Bernie Sanders

Brendan Smialowski/Getty Images

Political Science

Democrats Don’t Want to Nominate a Candidate Who Looks Like Bernie or Joe

In a new national survey, the party’s primary voters prefer female candidates of color in 2020. So why are two white guys in the lead?

Continue to article content

Gabriele Magni is a postdoctoral fellow at Princeton University.

Andrew Reynolds is a professor of political science at the University of North Carolina at Chapel Hill.

There are now, by our count, two dozen candidates in the race for the Democratic nomination for president, and it’s by far the most diverse field in modern American history. Two are black men; one is a black and Asian American woman; four are white women; there’s a Latino man, an Asian American man, a Pacific Islander woman and a white gay man.

And yet, two straight white men lead the polls, and 11 others are in the race, with more joining the field seemingly every week.

Story Continued Below

With shifting attitudes and demographics among the Democratic base, is a straight white man really what the party wants in this moment? We did some research, and it turns out that even among white male voters in the Democratic primary, the answer is no. In other words, Joe Biden and Bernie Sanders may be leading the polls despite their race and gender, rather than because of it.

Evaluating voter attitudes is tricky. All candidates present a bundle of characteristics that are often correlated. Biden is not only a white man, he is also older with a lot of political experience. Pete Buttigieg is gay, but he is also a white man, young and religious. For this reason, it can be hard to disentangle to what extent each candidate characteristic influences vote choice. And when you survey people, respondents often provide untruthful answers they consider to be socially acceptable. For instance, a voter may dislike a black or a gay candidate but be reluctant to openly admit that for fear of looking racist or homophobic.

To address these challenges, we engineered an experiment. In a large, nationally representative survey at the end of 2018, we presented respondents with pairs of hypothetical candidates and asked them to vote for their preferred ones. In the survey, we randomized candidate gender, sexual orientation, race and ethnicity, religion, education, age, health and political experience. Because of that randomization, we can estimate the independent effect of each candidate characteristic (for example, being gay) on vote choice while controlling for other attributes (such as being a young, white, religious, man).

What we found was that, at least in the 2020 Democratic presidential campaign, candidates are likely to be viewed more favorably for being nonwhite and nonmale—but not for being nonstraight.

Looking at all the Democratic voters in our sample, we found a strong preference for women over men (plus 6.8 percentage points) and for minority candidates: Compared with whites, black candidates received an advantage of 4.2 percentage points, Latinos 2.2 percentage points, Asians 0.4 percentage points and Native Americans 4.5 percentage points. Meanwhile, despite public opinion polling showing that most Americans today would be OK with a gay president, our study showed gay candidates face a slight penalty compared with straight ones (minus 1.2 percentage points).

Fascinating results emerged when we analyzed specific subsets of the Democratic primary electorate. Self-identified very liberal voters—a fourth of Democrats who voted in the 2016 primaries—showed a very strong preference for women candidates (plus 15 percentage points compared with men) and they also preferred gay over straight candidates (plus 2.3 percentage points). Democrats who never attend religious services—about a third in our sample and nationally—showed an even stronger preference for gay candidates: plus 5 percentage points. In contrast, moderate and conservative Democrats—who represent about 40 percent of the party’s primary voters—penalized gay candidates (minus 3.7 percentage points) but still preferred women to men (plus 6.1 percentage points).

Notably, women like female candidates more (plus 8 percentage points)—but so do men (plus 5.3 percentage points).

Black and Latino Dems—who, together, are about a third of the primary electorate—prefer candidates of their own race by a substantial margin (plus 14.8 and plus 7.3 percentage points, respectively, over whites). But even white voters show a preference for black and Latino candidates (plus 0.7 percentage points and plus 2 percentage points, respectively). Meanwhile, voters’ age and education seem to have little impact on their preferences for demographic attributes.

Wealthy voters, those with annual household income over $100,000, appear to prefer women over men (plus 7.8 percentage points). Perhaps unsurprisingly then, Kamala Harris led by a long way in big-donor fundraising after the first quarter of 2019, and the other five women on average pulled in slightly more than the average of all the men.

These survey results challenge existing studies on the role of identity and bias in candidate preferences. Past political science research has found that female candidates win at similar rates to male candidates, but that’s only because the women who enter races tend to be better candidates independent of gender. Based on our study, in this primary, women are actually preferred because of their gender, independent of other attributes. Similarly, when it comes to minority candidates, existing studies are on balance more negative and show that nonwhite candidates have often been penalized, especially by the white electorate. In 2019, however, most Democratic voters—including whites—say they would prefer minority candidates over their white counterparts. This is likely because of the shifting demographic makeup of the party’s primary voters as well as evolving attitudes.

Voters are eager for change, not only in terms of policy but also with regard to who represents them. Clearly, ideological and policy positions are also important, but our research shows that demographic characteristics can certainly affect how voters assess candidates.

So why then are two white men leading in the polls? Our findings give us probabilities based on the demographic traits of hypothetical candidates “all else being equal.” But in the real world, all else is not equal—the real race involves Biden, Harris, Sanders, Buttigieg, Elizabeth Warren and other candidates with specific histories, skills, foibles and policy positions. And right now, two white guys are in the lead, regardless of what Democratic primary voters prefer in the abstract.

It’s possible that, when presented with real names and backgrounds, Democrats stick with white men because of subconscious biases that affect their judgments of the character, history or “likability” of women and minority candidates. Or it could be that Democratic primary voters who prefer a nonwhite, nonmale candidate believe it’s safer to choose a white man for the general election. (Though a new study suggests that strategy could backfire, because a white male nominee could lower Democratic enthusiasm in the general election.)

Biden and Sanders are also the most well-known candidates, even as plenty of other white men are barely registering in the polls. Our study couldn’t account for the importance of name recognition, because all our hypothetical candidates were equally new to the voter. But when the field narrows, our survey suggests that a nonwhite or nonmale candidate is likely to emerge as a real contender—and when that happens, their identity will be a selling point rather than a hindrance.

Jump to sidebar section