When and Why Gender Differences Emerge for Sexual Harassment Behavior

by Jessica A. Kennedy, Laura J. Kray, and Michael Rosenblum

When and Why Gender Differences Emerge for Sexual Harassment Behavior

Image Credit |

We describe motives and identity of underlying sexual harassment behaviors in the workplace.

When former New York Governor Andrew Cuomo said he only meant to be playful and funny when he engaged in behavior that many staffers perceived as sexually harassing, he may have struck a chord. Survey data from Pew suggest that over half of U.S. adults think increased focus on sexual harassment has made it harder for men to know how to interact with women in the workplace, and over a third of men report fearing that false accusations of sexual harassment could be a major problem.

Related CMR Articles

“Changing the Narrative: Women as Negotiators—and Leaders” by Laura J. Kray & Jessica A. Kennedy. (Vol. 60/1) 2017.

“Leading through Negotiation: Harnessing the Power of Gender Stereotypes” by Laura J. Kray. (Vol. 50/1) 2007.

We hope that our new research helps to reduce these concerns by highlighting the role of specific motives in producing behaviors likely to be construed as sexual harassment. We conducted a series of studies with 2,667 participants to better understand when and why people engage in sexual behaviors in task settings. Attraction didn’t predict these behaviors in our studies or in prior research, but desire to appear powerful did.

The Motive of Harassers

In one study with 607 online participants (357 men, 250 women) who were on average 34 years old, people expected to get acquainted with a coworker of the other gender. We gave them some basic written information about the person before allowing them to select questions for conversation. Participants were given eight pairs of questions and could choose one from each pair to ask. One question in each pair was a normal interview question such as, “Have you ever had a workplace conflict?”, “Are you better at math or writing?”, and “Are you good at meeting deadlines?” The other question in the pair contained sexual innuendo, such as, “Have you ever had a workplace relationship?”, “Are you better at giving or receiving?”, and “Are you good at setting boundaries?” Our key dependent measure was how many of the sexually loaded questions participants chose to ask their new coworker. Our pretests suggested that people knew these questions would be offensive to others.

Our key experimental variable was the motive with which participants entered the get-acquainted task. Initially, we assigned people to one of four motive conditions. In the power motive condition, participants aimed to demonstrate they were more important in the workplace than their coworker. In the agentic motive condition, participants aimed to demonstrate an action orientation and effectiveness at pursuing their goals. In the affiliative motive condition, participants aimed to establish a positive relationship with their coworker. Finally, in the control condition, participants received no motive instructions. Psychologically, the power motive condition was identical to the agentic motive condition by producing intentions to pursue self-interest, achievement, image, and power. Conceptually, both the power motive and agentic motive conditions are similar because they focus on enhancing the self, making it bigger, stronger, and more impressive in terms of positive qualities. Therefore, we refer to these conditions together as the self-enhancement motive condition. Similarly, the affiliative and control conditions produced equivalent intentions to benefit the coworker and establish a positive relationship. We refer to these two conditions together as self-transcendence motive condition.


For Interpersonal Motive

Results were telling: Men selected more sexually harassing questions to ask when they aimed to self-enhance versus self-transcend. Women’s behavior did not vary by motive condition. A gender difference emerged only when self-enhancement motives were strong. Men and women behaved identically when they intended to build a relationship with the other person. Our results point to the value of relational intentions for eliminating gender differences and highlight the risk of self-enhancement motives for men. When men aim to establish their power over others or their power to achieve goals, sexually harassing behavior is a greater risk for them than for women. To minimize this risk, men might wish to focus on establishing positive relationships with others, a goal which helps them transcend image-related goals.


For Level of Power

In another study, we found that holding a low-power position activated desire for power more than holding a high-power position. When given the get-acquainted task in a laboratory setting, men and women were asked to select questions to ask a partner of the other sex. They briefly met a research assistant posing as another participant to enhance the believability of the task. We found that men tended to ask more of the sexual questions when they lacked power than when they possessed it. Women, by contrast, did not vary in their behavior by level of power. For men, desire for power corrupts, according to our data. Gender differences emerged only when power was low and power motivation was high.


The Mindset of Harassers

We also aimed to understand what people are thinking when they engage in these behaviors. After all, people are capable of regulating their behavior in line with internal standards of what is right and wrong. Prior research had shown the importance of moral identity, the tendency to define oneself in terms of traits like caring, fair, and honest, for predicting different types of ethical behavior and for explaining when gender differences in behavior emerge and disappear, so we measured how much people would feel good (or bad) to have these characteristics. We also measured a different type of identity (reminiscent of that described by Cuomo): how much people thought of themselves as natural flirts, as playful with the other sex, and as likely to be irresistible when they want something. We referred to people who agreed more strongly with statements like these as high in social sexual identity (“SSI”).

In one study, we asked 499 participants recruited online to imagine a person with each set of traits and report how much they wanted to be that type of person (or not) using a scale from 1 (not at all) to 7 (extremely). We then gave them a scenario describing their coworker and workplace and asked them to report how likely they would be to engage in a variety of sexual behaviors like, “Look at [name] in a sexually provocative way” and “Treat [name] as a confidant you can talk to about your sexual problems.” Here, moral identity strength predicted lower likelihood of these behaviors and SSI predicted a higher likelihood. In a second, related study, we experimentally induced stronger levels of these identities. With 733 participants recruited online, we randomly assigned people to one of three conditions: stronger moral identity, stronger SSI, or a control condition.  We strengthened each identity by giving them a list of adjectives related to it and then providing the same scenario and survey items as before. Here, moral identity didn’t consistently predict sexual behaviors, but SSI did.

We aimed to understand predictors of SSI strength. In later studies, we found that stronger power motivation strengthened SSI. A chronic gender difference also emerged. Men reported stronger SSIs than women, on average, perhaps because the measure captures a degree of sexual agency. Men, more than women, are socialized to initiate sexual relationships and are shamed less for sexual behavior, possibly leading them to internalize sexuality in their identities more than women do.

We next wanted to compare the ability of SSI to predict sexual behaviors in the workplace relative to other personality characteristics and attitudes that prior research would suggest are relevant. By doing so, we could better understand whether this type of identity is an important, new predictor of sexually harassing behavior or redundant with existing explanations. To do so, we collected data online from 394 people, measuring how likely they would be to engage in a variety of sexual behaviors at work. Some of these behaviors were the type likely to be potentially welcome (e.g., making someone feel attractive and desirable) and others were likely to be offensive and deemed sexual harassment (e.g., found a way to punish a coworker who was not sexually cooperative). We measured sexism, flirting styles, narcissism, and sense of personal power. We also measured and controlled for the tendency to answer questions in socially desirable ways. SSI strength predicted both types of sexual behavior, even when controlling for all the other variables. The result indicated that how we conceive of ourselves matters for sexual harassment. Men were again higher on SSI under baseline conditions than were women. Our research suggests that activating alternative selves or decreasing the centrality of sexual traits in the self-concept might help to mitigate sexually harassing behavior.

Does It Work?

Our studies measured how powerful people felt after they chose whether to ask their coworker sexual questions. We found no evidence that these behaviors promoted feelings of power. However, in a final study, we looked at how others perceived a subordinate man who asked sexual questions of his female boss (or didn’t). Asking these questions led people to perceive the subordinate man as more masculine and more powerful relative to normal, work-related questions.


Employers are still figuring out how to deal with sexual harassment. Given that many sexual harassment programs do more harm than good, it’s a thorny problem. Our work offers a number of insights to help people identify these tendencies within themselves and to hold themselves accountable before the widely feared career outcomes emerge. 

Practical Conclusions

  1. Desire for power, not attraction, is a red flag for sexual harassment. Men who fear socializing with women from work might rest easier knowing that it’s desire for power, not social connection or physical attraction that predicted offensive sexual behaviors in our studies. Paying attention to how much you desire power could help you avoid problematic flirting. If you feel you lack power, this could serve as a warning device for flirtation that might be asymmetrically desired.

  2. Beware rationalizing your behavior as “playful” and seek feedback when in doubt. People who see themselves as “big flirts” who know how to “turn on the charm” are the most at risk for engaging in offensive sexual behaviors in work settings. If you find yourself prone to rationalizing your behavior as “just flirting,” then you might have a blind spot as to behavior that is problematic and could escalate into a major career problem.

  3. Update beliefs about who flirts at work. It is time to update false beliefs about who flirts (and sexually harasses) at work. While men do it more under baseline conditions, gender differences disappear when men aim to create strong relationships with others and when the genders hold high levels of power. Gender differences in our studies were malleable by situation, not constant. Further, notions that subordinate women are likely to compensate for their low power by flirting don’t fit the facts. Low-power women engaged in less sexual behavior than low-power men and did not differ from high-power men and women in our studies.

  4. Be conscious about perceiving others’ power. Most of us wouldn’t wish to perceive sexual harassers as more powerful for doing so, but we found that it worked. Noting and self-correcting our impressions of people’s power might help to reduce incentives for men to sexually harass women at work.

  5. Transcending the power motive levels the playing field for each gender.  Organizations might wish to promote values that help people to self-transcend instead of trying to enhance their own power and image.  When people aimed beyond themselves, the genders acted identically.

Jessica A. Kennedy
Jessica A. Kennedy Jessica Kennedy is an Associate Professor of Management at Vanderbilt University. She holds a PhD from UC Berkeley’s Haas School of Business and a BS in Economics from The Wharton School of the University of Pennsylvania. Her research aims to understand conflicts between ethical values and career outcomes.
Laura J. Kray
Laura J. Kray Laura Kray received her PhD in Social Psychology from the University of Washington. Prior to joining Cal, she taught at the University of Arizona and completed a post-doctoral fellowship at Northwestern University. Kray studies the role of gender stereotypes and mindsets on workplace behavior, including negotiations and ethical decision-making.
Michael Rosenblum
Michael Rosenblum Michael Rosenblum is a Postdoctoral Researcher at NYU’s Stern School of Business with two interconnected streams of research. First, he investigates how inequality (i.e., gender, race, socioeconomic status) impacts interpersonal judgments. Second, he studies how political discourse shapes impressions of others and how our impressions shape how beliefs are communicated.


California Management Review

Berkeley-Haas's Premier Management Journal

Published at Berkeley Haas for more than sixty years, California Management Review seeks to share knowledge that challenges convention and shows a better way of doing business.

Learn more
Follow Us