The Washington PostDemocracy Dies in Darkness

We showed Trump voters photos of black and white Americans. Here’s how it affected their views.

Donald Trump speaks at a campaign rally in Charlotte. (Erik S. Lesser/European Pressphoto Agency)

Much has been written about Donald Trump’s improbable presidential candidacy. Many analysts believe that a key factor is prejudice among his overwhelmingly white supporters. Trump’s candidacy has few modern parallels in terms of the bluntness of its appeals to racial resentment, xenophobia and religious prejudice, as suggested by his willingness to question Barack Obama’s citizenship and propose religious tests for immigration to the United States. There is no shortage of evidence that Trump voters exhibit relatively high levels of prejudice toward African Americans, Muslims and others.

The science of Trump: Here’s how political science explains the rise of an unlikely candidate

But most of the evidence is observational, demonstrating a correlation between Trump support and animosity to minority groups. Those analyses don’t show that racial prejudice actually causes people to evaluate policies or make other political judgments differently.

In a new experiment, however, we provide some causal evidence. We show that white Trump supporters were more opposed to a mortgage assistance policy when they were experimentally induced to think of black rather than white Americans. Combined with the existing observational findings, this is strong evidence that racial animosity is indeed a key factor motivating Trump voters.

Our experiment: How we ‘cued’ race

We embedded our experiment in a recent Internet survey conducted by Survey Sampling International (SSI) for the Center for the Study of Political Psychology at the University of Minnesota.

Trump supporters already doubted women’s claims of sexual harassment, before they heard the infamous recording. Why would they change now?

Respondents were randomly assigned to see an image of either a black or a white man standing next to a foreclosure sign on the screen while they gave their opinions about a mortgage relief program. The picture was identical except for the race of the person in the photo.

Note that the racial cue in our experiment was subtle. We didn’t call attention to the photo, and respondents were never encouraged to think about it in the rest of the survey.

After seeing the image and reading text about recent “proposals to help people who are struggling with their mortgages and may lose their homes,” respondents answered several questions. We asked about their support for the mortgage relief program, whether the availability of mortgage relief would make them angry, and the extent to which they blamed potential beneficiaries of the program for their own plight.

Will Donald Trump cost Republicans the House? It’s very unlikely.

Though all respondents of various racial and ethnic groups were included in the survey, we focus only on the 746 white respondents in the sample. This is because there were almost no non-white Trump supporters.

How we know race shapes Trump supporters’ political attitudes

In the graph below, we present the percentage of respondents who said that they opposed the mortgage relief program, split by Trump support and which experimental condition they were assigned to.

On the left side, among respondents who said they did not support Trump and were assigned to white cue condition, 28 percent opposed the program. By contrast, 22 percent opposed it in the black cue condition.

But among Trump supporters, the relationship was reversed. Fifty-nine percent of respondents who saw the picture of the black man said they opposed the program, compared with just 51 percent among people assigned to the white condition. In other words, respondents were more hostile toward the program when they had seen the photo of a black man.

We found similar results when we looked at the effects of the racial cue on anger toward the mortgage policy. In the graph below, Trump supporters in the black condition were more likely to say they were more than “somewhat angry” about the program than fellow Trump voters in the white condition. At the same time, among those opposed to Trump, exposure to the black cue significantly decreased anger.

Finally, Trump supporters in the black condition (see the graph below) were more likely than their counterparts in the white condition to say that beneficiaries of the mortgage program were more than “somewhat to blame” for their financial situation.

On the other hand, whites who did not support Trump were significantly less likely to blame recipients after receiving the black cue.

It’s worth noting that all these findings hold up in more sophisticated analyses. When we control for age, income, sex, education, party identification, ideology, whether the respondent was unemployed, and perceptions of the national economy — other factors that might shape attitudes about mortgage relief — our results were the same.

In the end, our experiment shows that white Trump supporters are more likely to oppose government assistance when they are subtly led to think of African Americans rather than whites. Moreover, the prospect of such assistance makes them angrier and more likely to “blame the victim” when they have African Americans in mind. It appears clear that Trump’s supporters are driven, at least in part, by a distinctive pattern of racial animosity.

Matthew D. Luttig is a postdoctoral scholar in the department of political science at the University of Chicago. 

Howard Lavine is the Arleen Carlson Professor of Political Science and Psychology at the University of Minnesota and the author (with Marco Steenbergen and Christopher Johnston) of “The Ambivalent Partisan.”