Evidence from five studies focusing on different factors that influence interpretation of different behaviors constituting different forms of sexual harassment and gender harassment is presented. Six factors of influence were studied: (1) physical attractiveness and (2) quality of financial resources of the harasser, (3) gender and (4) attitudes of the harassee, (5) interpreting derogatory versus interpreting "flirtatious" behavior, and (6) interpreting actual harassment versus interpreting imagined harassment presented in a scenario (Note: With actual harassment, harassment that people actually and currently experience is meant, as compared to harassment that people imagine to experience or remember having actually experienced in the past).

In Study 1, women's interpretations of two forms of actual sexual and gender harassment were investigated under controlled laboratory conditions. The two types of harassment were explicitly derogatory toward the target or constituted sexual attention. These harassment types were operationalized as sexist jokes and, e.g., as remarks about physical appearance of the participant. In the control conditions, neutral jokes and remarks replaced harassing jokes and remarks. In addition, harasser attractiveness was varied. To study reactions to actual behavior, a new methodology was developed, the Reversed Computer Harassment Paradigm. In this paradigm, participants believed to be connected online to a real, but in fact computer simulated chat partner, who sent different materials via a chat line. Results of Study 1 show that participants evaluate non-harassers and attractive men and their behavior more favorably than harassers and unattractive men and their behavior. However, the attractiveness effect found on more indirect rating dimensions disappeared when participants were alerted to the sexually harassing quality of the behavior: Answering the direct question whether the behavior was sexually harassing, participants no longer judged the attractive harasser more favorably.

Study 2 was a part replication of Study 1: Female participants evaluated a scenario of Study 1, without attractiveness information, and indicated their imagined responses. Comparing Studies 1 and 2, results mirror a distinction found in previous research: When judging scenarios, participants overestimate assertive responses, compared with responses of survey participants who report reactions to past harassment experiences. This is only the second time this pattern was replicated in a study on actual harassment, and the first time replication was done under strict experimental control.

Study 3, in a scenario, disentangled the influences of physical attractiveness and quality of economic resources of the harasser, which were confounded in Study 1. In addition, predictions made by the evolutionary and the socio-structural perspective on sexual harassment were compared. Drawing on the notion of inclusive fitness, the (imagined) relationship between female participants and the harassed woman in the scenario was varied: The crucial comparison was between genetic plus social relatedness to the target (sister) and only social relatedness (female friend), or no relatedness (female stranger). Again, an attractiveness effect was found in the same direction as before, but economic resources of the harasser had no effect. Type of relationship had unexpected effects: Behavior toward a sister and a female friend was judged similarly.

In Study 4 and Study 5, men's interpretations of sexual and gender harassment were the focus. Study 4 closely resembled Study 1 in its methodology and is - to my knowledge - the pioneering study focusing on actual harassment of men under controlled laboratory conditions. Type of harassment and attractiveness of the female harasser was varied. Results resembled, but did not fully replicate, those of Study 1: Male participants, too, judged a non-harasser and her behavior more favorably than a harasser and her behavior. However, the attractiveness effect was only replicated as a tendency on a few response dimensions. Study 5 was a scenario version of Study 4 and closely resembled Study 2. Again, similarly to female participants, male participants heavily overestimated the proportion of assertive responses, compared with actual responses of participants in Study 4.

Together, these five studies constitute a series of experiments that add to the as yet meager knowledge-base about reactions to and interpretations of actual harassment of women and men under controlled laboratory conditions and that challenge common notions about how men and women experience sexual and gender harassment. In addition, they demonstrate the usefulness of the Reversed Computer Harassment Paradigm for studying actual harassment under full experimental control and in an ethically appropriate way.