Advertisement

Sex Roles

pp 1–15 | Cite as

Feeling Powerful but Incompetent: Fear of Negative Evaluation Predicts Men’s Sexual Harassment of Subordinates

  • Leah R. Halper
  • Kimberly Rios
Original Article

Abstract

Although research has examined the role of power in men’s likelihood of perpetrating sexual harassment against women, less is known about specific personality traits that might predict sexual harassment. Building upon theorizing that men are especially prone to engage in sexual harassment to the extent that their social status is threatened (Berdahl 2007a), we conducted three studies with samples of adults (Studies 1 and 3) and college students (Study 2). In Studies 1 and 3, we asked participants to indicate their likelihood of engaging in sexual harassment of subordinates across a variety of scenarios. In Study 2, we experimentally primed power and had participants choose to send either neutral or sexuality-related articles to an ostensibly real female participant. Results demonstrated that concerns about being perceived as incompetent (i.e., Fear of Negative Evaluation scores) positively predicted men’s sexual harassment of female subordinates. Among women, Fear of Negative Evaluation scores were unrelated to sexual harassment of male subordinates. Further, this relationship held controlling for Narcissism, Self-Esteem, and Generalized Self-Efficacy scores, suggesting that the fear that others would see oneself as incompetent was a better predictor of sexual harassment than one’s self-perceived incompetence. Implications for the relationship between power, personality, and sexual harassment, and for interventions designed to curb sexual harassment, are discussed.

Keywords

Sexual harassment Power Incompetence Gender harassment Fear of negative evaluation 

Power has long been theorized and shown to play a role in sexual harassment (Cleveland and Kerst 1993; Fineran and Bennett 1999), defined as behavior that derogates or demeans an individual based on his/her sex (Berdahl 2007a). However, the presence of a power dynamic in itself does not guarantee that sexual harassment will occur (Pryor et al. 1993). In light of theories that sexual harassment often stems from a desire to protect one’s social status (Berdahl 2007a, b), as well as findings demonstrating that power-holders are more aggressive toward subordinates the more incompetent they feel in their power (Fast and Chen 2009), we examine whether such feelings of incompetence can predict men’s sexual harassment toward women controlling for other individual difference predictors (e.g., narcissism).

Further, we seek to identify the source of felt incompetence—specifically, whether power-holders engage in sexual harassment because they perceive themselves to be incompetent (i.e., they are low in self-esteem (Rosenberg 1965) or low in self-efficacy (Chen et al. 2004)) or because they fear others see them as incompetent (i.e., they score higher in fear of negative evaluation; Leary 1983). Given that men (who hold a dominant position in the gender hierarchy) have the most to lose when their status is threatened and thus should be more motivated than are women to defend their status against such threats (Berdahl 2007a), we focus primarily on men’s sexual harassment of women in our studies. To demonstrate generalizability, we assess two forms of sexual harassment in our studies: quid pro quo (i.e., making work-related outcomes contingent on the receipt of sexual favors; Pryor et al. 1993) and gender harassment (i.e., behavior that creates a hostile environment for women, such as sexual jokes or sexist comments; Berdahl 2007a).

This particular subject is both timely and relevant. After Hollywood film producer Harvey Weinstein was accused by multiple women of sexual harassment in October 2017 (Harvey Weinstein timeline: How the scandal unfolded 2018), countless stories with titles such as “The (incomplete) list of powerful men accused of sexual harassment” (Criss 2017) surfaced in the news, as did accusations against several other prominent movie and/or television figures (e.g., Louis C.K., Matt Lauer). Some instances of Weinstein’s sexual harassment were audio recorded; during at least one recording, Weinstein cited his fame and power during attempts to coerce the women—many of whom worked on Hollywood sets for Weinstein’s company—to engage in sexual activity (Kantor and Abrams 2017). What factors might increase the likelihood that powerful individuals will sexually harass their subordinates?

Power and Sexual Harassment

For the purposes of the present research, power is defined as an individual’s ability to influence or produce an effect from other individuals (Popovich and Warren 2010) and to control another person’s resources and/or outcomes (Fast and Chen 2009). Because individuals with more power have the freedom and the resources for uninhibited actions and tend to be driven by rewards, higher power leads to approach-related behaviors whereas low power leads to inhibition-related behaviors (Keltner et al. 2003). For instance, supervisors who witness a subordinate performing poorly are more likely to offer to train or confront, and less likely to compensate for or ignore, the subordinate (Ferguson et al. 2010). Those in power are also more likely to approach rather than avoid risk (e.g., report having more unprotected sex; take a gamble in a decision-making game; share information in a negotiation that the other party might be able to use against them) and are less likely to perceive such risks as dangerous (Anderson and Galinsky 2006). Furthermore, power-holders are more likely to stereotype others (Fiske 1993) and are less likely to take others’ perspectives (Galinsky et al. 2006) because higher-power individuals are less dependent on outcomes from lower-power individuals than the reverse (Keltner et al. 2003). In sum, power-holders are more likely to take action, more likely to take risks, and less likely to consider others’ perspectives.

Given that powerful individuals are especially prone to behave in approach-oriented ways without regard for others, it is perhaps unsurprising that there is a clear power dynamic in sexual harassment (Fineran and Bennett 1999). Indeed, some work has suggested that power increases people’s (particularly men’s) tendencies to act on their sexual desires and impulses. In one set of studies, male participants scoring higher in sexual aggression were more likely to approach and flirt with their female coworkers, as well as to view them in a sexual way, when in a position of power (Bargh et al. 1995). Relatedly, powerful men are more inclined to interpret their female subordinates’ behavior as sexual to the extent that they possess a mating goal (e.g., they are single or their subordinate is romantically available; Kunstman and Maner 2011). Thus, power increases the likelihood that men motivated to seek sexual gratification (i.e., who are higher in sexual aggression or have mating goals) will engage in sexual harassment.

However, we argue that among power-holders, sexual harassment can stem not only from sexual desire, but also from the desire to protect one’s social status when this status is threatened. Our argument is based on theorizing that men’s sexual harassment of women represents an attempt to uphold the gender hierarchy (Berdahl 2007a, b; Connell 1987). Even though this theory allows for the possibility that women may also engage in sexual harassment (of men or of other women), men—who hold a dominant position in patriarchal societies—have more to lose from threats to their social status. Male-toward-female sexual harassment should therefore be more prevalent than female-toward-male sexual harassment because men are most motivated to maintain their powerful social status. In support of the theory that sexual harassment is largely a byproduct of motives to protect one’s threatened status, men in one set of studies were particularly likely to harass “uppity” women, who violated traditional gender roles (e.g., by holding a leadership position) as opposed to women who conformed to traditional gender roles (Berdahl 2007b). Similarly, in another study, when men were given false feedback that a woman had scored higher than they did on a masculine knowledge test, they asked more inappropriate and sexually offensive interview questions of female job applicants in a mock interview (Hitlan et al. 2009).

Feelings of Incompetence and Sexual Harassment

What dispositional factors are associated with feelings of threat to social status among powerful men? Some research shows that men are more likely to engage in sexual harassment the higher they score on measures of narcissism (i.e., an overly inflated sense of self-worth). Although this may be because people higher in narcissism have lower empathy for others and feel more entitled to pursue what they want (Bushman et al. 2003), an alternative explanation is that people higher in narcissism actually feel less competent and less secure in themselves (Zeigler-Hill 2006), and hence they may be more likely to compensate for their insecurity by sexually harassing their subordinates. However, to the best of our knowledge, the relationship between perceptions of incompetence and sexual harassment among power-holders has never directly been assessed. Doing so is important because it would corroborate the notion that sexual harassment is driven in part by harassers’ fears of being called out for ineffective leadership—in other words, that they harass not necessarily to seek sexual gratification, but rather to maintain their high-power position (Berdahl 2007a, b).

Some research has examined the link between power, perceived incompetence, and aggression more generally. In one set of studies, individuals who chronically felt, or were made to feel, both powerful and incompetent (e.g., by recalling a time during which they held power over another person but felt they did not have the skills or abilities to be effective in that situation) subsequently behaved more aggressively (e.g., by giving “another participant” louder sound blasts for not remembering information accurately; by sabotaging subordinates’ chances to win money) than did individuals who were made to feel powerless and incompetent (Fast and Chen 2009). Presumably, powerful but incompetent individuals aggressed against others to make themselves feel more secure in their own power; that is, they put their peers and subordinates down as a means of asserting their dominance and thus maintaining their powerful position. These effects emerged among men and women alike, possibly because the aggression assessed in the studies was general rather than gender-based. By contrast, because we measure sexual harassment specifically and draw upon the theory that male (more so than female) power-holders are motivated to protect their status when threatened (Berdahl 2007a, b), we expect our effects to be stronger among men than among women.

In addition to investigating sexual harassment (which we predict will be most pronounced among male power-holders who feel incompetent) rather than general aggression (the effects of which are not moderated by gender; Fast and Chen 2009), our studies extend beyond those of Fast and Chen (2009) by disentangling self-perceptions from others’ perceptions of incompetence. To date, it is unclear whether powerful men sexually harass subordinates when they harbor perceptions of themselves as incompetent or when they fear that others will perceive them as incompetent. These two possibilities are related (see Fast and Chen 2009), but we suspect that the fear of being seen as incompetent by others (i.e., possessing a high fear of negative evaluation) should be especially likely to predict our effects because such a fear, like power itself, directly involves interpersonal relationships. If male power-holders sexually harass in part to protect their dominant position in the gender hierarchy (Berdahl 2007a, b), then they may be motivated not so much by their own desires to view themselves as competent, but rather by their desires for other people to see them as competent. Thus, these individuals may believe they need to convince others that they are worthy of the power they wield—in this case, by exhibiting sexually aggressive behaviors toward women.

In our studies, we attempted to distinguish self-perceptions from concerns about others’ perceptions of incompetence by measuring individual differences related to each construct. We specifically compared self-esteem (i.e., one’s perception of his/her own self-worth; Rosenberg 1965) and generalized self-efficacy (i.e., one’s perception of his/her own competence in general; Chen et al. 2001) to fear of negative evaluation (i.e., worries about others’ perceptions of oneself; Leary 1983). By doing so, we sought to determine if other-perceived incompetence was a better predictor of sexual harassment among powerful men than self-perceived incompetence. Additionally, given that narcissism is also known to affect sexual harassment tendencies and relates to perceptions of oneself (Bushman et al. 2003), we included narcissism as a predictor in the current studies.

Overview of Studies

We conducted two studies with adult participants from Amazon’s Mechanical Turk (Studies 1 and 3) and one study with university students in the lab (Study 2). Mechanical Turk is a database consisting of hundreds of thousands of “workers,” who can be recruited to complete various tasks (including research studies) for small amounts of money. Mechanical Turk (MTurk) workers are typically more diverse in terms of age, gender, socioeconomic status, and political views than are university student samples (Buhrmester et al. 2011), making MTurk an ideal source of data collection for large correlational studies. We used a university student sample for Study 2 to maximize internal validity because that study involved an experimental prime.

In our studies, we either had all participants read vignettes in which they took the perspective of a person in power (Studies 1 and 3) or experimentally primed power (Study 2). We measured self-perceptions of competence (i.e., self-esteem and/or self-efficacy), fears that others will perceive oneself as incompetent (i.e., fear of negative evaluation), and narcissism. We then assessed intentions to engage (Studies 1 and 3) or actual engagement (Study 2) in sexual harassment, operationalized as either quid pro quo sexual harassment (Studies 1 and 3) or gender harassment (Study 2). Whereas we recruited only male participants in Study 1, we recruited both male and female participants in Studies 2 and 3 for comparison purposes.

Study 1

In Study 1, male participants completed measures of fear of negative evaluation, self-esteem, and narcissism. They then read ten vignettes that described a hypothetical situation in which they held power over a female subordinate and indicated their likelihood of coercing the subordinate to engage in sexual activity with them, a form of quid pro quo sexual harassment. We hypothesized that Fear of Negative Evaluation scores would positively predict male participants’ self-reported likelihood of engaging in quid pro quo sexual harassment (Hypothesis 1a). We also predicted that Fear of Negative Evaluation scores would predict male participants’ self-reported likelihood of engaging in quid pro quo sexual harassment controlling for Self-Esteem and Narcissism scores (Hypothesis 1b).

Method

Participants and Design

Participants were 273 adult male MTurk workers (Mage = 35.12, SD = 11.84, range = 18–79) who received $.50 as compensation; 194 (71.1%) identified as White/European American, 3 (1.1%) as Native American, 37 (13.4%) as Asian American, 18 (6.5%) as Black/African American, 19 (6.9%) as Hispanic/Latino, and 2 (.7%) as Middle Eastern. A minority (69, 25%) of the participants completed the study in the spring of 2015, and the remaining 204 completed it in the fall of 2016. However, the procedure and measures were identical across samples, and the results held after controlling for sample and its interaction with each individual-difference variable. There were also no main effects or interactions involving sample. We therefore collapsed the data from the two samples.

All participants were instructed to imagine themselves as the interviewer of a female job applicant (i.e., as holding a powerful role) in ten vignettes. Therefore, power was not experimentally manipulated. In soliciting participation, we told participants that they would answer questions about their personalities and attitudes in the study, as well as indicate how they would act in certain scenarios.

Materials and Procedure

Participants first completed the Fear of Negative Evaluation (Leary 1983), Self-Esteem (Rosenberg 1965), and Narcissism (Ames et al. 2006) scales in counterbalanced order. The 12-item Fear of Negative Evaluation scale, which assesses individuals’ concerns about being criticized or negatively judged by others, served as our measure of fears that others will see one as incompetent (see Fast and Chen 2009). An example item from the scale was: “I am frequently afraid of other people noticing my shortcomings.” Participants rated each statement on a scale from 1 (not at all characteristic of me) to 5 (extremely characteristic of me) (M = 2.62, SD = .93; α = .93). Scores across items were averaged so that higher scores indicate higher levels of fear of being negatively evaluated.

The 10-item Rosenberg Self-Esteem scale (e.g., “On the whole, I am satisfied with myself”) assessed participants’ perceptions of their global self-esteem (M = 3.81, SD = .79, α = .92). Participants responded to these items on 5-point scales from 1 (strongly disagree) to 5 (strongly agree). Scores across items were averaged so that higher scores indicate higher levels of self-esteem.

The Narcissism scale consisted of 16 pairs of items (e.g., “I think I am a special person” [coded 1] or “I am no better nor worse than most people” [coded 0]). Within each pair, participants were instructed to choose the item more characteristic of them (M = 4.77, SD = 3.82). One point was given for each response indicative of an overly inflated sense of self; thus, higher summed scores reflected higher levels of narcissism.

Next, participants completed Pryor’s (1987) Likelihood to Sexually Harass (LSH) scale (M = 1.86, SD = 1.05; α = .95). The LSH scale consists of 10 brief scenarios in which a male protagonist has the power to sexually exploit a female subordinate with impunity. Participants were asked to imagine themselves in the role of the protagonist in each scenario and to rate the likelihood that they would perform an act of quid pro quo sexual harassment. After averaging scores across the 10 vignettes, higher scores reflected a greater likelihood of sexual harassment. Past research has demonstrated that this quid pro quo scale is reliable, valid, and predictive of sexually harassing behavior (Pryor and Stoller 1994).

To provide an example, in one scenario, the male responder is asked to imagine that he is an executive in a large corporation, 42 years old, and secure in his job. The male responder is also asked to imagine he is interviewing candidates and an applicant named Michelle S. tells him during her interview that she desperately needs the job. The vignette also states that the male responder is attracted to Michelle S. The question assessing sexual harassment in this scenario is: “Assuming that you are secure enough in your job that no possible reprisals could happen to you, would you offer her the job in exchange for sexual favors?” rated on a scale from 1 (not at all likely) to 5 (very likely). A similarly worded question is used to measure sexual harassment in each of the other scenarios.

Results

Table 1 reports the zero-order correlations among study variables revealing that Fear of Negative Evaluation scores were significantly negatively correlated with Self-Esteem scores and unrelated to Narcissism scores among male participants. In addition, men’s higher self-reported likelihood to sexually harass a woman was associated with higher Fear of Negative Evaluation scores, lower Self-Esteem scores, and higher Narcissism scores.
Table 1

Correlations among study variables with male participants, Study 1

 

Correlations

Variables

1

2

3

1. Fear of negative evaluation

  

2. Self-esteem

−.53**

 

3. Narcissism

−.15

.22

4. Likelihood to sexually harass

.27**

−.17**

.24*

n = 273

*p < .05. **p < .01

We predicted that Fear of Negative Evaluation scores would be positively associated with likelihood of sexual harassment (Hypothesis 1a), even after controlling for Self-Esteem and Narcissism scores (Hypothesis 1b). To test these predictions, we performed a hierarchical regression analysis in which only scores on the Fear of Negative Evaluation scale were included in the first step of the model (total R2 = .073), and scores on the Self-Esteem and Narcissism scales were added to the second step of the model (total R2 = .165, ΔR2 = .092) (see Table 2). Given that Fear of Negative Evaluation scores positively and significantly predicted likelihood of sexual harassment with and without controlling for the other predictors, both Hypotheses 1a and 1b were supported.
Table 2

Summary of regression analysis for predicting likelihood of sexual harassment, Study 1

 

Step 1

Step 2

Variable

B

SE (B)

t

p

B

SE (B)

t

p

FNE

.304

.066

4.60

<.001

.312

.075

4.19

<.001

Self-esteem

    

−.063

.087

−.720

.472

Narcissism

    

.083

.015

5.42

<.001

Model

R2 = .073

R2 = .165

 Statistics

F(1, 271) = 21.19, p < .001

F(3, 269) = 17.66, p < .001

Change

 

ΔR2 = .092.

 Statistics

 

ΔF(2, 269) = 14.82, p < .001

n = 273. FNE Fear of Negative Evaluation

Discussion

Study 1 provided evidence for fear of being seen as incompetent by others (i.e., Fear of Negative Evaluation scores) as a predictor of sexual harassment intentions in male participants, independent of individual differences in self-esteem and narcissism (the latter of which also predicted likelihood of sexual harassment). These findings suggest that sexual harassment may not only come from a place of entitlement, but also from a place of insecurity. Additionally, the fact that fear of negative evaluation (as well as narcissism), but not self-esteem, predicted sexual harassment suggests that concerns about how others view one’s competence, rather than low self-esteem, are a driving force behind harassing behavior.

Study 2

There were three limitations of Study 1 that we aimed to address in Study 2. First, whereas the measure of sexual harassment used in Study 1 was self-reported, we employed a behavioral measure of harassment in Study 2. This measure, in addition to assessing actual behavior and thus reducing concerns about participants’ social desirability in reporting what they would hypothetically do, assessed gender harassment (i.e., creating a hostile environment for women) as opposed to quid pro quo sexual harassment. Second, Study 2 experimentally primed power, which was held constant in Study 1 (i.e., because all participants imagined themselves as holding high-power positions in the sexual harassment scenarios). Finally, in Study 2, we included women in the sample as a comparison group to test whether the interaction between power and fear of being perceived as incompetent by others on likelihood of behavioral harassment would be especially pronounced for men.

The behavioral task we used in Study 2 involved the opportunity to send either sexuality-related or neutral (unrelated to sexuality) articles to a purportedly real female participant in another room whom participants thought would need to read and memorize parts of the articles as part of the study. This paradigm is similar to the paradigm developed by Maass and colleagues (Maass et al. 2003), who examined men’s decisions to send pornographic pictures to women over the computer as a form of sexual harassment. However, we used article titles and short descriptions (blurbs), rather than pornographic pictures, to minimize ethical concerns. Unlike the behavioral intentions assessed in Study 1, this behavior was not quid pro quo sexual harassment, but rather gender harassment, which rarely takes the form of a sexual proposition. Gender harassment, a form of hostile environment harassment, refers to verbal behavior that is hostile or degrading toward individuals in the workplace, such as sexual or sexist jokes (Berdahl 2007a). We hypothesized that the relationship between power, fear of negative evaluation, and sexual harassment would be moderated by the participants’ gender, such that fear of negative evaluation would positively predict sexual harassment among men, but not among women, primed with power (Hypothesis 2).

Method

Participants and Design

Fully 159 undergraduate students at a large U.S. Midwestern university participated in the study in exchange for credit in their psychology courses. Participants were randomly assigned to either the high or low power condition; thus, the study followed a 2 (Participants’ Gender: male vs. female) × 2 (Power: high vs. low) design, with scores on the Fear of Negative Evaluation scale as a continuous moderator. Fifteen participants (9.4%) were excluded because they reported suspicion that their unseen partner, “Anna K.,” was not real, leaving 144 individuals in the final sample of 59 (41%) men and 85 (59%) women. Participants ages ranged from 18 to 22 years-old (Mage = 18.93, SD = .94). In terms of self-identified race and ethnicity, there were 116 (80.6%) White/European American, 3 (2.1%) Asian American, 2 (1.4%) Native American, 12 (8.3%) Black/African American, 9 (6.3%) Hispanic/Latino, and 2 (1.4%) Middle Eastern participants. Cell sizes for the final sample ranged from 25 (men in the low-power condition) to 45 (women in the low-power condition) and were consistent with recommendations that experiments recruit at least 20 participants per cell (Simmons et al. 2011). Furthermore, an a priori power analysis, conducted using G*Power (Faul et al. 2007), revealed that 128 participants were needed to obtain a medium-sized effect (f = .25) with 80% power. Male and female participants did not differ significantly from one another in terms of age (p = .070) or race/ethnicity (p = .156).

Materials and Procedure

Participants were told that the study consisted of two unrelated parts. The first part would involve answering some questions about themselves. The second part was designed to identify article titles and article blurbs to be used in future studies about learning and cognition.

At the beginning of the study, participants completed the 12-item Fear of Negative Evaluation scale (M = 3.15, SD = .75, α = .89), the 10-item Self-Esteem scale (M = 3.73, SD = .66, α = .90), and the 16-item Narcissism scale (M = 5.34, SD = 2.87, α = .64) used in Study 1, with each measure scored the same as in Study 1. Participants then were randomly assigned to recall and write about an instance in which they held power over another person (high-power condition) or an instance in which another person held power over them (low-power condition). This prime was adapted from previous research on power and produces effects similar to those obtained when actually placing participants in a high-power role (Galinsky et al. 2003). Immediately after the power prime, individuals indicated how powerful they felt relative to the individual about whom they wrote in their essay prompt on a scale from 1 (the other person had much more power than I did) to 7 (I had much more power than the other person did). This question served as the manipulation check for power.

As a measure of behavioral sexual harassment, participants were given the opportunity to send five articles to another unseen participant in the study. Each participant was told that they would choose between five different pairs of articles (based on the article titles and short blurbs) to send to a purportedly real individual named “Anna K.” who was in another room and with whom they had been randomly paired. In each pair, there was a sexuality-related article and a neutral article, determined by the pilot test reported in the following. To increase the believability of the cover story, they were told that “Anna K.” would need to read, memorize parts of, and answer questions about the articles they had chosen for her. The number of sexuality-related articles sent (of 5) was the dependent measure in our study (M = 2.67, SD = 1.35, range = 0–5).

Because Anna K. was not a real individual and participants did not actually send article titles or blurbs to anyone, we asked three questions at the end of the study to probe for suspicions. Participants’ data were only eliminated from the study if the participant guessed the real purpose of the experiment and/or admitted suspicion about the existence of Anna K.

Article Pair Pilot Study

The sexuality-related and neutral articles were tested in a pilot study of 87 undergraduate students (69, [79.3%] female, Mage = 18.91, SD = 1.23, range = 18–24) from the same university at which Study 2 was conducted. Participants’ gender produced no significant effects and thus will not be discussed further. Participants were asked to read ten different headlines and short summaries/blurbs of actual news articles and to rate the articles on a variety of dimensions. Five of the articles were sexuality-related (e.g., “Does a more equal marriage mean less sex?”), and five of the articles were on neutral (i.e., unrelated to sex or sexuality) topics (e.g., “Arctic Shipwreck Solves 170-Year-Old Mystery”). (The titles and blurbs for all ten articles can be found in the Online Supplement.)

Participants rated each article on a 9-point scale from 1 (not at all) to 9 (extremely) according to (a) how offended they thought a woman would be if a man sent the article to her, (b) to what extent they thought a woman would consider it sexual harassment if a man sent the article to her, (c) how offended they thought a woman would be if another woman sent the article to her, and (d) to what extent they thought a woman would consider it sexual harassment if another woman sent the article to her. Participants’ responses to the first two questions (i.e., about a woman being sent the article by a man) were averaged together for each article, as were their responses to the last two questions (i.e., about a woman being sent the article by another woman) to create perceived sexual harassment scores for each article and sender gender (rs > .36, ps < .001). (The results were identical to those we report here when the “offended” and “sexual harassment” questions were analyzed separately rather than combined.)

A principal axis factor analysis on the five neutral and five sexuality-related articles sent by a man, using varimax rotation, revealed that the articles loaded onto two separate factors with eigenvalues of at least 1. The first factor (eigenvalue = 3.31) contained the five neutral articles, each of which had a loading of at least .42 on that factor and a loading of .14 or less on the other factor. The second factor (eigenvalue = 2.86) contained the five sexuality-related articles, each of which had a loading of at least .55 on that factor and a loading of .21 or less on the other factor. An identical analysis on the ten articles sent by a woman similarly revealed two factors with eigenvalues of at least 1. The first (eigenvalue = 4.08) contained the five neutral articles, each of which had a loading of at least .50 on that factor and a loading of .08 or less on the other factor. The second (eigenvalue = 2.57) contained the five neutral articles, each of which had a loading of at least .51 on that factor and a loading of .25 or less on the other factor. Thus, for the following analyses, we combined the perceived sexual harassment scores for the five neutral articles and the five sexuality-related articles together to create two separate composites for each sender gender.

A repeated measures ANOVA with article type (sexuality-related vs. neutral) and sender gender (male vs. female) as within-participants factors and the perceived sexual harassment composite as the dependent variable revealed that participants thought a woman would perceive greater sexual harassment if a man (M = 3.34, SD = 1.15) (compared to another woman; M = 2.43, SD = 1.01) sent her the articles, F(1, 86) = 213.20, p < .001, ηp2 = .71, and that participants thought a woman would perceive greater sexual harassment if she were sent the sexuality-related (M = 4.53, SD = 1.63) (compared to neutral; M = 1.25, SD = .53) articles, F(1, 86) = 375.68, p < .001, ηp2 = .81.

However, these effects were qualified by an Article Type x Sender Gender interaction, F(1, 86) = 197.78, p < .001, ηp2 = .70. Although participants thought a woman would perceive greater sexual harassment from the neutral articles if sent by a man (M = 1.29, SD = .58) than by another woman (M = 1.20, SD = .48), t(86) = 4.41, p < .001, d = .17, this difference was much stronger for the sexuality-related articles (male sender: M = 5.40, SD = 1.72; female sender: M = 3.66, SD = 1.54), t(86) = 14.53, p < .001, d = 1.07. (Notably, although the means for perceptions of sexual harassment from the neutral articles were significantly different, they were less than half a point apart and very close to the bottom of the scale. This difference may have been an artifact of participants simply believing that women would be more likely in any context to perceive sexual harassment from a man rather than from another woman.) Furthermore, whereas participants rated the sexuality-related articles as significantly lower than the midpoint of perceived sexual harassment (5) if sent by a woman, t(86) = −8.14, p < .001, d = 1.76, they rated the sexuality-related articles as significantly higher than the midpoint of perceived sexual harassment if sent by a man, t(86) = 2.19, p = .032, d = .47.

To confirm that the neutral articles were not more interesting-sounding than the sexuality-related articles (or vice versa), participants in the pilot study also indicated how interesting the article titles and blurbs seemed on a 1 (not at all interesting) to 9 (extremely interesting) scale. On average, the five neutral articles (M = 4.65, SD = 1.43) were rated as no more interesting than the five sexuality-related articles (M = 4.36, SD = 1.54), t(86) = 1.36, p = .178, d = .20.

In this pilot study, unlike in our experiment, we did not provide any context for why a woman would be sent each article. However, we see the lack of context as a strong test of our hypothesis that participants would think women would perceive the sexuality-related articles as sexual harassment (specifically, gender harassment) when sent by a man. Even if some participants had made the most charitable attribution for why a man would send sexuality-related articles to a woman (e.g., because he wanted to convey that he felt the articles were offensive), they still construed doing so as sexual harassment, more than the midpoint and more than if another woman had sent the articles.

Results

Manipulation Check

An independent-samples t-test revealed that participants in the high-power condition (M = 6.00, SD = .92) felt significantly more powerful in the situation they recalled than did participants in the low-power condition (M = 1.79, SD = .96), t(142) = 26.86, p < .001, d = 4.48. Thus, our power prime was successful.

Behavioral Sexual Harassment

We predicted that male participants in the high-power condition, but not in the low-power condition, would send significantly more sexuality-related articles to the female target as their concerns with being perceived as incompetent (i.e., scores on the Fear of Negative Evaluation scale) increased. However, we predicted that female participants would not demonstrate any relationship between power or Fear of Negative Evaluation scores on sexuality-related articles sent. To test these hypotheses, we regressed number of sexuality-related articles sent to “Anna K.” onto gender (0 = male, 1 = female), power condition (0 = low, 1 = high), Fear of Negative Evaluation scores (mean-centered to reduce multicollinearity), and all two- and three-way interaction terms. The first step of the regression contained the main effects, the second step contained all possible two-way interactions, and the third step contained the three-way interaction term (Cohen et al. 2003). See Table 3 for the regression model (total R2 = .093).
Table 3

Summary of regression analysis for predicting behavioral sexual harassment, Study 2

 

Step 1

Step 2

Step 3

Variable

B

SE

t

p

B

SE

t

p

B

SE

t

p

Gender

.399

.229

1.74

.084

.521

.349

1.49

.138

.630

.348

1.81

.073

Power

−.273

.223

−1.22

.224

−.169

.362

−.467

.641

−.022

.364

−.062

.951

FNE

.206

.150

1.37

.173

.073

.322

.227

.821

−.465

.403

−1.16

.250

Power x Gender

    

−.178

.468

−.380

.704

−.292

.465

−.629

.531

FNE x Power

    

.338

.305

1.11

.269

1.34

.550

2.44

.016

FNE x Gender

    

−.059

.333

−.177

.859

.687

.474

1.45

.150

Gender x Power x FNE

        

−1.43

.657

−2.18

.031

Model statistics

R2 = .052

F(3, 140) = 2.55, p = .058

R2 = .061

F(6, 137) = 1.49, p = .188

R2 = .093

F(7, 136) = 1.99, p = .061

Change statistics

    

ΔR2 = .009

ΔF(3, 137) = .448, p = .719

ΔR2 = .032

ΔF(1, 136) = 4.75, p = .031

n = 144. FNE Fear of Negative Evaluation

There were no lower-order effects (R2 = .052 for Step 1, .061 for Step 2; ps > .084) (see Table 3). However, the predicted Gender x Power Condition x Fear of Negative Evaluation scores interaction was significant (p = .031, ΔR2 = .032; see Fig. 1). Supporting Hypothesis 2, Power Condition and Fear of Negative Evaluation scores interacted to predict number of sexuality-related articles sent among male participants (b = 1.34, SE = .55), t(136) = 2.44, p = .016, but not among female participants (b = .09, SE = .36), t(136) = .25, p = .802. Specifically, male participants in the high-power condition sent significantly more sexuality-related articles to “Anna K.” as their Fear of Negative Evaluation scores increased (b = .88, SE = .37), t(136) = 2.34, p = .021, whereas there was no relationship between Fear of Negative Evaluation scores and number of sexuality-related articles sent to “Anna K.” among male participants in the low-power condition (b = −.47, SE = .40), t(136) = −1.16, p = .250.
Fig. 1

Number of sexuality-related articles sent as a function of participants’ gender: a male vs. b female, power condition (low vs. high), and Fear of Negative Evaluation scores plotted at 1 SD above and below the mean in Study 2. Only the solid black line indicates a significant relationship

As in Study 1, we also assessed scores on the Self-Esteem and Narcissism scales as potential predictors of male participants’ sexual harassment tendencies under conditions of power. However, neither the Power x Gender x Self-Esteem scores interaction (p = .867) nor the Power x Gender x Narcissism scores interaction (p = .823) was significant. As a note, in the Power x Gender x Self-Esteem analysis, the only significant effect to emerge was that participants sent fewer harassing articles to “Anna K.” the lower their Self-Esteem scale scores were (b = −.36, SE = .17), t(136) = −2.12, p = .036. In the Power x Gender x Narcissism scores analysis, the only significant effect to emerge was that participants sent more sexuality-related articles to “Anna K.” the higher their Narcissism scale scores were (b = .11, SE = .04), t(136) = 2.80, p = .006.

Discussion

Using a different type of sample from Study 1 (i.e., college students instead of MTurk workers) and an experimental power prime, the results of Study 2 further demonstrated that power and perceived fear of being perceived as incompetent by others (i.e., scores on the Fear of Negative Evaluation scale) predicted behavioral sexual harassment for male participants but not for female participants. Therefore, Study 2 adds to the previous study, which showed that fear of being seen as incompetent predicts greater self-reported likelihood of engaging in sexual harassment toward subordinates. Specifically, Study 2 employed a method in which individuals had the opportunity to choose between sending sexuality-related and neutral articles to a purportedly real female participant. In contrast with the quid pro quo likelihood of sexual harassment scale in Study 1, actually sending sexuality-related articles to a woman constitutes gender harassment, a form of hostile environment harassment in which individuals communicate with or around the target in an offensive manner based on the target’s gender.

In Study 2, Fear of Negative Evaluation scores significantly moderated the effects of power and gender on behavioral sexual harassment, but Self-Esteem and Narcissism scores did not. These results, along with the results of Study 1, support our argument that powerful men are most likely to sexually harass when they are concerned about being perceived by others as incompetent, and not necessarily when they have weaker self-esteem or when they have stronger feelings of grandiosity (i.e., narcissism) about themselves.

Study 3

In a third study, we aimed to address limitations of the previous two studies. First, we included self-efficacy as a potential moderator that more specifically assessed self-perceived competence rather than global self-esteem. Generalized self-efficacy is similar to self-esteem in that it is a global measure of one’s self-worth. However, these concepts are theoretically and practically distinct. Self-esteem is a feeling of general self-worth (i.e., an affective quality), whereas generalized self-efficacy is a perception of one’s general day-to-day competence for most tasks encountered (Chen et al. 2004). Thus, self-efficacy is a closer parallel to fear of being negatively evaluated by others (in that both focus on competence, although self-perceived versus other-perceived competence, respectively) than is self-esteem.

Second, although Study 2 included female participants, they were instructed to send articles to another woman. It could thus be argued that Study 2 was not a strong test of whether men were more likely to engage in sexual harassment than were women because most women in the study were presumably heterosexual and likely hold cultural stereotypes about male perpetrators and female targets. Consistent with the latter speculation, the pilot test reported in Study 2 revealed that women were less inclined to think that sending sexuality-related articles to “Anna K.” (i.e., woman-toward-woman harassment) would be construed as sexual harassment at all. To address this concern, heterosexual female participants in Study 3 responded to questions about whether they would sexually harass a male subordinate. We hypothesized that the relationship between Fear of Negative Evaluation scores and sexual harassment would be moderated by participants’ gender, such that higher scores on the Fear of Negative Evaluation scale would predict a greater likelihood of sexual harassment among male but not female participants (Hypothesis 3a). We further expected that the effects predicted in Hypothesis 3a would hold after controlling for scores on the Self-Efficacy and Narcissism scales and their interactions with participants’ gender (Hypothesis 3b).

Method

Participants and Design

Fully 202 individuals were recruited from MTurk and were paid $.50 in exchange for participating. Because we wanted to ensure that the gender of the subordinate in each scenario corresponded to the gender to whom participants were sexually attracted, we excluded five participants whose sexual orientation was gay or lesbian, leaving 197 individuals in the final sample, of whom 107 (54.3%) were female. Participants ages ranged from 19 to 73 years-old (Mage = 35.04, SD = 10.01). In terms of self-identified race and ethnicity, there were 150 (76.1%) White/European American, 1 (.5%) Native American, 14 (7.1%) Asian American, 14 (7.1%) Black/African American, 15 (7.6%) Hispanic/Latino, and 3 (1.5%) Middle Eastern participants. We did not ask about sexual orientation in the previous studies.

Male and female participants did not differ significantly from one another in terms of race/ethnicity (p = .54). On average, male participants (Mage = 32.86, SD = 9.53, range = 19–73) were significantly younger than female participants (Mage = 36.87, SD = 10.08, range = 22–65), t(195) = −2.87, p = .005, d = .42. However, the predicted interaction between Participant Gender and Fear of Negative Evaluation scores held when controlling for participants’ age as a factor in the design (b = −.32, SE = .13), t(190) = −2.42, p = .017, and the three-way Participant Gender x Fear of Negative Evaluation x Participant Age interaction was not significant (p = .076). Therefore, we will not discuss participants’ age further.

As in Study 1, all participants imagined themselves in a high-power role (i.e., in which they could potentially sexually harass a subordinate).

Materials and Procedure

The procedure was identical to that of Study 1, but with two modifications. First, although male participants received the same ten scenarios as in Study 1, they were modified for female participants so that the subordinate was male. For example, in the scenario in which participants imagined that they were an executive in a large corporation and were attracted to a candidate they were interviewing for a job, the candidate was named “Michelle S.” (for male participants) or “Michael S.” (for female participants). Participants’ responses to the scenarios were averaged to form an index of quid pro quo sexual harassment, with higher scores reflecting a greater likelihood of sexual harassment (M = 1.60, SD = .94; α = .96).

Second, as in Study 1, participants completed the 12-item Fear of Negative Evaluation scale (averaged together with higher scores reflecting a stronger fear of negative evaluation; M = 2.79, SD = .96; α = .94) and the 16-item Narcissistic Personality Inventory (summed together with higher scores reflecting stronger narcissism; M = 4.08, SD = 3.63). However, instead of completing a measure of self-esteem, they completed the 10-item Generalized Self-Efficacy scale, which assesses participants’ confidence in their ability to accomplish different tasks (e.g., “I can usually handle whatever comes my way”; Schwarzer and Jerusalem 1995). Participants’ scores were averaged together so that higher scores reflected greater self-efficacy (M = 3.09, SD = .51) on a 5-point scale from 1 (Not at all true) to 5 (Exactly true) (α = .89).

Results

The zero-order correlations among all measured variables can be found in Table 4. Among all participants, Fear of Negative Evaluation scores were negatively correlated with Self-Efficacy scores, and likelihood of sexual harassment was positively correlated with Narcissism scores and negatively correlated with Self-Efficacy scores.
Table 4

Correlations among study variables, Study 3

 

Correlations

Variables

1

2

3

1. Fear of negative evaluation

  

2. Self-efficacy

−.24**

 

3. Narcissism

−.02

.12

4. Likelihood to sexually harass

.07

−.20**

.42**

n = 197

*p < .05. **p < .01

We predicted that fear of being seen as incompetent by others (i.e., Fear of Negative Evaluation scores) would positively predict likelihood of sexual harassment among male (but not female) participants (Hypothesis 3a), even when taking into account differences in Narcissism and Self-Efficacy scores (Hypothesis 3b). To test Hypothesis 3a, we conducted a hierarchical multiple regression analysis in which we regressed Likelihood of Sexual Harassment onto Fear of Negative Evaluation scores and Participant Gender (0 = male, 1 = female) in the first step (R2 = .131), and the two-way interaction term in the second step (Cohen et al. 2003; total R2 = .152, ΔR2 = .020) (see Table 5). Female participants reported being less likely to sexually harass their subordinates than did male participants; the overall effect of Fear of Negative Evaluation scores was not significant. As predicted, the interaction between Participant Gender and Fear of Negative Evaluation scores was significant. Male participants reported a greater Likelihood of Sexual Harassment as their Fear of Negative Evaluation scores increased (b = .30, SE = .11), t(193) = 2.82, p = .005, but this relationship did not emerge among female participants (b = .01, SE = .08), t(193) = .11, p = .911. Hypothesis 3a was thus supported.
Table 5

Summary of regression analysis for predicting likelihood of sexual harassment, Study 3

 

Step 1

Step 2

Variable

B

SE (B)

t

p

B

SE (B)

t

p

Gender

−.673

.127

−5.30

<.001

.128

.392

.068

.745

FNE

.119

.066

1.80

.073

.300

.107

2.82

.005

Gender x FNE

    

−.291

.135

−2.16

.032

Model

R2 = .131

R2 = .152

 Statistics

F(2, 194) = 14.65, p < .001

F(3, 193) = 11.50, p < .001

Change

 

ΔR2 = .020,

 Statistics

 

ΔF(1, 193) = 4.65, p = .032

n = 197. FNE Fear of Negative Evaluation

To test Hypothesis 3b, we conducted a hierarchical multiple regression analysis in which we regressed Likelihood of Sexual Harassment onto Fear of Negative Evaluation scores, Narcissism scores, Self-Efficacy scores, and Participant Gender (0 = male, 1 = female) in the first step (R2 = .27), and the interactions between participant gender and each of these variables in the second step (total R2 = .303, ΔR2 = .029) (see Table 6). Overall, female participants reported being less likely to sexually harass their subordinates than did male participants. In addition, Narcissism scores were positively related, and Self-Efficacy scores were negatively related, to Likelihood of Sexual Harassment. There was not a significant overall effect of Fear of Negative Evaluation scores. Critically, however, Participant Gender and Fear of Negative Evaluation interacted to predict likelihood of sexual harassment. Simple slopes analyses revealed that, as predicted, Fear of Negative Evaluation scores positively predicted likelihood of sexual harassment among male participants (b = .25, SE = .10), t(189) = 2.42, p = .016, but not among female participants (b = −.05, SE = .08), t(189) = −.69, p = .492. There were no significant interactions between Participant Gender and any of the other individual difference variables (Narcissism or Self-Efficacy scores; ps > .40) on likelihood of sexual harassment. Therefore, Hypothesis 3b was supported.
Table 6

Summary of regression analysis for predicting likelihood of sexual harassment, Study 3

 

Step 1

Step 2

Variable

B

SE (B)

t

p

B

SE (B)

t

p

Gender

−.396

.125

−3.16

.002

−.206

.899

−.229

.819

FNE

.059

.063

.941

.348

.250

.103

2.42

.016

Self-efficacy

−.369

.120

−3.09

.002

−.426

.161

−2.64

.009

Narcissism

.098

.017

5.75

<.001

.099

.022

4.44

<.001

Gender x FNE

    

−.304

.130

−2.34

.020

Gender x Self-efficacy

    

.199

.239

.833

.406

Gender x Narcissism

    

.010

.034

.290

.772

Model

R2 = .273

R2 = .303

 Statistics

F(4, 192) = 18.07, p < .001

F(7, 189) = 11.72, p < .001

Change

 

ΔR2 = .029,

 Statistics

 

ΔF(3, 189) = 2.63, p = .051

n = 197. FNE Fear of Negative Evaluation

Discussion

There were two main contributions of Study 3 beyond Studies 1 and 2. First, Study 3 was a replication of Study 1, but with female participants as well as male participants. Female participants read and responded to the same ten sexual harassment scenarios as male participants, but the names of the subordinates in the vignettes were altered appropriately so that the subordinates were male when the participant was female. This change thus provided a direct test of whether male participants were more likely than female participants to sexually harass subordinates of the gender to whom they were sexually attracted.

The second contribution was the measurement of self-efficacy rather than self-esteem, which solidified support for Hypotheses 1b and 3b that fear of others’ negative evaluations of one’s competence would be associated with sexual harassment controlling for self-perceived competence. Given that the Generalized Self-Efficacy scale directly assesses perceptions of one’s own competence (in contrast to the Rosenberg Self-Esteem scale, which assesses one’s global self-esteem), Study 3 provided a stronger test of whether other-perceived incompetence trumps self-perceived incompetence in predicting sexual harassment. Importantly, Fear of Negative Evaluation scores predicted sexual harassment when controlling for Self-Efficacy and Narcissism scores among male participants, but not among female participants.

General Discussion

Who are the people especially likely to engage in sexual harassment of their subordinates? Although prior research suggests that men who are more motivated to seek sexual gratification (e.g., who are higher in sexual gratification or have mating goals) become more inclined toward sexual harassment when in positions of power (Bargh et al. 1995; Kunstman and Maner 2011), less is known about whether dispositional predictors associated with motives to protect one’s threatened status may also predict men’s sexual harassment of female subordinates (see Berdahl 2007a, b). The current studies fill this gap in the literature by demonstrating a critical determinant of male power-holders’ tendencies to engage in sexual harassment: the fear of being perceived as incompetent in one’s power. Although this fear of incompetence among powerful individuals has been linked to general aggression (Fast and Chen 2009), the relationship among power, fear of incompetence, and general aggression is not moderated by participants’ gender. By contrast, we focus on sexual harassment, in which men have been theorized to be more motivated to engage due to a need to preserve their dominant social status (Berdahl 2007a, b). As such, we show that fear of being perceived as incompetent predicts greater sexual harassment in male, but not in female, participants. We also differentiate between self-perceptions and concerns about others’ perceptions of incompetence in predicting sexual harassment. In finding that the former plays more of a role in sexual harassment than the latter, we provide support for the theory that powerful men are especially inclined to sexually harass to the extent that they worry their incompetence will be uncovered by others (Berdahl 2007a, b).

We replicated our effects across three studies, using different samples (i.e., adults in Mechanical Turk in Studies 1 and 3; college students in Study 2), different operationalizations of power, and different operationalizations of sexual harassment, which included both quid pro quo sexual harassment (i.e., offering to provide work-related favors in exchange for sexual favors; Pryor 1987) and gender harassment (i.e., behavior that creates a hostile environment for the target based on his/her gender; Berdahl 2007a). Specifically, we held power constant and assessed self-reported likelihood of sexual harassment in Studies 1 and 3, whereas we primed power and assessed behavioral sexual harassment in Study 2. Furthermore, in Study 3 we demonstrated that the relationship between fear of being perceived as incompetent (i.e., Fear of Negative Evaluation scores) and sexual harassment of subordinates whose gender matches the gender to which participants are attracted holds among male, but not among female, participants. The variability in our procedures therefore helps to increase the generalizability of the present results across different paradigms and contexts.

Practice Implications

Much previous research on sexual harassment has focused on characteristics of the targets, rather than on the perpetrators, of harassing behaviors. In other words, research on sexual harassment tends to center on how women perceive harassment (e.g., Rotundo et al. 2001), cope with harassment (e.g., Gutek and Koss 1993; McLaughlin et al. 2017), and whether and how women report harassment (e.g., Bergman et al. 2002; Fitzgerald et al. 1995). Though such outcomes are undoubtedly important to study, understanding how to prevent sexual harassment from happening in the first place is also critical to improving women’s experiences at work. Our studies contribute to this understanding, along with a small but growing body of other work on the factors that increase likelihood of sexual harassment (e.g., Bargh et al. 1995; Bushman et al. 2003; Hitlan et al. 2009; Kunstman and Maner 2011; Pryor 1987; Williams et al. 2017). Thus, our studies shed light on when sexual harassment may occur, an area in which research is sorely lacking (Quick and McFadyen 2017).

More specifically, the present findings uncover a novel dispositional predictor of when power-holders (especially men) are prone to engage in sexual harassment. That is, we consistently found that the fear of being seen as incompetent predicts sexual harassment among male power-holders. In so doing, our findings corroborate the theory that sexual harassment is, in part, a byproduct of feelings of threat to, and motives to maintain, one’s social status (Berdahl 2007a, b). Further, the fact that we obtained similar results on measures of gender harassment as well as quid pro quo sexual harassment suggests that men sexually harass women not necessarily because they seek sexual gratification, but rather because their insecurity about being perceived as incompetent prompts them to want to undermine women’s position in the social hierarchy. If our results simply reflected motives for sexual gratification, we would not have replicated them on the measure of gender harassment in Study 2 because this measure did not involve seeking sexual favors. In practice, these findings underscore the importance of construing sexual harassment more broadly (i.e., not just in terms of solicitation of sexual favors; see Berdahl 2007a), and of maintaining organizational cultures that do not foster feelings of insecurity.

Our results also extend and have implications for recent research demonstrating an effect of chronic low power on sexual harassment. Whereas this recent research showed that powerless individuals placed into high-power situations are particularly likely to engage in harassment (Williams et al. 2017), we focused on powerful individuals who are insecure about others’ perceptions of them. However, one commonality between our studies and those of Williams et al. (2017) could be that sexual harassment is triggered by a “mismatch” between one’s current state of power (or lack thereof) and one’s feelings. Whether this mismatch results in asserting the power that one has been chronically denied (as in the Williams et al. 2017, studies) or asserting the power that one perceives others think s/he is incapable of holding, it bears nefarious consequences for men’s treatment of women in the workplace.

Limitations and Future Research Directions

One limitation of our studies is that the means and standard deviations of the Likelihood to Sexually Harass scale (i.e., LSH scale) were relatively low. These low means and standard deviations likely reflect social desirability—the tendency to want to respond to surveys in socially acceptable ways (Fisher 1993). That is, very few people want to openly admit to having sexually harassed someone else. The low means and standard deviations on the LSH scale is a limitation for Studies 1 and 3; however, in Study 2, we measured actual behavior rather than relying on self-reports, which supports the overall validity of our set of studies. Furthermore, even though social desirability tendencies likely lowered the overall means on our self-reported dependent measures, we do not have a theoretical reason to believe that social desirability differed as a function of fear of being perceived as incompetent. If anything, social desirability might negatively predict fear of being perceived as incompetent, which would have worked against our findings that such fears were associated with a greater likelihood of sexual harassment (i.e., a socially undesirable behavior) among power-holders.

Importantly, the behavioral harassment study was also the study that demonstrated gender harassment. Although there are some theoretical distinctions between gender and quid pro quo sexual harassment (Berdahl 2007a; Fitzgerald et al. 1995), both types of harassment bear negative consequences for the target. However, it will be important for future research to confirm that such variables can also predict quid pro quo harassment in a behavioral study. Research ethics will need to be carefully considered when designing such a study.

Another limitation of the present research is that Study 2 had a lower-than-ideal number of male participants (n = 59) in the final sample, due to the disproportionately low male-to-female ratio in many college psychology courses. However, the non-college samples in both Studies 1 and 3 were more gender-balanced, which increases our confidence in the overall robustness and replicability of our findings.

Future research should examine how to reduce acts of sexual harassment. In showing that fear of being seen as incompetent increases power-holders’ likelihood of harassment, our studies suggest that employees’ perceptions of how others see them may be important. In particular, fostering a work environment in which employees make it clear that they recognize one another’s areas of strength and competence may help to reduce fears of negative evaluation. By contrast, organizational cultures that perpetuate insecurities (e.g., through frequent gossip, creating vast psychological distance between supervisors and employees, or instilling a sense of zero-sum competition) may as a consequence increase instances of sexual harassment, especially when such insecurities are present in supervisors. Additionally, it is possible that sexual harassment training programs would help attenuate the relationship between fear of negative evaluation and harassment tendencies. Such programs may increase awareness among individuals, regardless of their concerns about being seen as incompetent, about what does and does not constitute appropriate behavior in the workplace.

The topic of when gender does and does not influence power-holders’ sexual harassment tendencies is also worthy of future investigation. Although we hypothesized and found that the relationship between fear of being seen as incompetent and sexual harassment was only present among male power-holders, participants’ gender did not moderate the results of the Williams et al. (2017) studies in which being low in chronic and high in acute power increased the likelihood of sexual harassment. One possible reason is that in our studies, participants who worried that others saw them as incompetent were faced with the threat of losing the power they currently had. By contrast, chronically low-power participants in the Williams et al. studies newly found themselves in a powerful position. Some research has demonstrated that the prospect of losing social status has stronger psychological consequences than does gaining social status (Pettit et al. 2010). Given this research, as well as the aforementioned theorizing that men should be more motivated to protect their dominant position than women (Berdahl 2007a, b), men may experience more adverse reactions than women do to potential status losses, especially when they have the opportunity to react by sexually harassing their subordinates. However, occupying a high-power position to which one is not accustomed, as in the Williams et al. studies, may affect men and women similarly.

Conclusions

In sum, our research examines a new individual-difference variable related to sexual harassment among male power-holders: fear of being perceived as incompetent. Across three studies, we have shown that powerful men are particularly inclined to perpetrate sexual harassment behavior to the extent that they score higher in Fear of Negative Evaluation. It is our hope that our research will inspire further investigation into when sexual harassment of female subordinates is most likely to occur. In the case of the numerous high-profile men who have been accused or convicted of sexual harassment, perhaps it is not necessarily that these individuals were granted too much power. Instead, their behaviors may have stemmed more from a belief that others found them ill-suited for their dominant position and hence from a fear that others would consider them incapable or undeserving of power.

Notes

Compliance with Ethical Standards

The material in this paper has not been previously published and is not under consideration for publication elsewhere. All who participated provided informed consent. The paper complies with APA ethical standards, and all authors have seen and approved of the paper in its current form. All research in this manuscript was completed at Ohio University although the first author has since changed her affiliation.

Supplementary material

11199_2018_938_MOESM1_ESM.docx (16 kb)
ESM 1 (DOCX 15 kb)

References

  1. Ames, D. R., Rose, P., & Anderson, C. P. (2006). The NPI-16 as a short measure of narcissism. Journal of Research in Personality, 40(4), 440–450.  https://doi.org/10.1016/j.jrp.2005.03.002.CrossRefGoogle Scholar
  2. Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk-taking. European Journal of Social Psychology, 36(4), 511–536.  https://doi.org/10.1002/ejsp.324.CrossRefGoogle Scholar
  3. Bargh, J. A., Raymond, P., Pryor, J. B., & Strack, F. (1995). Attractiveness of the underling: An automatic power→ sex association and its consequences for sexual harassment and aggression. Journal of Personality and Social Psychology, 68(5), 768–781.  https://doi.org/10.1037/0022-3514.68.5.768.CrossRefPubMedGoogle Scholar
  4. Berdahl, J. L. (2007a). Harassment based on sex: Protecting social status in the context of gender hierarchy. Academy of Management Review, 32, 641–658.CrossRefGoogle Scholar
  5. Berdahl, J. L. (2007b). The sexual harassment of uppity women. Journal of Applied Psychology, 92, 425–437.  https://doi.org/10.1037/0021-9010.92.2.425.CrossRefPubMedGoogle Scholar
  6. Bergman, M. E., Langhout, R. D., Palmieri, P. A., Cortina, L. M., & Fitzgerald, L. F. (2002). The (un) reasonableness of reporting: Antecedents and consequences of reporting sexual harassment. Journal of Applied Psychology, 87(2), 230–242.  https://doi.org/10.1037//0021-9010.87.2.230.CrossRefPubMedGoogle Scholar
  7. Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.  https://doi.org/10.1177/1745691610393980.CrossRefPubMedGoogle Scholar
  8. Bushman, B. J., Bonacci, A. M., Van Dijk, M., & Baumeister, R. F. (2003). Narcissism, sexual refusal, and aggression: Testing a narcissistic reactance model of sexual coercion. Journal of Personality and Social Psychology, 84(5), 1027–1040.  https://doi.org/10.1037/0022-3514.84.5.1027.CrossRefPubMedGoogle Scholar
  9. Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a new general self-efficacy scale. Organizational Research Methods, 4, 62–83.  https://doi.org/10.1177/109442810141004.CrossRefGoogle Scholar
  10. Chen, G., Gully, S. M., & Eden, D. (2004). General self-efficacy and self-esteem: Toward theoretical and empirical distinction between correlated self-evaluations. Journal of Organizational Behavior, 25(3), 375–395.  https://doi.org/10.1002/job.251.CrossRefGoogle Scholar
  11. Cleveland, J. N., & Kerst, M. E. (1993). Sexual harassment and perceptions of power: An under- articulated relationship. Journal of Vocational Behavior, 42(1), 49–67.  https://doi.org/10.1006/jvbe.1993.1004/.CrossRefGoogle Scholar
  12. Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum Associates, Inc..Google Scholar
  13. Connell, R. W. (1987). Gender and power. Stanford, CA: Stanford University Press.Google Scholar
  14. Criss, D. (2017, Nov 1). The (incomplete) list of powerful men accused of sexual harassment after Harvey Weinstein. CNN. Retrieved from cnn.com.
  15. Fast, N. J., & Chen, S. (2009). When the boss feels inadequate: Power, incompetence, and aggression. Psychological Science, 20(11), 1406–1413.  https://doi.org/10.1111/j.1467-9280.2009.02452.x.CrossRefPubMedGoogle Scholar
  16. Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191.  https://doi.org/10.3758/bf03193146.CrossRefPubMedGoogle Scholar
  17. Ferguson, A. J., Ormiston, M. E., & Moon, H. (2010). From approach to inhibition: The influence of power on responses to poor performers. Journal of Applied Psychology, 95(2), 305–320.  https://doi.org/10.1037/a0018376.CrossRefPubMedGoogle Scholar
  18. Fineran, S., & Bennett, L. (1999). Gender and power issues of peer sexual harassment among teenagers. Journal of Interpersonal Violence, 14(6), 626–641.  https://doi.org/10.1177/088626099014006004.CrossRefGoogle Scholar
  19. Fisher, R. J. (1993). Social desirability bias and the validity of indirect questioning. Journal of Consumer Research, 20(2), 303–315.  https://doi.org/10.1086/209351.CrossRefGoogle Scholar
  20. Fiske, S. T. (1993). Controlling other people: The impact of power on stereotyping. American Psychologist, 48(6), 621–628.  https://doi.org/10.1037//0003-066x.48.6.621.CrossRefPubMedGoogle Scholar
  21. Fitzgerald, L. F., Swan, S., & Fischer, K. (1995). Why didn't she just report him? The psychological and legal implications of women’s responses to sexual harassment. Journal of Social Issues, 51(1), 117–138.  https://doi.org/10.1111/j.1540-4560.1995.tb01312.x.CrossRefGoogle Scholar
  22. Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of Personality and Social Psychology, 85(3), 453–466.  https://doi.org/10.1037/0022-3514.85.3.453.CrossRefPubMedGoogle Scholar
  23. Galinsky, A. D., Magee, J. C., Inesi, M. E., & Gruenfeld, D. H. (2006). Power and perspectives not taken. Psychological Science, 17(12), 1068–1074.  https://doi.org/10.1111/j.1467-9280.2006.01824.x.CrossRefPubMedGoogle Scholar
  24. Gutek, B. A., & Koss, M. P. (1993). Changed women and changed organizations: Consequences of and coping with sexual harassment. Journal of Vocational Behavior, 42(1), 28–48.  https://doi.org/10.1006/jvbe.1993.1003.CrossRefGoogle Scholar
  25. Harvey Weinstein timeline: How the scandal unfolded (2018, Feb 12), BBC News. Retrieved from bbc.com/news.
  26. Hitlan, R. T., Pryor, J. B., Hesson-McInnis, M. S., & Olson, M. (2009). Antecedents of gender harassment: An analysis of person and situation factors. Sex Roles, 61(11–12), 794–807.  https://doi.org/10.1007/s11199-009-9689-2.CrossRefGoogle Scholar
  27. Kantor, J., & Abrams, R. (2017, Oct 10). Gwyneth Paltrow, Angelina Jolie and others say Weinstein harassed them. The New York Times. Retrieved from nytimes.com.
  28. Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110(2), 265–284.  https://doi.org/10.1037/0033-295x.110.2.265.CrossRefPubMedGoogle Scholar
  29. Kunstman, J. W., & Maner, J. K. (2011). Sexual overperception: Power, mating motives, and biases in social judgment. Journal of Personality and Social Psychology, 100(2), 282–294.  https://doi.org/10.1037/a0021135.CrossRefPubMedGoogle Scholar
  30. Leary, M. R. (1983). A brief version of the Fear of Negative Evaluation Scale. Personality and Social Psychology Bulletin, 9(3), 371–375.  https://doi.org/10.1177/0146167283093007.CrossRefGoogle Scholar
  31. Maass, A., Cadinu, M., Guarnieri, G., & Grasselli, A. (2003). Sexual harassment under social identity threat: The computer harassment paradigm. Journal of Personality and Social Psychology, 85(5), 853–870.  https://doi.org/10.1037/0022-3514.85.5.853.CrossRefPubMedGoogle Scholar
  32. McLaughlin, H., Uggen, C., & Blackstone, A. (2017). The economic and career effects of sexual harassment on working women. Gender and Society, 31(3), 333–358.  https://doi.org/10.1177/0891243217704631.CrossRefPubMedGoogle Scholar
  33. Pettit, N. C., Yong, K., & Spataro, S. E. (2010). Holding your place: Reactions to the prospect of status gains and losses. Journal of Experimental Social Psychology, 46(2), 396–401.  https://doi.org/10.1016/j.jesp.2009.12.007.CrossRefGoogle Scholar
  34. Popovich, P. M., & Warren, M. A. (2010). The role of power in sexual harassment as a counterproductive behavior in organizations. Human Resource Management Review, 20(1), 45–53.  https://doi.org/10.1016/j.hrmr.2009.05.003.CrossRefGoogle Scholar
  35. Pryor, J. B. (1987). Sexual harassment proclivities in men. Sex Roles, 17(5–6), 269–290.  https://doi.org/10.1007/bf00288453.CrossRefGoogle Scholar
  36. Pryor, J. B., & Stoller, L. (1994). Sexual cognition processes in men high in the likelihood to sexually harass. Personality and Social Psychology Bulletin, 20(2), 163–169.  https://doi.org/10.1177/0146167294202003.CrossRefGoogle Scholar
  37. Pryor, J. B., LaVite, C. M., & Stoller, L. M. (1993). A social psychological analysis of sexual harassment: The person/situation interaction. Journal of Vocational Behavior, 42(1), 68–83.  https://doi.org/10.1006/jvbe.1993.1005.CrossRefGoogle Scholar
  38. Quick, J. C., & McFadyen, M. (2017). Sexual harassment: Have we made any progress? Journal of Occupational Health Psychology, 22(3), 286–298.  https://doi.org/10.1037/ocp0000054.CrossRefPubMedGoogle Scholar
  39. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.CrossRefGoogle Scholar
  40. Rotundo, M., Nguyen, D. H., & Sackett, P. R. (2001). A meta-analytic review of gender differences in perceptions of sexual harassment. Journal of Applied Psychology, 86(5), 914–922.  https://doi.org/10.1037//0021-9010.86.5.914.CrossRefPubMedGoogle Scholar
  41. Schwarzer, R., & Jerusalem, M. (1995). Generalized Self-Efficacy Scale. In J. Weinman, S. Wright, & M. Johnston (Eds.), Measures in healthy psychology: A user’s portfolio: Casual and controlled beliefs (pp. 35–37). Windsor, UK: NFER-NELSON.Google Scholar
  42. Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science, 22(11), 1359–1366.  https://doi.org/10.1177/0956797611417632.CrossRefPubMedGoogle Scholar
  43. Williams, M. J., Gruenfeld, D. H., & Guillory, L. E. (2017). Sexual aggression when power is new: Effects of acute high power on chronically low-power individuals. Journal of Personality and Social Psychology, 112(2), 201–223.  https://doi.org/10.1037/pspi0000068.CrossRefPubMedGoogle Scholar
  44. Zeigler-Hill, V. (2006). Discrepancies between implicit and explicit self-esteem: Implications for narcissism and self-esteem instability. Journal of Personality, 74(1), 119–144.  https://doi.org/10.1111/j.1467-6494.2005.00371.x.CrossRefPubMedGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Department of PsychologyOhio UniversityAthensUSA
  2. 2.Office of Student Life, Center for the Study of Student LifeThe Ohio State UniversityColumbusUSA

Personalised recommendations