Abstract
Recent high-profile cases of sexual harassment focused the spotlight on inappropriate workplace behavior. Much of the prior research on sexual harassment focuses on organizational culture, what organizations can do to create harassment-free environments, and to increase reporting when it does occur. Less work explores what happens in the actual harassment situations, or how the immediate responses to the incivility affect future interactions. This study seeks to fill that gap by exploring effectiveness of the message responses used by female targets of sexual harassment by male harassers to curtail future harassment in the workplace. We also explore how the target’s responses affect bystanders’ perceptions of her communication effectiveness and her future potential of being promoted. Data were gathered from workers with an average of 12 years of work experience. Using a variety of sexual harassment scenarios developed for this study, we found that assertive responses were considered the most effective in supporting a positive image of the target and avoidance was the least effective. To curtail future harassment, assertiveness and assertive-empathetic responses were perceived as effective strategies. In terms of maintaining the prospects for future promotions, participants again rated assertiveness as the best strategy for the target to employ. Across all scenarios, avoidance was a poor strategy. Beyond several interesting research findings, the scenarios provide materials that could be modified for use by those who are trainers dealing with sexual harassment or could be used as a foundation for more advanced research regarding sexually harassing messages.
Keywords
The Harvey Weinstein scandal launched the #MeToo Movement that has attracted international attention. Targets of sexual harassment began speaking about the abuse they faced ranging from inappropriate comments and sexist workplaces to expectations and pressures for sexual activities. Targets implicated high-profile corporate leaders from companies such Under Armor, Fox News, Google, Sherpa Capital, and Uber. The accused face public scrutiny, career loss, and criminal charges. Yet, there are grave consequences for the reporters of harassment, as well: their credibility is often questioned, they are seen as unpromotable, or the scrutiny is so intense, they leave their jobs (Hart, 2019; Jones & Wade, 2020; Monroe, 2019). In a survey done in the months immediately following the rise of the #MeToo movement, Atwater et al. (2019) found that 16% of men in their study said they will be less likely to hire women following #MeToo with 41% of respondents indicating that men will be reluctant to be in a room alone with a woman. Zugelder (2019) argues that while most Americans believe more women will report harassment, a sizable number also believe that fear of false reports will decrease men’s mentorship of women—potentially setting gender equality in the workplace back. This type of backlash and the unintended consequences further affect how women may respond when harassed.
Even prior to #MeToo, research indicated that in everyday situations, targets have been unlikely to follow through with official reporting for various reasons including the stigma of being a victim (Hart, 2019; McDonald, 2012), the overall distrust in the organizational system (McDonald, 2012), the fear of retaliation (Bergman et al., 2002; Cortina & Magley, 2003; McDonald, 2012), and concerns for long-term damage to their careers (Hart, 2019). Historically, only 5% to 30% of targets will report sexual harassment (Johnson et al., 2016; McDonald, 2012). The consequences of reporting may be too high for some targets. Although targets can choose to report or not, once faced with harassment, they must respond in some way.
Despite a few early scholars such as Bingham (1991), Gruber and Smith (1995), and Maypole (1986) addressing communication strategies for responding to sexual harassment, recent research has not provided commensurate attention to appropriate, immediate communication messages to end that unwanted behaviors. Moreover, little contemporary research has considered the message responses of targets of sexual harassment and the effects of those responses. This study places the spotlight on target responses to sexual harassment and the resulting impact. That is, how those responses affect perceptions of the harassment target, how the responses affect perceptions of ending harassment, and how the target’s response ultimately affects others’ perceptions of her promotability.
Defining Sexual Harassment
According to the three-component model of sexual harassment advanced by Fitzgerald et al. (1995) sexual harassment has three primary forms. The first is known as quid pro quo, which aligns with the legal definition of sexual coercion, in which the harasser seeks sexual favors for career-related benefits. The second, is unwanted sexual attention including inappropriate sexual commentary, sexual advances, and inappropriate touching without the implication of reciprocal favors. The third form is also known as gender harassment, which involves targeting victims based on their gender category (typically women). Gender harassment entails insulting or demeaning targets, telling inappropriate jokes, displaying sexual material, and gender-based hazing. The latter two are often classified as creating a hostile work environment (Fitzgerald et al., 1995).
Sexual harassment varies on a continuum from explicit messages making clear the expectation of sexual behaviors as a condition of employment to uncomfortable sexual innuendos from colleagues who have no power to reward or punish their peers. Despite the varied forms of sexual harassment, Sojo, Wood, and Genat’s (2016) study demonstrated that all forms of sexual harassment have profoundly negative consequences for targets.
Social Confrontation and Politeness
Newell and Stutman (1988, 1989; Stutman & Newell, 1990) noted that life is filled with problematic situations that necessitate analyzing various situations and outcomes in response to social norms and breeches of relationship expectations. Those disruptions require some form of social resolution. Vollbrecht, et al. (1997) argue that these disruptions often require some form of response, but that the target feels undue pressure to respond in socially appropriate ways to avoid further social violations (negative consequences) by her or his communication strategies.
In choosing among response options, the target is constrained by what Brown and Levinson (1987) term “politeness” which is the need to maintain social order by supporting the impressions that others are attempting to manage. Drawing on Goffman (1967), Brown and Levinson (1987) argue that to maintain politeness, communicators must attend to the positive face of others: the identity that that others are crafting to enable them to perform their social role. Vollbrecht, et al. (1997) further suggest that targets must balance the concern for others’ face against maintaining their own. Yet preserving the face of the harasser may limit the effectiveness of the message to end sexually offensive behaviors. These multiple concerns may be, and often are, in conflict.
Brown and Levinson’s (1987) politeness theory suggests that people usually try to minimize damage to other people’s identity presentation when presenting potentially challenging information, but there are exceptions where individuals are strategically rude. Politeness theory considers power differences between the speaker and hearer, the social distance, and the potential for damage. Brown and Levinson (1987) note that people often choose one of the following five strategies: refrain from initiating the face-threatening act (avoidance), going off-record (indirectly noting the issue or hinting), negative politeness (apologizing for infringing on the autonomy of the hearer), positive politeness (providing a compliment or act of kindness to accompany the threat), and going bald on record (being direct). Yet in the realm of sexual harassment other factors are involved in the choice of communication strategies (Bingham, 1991; Bumiller, 1988; Hesson-McInnis & Fitzgerald, 1997) such as ensuring the behavior ceases.
Bingham (1991) noted that effectively deflecting sexual communication at work is fraught with multiple problems including balancing the multiple goals of organizational success, personal respect, and relational maintenance. Bingham details the multiple levels of conflict and incompatible goals that may occur as a target considers how to respond. For instance, a target might be concerned that addressing the harassment might result in hurt feelings or retaliation, yet by failing to address the behavior might result in continued harassment. In addition, studies have shown that some responses, including the much endorsed “assertive” response can further damage the target. Hesson-McInnis and Fitzgerald (1997) found that women’s assertive responses to harassment often escalate harassment. Retaliation could include a female subordinate being blocked from regular promotions. Conversely, even acknowledging the act of harassment by targets means admitting that they have become targets (Bumiller, 1988). Thus, responding to sexual harassment often poses a no-win situation, or a classic double-bind (Cortina & Magley, 2003). These issues led us to consider three research questions regarding the message choices that victims make in response to sexual harassment and some of the resulting implications.
Method
This study employed four hypothetical scenarios designed specifically for this study to explore the evaluations of responses to varying types of sexually harassing comments that targeted a female employee. Although males are targets of sexual harassment, the conditions, responses, and outcomes are often quite different (Scarduzio et al., 2018). For this study, we focus on female targets with male harassers.
These four scenarios represent one form of treatment or independent variable. The first scenario involves a manager making a very blatant suggestion of showing some skin that would not be appropriate. This second was a harassing message from a manager containing a clear quid pro quo situation. Both were severe scenarios. Moderate scenarios were those that involved joking or innuendo that allowed the perpetrator of equal power, a coworker, the ability to deny intent, whereas in the severe scenarios, harassing behavior was direct and less ambiguous (see Jacobson & Eaton, 2018).
Each respondent received only one of the four scenarios. We recognize that type of scenario does not predict the long-term damage or impact from the harassment, and that scenarios often deemed minor by observers can have devastating, long-term impact for the target (Scarduzio et al., 2018).
The second independent or treatment variable that was manipulated involved the type of message responses made by targets in the four scenarios. Each respondent only received one response in one of the four scenarios. Although qualitative work is often used for exploratory work, Franklin (2005) argues that with a wide variety of instruments available to researchers, and theoretical suggestions that point the way to connections, quantitative work can often help scholars explore relationships among and between variables in efficient and insightful ways. Guided by prior work on sexual harassment and communicative competence, we developed scenarios that capture some of the larger issues entailed in harassing comments and responses to begin understanding the importance of responses to such acts. As with qualitative exploratory research, subsequent studies can begin to refine these categories even further.
Participants
Working adults, across a variety of occupations, were used in this study because we were interested in participants who had been in the workforce and might be more familiar with the social and organizational experiences than student samples. A log-rolling technique was used to recruit participants. Students in a research methods course, at a large Midwestern university, recruited participants to complete a questionnaire with the following conditions: participants had to be 18 years old or older, could not be college students, and had to be employed for at least 20 hours per week. A total of 283 (158 women, 125 men) completed the survey, nearly all were from the Midwest. The collection resulted in 290 returned questionnaires, but only 283 had sufficient data for most of the analyses conducted in this study. A few cases with missing data resulted in some analyses being conducted with as few as 278 participants. Participants averaged 32.5 years old with a median age of 25. On average, participants had worked 12.1 years with an extremely high standard deviation of 11.4 years that indicated a high diversity in work experience. The sample was 76% White non-Latino, 14% African American, 3% Asian, 2% Latino/Latina, 3% were other, and 10 persons did not identify ethnicity.
Procedures
After obtaining institutional review board approval, students in a research methods course were instructed to disseminate the hypothetical experimental situations to participants based on the above participant criteria.
Participants were presented with a study packet that began with demographic information, as well as a consent form. After consenting, they were presented with the study instructions and pictures of two persons who worked together, Susan and Tom. Because of the number of cells required to cross all the conditions and message types, only pictures of a white male and white female were used in this initial study. This limitation is addressed later.
The pictures were chosen from the publicly available website areyouhotornot.com . They were selected because Susan and Tom appeared equivalent in age, and both were rated as similar in attractiveness (7.0 on a 10-point scale) by thousands of raters.
To ensure that participants adequately read and understood the scenario, they were asked to answer the question, “In your own words, describe what happened in the above scenario.” The participants who failed to answer the question, or did not properly represent the situation, were excluded (n = 19).
Then participants were given one of four hypothetical scenarios in which Tom said something inappropriate to Susan. In two scenarios, he was presented as a boss, and in two others, he was presented as a coworker. The intent was to create four scenarios that varied in their level of sexual harassment. It was not to directly test the power differential of the power of a boss in contrast to a colleague. In the first of the boss situations, Tom led Susan to a storage room, noted the lack of cameras and suggested (nonverbal threat), “This is a good place to show me some skin since it is the only place without security cameras.” In this scenario, Susan was being harassed, and could be fearing sexual assault. This was the most explicit message and expected to be the most sexually harassing. The second boss scenario asked Susan for a sexual favor in a veiled way, “I’ll promote you when you give me something the guys can’t give me.” It was a more indirect message, but it suggests a quid pro quo. This is consistent with what Gruber (1992) called “sexual bribery.” In this situation, though covert, the implication is that Susan’s career success hinges on her sexual activity with the boss. In the coworker scenarios, Tom makes an inappropriate personal remark about Susan’s appearance, “Your summer wardrobe is much more flattering than your winter one. It’s too bad that winter clothing doesn’t show off more of your better features.” This message does not suggest a quid pro quo. It can create a hostile work environment. It was expected to be third in perceived sexual harassment. In the last scenario, Tom tells an off-color sexual joke: I heard a funny joke this weekend. A man walks into a hotel lobby. He wants to ask the clerk a question. As he turns to go to the front desk, he accidentally bumps into a woman, and as he does his elbow goes into her breast. They are both startled, and he says, “Ma’am if your heart is as soft as your breast, I know you’ll forgive me.” She replies, “If your penis is as hard as your elbow, I’m in room 1221.”
This use of humor was designed to be even more indirect harassment than the third scenario. Research indicates that sexual jokes in the work place are not simply inappropriate, but present a form of sexual harassment that is hard to confront because it is often excused as just playing around despite the anxiety and negativity they create (Clason, 2019).
For each of the four scenarios, message responses were created for Susan who was the target of Tom’s sexual harassment in each of the four scenarios. Message response types were the second independent or treatment variable that was manipulated. These messages were designed to be consistent with Bingham’s (1991) typology of four types of responses that arise from the assertiveness literature: assertive, nonassertive (ignoring the comment), aggressive, and assertive-empathic. Assertive responses are those that address inappropriate or unwanted ways in direct, forthcoming, and honest ways that make the violation clear. Assertive-empathetic responses add the concern for the offenders’ feelings or face needs. Nonassertive responses hide the targets’ emotions and typically involve silence or avoidance. Aggressive responses address the behavior in honest and direct ways, but also entail attacking or insulting the offender.
Some literature suggests humor might be an effective response. Opinions are mixed with some suggesting humor might be a way of minimizing the impact of a transgression—or even normalizing it (see Clason, 2019; Dougherty & Smythe, 2004). So, we added a humor response resulting in a total of five message response types. In the humor response conditions, the target makes the violation obvious, but does so in a way the offender can pretend that he or she is part of a mutually positive relationship; so, the offense may be viewed as “no big deal.”
Significance tests compared individual response categories using the post hoc least significant difference (LSD) procedure available in the univariate analysis of variance (ANOVA) procedure in SPSS. Crossing these two independent variables in a univariate ANOVA resulted in 20 cells. The ANOVA design allowed us to assess whether there were any significant interactions among the scenarios and the message response categories. We also checked for significant gender effects based on the reported gender of the participants.
Significant ANOVAs were followed by Fisher’s LSD tests that balance the power of the test while controlling the familywise error rate (Hayter, 1986). Levin et al. (1994) argue that many traditional post hoc procedures are overly conservative and reduce the power of the test to control familywise error rate. Fisher’s LSD procedure balances statistical power and controls the familywise error rate substantially over doing the same number of comparisons using t tests at the .05 level.
The following examples explain how the five messages categories were instantiated within the summer wardrobe comments: Susan makes eye contact with Tom and replies, “Wow, I’m a bit startled that you would say something like that to me. Your making comments about how my clothes look on me is completely inappropriate.” (assertive) Susan sighs, looks Tom in the eye, and responds, “I have liked working with you and have found you to be an understanding colleague. So, I hope you don’t take offense by my saying that your comment was out of line. I would like it if you would refrain from making comments such as that in the future” (assertive-empathetic). Susan looks down, keeps working on the project in front of her and waits for her colleague to leave the office (nonassertive/ignoring). Susan sternly responds, “You have a lot of nerve you jerk. I refuse to believe that you don’t know that your comment is inappropriate and rude. Even thinking such things is blatantly wrong and you need to learn to keep comments like that to yourself if you want to succeed in this business” (aggressive). Susan laughs and replies, “I can’t believe that someone who would wear a tie like that would comment on my wardrobe” (humor).
The perceptions of these response message categories were rated as high, neutral, or low. Neutral was a response within the confidence range of the midpoint on the scale that was scored as 4.0. The instrument clearly indicated that 4.0 was neutral for all dependent variables. High responses on each dependent variable were values exceeding the confidence interval range for the neutral point of 4.0. Low values were those below the confidence interval range of the neutral category.
Dependent Variables
Sexual Harassment Judgment Scale
This scale consisted of three items with an alpha reliability of .93. It included items “Tom’s behavior constitutes sexual harassment,” “Tom’s actions were sexually harassing,” and “Tom sexually harassed Susan.” The scale had a maximum value of 7.0 representing high perceived harassment and a value of 1.0 being low perceived harassment. The scale neutral point was 4.0 as it was on all of the scales. It was used to assess whether the different scenarios were judged as being sexually harassing.
The Threat-Intimidation-Harm Scale
This scale contained five items with an alpha reliability of .82. It varied from a low of 1.0 to a high of 7.0. Items included the following: Tom’s actions were harmful, Tom harmed Susan, Tom’s actions threatened Susan, Tom intimidated Susan, and Tom said threatening things to Susan. This was a second dimension of sexual harassment measuring the degree of threat, intimidation, and harm each scenario was judged to contain.
Perceptions of Target’s Response Effectiveness
The Target’s Response Effectiveness variable was a three-item scale created to measure the perception of the effectiveness of a message issued by the target of an incident involving sexual harassment in one of the four scenarios previously discussed. Scores ranged from 1.0 representing the lowest level of effectiveness to 7.0 indicating the highest level of effectiveness. Those items included items such as “In my opinion Susan handled things in an effective way.” As well as, “I think Susan handled the situation in a good way,” and, “Overall I think Susan responded to Tom in an effective way.” The scale had an alpha reliability of .92. This scale served as the dependent variable for Research Question 1.
Harassment Cessation Effectiveness Scale
This three-item measure tapped whether Susan’s responses would halt future inappropriate comments. It had an alpha reliability of .92. The items comprising this message are as follows: “Her response will effectively stop comments like this from Tom in the future,” “Before making troublesome statements, Tom will think about his words,” and “Tom will think twice before making comments like this to Susan the next time.” Scale scores ranged from 1.0 representing the lowest level of the likelihood the message would curtail future sexual harassment to 7.0 indicating the highest level of likely curtailment. The Harassment Cessation Effectiveness Scale was the dependent variable for Research Question 2.
Perception of Target’s Promotability
The mean perception of promotability (POP) scale had an alpha reliability of .72. The scale included items, “I think her reaction will reflect poorly on her when she applies for a promotion (reverse coded),” “With responses like this, Susan should not expect to receive promotions (reverse coded),” and “I think Susan is likely to be promoted in the organization.” The POP scale became the dependent variable for Research Question 3.
Results
The results begin with the macro-level findings of scenario development and then move to results at the message category level. The ANOVA using the scenarios as the independent variable and perceived sexual harassment as the dependent variable revealed significant differences across the four scenarios, F(3, 276) = 31.01, p = .001, R2 = .26.
Scenarios 1 and 2 were expected to be perceived as significantly more sexually harassing than Scenarios 3 and 4 because they involved greater intensity and conformed to traditional understandings of sexual harassment statutes. The pattern of the means was in the predicted direction. The LSD post hoc test did reveal a significant difference between Scenario 1 of Showing Skin with a mean of 6.41 (SD = 1.00) and Scenario 2 dangling the issue of earning a promotion (M = 5.86, SD = 1.54). These results were considered high because the values were beyond the confidence interval of neutral responses. Scenario 1 also was significantly greater than Scenario 3 expressing the desirability of a more revealing summer wardrobe (M = 5.26, SD = 1.50) and humorous Scenario 4 (M = 4.11, SD = 1.81). Scenario 3 was also considered high because it exceeded the confidence interval of the neutral zone. These results were in line with expectations. Scenario 2 also was significantly greater than Scenarios 3 and 4. Scenario 3 was expected to be judged as more sexually harassing than the Scenario 4 that used a sexual joke. As predicted, Scenario 3 was judged as significantly more sexually harassing than Scenario 4. There were no significant gender differences by scenario or a scenario by gender interaction.
A second way of looking at sexual harassment was the degree of perceived threat, intimidation, and harm represented by each scenario. The overall ANOVA indicated a significant difference on the Threat-Intimidation-Harm Scale across the scenarios, F(3, 278) = 20.07, p = .001, R2 = .20. There was no significant gender effect or scenario by gender interaction for the Threat-Intimidation-Harm Scale. The results had similarities to judgments of sexual harassment across the four scenarios, but contained some notable differences.
A notable difference was that the mean of 5.22 (SD = 1.20) for Scenario 1 did not differ significantly from Scenario 2’s mean of 5.25 (SD = 1.31). So, the perceived degree of harm was about the same for Scenarios 1 and 2. Scenario 1 did differ significantly from Scenario 3 (M = 4.52, SD = 1.21) and Scenario 4 (3.81, SD =1.42). Scenario 2 also differed significantly from Scenarios 3 and 4 as expected. Scenario 3 with a mean of 4.52 was significantly above the neutral point and was significantly higher than Scenario 4 with a mean of 3.81 that was within the neutral range. Although the patterns for these two dependent variables are mostly similar, all means dropped for each scenario for the perceived degree of threat, intimidation, and harm.
The important point to see here is that a scenario can be viewed as sexually harassing, but the degree of threat, intimidation, or harm appears to vary as a separate dimension of intensity. Even though the scenario has been judged to clearly be sexually harassing, the level of threat, intimidation, and harm can vary.
The central three research questions shift to message responses to sexual harassment and their impacts on victims using those responses. These are important issues to address the double bind in which many victims find themselves. They may not be only a victim of sexual harassment, but also victims of reactions to strong responses to them. As previously discussed, there are many well-documented cases of negative consequences for victims of sexual harassment drawing attention to those events.
To begin, we asked whether the different message response strategies would lead to differences in the perceptions of the perceived effectiveness of Susan as the responder to Tom’s sexual harassment. That is, did respondents perceive that she handled the harassment well? First, it was necessary to determine whether there were any differences of perceived effectiveness across the four scenarios that could create possible interaction effects. No a priori reasons led to predicted differences across the scenarios. There were no significant differences across the scenarios for the perceived effectiveness of the target’s (Susan’s) response to the perpetrator (Tom), F(3, 274) = 2.08, p = .10, R2 = .00. No significance in this case is a good thing because it makes interpretation more straightforward.
Differences of perceptions of Susan’s effectiveness in her response were examined to answer Research Question 1. This question considered whether the different responses to harassment would play a role in the perceived effectiveness of the harassed, Susan. More specifically, would some responses by Susan to Tom’s sexually harassing communication be perceived as enhancing or diminishing Susan’s image of being effective in her response? Response message categories did result in different perceptions of effectiveness across the five message categories, F(4, 273) = 20.49, R2 = .23, p = .001. Table 1 provides the results.
Perceived Effectiveness of Targets of Sexual Harassment by Their Use of Message Response Categories to Confront Sexual Harassment Incidents.
Note. N = 278. Scale ranged from a low of 1.0 to a high of 7.0 with 4.0 being the neutral point. Significance tests use Fisher’s least significant difference procedure.
The most notable result was for assertive responses with a mean of 5.69 (SD = 1.35) in establishing the image of Susan as communicatively effective. It was significantly greater than all other categories except the second-best response category of assertive-empathetic (M = 5.57, SD = 1.52). Assertive-empathetic did not differ significantly from aggressive messages (M = 5.05, SD = 1.71). All three of these message response categories were significantly above the neutral range. So, even aggressive messages were viewed as contributing to a positive image of Susan as being effective. Aggressive messages were significantly higher than humorous responses (M = 4.37, SD = 1.82) that fell in the neutral range. All response categories judged Susan as more communicatively effective than avoidant messages (M = 3.34, SD = 1.61).
The second question addressed the degree to which the responses would be judged as likely to stop or curtail the sexual harassment in the future. The response messages were significantly different in the perception of their ability to stop or curtail future sexual harassment, F(4, 277) = 39.93, p = .001, R2 = .36. There was no scenario-message category interaction. Table 2 breaks down the specific message categories.
Perceived Effectiveness of Targets of Sexual Harassment by Their Use of Message Response Categories to Stop or Curtail Sexual Harassment.
Note. N = 282. All conditions were significantly different above or below the neutral point. Scale ranged from a low of 1.0 to a high of 7.0 with 4.0 being the neutral point. Significance tests use the Fisher’s least significant difference procedure.
Assertive-empathetic (M = 4.95, SD = 1.27) and assertive messages (M = 4.92, SD = 1.21) led the message response categories as perceived as best in curtailing future sexual harassment. Yet aggressive messages (M = 4.70, SD = 1.29) were not far behind. This top group did not significantly differ from each other. Yet the means of this group were only modestly above the neutral zone even though they were statistically significant. The humor message strategy (M = 3.07, SD = 1.35) was significantly below the neutral point of the scale. Avoidant messages (M = 2.63, SD = 1.17) were also well below the neutral point and were judged as the worst message strategy for curtailing future sexual harassment.
The final research question considered which message categories created the perception that Susan would be promotable after using them. Some message strategies could be seen by some as “poisoning the well.” Such messages create a desired immediate impact, but create long-term negative consequences. One consequence of an aggressive response to sexual harassment is that the target may gain a reputation that limits future promotability. The POP scale was used to assess the perceived likelihood of future promotions. There were no perceived significant differences of the future promotability of Susan across the four scenarios, F(3, 262) = 2.23, p = .09. However, there were significant differences of perceived promotability across different message categories, F(4, 275) = 5.45, p = .001. R2 = .07. So, using some message categories was viewed as having more potential for promotion than others. The specific categories of messages responses appear in Table 3.
Perceived Promotability of Targets of Sexual Harassment by Response Message Category Used to Confront Perpetrator.
Note. N = 280. Scale ranged from a low of 1.0 to a high of 7.0 with 4.0 being the neutral point. Significance tests use the Fisher’s least significant difference procedure.
Assertiveness (M = 5.11, SD = 0.91) was perceived as best response category to preserve promotability, but it was not significantly different from assertive-empathetic (M = 4.89, SD = 0.95). Aggressive responses (M = 4.61, SD = 1.20) created some perceived degree of acceptability. Even the humor (M = 4.51, SD = 1.14) message response strategy was significantly above the neutral point of 4.0 of the scale in supporting the promotability of Susan. Only the avoidant strategy (M = 4.24, SD = 1.14) was in the neutral range.
Discussion
The results of this study help us begin to map the terrain of responses to outcomes of sexual harassment. They also provide potentially useful directions for many applied communication practitioners who are involved in sexual harassment training. For instance, the results highlight the productive possibilities for message categories of assertiveness and assertive-empathetic as productive responses to sexually harassing behavior. The effectiveness of assertive-empathetic strategies was a surprise in this study, but it is consistent with politeness theory. This is an interesting result because assertive-empathetic is often seen as being less direct. To the contrary, assertive-empathetic messages can address the issue at hand, but also demonstrate regard for the relationship.
Perhaps assertive-empathetic messages tap into the more diplomatic side of dealing with problematic behaviors in the workplace, while also demonstrating respect for face saving and relational maintenance. It is also consistent with the findings of Herovic et al. (2019) who found that the targets of harassment in their study of young adults often have a preexisting relationship with the harasser that they might not want to disrupt, and further, that confronting the harassment might disrupt the relationships with others in the organization.
Alternatively, this could be a gendered expectation where women (the targets in our scenario) are expected to respond in demure rather than more direct ways. Thus, this type of response should be explored in more detail in future research in relation to gender roles and social expectations. Furthermore, we need to better understand the conditions where assertive-empathetic responses would be preferred over pure assertiveness, particularly when we have now entered a period when women are empowered to speak up and speak out about harassment without express concern for the harasser. It is possible that assertiveness may be more likely to stop harassment, but less positive for the career and perceptions of the target. That issue deserves more attention in future research given that when confronted with harassment, targets do consider multiple factors such as stopping the harassment, as well as how their behavior might affect their own career trajectories.
The answers may be further complicated by organizational-specific factors, such as the differential impact of varying organizational cultures. For example, Clason’s (2019) recent research indicates the acceptability of “joking” in traditionally male-dominated industries, and Reiter (1991) noted that sexual harassment disguised as flirting is prevalent in the food industry which is often staffed with young, temporary workers. Thus, the organizational culture may influence the ratings of responses.
Most clear in this study was that avoidant behaviors had few positive outcomes. Staying silent is no longer seen as acceptable or beneficial. Inappropriate comments, no matter how harassing or severe, need to be addressed. In many ways this is a double-edged sword—or putting the onus on the target to maintain her or his sense of credibility and stop the harassment. Doing nothing is not seen as effective. Thus, more attention needs to be given to specific message behaviors that are considered avoidant. Not clearly addressed in this study are the message behaviors in response to a sexually harassing situation that make a person look weak or ineffective. Could some forms of humor be viewed as an avoidance strategy versus an assertive-empathetic strategy? Can humor be considered aggressive? Future research could look at it as a component or delivery method of multiple strategies, and further explore if humor dilutes or strengthens messages.
The scenarios in this study provide an approach that may lead to more productive discussion in sexual harassment awareness training than is currently being done in many corporate and organizational contexts. Many of the heavily-content laden training programs to curtail sexual harassment or sexual abuse have not been effective (Daniel et al., 2019; Dobbin, & Kalev, 2017 ; Kuchulis, 2019). Some issues may be occurring because many organizations have mandatory Title IX training, but much of it is now being delivered through the internet because face-to-face training costs are too high. Moreover, those programs rely on institutional reporting rather than providing targets immediate, empowering tactical responses that can curtail harassment before it escalates. Perhaps micro-level response training would empower bystanders, as well—though that needs separate research. Even if someone reports harassment, others may see lack of immediate bystander confirmation as problematic silence; and thus, further victimize the target.
Although the message categories need much more development than is provided in this initial study, the messages provided in this research are a beginning point for providing insight into training in resistance to sexual harassment. Training programs seem to focus on the perpetrators of blatant sexual harassment with little attention to the message strategies that are best used in resistance to sexual harassment, which can be subtle or even framed as humor. Future research and development should examine proportionate message responses. Moreover, the scenarios in this study did not consider the issue of public disclosure of sexually harassing behavior. These scenarios in this study were between two people. Threats of exposure to the public or important third parties (bystanders) are another avenue to be explored when sexual harassment becomes a chronic or serial problem.
This study has much value, yet there are limitations that warrant attention. It utilized photographs of upwardly mobile White people—one male and one female. Many other variations need to be tried in our more diverse environments such as reversing the gender roles, looking at same-sex sexual harassment, and various ethnic pairings. Other covariates certainly exist such as the position in the hierarchy of an organization that a person holds. Getting access to samples of upper-level managers to participate in such research would be wonderful, as would be the ability to explore those who identify as harassers or targets.
A surprising result was the failure to find significant differences between males and females in perceptions of the scenarios and the message categories used. Maybe this was a result of the participant population accessed. The recent study by Clason (2019) suggests that humor may be viewed very differently by males in industrial settings where those contexts are clearly male dominated. Yet, in this study, males and females reported perceptions of sexual harassment significantly above the neutral point occurring in at least three of the four scenarios. Other methods may be needed to understand gender differences in the perceptions of sexual harassment occurring in the workplace.
This study attempts to move beyond further victimization of those who are sexually harassed to think about that which can be done with messages to effectively resist such situations. Baugh (1997) found that victims often receive blame from other people for the incident. The view that the victim overreacted to the harasser’s comment, or that a woman was asking for the harassing treatment by the clothes that the person was wearing is often reported. That standard sets the belief that victims are to accept these experiences, and if they cannot, they are viewed as less competent workers and should not be in the workforce (Baugh, 1997).
Summary
There are some clear takeaways from this study for those doing sexual harassment training. First, the scenario method has potential as a training tool to distinguish different types of sexual harassment that occur to further highlight that not all sexual harassment is coercive, blatantly aggressive, or overtly threatening. Some harassment can be disguised as good-natured humor reinforcing gender discrimination. Second, several new scales were developed for this study regarding elements of sexual harassment that had good to excellent reliability. These brief scales could be used in training sessions and future research. Third, training regarding the issue of sexual harassment needs to focus on messages that may effectively resist sexual harassment without derailing a professional career path. Although women are unlikely to formally report sexual harassment, we can empower them with strategies to respond in the most productive ways possible. Finally, the power of assertive messages and the lack of power of avoidance messages in response to sexual harassment need to be highlighted. On the level of message categories, the superiority of assertive messages in response to a sexually harassing male was found in terms of enhancing the perception of the responding female employee while maintaining her ability to be promoted. On the other hand, avoidance tactics failed to provide positive outcomes. So, staying silent communicates.
Conclusion
Long-time feminist advocate and legal scholar Catherine MacKinnon (1979) sees sexual harassment from a longer arc of time that began with her landmark book Sexual Harassment of Working Women. Recently, MacKinnon (2018) remarked in the New York Times, “#MeToo has done what the law could not.” Further she said, “Powerful individuals and entities are taking sexual abuse seriously for once and acting against it as never before.” Maybe now the time has come to advance the study of effective communication practices in combatting sexual harassment in the moment that it occurs. We recognize that there are cultural and resource limitations that make responding challenging, but as we begin to take sexual harassment more seriously as a culture, we might also want to explore how to reduce escalation and empower women and allies to change the organizational culture, itself. This study is one beginning point for those wishing to take a more message-centered approach in the realm of sexual harassment training that teaches effective and ineffective responses. We believe it provides a viable model for a more empowered approach for women and others facing the challenges of sexual harassment. This also provides organizations an opportunity to train and empower targets to address harassment as it occurs, which can also contribute to a larger cultural shift.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
