New Evidence on Elections and Norm Change

Social norms – or informal rules for behavior that is acceptable in a group or society – can be powerful bulwarks against Dangerous Speech, but they can change quite suddenly. At the Dangerous Speech Project, we have noticed this in many countries – informally. Now there is new evidence from an academic study that an election can provoke sudden shifts in norms, especially when a candidate expresses extreme views and receives widespread support among nearby voters. Donald Trump’s election, a team of economists found, made it more socially acceptable to express xenophobia, in strongly pro-Trump areas in the United States.

In a working paper titled From Extreme to Mainstream: How Social Norms Unravel Leonardo Bursztyn, Georgy Egorov, and Stefano Fiorin “argue that aggregators of private opinions in a society, such as elections, might lead to updates in individuals’ perceptions of what people around them think, and thus induce fast changes in the social acceptability of holding certain opinions and in the likelihood that these opinions are publicly expressed.”

In their first experiment, the authors surveyed residents of states where Trump’s victory seemed all but certain, in the two weeks before the election. They offered participants money in exchange for authorizing the researchers to make a donation to an anti-immigration advocacy group – either in the participant’s name or anonymously. Those who were told they lived in a state with a 100% chance of voting for Trump were 12% more likely to agree to attach their names to a donation – a public way of expressing their views – than the control group. The experiment was repeated after the election, and the authors found another 14% increase in willingness to donate publicly after Trump’s victory. In both cases, the same proportion of people were willing to donate anonymously in the treatment and control groups. This suggests, importantly, that the experiment did not influence privately held beliefs, only one’s willingness to express those beliefs publicly.

In a second experiment, the authors found that participants with tolerant political views were willing to share a greater portion of money with an intolerant Swiss man if they were also told that a majority of Swiss people shared the man’s intolerant views. The authors interpret these results to mean that people with a tolerant position are more likely to forgive an intolerant view when they perceive that this would mean conforming with the majority’s view. They are less forgiving when they don’t see such a strategic reason (i.e. most people are not intolerant, or the intolerant view was expressed anonymously.)

This working paper has its limitations – for example it tests interactions with a stranger online, not in-person dialogue with a known person, but it offers valuable evidence in support of a theory that is central to the idea of dangerous speech – that the public expression of extreme views by influential people can quickly change social norms, and that ideas once confined to the fringes of society can quickly come to seem acceptable or even strategic.