There are many factors which can determine why people vote for particular political candidates over others – party affiliation, race, and gender can all be important. In new research, Morten Hjortskov and Simon Calmar Andersen examine how voters order the preference of what they look for in a candidate. They find that candidates’ gender and race can be important to voters, but only when voters are first satisfied with their party affiliation. 

Politicians, media, and commentators often discuss the significance of a political candidate’s gender and/or race in relation to electability and party strategy. For instance, when Joe Biden and the Democratic Party selected Kamala Harris as the vice presidential nominee in 2020, it was viewed as a departure from prioritizing geographical balance across the country as the primary strategic consideration in favor of achieving gender and racial balance on the ballot.

However, research on which candidates voters choose is surprisingly mixed on these two characteristics. Some studies find that the candidate’s gender matters, some not. Some find that the race of the candidate matters, some not. And some find that the influence is determined by the voter’s party affiliation.

The importance of one specific characteristic for voters

In new research, we look at the question from a slightly different angle. Most studies assume that voters consider several characteristics and policy preferences of the candidates on equal terms, so to speak. While they will not put equal weight on each characteristic, they will make trade-offs between them. The chosen candidate may not have the ideal makeup of characteristics and policy preferences, but they will hold a combination that is optimal given the available information when comparing the candidates on the ballot. The voter is maximizing utility given the circumstances they face.

But is this how people actually make up their minds about political candidates? Not always, in our view. Sometimes, people stand firm on their beliefs. Or, perhaps, their belief (in singular). This inclination for one characteristic is called lexicographic preference. When we apply it to people’s choices about candidates, it means that they have one characteristic or policy preference, and only one, that they focus on when making their choice on who to vote for. In a polarized US, this characteristic could be the party affiliation of the candidate.

If people have this type of preference, they do not trade it off and substitute it for some other characteristic. If the candidate lacks a certain characteristic, which is a must for the voter, then no amount of “correct” policy opinions or personal characteristics can sway them to vote for this candidate. There is no trading off and no substitution. This type of preference can make voting easier: you just need to look for one factor in a candidate. If the candidate does not have it, then they are rejected, and other characteristics do not come into play.

How voters order what is important to them in a candidate

Of course, this is also a simplification. Voters do consider more than one characteristic when voting. Our contention is, however, that they do it in a certain order. There are first-order preferences, like party affiliation of the candidate, and there are second-order preferences, which could be characteristics such as gender and race. What is crucial is that only when the first-order preference is tied will the voter consider second-order preferences like gender and race. For example, when choosing between two candidates from the same party in a primary or in “top-two” primary elections, gender and race might have an impact. When the choice is between two candidates from different parties, party affiliation dominates gender and race. For example, if the first-order preference is for a woman candidate, and there are two women on the top of the ballot, then the voter’s party affiliation will be the deciding factor.

It might not seem like a big deal, but we think it is. Studies of gender and race have often analyzed either within or between parties (or with no information of party affiliation at all) when concluding whether these two characteristics matter. The idea, of course, is that when using this type of preference as a lens to understand people’s choices about candidates, we would mostly expect second-order preferences like gender and race to matter when party affiliation is absent or the same.

“Voting in June 8th, 2010 California prim” (CC BY-NC-ND 2.0) by Steve Rhodes

Comparing candidates’ attributes effects on voter support

In our study, we use a special kind of experiment to identify the potential effect of second-order preferences: conjoint experiments. In these, the voter is presented with two made-up candidates with a set of characteristics like gender and race. They are asked to choose between these two, and the influence of the various attributes can be determined afterwards. Because the attributes vary across respondents, some are presented with two candidates with different party affiliations while others see two candidates with the same party affiliation.

Our results show that gender and race have more influence when party affiliation of the candidates is the same, indicating lexicographic preferences. So, voters seem to take gender and race into account when voting. It mostly happens, however, when they have checked the party affiliations of the candidates.

It is important to note that we are not saying that voters do not substitute between any characteristics when making their choices about who to vote for. It is entirely possible that they do. Our claim is that it mostly happens within the levels in the decision hierarchy – not between the first-order preference and the next order preferences. It is also possible that there are more than just two levels of preferences. Indeed, we cannot say what exactly makes a first-, second-, or third-order preference. We had an idea that since there seems to be a great deal of polarization in the US, party affiliation is a candidate for a lexicographic, first-order preference that should be more important than gender and race in candidate choice. Our results seem to point in this direction.

This raises the question about the implications of such knowledge about how voters choose political candidates. When political strategists focus on gender and race representation on the ballot, it looks as if it is a valid concern. But mostly when the choice happens within party, like in primaries and “top-two” elections.

• This article is based on the paper, ‘Lexicographic Preferences in Candidate Choice. How Party Affiliation Dominates Gender and Race’, in the British Journal of Political Science.
• Please read our comments policy before commenting. 
• Note: This article gives the views of the author, and not the position of USAPP – American Politics and Policy, nor the London School of Economics. 
• Shortened URL for this post: https://wp.me/p3I2YF-e80