Abstract

harassment. Suppose each male employee comes to prefer to follow this rule, on condition that he expects his colleagues both to follow it (empirical expectation) and to believe he in turn should follow it (normative expectation). The rule then functions as a social norm; it is upheld as long as the male employees’ (empirical and normative) expectations are upheld. Arguably, one of the functions of #MeToo was to make salient such a non-harassment norm.

However, when norm violations – some male employees “slip back” into old macho habits – are observed by other male employees, their empirical expectations may falter. Moreover, when such violations are not met with protest, their normative expectations may falter. The social norm risks breaking down, bringing back the toxic work-environment. Yet protesting is costly for victims and bystanders, who are hence caught in a collective dilemma of their own. This dilemma can be solved, analogously, by a speaking-up norm. Arguably, another function of #MeToo was to make salient this norm, too.

This game-theoretic analysis of the dynamics of #MeToo can be further enhanced by considering Bicchieri’s recent findings on norm-nudging: influencing expectations via (general) social information. By communicating widespread condemnation of sexual harassment, campaigns like #MeToo raise normative expectations of non-harassment. However, by simultaneously communicating frequent failures of norm-following, empirical expectations may be lowered. Yet when there is incongruence between normative and empirical information, studies show that the latter trumps the former. In other words, social information transmitted under hashtags like #YesAllWomen carries a risk of undermining the non-harassment norm. This analysis throws new light on criticised hashtags like #NotAllMen: by emphasising norm-following, they may counteract the incongruence, thereby contributing to the reinforcement of the norm.

Further, regarding the speaking-up norm, a norm-nudging perspective reveals an interesting contrast: by showcasing widespread condemnation of sexual harassment, #MeToo raises empirical expectations of speaking up. However, the widespread backlash that followed (defamation suits, public victim shaming) communicated condemnation of speaking-up, lowering normative expectations. Given this incongruence, one should again expect empirical information to trump normative information. However, studies suggest that there may be another mechanism in play: when social information conflicts with people's self-interest, they may exploit any “wiggle-room” in the information, interpreting the social norm as weak, to avoid conforming to it. This analysis suggests that the backlash against #MeToo should not (only) be understood as raising the costs of speaking up, but (also) as undermining the speaking-up norm itself.

The paper concludes by discussing some upshots of these analyses for addressing the persistent problem of sexual harassment.