Anatomy of a Trump voter: How racism propelled Trump to the Republican nomination

Data shows racism is a powerful predictor of Trump support — far more than economic concern

Published July 23, 2016 10:30AM (EDT)

The Republican National Convention in Cleveland, July 19, 2016.   (Reuters/Brian Snyder)
The Republican National Convention in Cleveland, July 19, 2016. (Reuters/Brian Snyder)

This week, Donald Trump formally accepted the Republican nomination, and many narratives have emerged about what led to his victory and what his supporters demand. But current polling makes that difficult: by now, Republicans are increasingly settling behind the Trump as the nominee. What actually determined support or opposition to Trump early in the primary when it mattered?

To explore what sets Trump supporters apart from the rest of the GOP, I used the ANES 2016 pilot study (discussed here at more length). The survey took place in late January, when the GOP primary was still contested, so it’s an ideal survey to test the determinants of Trump support. I explore only Republican respondents, to see what parts of the GOP base that Trump appealed to. At the time of the survey, 56 percent of those who identified as Republican supported a candidate that was not Trump, and 41 percent supported Trump (he had support from a plurality of Republicans). A negligible percentage of Republicans chose no candidate. I find that racism is a powerful predictor of Trump support, while other suggested motivating factors simply don’t differentiate Trump supporters from other Republicans.

Trump Supporters Are Whiter

To begin, I explored some basic demographic differences between the Republicans that supported Trump and those who rejected him. (My analysis here only explores the attitudes of Republicans, for a more general analysis of Trump support, see here.) In the ANES dataset, 95 percent of Republicans who supported Trump were white, compared with 87 percent of non-Trump supporters. In all, 90 percent of Republicans were white, compared with 61 percent of Democrats. Trump supporters leaned somewhat older: 23 percent of Trump supporters were older than 70, compared to 12 percent of supporters for other Republicans (age was not a strong predictor of Trump support). In addition, 14 percent of Trump’s supporters lacked a high school degree (compared with 8 percent of non-Trump Republicans). Sixty-six percent of Trump supporters had not completed a 2 or 4 year degree, compared with half of non-Trump Republicans. There weren’t any major differences across income, as other research has found.

Trump Supporters Buy Into the Politics of Resentment

In previous work with Jason McDaniel and Philip Cohen, I showed that Trump supporters were more likely to endorse racial resentment and stereotypes. For this analysis, McDaniel and I examined the stereotype variables slightly differently, examining stereotyping both as a scale, and stereotypes towards particular groups independently. Both stereotypes and resentment were powerful predictors of support for Trump in the primary, among white Republicans and Independents. At the lowest level of resentment and stereotype, probability of support for Trump was less than 10 percent. At the highest level of resentment, the probability of support for Trump was 50 percent. At the highest level of anti-Black stereotyping, the probability of Trump support is nearly 80 percent. I examined Muslim and Hispanic stereotyping and found similar results: At the lowest level of stereotyping, the probability of support for Trump was below 10 percent, at the highest level, the probability of support was more than 70 percent.


To examine the data in another way, I compared Republicans (of all races) who supported Trump to those who did not on the stereotyping questions. As a baseline, I also included stereotyping among all respondents (of any party). As the results show, Trump supporters were more likely to endorse stereotypes for all groups than non-Trump supporters. The stereotyping measure was generated by taking respondents views of how well the words “violent” and “lazy” described black, Hispanic and Muslim people, and subtracted them from whites. For instance, 76 percent of Republicans who support Trump say Muslims are more violent than whites, compared with 60 percent of non-Trump Republicans and 44 percent of all respondents. (Forty-one percent of non-Trump Republicans say Obama is a Muslim, compared to 70 percent of Trump supporters).


In an early examination of ANES, political scientists Michael Tesler and John Sides showed that whites who believed that whites faced discrimination or worried about losing jobs to people of color were more supportive of Trump. I explored two variables that stood out to me as interesting. The first asked respondents how important it was for white people to work together to change laws unfair to whites. The second asked how likely it is that whites are losing jobs to people of color. Responses to both of these questions indicate the beleaguered racial attitudes that motivate white nationalist movements. I examined these variables using a model that controlled for ideology, family income, education, age and gender (these two questions were only asked to whites). Again, examining only Republicans, endorsing these views increased the probability that a respondent would support Trump.


Previous work by myself and McDaniel, as well as others, has shown that “economic anxiety” simply isn’t sufficient to explain the rise of Trump. Young people are among the most economically insecure group, yet few are rushing to Trump. People of color are far more likely to be working class, and yet Trump’s support is almost exclusively white. It’s important to talk about economic and social inequality, but what motivates Trumpism is racism.  Indeed, there was virtually no difference between Republicans who supported and rejected Trump in views of government services and spending. In the models I examined, free trade didn’t increase the probability Trump support, nor did perceptions of the economy. Trump supporters were no more likely to pick Social Security as one of their top four issues. Support for the minimum wage did increase Trump support, but support for campaign finance restrictions did not. None of the questions related to opportunity or upward mobility increased the probability of Trump support. In a previous exploration, McDaniel didn’t find strong results examining the terrorism questions, after controlling for other factors.

Trump Supporters See More Discrimination Against Whites

Immigration has also been noted as a factor driving support for Trump, and ANES data provide support for this. While 37 percent of Republicans who do not support Trump say that immigration should be decreased, 60 of Republicans who support Trump say the same. Trump supporters are also more likely to prioritize immigration. (16 percent rated it as their most important issue, compared to 4 percent of non-Trump Republicans.) While 69 percent of Trump supporters said they “oppose a great deal” accepting Syrian refugees, 49 percent of non-Trump Republicans did. Trump supporters are more likely than other Republicans to say that whites face more discrimination (18 percent to 32 percent). Trump supporters are also more likely to say the government favors black people over whites (51 percent to 61 percent). This may stem from the fact that white Trump supporters are far more likely to say their racial identity is “extremely” important to them (40 percent compared to 15 percent). 



Few things have been more frequently discussed in the recent months than the rise of Trump. Many commentators have noted the key political and institutional dynamics that gave Trump a path to the nomination (as I did here). Others have noted the ways in which the GOP set itself up for a fall by pandering to the donor class and delegitimizing political institutions (I discussed these here). These are all true. But what sets Trump supporters apart from the people who supported other candidates is their endorsement of vile stereotypes about people of color. Anti-black and -Muslim stereotypes are the strongest predictor of Trump support I was able to find — far more powerful than economic factors.

Worryingly, research shows that when people support a candidate or party, they tend to move closer to the views of the candidate or party they support. By tacitly accepting Trump’s racism, the GOP has empowered white supremacists and widened the boundaries of racism that their party will tolerate. That will leave a lasting negative impact on American politics.


By Sean McElwee

Sean McElwee is founding executive director of Data for Progress. He tweets at @seanmcelwee.

MORE FROM Sean McElwee

Related Topics ------------------------------------------

2016 Donald Trump Editor's Picks Elections 2016 Immigration Republican Party Rnc 2016