Power and the confrontation of sexism: the impact of measured and manipulated power on confronting behavior

ABSTRACT The confrontation of prejudicial acts and comments promotes multiple benefits, most notably the prevention of future prejudicial remarks and the reduction of stereotype use. Research, however, consistently shows low rates of confronting prejudice, particularly regarding sexism. Here, we examine whether personal sense of power, known to increase action and activate the behavioral approach system, increases the likelihood of confronting a sexist remark. In Study 1, we demonstrate that for both women and men, self-reported power is associated with a higher frequency of confronting sexism. In Study 2, we manipulate women’s sense of power (i.e., high power, low power, or control) and subsequently present an opportunity to confront a sexist remark. Results show that women primed to feel powerful were more likely to confront the sexist remark and expressed greater disagreement with the comment, compared to women primed to feel powerless. Implications for the confronting literature and behavior are discussed.

Sexist comments and behaviors are pervasive, with women reporting explicit incidents of sexism as frequently as once or twice a week (Hyers, 2007;Swim et al., 2001).These prejudicial acts, in turn, have detrimental effects on mental and physical health with research linking experiences of prejudice to higher rates of depression, lower self-esteem, and greater negative emotionality (Pascoe & Smart Richman, 2009;Wang et al., 2011).As such, it is important to investigate means of curtailing the use and expression of prejudice.One such effective and enduring means of countering sexist acts is confronting, defined as "verbally or nonverbally expressing one's dissatisfaction with prejudicial and discriminatory treatment to the person who is responsible for the remark or behavior" (Shelton et al., 2005, p. 67).While efficacious in reducing future stereotype use and prejudiced comments (Chaney & Sanchez, 2017;Czopp & Monteith, 2003), it is uncommon for people to decide to confront another person's prejudiced behavior, even though individuals report recognizing the prejudice and stating they would confront in a hypothetical scenario (Ayres et al., 2009;Brinkman et al., 2011;Kawakami et al., 2019;Shelton & Stewart, 2004;Swim & Hyers, 1999;Woodzicka & LaFrance, 2001).
We posit that one factor important to understanding when an individual chooses to confront is an individual's sense of power.Indeed, previous research notes that power is associated with a host of behavioral and cognitive effects that increase the likelihood an individual chooses to act (Guinote, 2017;Keltner et al., 2003).Here, we apply research and theory on the psychological processes of power to confronting behaviors, examining whether there is an association between individuals' personal sense of power and the frequency of confronting behavior, and whether priming power increases the likelihood that women will confront a sexist remark.Thus, the current research advances our understanding of the impact of power on behavior and seeks to examine how we can increase individuals' willingness to enact a potent prejudice reduction strategy -the confrontation of prejudice.

Confronting as a prejudice reduction strategy
As a prejudice reduction strategy, confronting instances of bias and prejudice is highly effective (Barreto & Ellemers, 2015;Burns & Monteith, 2018;Chaney et al., 2015), mitigating individuals' subsequent use of prejudicial statements and stereotypes (Czopp & Monteith, 2003;Mallett & Wagner, 2011) up to one week past the initial confrontation (Chaney & Sanchez, 2017).Furthermore, evidence showcases that confronting one form of bias (e.g., a racist stereotype) can reduce the future use of stereotypes related to other groups (e.g., gender stereotypes), highlighting the breadth of this prejudice reduction strategy (Chaney et al., 2021).Confronting is also associated with a host of other benefits such as exposing prejudice that otherwise may have gone unnoticed (Becker & Swim, 2011), engendering greater feelings of satisfaction (Dickter, 2012), and reducing guilt for the individual who confronts (Shelton et al., 2005).Witnesses of confrontations also report greater norms of egalitarianism (Blanchard et al., 1994), higher perceived offensiveness of the comment (McClelland & Hunter, 1992), and decreased self-reported sexist attitudes (Boysen, 2013).Overall, confronting is a potentially powerful prejudice reduction strategy with positive effects for the confronted, confronters, and witnesses.
Yet, a consistent finding across the literature is that confronting prejudice is infrequent (Ayres et al., 2009;Brinkman et al., 2011;Shelton & Stewart, 2004;Swim & Hyers, 1999;Woodzicka & LaFrance, 2001).Indeed, Swim and Hyers (1999) found that only 15% of women directly confronted a man who asked a sexist question in a mock job interview, despite many identifying the comment as sexist and expressing a desire to confront.More recent findings also support this low rate of direct confronting of sexism, with only 13% of participants confronting an instance of sexism immediately after it occurred (Woodzicka & Good, 2020).Researchers have identified multiple barriers that reduce the likelihood that a sexist remark will be confronted.Beyond the need to identify the situation as sexist and interpret it as a moment when action must be taken (Ashburn-Nardo et al., 2008), individuals must also weigh the potential benefits and costs to speaking up.Indeed, multiple findings suggest that individuals, after confronting a prejudicial comment, face more negative evaluations such as being labeled a complainer (Kaiser & Miller, 2001) and receive more negative evaluations from others (Dodd et al., 2001), especially when these comments relate to a positive stereotype (Alt et al., 2018).This cost-benefit analysis, in turn, plays a key role in decisions to confront sexism, with existing evidence showing that women who associate greater benefits, and fewer costs, are more likely to confront sexism in their everyday lives (Good et al., 2012).Overall, individuals often recognize and are motivated to confront sexism, however barriers exist which prevents them from confronting.

Power and the confrontation of sexism
Thus far, researchers have examined rates of confronting (Swim & Hyers, 1999), potential benefits of confronting (Czopp & Monteith, 2003;Czopp et al., 2006), and reasons why individuals choose not to confront (Good et al., 2012).A growing area of interest seeks to examine means of increasing confronting behavior, particularly as it relates to actual, as opposed to hypothetical, confrontations of prejudice (Mallett & Monteith, 2019;Rattan & Dweck, 2010;Szekeres et al., 2019).Here, we posit that one strategy to enhance confronting behavior is by manipulating a potential confronters' sense of power, that is, the state or feeling that one has some resource or control over another.
Prior to considering the role of power in confronting behavior, it is important to first reflect on the interplay between power, gender, and sexism.In society, gendered power inequalities persist, with women experiencing low representation in positions of power within education, employment, and politics (Milazzo & Goldstein, 2019).For instance, women are less likely to become leaders than men, which sets the prototype for leadership as male (Badura et al., 2018).This outcome impacts the perception of women negatively by making then appear less competent, and thus less likely to be elected to positions of high power, forming a vicious cycle (Esteve-Volart & Bagues, 2012).This cycle subsequently perpetuates sexism.Seeing women in positions of low power creates the stereotype of lower competency, which creates the expectation that women should stay in certain job positions and behave in certain ways.Women who conform to expectations may be perceived positively yet are still seen as inferior and may experience instances of benevolent sexism as a result (Glick & Fiske, 2001).Thus, sexism can lead to gendered power disparities and vice versa.While these societal and structural links between power and gender are important, here, we examine whether interpersonal power may be harnessed to combat sexism.
Power, at the interpersonal level, is associated with a host of behavioral and cognitive effects organized around a general principle that power increases activation of the behavioral approach system (Guinote, 2017;Keltner et al., 2003).For instance, power induces individuals to take action and reduces individuals' cognitions regarding the constraints imposed by a situation (Galinsky et al., 2003;Guinote, 2007).In prior work, individuals primed to feel powerful showed greater approach motivations via a reduced inhibition, fewer cognitions about social constraints, and a diminished influence of others' opinions (Galinsky et al., 2008;Whitson et al., 2013).Individuals primed to feel high power are also more likely to take action such as removing an annoying stimulus from their environment (Galinsky et al., 2003) and generate fewer "prefactuals" or "what-ifs" when making action-oriented decisions (Scholl & Sassenberg, 2015).Furthermore, individuals in power are more likely to express their personal opinions and attitudes (Berdahl & Martorana, 2006), demonstrate lower levels of conformity (Galinsky et al., 2008;Hays & Goldstein, 2015), and confront instances of workplace incivility (Hershcovis et al., 2017) compared to those not in power.Generalized across these effects, power is associated with greater approach compared to avoidance behaviors, particularly when taking action aligns with one's goals and motivations (Guinote, 2017).
Applied to the confrontation of sexism, these findings suggest that power may be uniquely associated with confronting behaviors, such that high-power individuals are more likely to take action -confronting and expressing their disagreement with a sexist comment.Indeed, recent work has investigated various constructs associated with power in relation to confronting such as status, social network centrality, and authority (Ashburn-Nardo et al., 2014;Brands & Rattan, 2020;Gervais et al., 2010).For instance, in a set of studies, Ashburn-Nardo et al. (2020) demonstrated that individuals perceived to have greater authority were viewed as having more responsibility to confront, although putting participants in a position of authority did not increase individuals' actual confronting behaviors.In addition, perpetrator status (i.e., boss or coworker), an indication of power, affected participants' confronting intentions in a hypothetical scenario, with participants less likely to confront the boss compared to the coworker after they made a sexist remark (Ashburn-Nardo et al., 2014).Collectively, these studies highlight how our social position and roles impact our willingness to confront prejudice, broadly suggesting that the conferral of higher social positions increases the likelihood of confronting behavior.To the best of our knowledge, however, no study has directly examined how power impacts individuals' frequency and willingness to confront sexist comments; thus, here we answer calls (Mallett & Monteith, 2019) to examine whether power affects confronting behaviors.

The present research
In the present research, we investigated the effects of power on confronting behavior.First, we examined whether women (Study 1a) and men's (Study 1b), self-reported personal sense of power was positively associated with confronting behaviors in their daily lives.In Study 2, we tested whether priming power influenced confronting behavior by investigating whether women primed with high power confronted, and expressed greater disagreement, with a sexist remark, compared to those primed with low power or those in a control condition.Data and materials for all studies are available: https://osf.io/tuknq

Study 1a and 1b
Across two studies, we tested whether self-reported personal sense of power was associated with selfreported frequency of confronting sexist acts for women (Study 1a) and men (Study 1b).In addition, we combined these two datasets to assess whether participant gender moderated the association between personal sense of power and the frequency of confronting sexist acts.Overall, we hypothesized that individuals who reported higher levels of personal sense of power in their everyday lives would also report higher frequencies of confronting sexism directed toward women.If true, this would reflect evidence of an association between power and confronting sexism.

Participants
We calculated the required sample size using G*Power (3.1.9.4) with our power = .95,α = .05,and an anticipated correlation of 0.2 (Faul et al., 2009).This indicated we would need a total sample size of 319 participants which we increased to 350 to offset the possibility of participant attrition.We used this sample size for both Study 1a and Study 1b.
For Study 1a, we recruited 350 women on MTurk (M age = 35.81,SD age = 11.84;5.7% African American, 9.4% Asian American, 3.4% Latinx American, 0.3% American Indian or Alaska Native; 74.3% White; 3.4% Multi-racial, 3.4% Other).We limited study recruitment to women by using a prestudy survey whereby participants who did not indicate their gender identity as woman were not allowed to continue.Even with this check, two participants identified as men in the post-study demographics and were excluded from analyses.This left a total sample size of 348.
For Study 1b, we recruited 350 men (M age = 39.50,SD age = 12.33; 5.8% African American, 12.9% Asian American, 4.7% Latinx American, 0.3% American Indian or Alaska Native; 71.6% White, 3.9%, Multi-racial, 0.9% Other) on CloudResearch, a panel-based online recruitment website (Litman et al., 2017).CloudResearch's panels specifically recruited only men for the survey.We excluded five participants who failed a manipulation check and three participants who had incomplete data, leaving a total sample size of 342.For both studies, written informed consent was obtained from all participants and all procedures were approved by the primary author's Institutional Review Boards.

Materials and procedure
For Study 1a, participants responded to questions assessing the frequency of confronting sexism in their daily lives followed by a survey on self-reported personal sense of power.We also assessed participants' frequency of experience with sexism and chronic perceptions of sexism.For all measures we created a mean score.Lastly, participants reported demographic information and were debriefed.
For Study 1b, participants completed the frequency of confronting sexism, self-reported personal sense of power, and the frequency of experience with sexism scales.We did not include the chronic perceptions of sexism given the specific nature of the questions related to women's experiences.In this study, the order of the scales was randomized. 1 As a slight modification, for the frequency of confronting sexism and frequency of experience with sexism we made clear to participants that the sexism was directed toward women (as opposed to interpreting the questions to assume the sexism was directed toward men).For all measures we created a mean score.Lastly, participants reported demographic information and were debriefed.

Frequency of confronting sexism
The Frequency of Confronting Sexism scale (Good et al., 2012) consists of eight items (Study 1a: α = .82;Study 1b: α = .90).Sample items include: "When you have experienced sexism, how often have you confronted the sexist person?,""How often have you told someone to stop when they are telling a sexist joke or making fun of you because of your gender?."In Study 1b sample items include: "When you have witnessed sexism directed towards women, how often have you confronted the sexist person?,""How often have you told someone to stop when they are telling a sexist joke or making fun of women?." Participants responded using a 7-point Likert scale (1 = never; 7 = every time).

Personal sense of power
The Personal Sense of Power scale (Anderson et al., 2011) consists of eight items (Study 1a: α = .92;Study 1b: α = .94)assessing participants' degree of power, sample items include: "I can get people to listen to what I say.,""I think I have a great deal of power."Participants responded using a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).

Frequency of experiencing sexism
The Frequency of Experienced Sexism scale (Good et al., 2012) consists of nine items (Study 1a: α = .94;Study 1b: α = .95).Sample items include: "How often have you been harassed because of your gender?," "In the past year, how often have you experienced sexism?."In Study 1b sample items include: "In the past year, how often have you witnessed sexism directed towards women?,""In the past year, how often have you seen a woman be harassed because of their gender?." Participants responded using a 7-point Likert scale (1 = never, 7 = every day).

Chronic perception of sexism
The Chronic Perception of Sexism scale (Townsend et al., 2011) consists of four items (α = .82).Sample items include: "My gender group is discriminated against," "I experience discrimination because of my gender."Participants responded using a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).

Results
To examine whether participants' personal sense of power predicted frequency of confronting we conducted regression analyses (see, Table 1 for means, standard deviations, and correlations).For Study 1a, we regressed Frequency of Confronting Sexism on Personal Sense of Power.The overall model was significant, R 2 = .05,F(1, 346) = 17.32, p < .001,such that women who had higher personal sense of power reported confronting sexism more frequently, b = 0.21, t(346) = 4.16, p < .001.Results from Study 1b revealed a similar finding, R 2 = .03,F(1, 340) = 9.76, p = .002,such that men who had higher personal sense of power reported confronting sexism (directed toward women) more frequently, b = 0.17, t(340) = 3.13, p = .002(see, Figure 1).Collectively, these findings reveal that for both women and men, personal sense of power was a statistically significant predictor of their frequency of confronting sexism.We also ran additional models where we included Frequency of Experiencing Sexism (Study 1a and 1b) and Chronic Perceptions of Sexism (Study 1a only) as predictors, along with Personal Sense of Power, to determine whether power predicted confronting sexism over and above these measures.For Study 1a, results showed that Frequency of Experiencing Sexism was a significant predictor, b = 0.37, t (329) = 5.46, p < .001,while Chronic Perceptions of Sexism was not, b = −0.08,t( 329 In addition, we combined the two datasets to examine whether gender moderated the relationship between personal sense of power and confronting sexism.As a note of caution, we did not originally design these studies to compare across these two groups as the time of data collection differed between the two groups (2018 for women, 2020 for men).For this analysis, we dummy coded Participant Gender (0 = man, 1 = woman) and used this variable to create an interaction term with Personal Sense of Power.Results revealed no significant interaction, b = 0.04, t(686) = 0.53, p = .600,indicating that Personal Sense of Power operated similarly for men and women.Specifically, when we regressed Participant Gender and Personal Sense of Power (with no interaction term) on Frequency of Confronting Sexism there was a significant effect of Participant Gender, such that women reported confronting sexism more often than men, b = 0.55, t(687) = 6.18, p < .001,while Personal Sense of Power was still a significant predictor, b = 0.19, t(687) = 5.10, p < .001.

Discussion
Studies 1a and 1b revealed that for both women and men, higher personal sense of power predicted greater frequency of confronting sexism in their daily lives.This association remained even after including frequency of experiencing sexism (for women and men) and chronic perceptions of sexism (for women).Interestingly, we found that this association between power and confronting sexism for both women and men, a point we return to later in the General Discussion.Overall, Studies 1a and 1b provide evidence to the assertion that power is associated with the confrontation of sexism.

Study 2
In Study 1 we found an association between power and the frequency of confronting sexism -the more powerful one felt in their daily lives, the more they reported confronting behaviors.This initial evidence supported our hypothesis that power is related to confronting; however, it was correlational and reliant on retrospective self-report.Thus, in Study 2 we experimentally primed participants' sense of power and subsequently provided them with an opportunity to confront a sexist comment.Given research on power and its association with taking action, as well as activation of the behavioral approach system (Guinote, 2017), we hypothesized that women who were primed to feel high power would be more likely to confront, and would express greater disagreement, with a sexist comment compared to women who were primed to feel low power.We also added a control condition allowing us to test the directionality of our effects.
In order to unobtrusively prime participants' sense of power, we randomly assigned participants to write about a time they felt high power, low power, or what they did the day before (i.e., control).This manipulation has been successful in eliciting participants' feelings of power (Galinsky et al., 2003;Scholl & Sassenberg, 2015;Whitson et al., 2013).In a subsequent, yet ostensibly unrelated study, we created an opportunity to confront a sexist remark by employing an online communication paradigm that has been successfully used in previous confronting research (Czopp et al., 2006;Kroeper et al., 2014).Specifically, participants interacted with a confederate through an online instant messaging system (i.e., Skype).During the interaction, the confederate made a sexist remark to which participants were allowed to respond.This format allowed us to ensure all participants saw the same sexist remark, delivered in a uniform manner, and that all participants had an equal opportunity to respond to the comment.

Participants
For Study 2, an exact effect size was unknown as this was the first study to examine priming power and the confrontation of sexism.We, however, sought to recruit at least 60 participants per condition.In total, 186 female participants were recruited for this study from a large U.S. public university on the west coast.We excluded participants who indicated suspicion that they were not talking to another participant (n = 6) and participants who, according to experimenters, had made an error (e.g., incorrectly followed procedure, n = 4).This left a total of 176 (M age = 20.70,SD age = 3.34, 32.4% East Asian, 14.8% White, 13.1% Southeast Asian, 10.2% Latina, 10.2% South Asian, 9.1% Multi-racial, 2.3% Black).Written informed consent was obtained from all participants and all procedures were approved by the Institutional Review Board.

Procedure
Upon arriving to the lab participants were informed that they would be interacting with another participant in a study examining online communication and decision-making.This procedure was employed so that participants never interacted with their partner face-to-face at any point in the study and only interacted via online chat.When participants arrived in the lab, the experimenter informed the participant that they had received an e-mail saying that their interaction partner was delayed.At this point, participants were offered to participate in a separate five-minute study investigating the effects of writing on emotional outcomes for an additional $1 compensation.This procedure was used to convince participants they were taking part in two separate experiments, so they would not suspect that both studies were connected.
After agreeing to participate in the writing study (all participants agreed), the experimenter opened a Qualtrics survey.The survey randomly assigned participants to one of three conditions, high power, low power, or control.In the high-power condition, participants wrote via an open textbox embedded in the survey, about a time when they had power over another person.In the low power condition, participants wrote about a time when someone had power over them.In the control condition, participants were asked to recall what they did the previous day (see Galinsky et al., 2003).To ensure participants' attention we asked participants to write a minimum of 500 characters, approximately 100 words.Following the writing task, participants completed the Positive and Negative Affective Schedule (PANAS; Watson et al., 1988), where they rated various emotional words (e.g., happy, upset, etc.) on a five-point Likert scale (1 = very slightly or not at all, to 5 = extremely) in which "powerful" and "strong" were embedded.The PANAS served as both a manipulation check on power as well as an aid to maintain the study cover story.
Approximately three minutes after participants began the power manipulation, the experimenter, in a loud voice, ostensibly greeted the other participant ("Hello, are you here for the online interaction study?") and opened and closed a separate lab door to simulate them entering an adjacent room.In actuality, there was no other participant.After participants completed the writing study, they were told that their partner for the original study had arrived.Participants were then given a packet that contained instructions informing participants that they and their partner would read a series of scenarios and discuss each scenario via Skype's instant messenger program (a program that allows exchanges of written messages in real-time).For this exchange, the experimenter acted as the partner and completed the chat interaction with scripted responses.For each scenario, participants first read brief information about two individuals (e.g., names, hobbies, age, and academics) and then made decisions about which individual should perform one of two tasks.Participants were told they were randomly selected to always respond to their partner's decision such that the participant could agree or disagree with their partner's decision and offer a reason for the agreement or disagreement.This procedure allowed all participants a chance to respond to the sexist comment.Participants first read a filler scenario in which a man (Alex) and a woman (Sarah) had to be assigned to tasks of "writing a class paper" and "making a class presentation."This served as a practice of the chat procedure for the participant.The scenario of interest always came next and was the last interaction.In the scenario, modeled after Dodd et al. (2001), the participant's partner (i.e., the experimenter) was tasked with assigning a man (Ryan) and a woman (Lauren) to the tasks of "setting up the tent" and "preparing the meal."In response to this scenario, the experimenter, playing the role of the other participant, typed, "Lauren should take care of the cooking because she is a woman."The participant then typed their response, after which the interaction ended.We recorded participants' typed responses to this comment for coding later.After completing the online interaction, participants were instructed to precede to an online survey where they were presented with open text boxes.In these text boxes participants responded to questions regarding what they thought the hypothesis of the study was (no participants correctly identified that the two studies were linked) and if they had anything they wanted to let the experimenter know (at this point six participants stated they did not think they were speaking with a real participant, these participants were excluded). 2Following the online survey, participants provided demographic information (e.g., age and race/ethnicity), were fully debriefed, and compensated with a small monetary amount.

Coded reponses
Aligning with previous research on coding participants' responses to prejudicial comments (Lee et al., 2012), participants' responses to the sexist comment were independently coded by two research assistants (RA), who were blind to hypothesis and condition.The RAs rated the degree to which the participant confronted the sexist remark on a five-point scale: 1 = she made a similar comment, endorsing what the person said, 2 = she showed some agreement with the comment, 3 = she did nothing, ignored the comment, 4 = she showed some disagreement with the comment 5 = she verbally reprimanded the person by telling them that they were wrong or that they were offended.In addition, the RAs rated the degree to which participants expressed disagreement with the sexist remark (1 = not at all disagreed, 7 = very strongly disagreed).These measures were moderately correlated, r(176) = 0.68, p < .001.Discrepancies reflecting a difference score of two or more were discussed with the RAs and the first author until consensus was reached. 3Interrater reliability, assessed via the intraclass correlation coefficients, was high for confronting score (ICC = .89)and disagreement score (ICC = .93).

Manipulation check
A one-way analysis of variance (ANOVA) was conducted to test whether participants' mean score for the items "powerful" and "strong" (r = .66)differed based on the power conditions.Results revealed a significant difference across the three conditions, F(2, 171) = 4.26, p = .016,η 2 p = .05.Planned contrasts showed that women in the high power condition felt more powerful (M = 2.70, SE = 0.14) compared to those in the low power condition (M = 2.27, SE = 0.14) and those in the control condition (M = 2.16, SE = 0.14), Fs(1, 171) = 4.72 and 7.69, ps > .031,η 2 p s = .03and .04,respectively.

Confront and disagreement score
We tested the hypothesis that the high power prime, compared to the low power prime and control condition, would increase participants' confrontation of the sexist remark and lead to greater expression of disagreement by conducting two between-subjects ANOVAs.First, for the confront score, results revealed significant differences across the three conditions, F(2, 173) = 3.32, p = .039,η 2 p = .037.Planned contrasts (high power compared to low power) revealed that women who were primed to feel high power had higher confront scores (M = 4.07, SE = 0.10) than women who were primed to feel low power (M = 3.69, SE = 0.10), F(1, 173) = 6.18, p = .014,η 2 p = .03.Additionally, planned contrasts comparing high power to the control condition (M = 3.79, SE = 0.11), while trending in the hypothesized direction, were not statistically significant, F(1, 173) = 3.36, p = .069,η 2 p = .02.There was no significant difference between low power and control, F(1, 173) = 0.46, p = .501,η 2 p = .003(see, Figure 2).

General discussion
Across three studies we investigated whether measured (Studies 1a and 1b) and manipulated (Study 2) power was associated with a greater likelihood of confronting sexist acts and remarks.Studies 1a and 1b provided initial evidence for such a relationship, demonstrating that women and men who reported higher personal sense of power also reported more confronting sexism in their everyday life.In Study 2, we experimentally manipulated power to examine its effects on actual confronting behavior.Results revealed that women primed to feel high power, compared to low power, were more likely to confront and expressed greater disagreement with a sexist statement.Additionally, for confronting behavior, while not statistically significant, results trended in the direction that high power primes, compared to a control prime, led to increased confronting behavior.Together, these results suggest that power is positively associated with, and increases, the likelihood of confronting and expressing disagreement with sexism remarks and behaviors.These studies highlight one potential strategy to overcome the common finding that prejudice frequently goes unchallenged, especially when it comes to real-world behavior (Ayres et al., 2009;Brinkman et al., 2011;Shelton & Stewart, 2004;Swim & Hyers, 1999;Woodzicka & LaFrance, 2001).The fact that power is associated with confronting behavior and can increase its likelihood opens access to the numerous benefits of confronting behaviors such as the enduring reduction of future prejudicial remarks (Chaney & Sanchez, 2017;Chaney et al., 2021;Czopp & Monteith, 2003;Mallett & Wagner, 2011), boosts to confronters' satisfaction (Dickter, 2012), and protection against psychological distress (Sanchez et al., 2015).Furthermore, by increasing feelings of autonomy and decreasing rumination after experiencing discrimination, confronting behaviors may promote mental and physical health (Chaney et al., 2015).Given the regular expression of sexist comments and actions (Hyers, 2007;Swim et al., 2001), it is important to investigate means of increasing their confrontation, with our studies suggesting power as a potential tool.One beneficial outcome is that confronting behaviors may start a cycle of active resistance, leading to greater empowerment (Chu, 2017) and other actions against forms of prejudice and discrimination.Still, it is also important to consider the costs of confronting behaviors, since depending on the situation these costs may not solely be reputational (Kaiser & Miller, 2001) but may also be physical (e.g., personal safety), mental-health related (e.g., personal safety, stress of being the only individual challenging prejudice) and material (e.g., loss of a job).
Our findings also align with and extend the broader literature on power and social behavior, particular regarding the effects of power on action and the behavioral approach system (Cho & Keltner, 2020;Guinote, 2017).As found in Study 2, priming power prompted action in a manner beneficial and critical to the reduction of prejudicial expressions, reflecting the effect of power on action beyond previously demonstrated simple circumstances (e.g., moving an annoying fan from one's environment, Galinsky et al., 2003).We believe it is important to understand how power may underlie potential strategies to alleviate and address social inequality, such as recognizing unfairness in interactions (Sawaoka et al., 2015), mitigating the effects of stereotype threat (Van Loo & Rydell, 2013), and protecting or reducing the possible costs of whistleblowing or moral objections (Kundro & Rothbard, 2022;Rehg et al., 2008).Clearly power may motivate a range of activities and psychological processes that could be beneficial and our study examined one critical action -the act of confronting bias.
Along these lines, it is important to note that we investigated only one form of power (i.e., feeling powerful at the individual level), using a well-validated but transient power prime (Galinsky et al., 2003).This act of priming power to motivate action is useful because it is a relatively simple task and could be implemented in a variety of real-world settings (e.g., workplace, schools, etc.), possibly even in the moment, or soon after, a prejudicial comment is said.It is also possible that other forms of priming power exist that could lead to greater confronting behaviors such as increasing autonomy or providing greater control over communal resources or space (Lammers et al., 2009).We believe that these other forms would be equally valid as the psychological processes associated with power motivating action would remain (Cho & Keltner, 2020).
One additional consideration is the link between structural forms of power (e.g., being a leader or decision-maker, higher status in society at large) and one's effectiveness and willingness to confront bias.For instance, individuals with high organizational power or esteemed reputations such as supervisors or leaders could be potent sources to enact long-lasting change by affirming egalitarian norms through confronting behavior (see, Paluck et al., 2016).Conversely, when people in power make prejudicial comments, power may act as a barrier to confronting, as the social costs and one's lower relative status stymies action (Ashburn-Nardo et al., 2014).While this paper was focused on individuals' personal sense of power, we speculate that having actual structural power may lead to the ability to make policy choices that impact the institutional (e.g., societal inequalities such as the gender pay gap) as well as interpersonal levels (see, Brown et al., 2021).For example, auditing payroll for gender pay gaps and issuing corrections if they exist would be an effective policy decision to combat institutionalized pay disparities, and such a decision likely could only be enacted by someone with institutional power.Overall, understanding how power operates at both the individual, organizational, and societal level is critical to addressing confronting behaviors and gender bias.
We believe our findings point to multiple interesting and valuable future directions.In particular, it would be productive to investigate whether priming power also affects the confrontation of prejudicial expressions directed at other subordinate groups such as racial or sexual minorities.Indeed, confronting prejudicial comments in these domains reduces future biased remarks (Czopp & Monteith, 2003) but also occurs at relatively low rates (Dickter, 2012;Kroeper et al., 2014).Additionally, it would be important to investigate whether power affects the confrontation rates of non-target confronters (e.g., allies).This is a critical group as some evidence suggests they face fewer social costs to confronting prejudice and that their confrontations are equally, if not more, effective (Gulker et al., 2013).Indeed, in Study 2 we only examined power as it relates to women's confronting behaviors and future research would benefit from examining whether the effect of power also emerges when it comes to men's confronting behavior.We hypothesize that since power leads to action that aligns with prior beliefs (Guinote, 2007), priming power would only spark action for men who believe in and value gender equality.Lastly, our sample for Study 2 was largely focused on university-aged women, thus to improve generalizability, future work could recruit a sample more reflective of the broader population.Despite these limitations, we hope that power may be a means of increasing the rates of confronting, as well as other prejudice reducing behaviors (e.g., taking social action), conveyed in response to prejudice across multiple domains, leading to its implementation as a broader strategy against multiple forms of bias.
While our overall framework follows the behavioral approach system which relates power to taking action (Cho & Keltner, 2020), we do not specifically identify the exact causal mechanism through which power promotes confronting behaviors.Indeed, the vast literature on power suggests a multitude of means by which power could influence decisions to confront, which may be operating separately, interactively, or collectively.For example, power reduces the number of perceived constraints on action (Whitson et al., 2013), promotes less conformity (Hays & Goldstein, 2015), increases goal-related cognitions (Slabu & Guinote, 2010), increases risk-taking (Anderson & Galinsky, 2006) and promotes faster decision-making (Guinote, 2007).Research also shows that power can even affect women's identification with their gender group (Vial & Napier, 2016).It is likely the case that in different situations, and for different people, power will operate through different mechanisms.For instance, when in a situation where multiple women are present but all choose not to confront an offensive comment, power may influence action through reducing conformity.When faced, however, with a transient interaction (e.g., overhearing a comment in a public venue) faster decision-making and perceiving lower costs and more benefits to confronting may be the means by which power increases confronting behavior.Overall, future research should investigate the different mechanisms through which power may affect confronting behaviors and the role of situational and individual differences.

Conclusion
Broadly, this research answers a call to action whereby psychologists utilize psychological principles and findings to develop novel solutions to social issues of oppression, prejudice, and discrimination (Cohen, 2011).In the studies presented here, we applied research findings on power to address the issues surrounding the confrontation of sexism.We demonstrate that power is associated with a greater likelihood of confronting sexism and that priming power can encourage the confrontation of sexist individuals.

Notes
1.No significant order effects across the seven different order combinations emerged, ps >.284.2. Participants also responded to a battery of questionnaires including a gender group identity scale, The Ambivalent Sexism Inventory (Glick & Fiske, 1996), The Modern and Old-Fashioned Sexism Scales (Swim et al., 1995), the Frequency of Confronting scale, the Chronic Perception of Sexism Scale, and Perceived Costs and Benefits of Confronting (Good et al., 2012).While not central to our hypotheses, we included these questionnaires as potential moderators of our effects.Of these measures, the only significant interaction to emerge was with the benevolent sexist attitude sub-scales of Ambivalent Sexism Inventory, which is described in the Online Supplemental Material.3.This amounted to 19 out of 176 responses for the confront coding and 16 responses for the disagreement coding.
) = −1.26,p = .210.Still, Personal Sense of Power remained a significant predictor, b = 0.28, t(329) = 5.42, p < .001.For Study 1b, results showed that Frequency of Experiencing Sexism (directed toward women) was a significant predictor, b = 0.37, t(339) = 6.96, p < .001,while Personal Sense of Power remained a significant predictor, b = 0.18, t(339) = 3.56, p < .001.These results indicate that for both men and women Personal Sense of Power remains a significant predictor of their Frequency of Confronting Sexism over and above their Frequency of Experiencing Sexism.

Figure 1 .
Figure 1.Results from Study 1 with self-reported personal sense of power predicting frequency of confronting behaviors.

Figure 2 .
Figure 2. Mean confront score by power condition for Study 2. Error bars represent standard errors.*p < .05.
Cheryl L. Dickter -is a Professor in the Psychological Sciences Department at William & Mary.Her research examines how people perceive individuals of different social groups as well as the factors involved in the confrontation of prejudice.Margaret J. Shih is the Neil Jacoby Chair Professor of Management at the UCLA Anderson School of Management.Her