The energy around Vice-President Kamala Harris’s campaign is undeniable. She easily secured the Democratic nomination, record-breaking Zoom fundraisers in support of her campaign have drawn hundreds of thousands attendees, and volunteers are signing up at eye-popping rates. Her rally crowds vastly overshadow the Republican ticket’s. And yet, when deciding who’d be her running mate, Harris seemed to seriously consider only white men for the role, a potential response to some voters’ concerns around her history-making candidacy and fears that America simply may not be ready to elect a woman, especially a woman of color, to the highest office in the land.
Political scientist Sanne van Oosten strongly disagrees with those concerns. As a postdoctoral researcher at the University of Oxford, she was part of a team that published a research paper last year finding that gender and race are less influential in voters’ perceptions of candidates than people may think. “I assumed that there was a whole lot of discrimination from voters for politicians of color,” she tells me. “I thought, This discrimination by voters is really hurting our representative democracy, which is one of our biggest assets. And if it’s not representative, then what is it worth? But that’s not what I found.” In fact, the data analysis, which pulled from 43 published studies, many of them U.S.-based, showed that a candidate’s race and ethnicity doesn’t negatively influence voting behaviors, and that some voters slightly prefer female candidates over male candidates. I spoke with van Oosten about what’s wrong with the assumption that candidates from marginalized groups are less electable and how it may affect Harris’s bid for the White House.
Before we unpack your research, can you elaborate on the idea that people won’t elect lawmakers from marginalized communities? A common perception, at least in the U.S., is that voters are less likely to vote for a woman or a person of color than a white man.
There is a lot of discrimination in the world, and people of color who are politicians in our democracies have it much harder. I assumed this was something that was coming from voters, but of course, there are so many different steps in the process of getting elected — the actual voting is only one of those steps. The discrimination takes place elsewhere. It goes to show how important it is that we have diversity in our democracies, in our parliaments, in our presidencies, because voters also want this. I’m talking about the average voter, and that’s an important distinction because there are very, very racist and sexist voters out there. But on average, voters aren’t biased against these politicians at a first glance. They can be made biased through a campaign, but their first impression isn’t a negative one.
Tell me about the meta-analysis you conducted and what you expected to find.
When you survey data, you have to know what kind of an effect you expect to find so that you can calculate how many respondents you need to field. I expected there to be a substantial effect size showing discrimination. So I ran my own candidate experiments, showing hypothetical politicians to respondents in a survey. These hypothetical politicians varied in age, gender, policy positions, race, ethnicity, religion. Then you let a respondent read this profile of a politician, and you just ask them, “Would you vote for this person?” Or you show them two politicians and ask, “Which one would you vote for?” But also I wanted to see what previous research had done. The outcomes of that previous research were actually kind of mixed, so I went to their original data sets and reran the analysis. Meta-analyses are seen as the gold standard because it doesn’t get more robust than this — you’re not just doing your own research, you’re synthesizing all of the research that has been done.
What were your results?
I started off studying gender, and I found that voters are significantly more positive about women politicians than about men politicians. I also found that there’s no significant difference in how voters assess politicians of color and white politicians. I was a bit surprised when I found these outcomes because I expected there to be a lot of sexism and racism among voters. Some of my colleagues thought I had done it wrong, especially for the gender outcome. So I went back and I did it again. Then some colleagues came out with a similar paper on gender and found the same results. So I thought, Okay, if there are two groups of researchers that are doing this really robust research and they both find the same outcome, then it should really be right. My paper also went through rigorous peer review. This is not my opinion, this is the outcome of a whole lot of data.
I don’t want to say that politicians of color do not face any racism, or women politicians don’t face any sexism. They do. But it’s not the first reaction from voters. This initial reaction from voters is actually quite positive. People on average want diversity. There are people on the flanks who are extremely negative about politicians of color or women politicians, especially women politicians of color, and they can really ruin it for a lot of people. But this isn’t what’s happening.
I also looked at in-group voting. Those findings were only based on the U.S. studies, because those were the only groups big enough to aggregate the data for. Kamala Harris could benefit very much from in-group voting, depending on whether people see her as their in-group. Black citizens in the U.S. are much more likely to vote for a Black candidate, same for Latinx citizens and Asian citizens, to a slightly lesser extent. White citizens in the U.S., however, are not significantly more likely on average to vote for white politicians. The meta-analysis also finds that there’s an overall positive effect size of whether people want to vote for a woman. Women are more likely to vote for their in-group and, though slightly less than them, male voters are also more likely to vote for women.
What would you say were some of the limitations of the meta-analysis? Any caveats we should consider?
Our findings are really about that first impression, and in actuality, especially in presidential elections, people don’t only have a first impression. They go through a campaign with a candidate. There are a lot of elections where people vote based on their first impression — we call these low-information elections, where people don’t really know who they’re voting for and don’t follow the news in that particular election. But with the U.S. presidential election, this is definitely not low information, people are following this stuff.
A classic criticism for this kind of research is social-desirability bias. We call it SDB — we talk about it so much in the field that we have an acronym for it— and it means that respondents know that they’re not allowed to be sexist and racist, so they answer in a way to hide that. That could be happening in this research, but on the other hand, I ran the same candidate experiments in Europe specifically looking at the Netherlands, Germany, and France. There was no sign of discrimination against politicians from ethnic minority groups, but there were massive, just unheard of, effect sizes of discrimination against Muslim politicians. That shows that in these kinds of experiments, people are willing to disclose prejudice. It’s not like this research cannot find racism.
Can you talk more about the impact of campaigning in how we perceive female candidates and candidates of color?
The goal of campaigns is to change voters’ ideas. Politicians and party strategists will do anything they can think of to get people to change their mind. If they can use racist and sexist stereotypes, they will. An example that I always go back to is that Donald Trump’s first step in politics was his birther campaign, where he was trying to make people doubt whether President Obama was really from the U.S. Then you also have all the hate campaigns against Hillary Clinton, making her seem untrustworthy and evil. There’s also J.D. Vance calling out Kamala Harris as a “childless cat lady.” I don’t know if that’s very effective campaigning, by the way, but he’s trying to mobilize these racist and sexist stereotypes.
But if we are aware of how racism and sexism are going to be mobilized in a campaign, it can have less of an effect. People will watch out for it and recognize it when it hits them in the face.
Knowing all of this, why do you think we have comparatively fewer women and people of color in office?
One reason is assumed prejudice. Party selectors — people who work in parties and who make decisions about who’s going to be selected for office, who gets support from within the party — assume that there is a lot of prejudice, and that makes them a little wary to select someone they think is risky. For example, there was so much talk about skipping over Kamala Harris and choosing a white man to replace Joe Biden. People thought, Well, a woman of color can’t do it, but I don’t think that’s necessarily true. We call this strategic discrimination. People engaging in strategic discrimination are not necessarily racist or sexist themselves, they just think that others have these biases. This assumed prejudice has been holding people back. We also talk about supply and demand. A lot of people say, “Oh, we want to put a woman in office, but women just aren’t applying.”
So what do you think about Kamala Harris’s candidacy? Does your research support the idea that she can win?
Definitely. The last couple of weeks have shown that there’s so much positivity. It could be the honeymoon phase, but she’s getting record funds and support. It looks like people are getting really enthusiastic about her. She can win and her race, ethnicity, and gender are no impediment to her winning, at least at first glance. These things could be used against her, but at first glance, people are not scared of voting for a woman of color.
This interview has been edited and condensed for length and clarity.
