Evidence of unintended consequences

Because no social action takes place in a vacuum, there will always be repercussions that were unanticipated. These unanticipated outcomes can be desirable or undesirable, but here we concern ourselves only with the latter. Specifically, a number of studies have documented how sexual harassment prevention training (SHPT) can backfire, which refers to when an intervention affects the intended outcome in an undesirable direction (Leslie, Reference Lesliein press). For example, in Bingham and Scherer’s (Reference Bingham and Scherer2001) study of 530 university employees (faculty and staff), half of the sample was randomly assigned to attend a 30-minute SHPT, then they responded to a survey one week later. As expected, employees who attended the training were significantly more knowledgeable compared to those who did not attend. However, counter to expectations, male trainees were more likely than male nontrainees and female trainees to blame the victim in hypothetical scenarios portraying sexual behavior at work, less likely to view sexual coercion of a subordinate as harassment, and less likely to report sexual harassment. In a survey of employees, Dobbin and Kalev (Reference Dobbin and Kalev2017) found similar results. Although most people who participated in SHPT could better define and identify sexual harassment, a subset of their sample—men who scored more highly on a measure of likelihood to harass women—became more dismissive in their attitudes about harassment, believing that the severity of the problem was exaggerated.

In a third study, Tinkler, Li, and Mollborn (Reference Tinkler, Li and Mollborn2007) sampled 66 undergraduate male students to examine the effects of reading a university’s sexual harassment policy on one’s explicit and implicit beliefs about status differences across genders. Reading the policy had no effect on one’s explicit beliefs, relative to a control group that did not read the policy, but it did change men’s implicit beliefs in the opposite of the intended direction. That is, in an implicit associations test, participants who read the sexual harassment policy more closely linked “male” with “high status” and “female” with “low status” than did participants in the control group. In a related study by Tinkler (Reference Tinkler2013), she found that after undergraduate students watched a video on preventing sexual harassment, participants—men and women—who strongly adhered to gender norms only strengthened their implicit attitudes about gender stereotypes following the intervention.

Explaining unintended consequences

One problem with SHPT programs is that they can inadvertently threaten important identities (Petriglieri, Reference Petriglieri2011). Evidence for this was found in a qualitative study involving observation of multiple SHPT sessions at a university and interviews with students and the training facilitators (Tinkler, Reference Tinkler2012). The author noted that SHPT often polarizes men and women by making gender the most salient component of people’s identity. As a result, when dissecting case studies and role play activities, conversations between participants tended to rely upon gender stereotypes to interpret ambiguous situations (Ridgeway & Smith-Lovin, Reference Ridgeway and Smith-Lovin1999), such as women as passive and men as aggressive.

In addition to observing that one’s gender identity became highly salient, Tinkler (Reference Tinkler2012) found that both men and women reported feeling threatened in two senses. First, participants expressed concern that “normal” interactions between men and women were becoming prohibited. For example, some men perceived that behaviors that they had considered “friendly” were being discussed in the SHPT as potentially unwelcome attention. Thus, training exercises intended to increase awareness of the breadth and pervasiveness of sexual harassment can have the unintended consequence of increasing men’s sense of threat and vulnerability. Second, men and women reported feeling worried that their relative status might be at stake. Women, for instance, were reluctant to report incidences of sexual harassment as such because they feared being labeled a victim or as oversensitive.

A recent quantitative study corroborated the findings of Tinkler (Reference Tinkler2012) using a sample of 154 university employees (Rawski, Reference Rawski2016). Rawski (Reference Rawski2016) found that trainees who felt more threatened before and during the SHPT were more likely to have negative attitudes about the training, to avoid “oversensitive” employees, to engage in sex-based interactions (e.g., flirt with a coworker, tell a sexual joke). Trainees who felt more threatened also had decreased intentions to share what they learned about sexual harassment policies with others and were less motivated to attend a future SHPT. In short, SHPT may have polarizing effects: It may be least effective for the people who could benefit from it the most, while those that are most receptive to it may be the least likely to need it in the first place (Rawski & Workman-Stark, Reference Rawski and Workman-Stark2018).

Preventing unintended consequences

So, what can be done to prevent these unintended consequences from occurring? If we begin from the premise that these unintended consequences emerge from a threat to one’s identity and the consequent defensiveness, a set of promising solutions can be found in work on self-affirmation (Sherman & Cohen, Reference Sherman, Cohen and Zanna2006; Steele, Reference Steele and Berkowitz1988). The theory of self-affirmation suggests that the overarching goal of the self is to protect its integrity or coherence (see also Dweck, Reference Dweck2017). When this sense of integrity is threatened, people may protect the self through defensiveness (e.g., denial, distancing), which diminishes the threat. However, an alternative approach to managing threats is affirming parts of the self that do not feel threatened.

Bystander interventions of workplace harassment (Banyard, Plante, & Moynihan, Reference Banyard, Plante and Moynihan2004; Bowes-Sperry & O’Leary-Kelly, Reference Bowes-Sperry and O’Leary-Kelly2005), for instance, do just that. Rawski (Reference Rawski2016) pointed out that in a two-person interaction, “if one party ‘takes’ the role of ‘victim,’ she simultaneously ‘makes’ the role of ‘harasser’ for her interactional partner” (p. 53). However, bystander interventions offer a third option for an identity to assume. Bystander interventions call upon observers of sexual harassment to intervene and report incidences of sexual harassment. Early indications show that bystander intervention training is a promising approach (e.g., Cares et al., Reference Cares, Banyard, Moynihan, Williams, Potter and Stapleton2015; Potter & Moynihan, Reference Potter and Moynihan2011). Seen through the lens of self-affirmation theory, bystander interventions allow people to cope with the threats to their identity brought about by, for example, considering how one might have made a coworker uncomfortable with a sexual joke he told. Being given an opportunity to enact one’s values in a socially valued and constructive way (i.e., as a diligent bystander) provides self-affirmation and, as a result, allows him to address the threatening stimuli (i.e., “Did I make my coworker uncomfortable?”) with less defensiveness. Further, it frames men as not simply potential harassers but as allies to women (and other men) and guardians of honorable behavior.

A second example of a self-affirmation intervention is asking people to write a short essay about their core values. In these essays, participants write about when their highest-ranked value was especially important to them (e.g., Cohen, Aronson, & Steele, Reference Cohen, Aronson and Steele2000, Study 1) or how living out that value made them feel good about themselves (e.g., Sherman, Nelson, & Steele, Reference Sherman, Nelson and Steele2000, Study 2). Such exercises then allow the writers to challenge their own questionable attitudes or behaviors with less defensiveness. In the context of SHPT, this intervention might take the form of having trainees take a few minutes to write about their highest values before the training even begins. This enables trainees to come into the training more open to new points of view (Cohen et al., Reference Cohen, Sherman, Bastardi, Hsu, McGoey and Ross2007), and their focus is not just on defending oneself on the issue at hand (Cohen & Sherman, Reference Cohen and Sherman2014; Sherman & Hartson, Reference Sherman, Hartson, Alicke and Sedikides2011). Critically, these interventions should occur before a threatening stimulus is presented or shortly after. Once a person begins to engage in defensive rationalizations, self-affirmation interventions tend to be ineffective (Critcher, Dunning, & Armor, Reference Critcher, Dunning and Armor2010).