Intimidation and harassment have spiked throughout the United States since the recent presidential election. Women, people of color, immigrants, Muslims, Jews, and LGBTQ people—including many of our own students—report palpable fear. In the 10 days after Election Day, the Southern Poverty Law Center collected 867 reports of hateful intimidation and harassment (Southern Poverty Law Center 2016). On November 16, 2016, a man in Sarasota, Florida, reported being physically attacked by a person who said, “You know my new president says we can kill all you f-ggots now” (Masek Reference Masek2016). On November 17, 2016, a Puerto Rican family’s car was vandalized in West Springfield, Massachusetts, with the words “Trump” and “Go home” scratched into it (Yan, Sgueglia, and Walker Reference Yan, Sgueglia and Walker2016). In late November 2016, more than 10 mosques received letters saying that Trump will “do to you Muslims what Hitler did to the Jews” (Guerra Reference Guerra2016). In February and March 2017, Jewish cemeteries in St. Louis, Philadelphia, and Rochester, New York, were desecrated (Berlinger and Frehse Reference Berlinger and Frehse2017; Chokshi Reference Chokshi2017).

“I have experienced discrimination in my life, but never in such a public and unashamed manner,” reported one Asian American woman (Southern Poverty Law Center 2016, 4). This woman’s observation about the “public” and “unashamed” nature of hateful acts highlights the social processes that both drive and prevent hate speech and hate crime. To stop hate crimes, we must understand why perpetrators do what they do: the personal and social sources of their motivation. Research shows that potential perpetrators of hate crimes and bullying are actually quite conscious of the degree to which their community supports or condemns their actions.

For example, psychology experiments suggest that a person who hears racist or sexist jokes tolerates subsequent gender or racial discrimination to a greater degree. Ford and Ferguson (2004, 90–91) summarized multiple studies and found that racist or sexist humor “implies a change in the...norms in a given context that dictate appropriate actions.... [D]isparagement humor communicates a message of tacit approval or tolerance of discrimination.” It is interesting that what socially shared disparagement does not do is increase racist or sexist stereotyping: “[i]t does not appear that exposure to disparagement humor reinforces negative images of the targeted group.” Rather—and as Ford, Boxer, Armstrong, and Edel (2008, 168) also found—prejudiced humor creates a “local, prejudiced norm…. Sexist participants took advantage of the local prejudiced norm to release their prejudice against women without fear of disapproval from others.”

Even at more extreme levels, social scientists have identified a similar dynamic. Yanagizawa-Drott (Reference Yanagizawa-Drott2014) found that inflammatory messages played on a Rwandan hate radio station in 1994, which were aimed at motivating Hutus to murder their Tutsi neighbors, motivated more collective participation in the genocide when a larger proportion of people in a locality could receive the radio signal. Yanagizawa-Drott suggested that one of a few explanations for this effect is that widespread exposure to the hate radio resulted in greater awareness of community support of violence. In other words, it is possible that hate radio did not simply increase individual hatred but also made perpetrators believe that community members would support them.

These studies, set in disparate contexts, imply that potential perpetrators of hate crimes in the present-day United States are not necessarily “learning” hatred from Trump’s dehumanizing statements, including those about Mexican Americans (“They’re bringing drugs, they’re bringing crime, they’re rapists”), Muslim Americans (“There were people that were cheering on the other side of New Jersey where you have large Arab populations.... They were cheering as the World Trade Center came down”), disabled people (physically mocking disabled reporter Serge Kovaleski), and women (“Grab them by the p—y, you can do anything”) (Burns Reference Burns2015; Fahrenthold Reference Fahrenthold2016; Haberman Reference Haberman2015; Kessler Reference Kessler2015). Rather, potential perpetrators are encouraged to act by the fact that Trump garnered votes and now holds the highest office. They infer from this that they have a better chance of escaping social and legal sanction than before his election. According to this model of hate-crime motivations, prevention efforts must focus on convincing potential perpetrators that those in their community are opposed to this behavior.

Laboratory experiments on social confrontations of prejudice support the idea that hateful actions are socially mediated. Moreover, Czopp, Monteith, and Mark (Reference Czopp, Margo Monteith and Mark2006) found that being confronted by others does seem to change a person’s behavior. In their experiments, when white American students were confronted about answering questions in a racially stereotypical way, they responded with anger and irritation toward the person confronting them and negative affect (such as anger, disappointment, and guilt) toward themselves. However, following the confrontation, they were less likely to engage in stereotypes and report prejudiced attitudes. Confrontations that did not explicitly label the person’s behavior as racist (by saying that they should be “more fair” as opposed to “less prejudiced”) provoked less hostility in the study participants. However, both types of confrontation were effective for reducing subsequent stereotyping and prejudiced attitudinal reports. Czopp, Monteith, and Mark (2006, 799) wrote that “potential confronters may be willing to endure unpleasant interpersonal reactions if the confrontation will be ultimately successful in changing future behavior.”

Recently Munger (Reference Munger2016) tested different strategies for confronting white males who use racial slurs (specifically the n-word) against others on Twitter. He created Twitter “bots” with the identity of either white or black males, according to their cartoon avatar. The bots also varied in terms of the number of Twitter followers they had (a high number indicated popularity and high status). When Munger found a Twitter user who had used the n-word racial slur in a tweet, within 24 hours a bot would tweet back, “Hey man, just remember that there are real people who are hurt when you harass them with that kind of language.” Munger found that this confrontation, on average, reduced the Twitter user’s use of the n-word in the following weeks. However, this was mostly when the Twitter bot who confronted him was white and had numerous followers; black males and white males with few followers, on average, were not as successful.

In a related study, Paluck, Shepherd, and Aronow (Reference Paluck, Shepherd and Aronow2016) introduced an experimental peer-to-peer intervention to stop conflict and bullying in 28 public middle schools in New Jersey. On average, 26 students at each school were “seeded” by being encouraged and trained to publicly oppose conflict at their school. This intervention reduced disciplinary reports of peer conflict by roughly 30%. However, when the seeded students included “social referents”—students who were in the top 10% in terms of popularity—the intervention was even more effective. When 20% of a seed group was composed of social referents, disciplinary reports of peer conflict declined by up to 60%.

These studies suggest that some people are better than others at delegitimizing hatred and violence. These “elite influencers” are more likely to come from a community considered important by a potential perpetrator—whether their own racial community or their friendship group. Also, these influencers are more likely to have higher status—connected to many people within those networks.

Given that people’s actions depend on their awareness of the approval or disapproval of others around them, how should a campaign against hate crimes and harassment best be communicated? McAdams (Reference McAdams2015) argued that people’s choice of whether to obey a law depends on whether they think that other people also are obeying. In other words, laws affect human behavior not only through punishment but also by publicly communicating a social norm or standard of behavior. Shteynberg, Bramlett, Fles, and Cameron (Reference Shteynberg, Bramlett, Fles and Cameron2016) found experimental evidence for the importance of synchronous co-attention (listening to a message simultaneously with others). They asked people to watch a video of a speech given by US House Representative Rosa DeLauro arguing for the reauthorization of the Violence Against Women Act. When people were told that they were watching live with about 100 other people, subjects were more persuaded by the speech than when they were told that it was prerecorded and watched earlier by 100 other people. Importantly, believing that 90 to 100 people were watching simultaneously did not make a significant additional difference compared to believing that two or three people were watching simultaneously. Listening simultaneously with even only a few people seemed to be more effective because people processed the message in a deeper and more serious manner.

Other work underlines the importance of synchronized community attention but finds that the scale of attention matters. From this perspective, large numbers of people who assert values of inclusiveness and tolerance in a significant public way can change minds and behavior. Major media events or assemblies that create “common knowledge” of these values, which show each member of the community that other members share these values, would be most successful (Chwe Reference Chwe2013). Using this logic, a public service announcement during the Super Bowl that encourages people to report domestic violence would deter domestic violence better than a magazine advertisement. A potential perpetrator would infer that the millions watching the Super Bowl found domestic violence unacceptable and would be more likely to report offenders.

Hate-crime perpetrators often are thought to be isolated people with extreme personality traits. For example, it is commonly believed that school shooters tend to be loners; however, this is “a widespread but wrong or unverified impression” fostered by news coverage (O’Toole Reference O’Toole1999). This discourse discounts the social sources of the motivation of hate-crime perpetrators, particularly the effects that community or peer pressure can bring to bear.

If communities do not stand collectively against hate speech, potential perpetrators will feel increasingly emboldened. Schools, universities, and localities cannot simply “play defense” and wait for their members to be victimized. Reporting events after they occur is not enough. Potential perpetrators must understand clearly that everyone around them, regardless of their political views, believes that hate is unacceptable. Elite influencers in every community can broadcast this message. Standing with them, there is strength in numbers; as individuals and communities, we must come together to speak as loudly and publicly as possible. ^{Footnote 1}