clock menu more-arrow no yes mobile

Filed under:

White support for Donald Trump was driven by economic anxiety, but also by racism and sexism

People cheer as voting results come in at Republican presidential nominee Donald Trump’s election night event at the New York Hilton Midtown on November 8, 2016, in New York City.
People cheer as voting results come in at Republican presidential nominee Donald Trump’s election night event at the New York Hilton Midtown on November 8, 2016, in New York City.
Spencer Platt/Getty Images

In the wake of Donald Trump’s victory over Hillary Clinton last week, many are still struggling to understand what happened. Explanations of Trump’s support have largely been driven by exit polls, which often identify "working-class whites" or "non-college-educated whites" as the significant swing group that help propel Trump to victory. But while the exit polls can help us identify which groups were crucial, they are much less helpful for understanding why these groups voted the way they did.

This is largely because exit polls fail to ask questions that measure the key concepts that may have been in play in this election. As a result, we are left with competing narratives, with some reports suggesting that economic insecurity was the decisive factor in this election, and others highlighting the role of racism or sexism in driving voters toward Trump. The truth, however, is that there is no single cause of Trump’s success among whites. All three factors played an important role.

To demonstrate that this is the case, I draw on a nationally representative survey designed by Matthew MacWilliams, Tatishe Nteta, and myself and fielded by YouGov two weeks before the election (October 26 to 31). Fortunately, that survey includes the kinds of questions needed to measure the competing narratives about what drove the white vote.

To capture economic security, I use a question that asked individuals how satisfied they were with their own economic situation. For sexism, I used an item taken from a well-established battery of questions designed to measure the concept. The question asks respondents whether they agree or disagree that "women seek to gain power by getting control over men." And to capture racism, I use an item taken from a battery that measures the extent to which an individual acknowledges the existence of racism. This particular item asks respondents to indicate whether they agree or disagree that "[w]hite people in the US have certain advantages because of the color of their skin."

The first chart below shows how whites responded to these questions in our survey. Note that approximately one in four white Americans were "not at all satisfied" with their personal economic situation. Such economic dissatisfaction is presumed by many to be the chief explanation for the Trump vote. But there is also a significant percentage of white Americans who demonstrated a denial of racism on our survey – more than 40 percent disagree that white people have advantages in the US because of their skin color. About 40 percent of white Americans were also either in agreement or neutral in rating the statement that "women seek to gain power by getting control over men."

Distribution of responses to three key poll questions.

To get a sense of the relative importance of each of these factors in explaining the vote for Trump, I estimated a statistical model of these white respondents’ vote choices. In the model, I included standard demographic and political control variables, such as gender, age, education, income, ideology, and party identification. I then added all three questions described above into the model at the same time. This allows us to see, for example, whether views about racism are still important when one accounts for a respondent’s reported levels of economic dissatisfaction.

To show the comparative importance of each of these three measures, I plot the relationships between the items and predicted support for Trump in the graph below. These graphs show how changing an average white respondent’s answer on each of these questions would affect his probability of voting for Trump rather than Clinton.

It is clear from this graph that there is no simple single explanation for why whites supported Trump over Clinton. Indeed, each of these three explanations appears to be important in explaining the vote. Dissatisfaction with one’s own economic situation leads to a sizable increase in the probability of supporting Trump, but so too does increasing levels of denial that racism exists in America.

Sexism is also a powerful force, as those who are more in agreement that women seek to gain control over men were far more likely to support Trump. Notably, the impact of sexism appears to be unique to the 2016 election, as I have shown in an analysis of other polling data.

Ultimately, the competing narratives about why Trump performed so well among whites are not competing at all; they are complementary. To truly understand Trump’s success means acknowledging that economic insecurity was part of the story, but so too were racism and sexism. And in truth, it likely took all three factors to allow him to edge past Clinton for a narrow victory.

Brian Schaffner is a professor in the department of political science at the University of Massachusetts Amherst and a faculty associate at the Institute for Quantitative Social Science at Harvard University. His research focuses on public opinion, campaigns and elections, political parties, and legislative politics. He is the co-author of the book Campaign Finance and Political Polarization: When Purists Prevail, co-editor of the book Winning with Words: The Origins & Impact of Political Framing, co-author of Understanding Political Science Research Methods: The Challenge of Inference, and author of Politics, Parties and Elections in America (seventh edition). His research has appeared in more than 30 journal articles and has received more than $2 million in external funding.

Schaffner is also the founding director of the UMass Poll and a co-PI for the Cooperative Congressional Election Study.