistent with our hypothesis, women in the bystander sexism condition reported greater intergroup anger than intergroup fear, F(1,53) = 150.48), p < .001, ηp2 = .74.
  • Hypothesis 3a: Intergroup motivation across condition

We predicted that women in the bystander sexism condition would report greater intergroup motivations to move against and away from men than women in the control condition. In order to examine this hypothesis, we conducted a one-way MANOVA of the intergroup motivation to move against men and motivation to move away from men scales with condition as a between-subject factor. This analysis supported our hypothesis, demonstrating that women in the bystander sexism condition reported more intergroup emotions compared to women in the control condition, Wilks’ λ = .40, F(2, 97) = 73.91, p < .001. Follow-up univariate tests indicated that women in the bystander sexism condition reported both more intergroup motivation to move against men, F(1, 98) = 78.42, p < .001, and intergroup motivation to move away from men, F(1, 98) = 98.39, p < .001, compared to women in the control condition.
  • Hypothesis 3b: Intergroup motivation to move against men vs. motivation to move away from men in bystander sexism condition

We hypothesized that women in the bystander sexism condition would report greater motivation to move against men compared to motivation to move away from men. In order to examine this hypothesis, we conducted a one-way ANOVA among women in the bystander sexism condition using intergroup motivation (move against vs. move away from) as a within-subject factor. Contrary to our hypothesis, women in the bystander sexism condition reported less motivation to move against men compared to their motivation to move away from men, F(1,53) = 14.70, p < .001, ηp2 = .22.

Given that this finding was contrary to our hypothesis, we wanted to further examine whether this effect occurred only among women in the bystander sexism or if it occurred for women in the control condition as well. We conducted an additional one-way ANOVA among women in the control condition using intergroup motivation (move against vs. move away from) as a within-subject factor. This analysis reveals that women in the control condition demonstrate the same effect—they report a greater desire to move away from men than move against them, F(1,45) = 8.55, p < .01, ηp2 = .16. Thus, across both experimental conditions, women demonstrated greater motivation to move away from than against the outgroup of men.
  • Hypothesis 4: Mediation analysis of intergroup anger and motivation to move against men

In accordance with intergroup emotions theory (Smith 1993, 1999; Mackie et al. 2000), we hypothesized that intergroup anger, but not intergroup fear, would mediate the effect of condition on the motivation to move against men. We utilized procedures outlined in Preacher and Hayes (2008) to examine bootstrapping estimates of the indirect effects in a multiple mediator model. These procedures allow us to estimate the respective effect of both intergroup anger and intergroup fear simultaneously. Further, bootstrapping allows us to derive estimates of the indirect or mediated effects from a sampling distribution (Shrout and Bolger 2002), and it is generally preferred over the causal steps approach (i.e., Baron and Kenny 1986) when dealing with relatively small sample sizes.

Consistent with hypotheses, the indirect effect of condition on motivation to move against men through intergroup anger was significant, B = 2.13, SE = .36, p < .05, 95% BCa bootstrap CI: 1.41, 2.82, while the indirect effect through intergroup fear was not significant, B = −.15, SE = .12, n.s., 95% BCa bootstrap CI: −.41, .08. The direct effect of condition on motivation to move against men was not significant (B = .25, SE = .30, n.s.). Unstandardized path coefficients are presented in Fig. 1. In sum, intergroup anger, but not intergroup fear, mediated the effect of condition on motivation to move against men.
https://static-content.springer.com/image/art%3A10.1007%2Fs11199-009-9735-0/MediaObjects/11199_2009_9735_Fig1_HTML.gif
Fig. 1

Mediation analysis of condition on motivation to move against men. Path coefficients are unstandardized. **p < .01.

Additional Meditational Analyses

In addition to providing a direct test of this hypothesis, we also conducted an additional mediation analysis in order to provide additional, convergent support for our main prediction. Because we expected that intergroup anger would be the predominant intergroup emotional response to an instance of bystander sexism, we can also expect that intergroup fear should not mediate the effect of condition on the motivation to move away from men. Further, intergroup anger should not mediate the effect of condition on its opposite behavioral motivation to move away from men.

Consistent with hypotheses, neither the indirect effect through intergroup anger, B = .19, SE = .31, n.s., 95% BCa bootstrap CI: −.49, .77, nor intergroup fear, B = −.06, SE = .17, n.s., 95% BCa bootstrap CI: −.38, .29, mediated the effect of condition on motivation to move away from men. The direct effect of cat-call on motivation to move away from men was significant, B = 2.55, SE = .43, p < .001. Thus, this analysis demonstrates that neither intergroup emotion mediated the effect of condition on motivation to move away from men.

Finally, if our theorizing is correct and bystander sexism leads women to shift from an individual- to a group-level of self-categorization and experience emotions on behalf of their gender group, it should also be the case that women in the bystander sexism condition should not experience greater individual-based emotions than women in the control condition. A multivariate analysis of variance (MANOVA) supports this assumption. This multivariate analysis indicates that there were no differences in state negative affect based on experimental condition (Wilks’ λ = .94, F(3, 96) = 2.22, n.s.), and the follow-up univariate tests confirmed this (all ps > .15).

Discussion

The current study examined women’s group-based reactions to overhearing a cat-call remark. We drew on insights from social identity perspectives (Tajfel and Turner 1979; Turner et al. 1987) and intergroup emotions theory (Mackie et al. 2000; Smith 1993, 1999) to examine whether women experienced group-based emotions and behavioral tendencies in response to this instance of bystander sexism. Our results suggest a marginally significant trend for women’s gender identity to be more likely to become salient in instances of bystander sexism. Thus, women may be more likely to think about themselves in terms of their gender group identity and, therefore, react to an instance of bystander sexism based on their group-level concerns. Our results confirm that women experienced greater intergroup emotions (i.e., anger and fear) and motivations towards the outgroup of men (i.e., move against and move away from) when they were an imagined bystander to a cat-call comment. Consistent with our expectation that intergroup anger would be the predominant emotional response to bystander sexism, women in the cat-call condition reported feeling more anger towards men than fear. Further, the effect of intergroup anger mediated the relationship between condition and motivations to move against or oppose men, demonstrating that greater anger towards men accounted for women’s motivation to oppose them. However, intergroup fear did not mediate the relationship between condition and motivations to move away from men, a finding that lends additional support to our hypothesis that intergroup anger would be the predominant emotional response to bystander sexism. Together, these results provide new insight into the consequences of bystander sexism and the utility of conceptualizing sexist incidences from an explicit intergroup framework.

One unexpected finding in our data was that women who overheard the cat-call comment reported more motivation to move away from men relative to their motivation to move against men. However, this relative difference in motivations towards men occurred across both conditions, meaning that participants were more inclined to move away from the outgroup of men than to move against them regardless of what they heard a man say. We offer two possible explanations for these results. First, it may be the case that although the cat-call elicited more intergroup anger than fear, it was not sufficiently severe enough to produce marked increases in the desire to oppose men directly. Perhaps a more derogatory cat-call comment or a more severe sexist incident (e.g., direct sexual coercion by a coworker) would have produced greater motivations to oppose men compared to motivations to avoid them. Secondly, these results may simply demonstrate the tendency for group members to refrain from taking direct actions against offending outgroup members (e.g., Hyers 2007; Wright et al. 1990) or the tendency for women to refrain from engaging in confrontational behavior (e.g., Rudman 1999), perhaps due to gender role prescriptions (Henley 1977).

The current study highlights the impact of group identities in instances of prejudice. Previous work has focused on the impact of group identities from the perspective of the sexist perpetrator and has demonstrated that (male) identity concerns often lead men to enact sexist behavior (Hitlan et al. 2009; Maass et al. 2003; Pryor and Whalen 1997). From the perspective of the target, however, researchers have tended to emphasize the role of chronic group identification in affecting women’s individual reactions to sexism (Cameron 2001; McCoy and Major 2003). Thus, our study’s emphasis on the effect of situational salience of gender group identity is a new contribution and underscores the need for future research that examines how group identities are implicated in responses to sexist incidents.

Our results also provide new information regarding the psychological consequences of bystander sexism, an area of research that has received little empirical attention. Women experience a variety of negative consequences as the direct targets of sexism (e.g., Shelton and Stewart 2004; Swim and Hyers 1999), and the current study demonstrates that bystanders are affected by instances of sexism as well. Although women may not feel greater negative emotional reactions directed inward in response to bystander sexism, they may feel greater negative emotions directed outwards towards men. Thus, this study provides some evidence that the effects of sexism are not confined solely to the target of prejudice when the sexist event occurs in a public setting. Instead, our results indicate that for every woman who is a direct target of sexism, there may be several other women who witness the event and are also affected as bystanders.

By positioning the current work within an intergroup framework, this study also emphasizes the utility of exploring the nature of group-based reactions to sexism. Most extant work has focused on individual reactions to sexism such as individual level negative affect and self-esteem (e.g., Swim and Hyers 1999), an emphasis that may fail to identify how specific instances of sexism affect how men and women perceive and interact with each other more generally. The current study demonstrates that women’s emotions and motivations towards men become more negative when they are bystanders of sexism. These results not only demonstrate that women are affected as bystanders of sexism, but they also suggest that women’s feelings and behaviors towards all men can be affected by the actions of a single man. That is, the actions of one sexist man can impact how female bystanders may perceive and interact with other men. From this perspective, instances of prejudice negatively impact bystanders in both groups; female bystanders may react negatively towards men and male bystanders may be perceived negatively because of the actions of a single sexist man.

Our study also extends current theorizing about group-based emotions. Previous work on intergroup emotions has largely focused on how people perceive and react to threats from collective outgroups on one’s ingroup (e.g., Mackie et al. 2000), but has largely overlooked situations in which one’s ingroup is threatened by an individual outgroup member. Any situation in which the individuals involved perceive themselves to be acting on behalf of their group may implicate group identity and be perceived and reacted to in terms of that group identity (Tajfel and Turner 1986; Turner et al. 1987). In the case of sexism, for example, women’s social identity may become salient when men as a group threaten them (e.g., acknowledgement of pervasive gender discrimination) and when an individual man who is perceived to be acting on behalf of his gender threatens an individual woman (e.g., cat-call comments). However, this latter type of situation has not been examined within the context of intergroup emotions. By exploring the role of group-based emotions and behavioral tendencies in a context where an individual outgroup member threatens one’s ingroup, the present work extends the utility of intergroup emotions theory to new contexts that invoke social identity threats.

It is important that these results be interpreted with several limitations in mind. In the current study design, women were not physically present as bystanders in the sexist situation; rather, they imagined themselves in this situation via a video recording. Given that participants typically cannot accurately predict their reactions to sexist situations when they are not physically present in the situation (e.g., Shelton and Stewart 2004), current results may overestimate the anger or underestimate the fear women would experience had they overheard the cat-call remark in person. Thus, the ability of these results to generalize to women’s reactions in real world settings is limited. Further, gender identity was only marginally more salient in the cat-call condition compared to our control. Additional work that replicates this pattern at conventional levels of statistical significance is needed in order to draw firm conclusions about this effect. Finally, our reliance on an undergraduate sample of women limits our ability to generalize these findings across women, more broadly. Future research that examines these processes among middle- and late-adult aged women who, presumably, have had more chances to be bystanders to sexism could offer one method to address this concern.

Nonetheless, the results from this study point to several promising directions for future research in the areas of sexism and intergroup emotions. Our study only examined the effect of bystander sexism on two intergroup emotions—anger and fear. Future work may benefit by examining other relevant types of negative intergroup emotions such as disgust (Cottrell and Neuberg 2005). Further, given that cat-calls are relatively ambiguous events that could be interpreted as complimentary (Bowman 1993), examining positive emotional reactions may also be a fruitful area for future work. In addition to intergroup emotions and motivations, additional research examining bystanders’ likelihood of intervening in specific sexist incidents or likelihood of contributing to broader efforts to reduce sexism are also interesting areas for future work.

Ultimately, the current work underscores the notion that sexism can be bad for everyone. Women are obviously implicated because they often suffer direct negative consequences as targets of prejudice and, as the current work demonstrates, indirect consequences as bystanders. But sexism also harms men as well. Whenever a single man’s prejudiced actions make gender identity salient, male perpetrators can impact how women view and react to men more generally. From this perspective, sexist instances do not occur in a social vacuum wherein a single perpetrator and target interact. As numerous researchers have already demonstrated, sexual harassment in a work environment can negatively affect women as direct targets and bystanders (e.g., Fitzgerald 1993; Hitlan et al. 2006), can compromise the organizational climate, and can, ultimately, be financially costly to organizations (for a review, see Terpstra and Baker 1986). Thus, our study adds to the growing literature illustrating that individual incidences of sexism can have wide-ranging and deleterious consequences.

Acknowledgments

This work is based on the master’s thesis of the first author. We thank Sjoerd Pennekamp for his helpful comments on a previous draft of this manuscript, Jack Dovidio and Janet Barnes-Farrell for their guidance as thesis committee members, and Lindsay Aronheim, Matthew Barry, Daniel Butler, Elizabeth Fabrizi, Randi Ferguson, Nick Frogley, and Sarah Pennington for their help with data collection. Portions of this research were presented at the 6th and 7th Annual Meetings of the Society for Personality and Social Psychology.

Copyright information

© Springer Science+Business Media, LLC 2010