When Americans cast their vote for the next President of the United States, the candidates’ personalities may matter more to voters than their stances on international trade. In an Opinion published October 18 in Trends in Cognitive Sciences, Duke University researchers propose a new model for predicting voter choice that shows that identity can override policy issues when voting for a candidate. This model, which is based on cognitive and political science studies, aims to better predict election outcomes.
“One of the most striking features of elections is just how strongly people feel about the candidates, despite not being experts in all the issues in the election,” says Scott Huettel, a professor in the Department of Psychology and Neuroscience at Duke University, who co-wrote the paper with Libby Jenke, a graduate student in political science at Duke. “When we vote, that action can reinforce features of our own identity, such as our patriotism or our identification with a social group. We think that relying on identity allows people to make voting decisions in ways that aren’t predicted by older models.”
Traditional models assume that voters evaluate candidates’ policy positions and then choose the candidate whose positions most closely match their own, even though surveys have shown that very few voters can explain their own positions and the candidates’ positions accurately, even concerning major political issues like gun control or health care reform. These models still allow for fairly accurate voting predictions, but have limitations when used on a smaller scale (e.g., primary elections).
Jenke and Huettel address these limitations by using cognitive science to explore how identity and policy compete to shape voters’ choices. They identify three ways this conflict can play out:
1. The act of voting (identity) versus the consequences of the vote (policy)
Work in social and behavioral psychology has shown that when there is an immediate reward, people tend to devalue a task that has long-term benefits. A vote that reinforces a powerful personal identity carries immediate benefits (e.g., a sense of community with others who voted for a specific party), but could also cause us to ignore the long-term benefits we would receive if we voted on policy.
2. Social reputation (identity) versus own interests (policy)
Our social identity or reputation often plays a bigger role in driving our decisions than our own economic and personal interest. For example, when everybody you know votes for a particular candidate, you may feel compelled to vote for them to maintain your social identity.
3. What you value more (identity versus policy)
Neuroscience studies show that the way we track, evaluate, and learn about outcomes differs depending on whether a decision involves a social context, as it does when identity enters into voting decisions.
“We conceive of the balance between identity and policy as like turning a dial,” says Huettel. “When people think more about identity, like when considering an appeal to their social class or sense of self, then that diminishes the effect of policies upon their votes.”
Huettel and Jenke say that this presidential election is a good example. During the Republican Primaries, voters had 17 candidates to choose from. In order to stand out, candidates had to appeal to issues related to identity–like religion or race–as everyone belonged to the same political party and had generally similar policy positions. “These identities are less well connected to policy than party identity, which can’t be used in primaries,” says Jenke. “This explains why previous models, with their emphasis on policy, do better at predicting general elections than primaries.”
“In this election, identity has played a large role in both parties, as there are large swaths of the electorate who see their vote not just as an expression of policy positions, but as saying something about who they are, like saying, ‘I voted to protest corruption or I voted because I’m not part of the Washington Elite,'” says Huettel. “When they connect with a candidate, what are being reinforced are the identity variables and not the policy positions.”
Although the researchers emphasize that their work with this model needs to be tested and is not complete, they are optimistic that new research in cognitive science will continue to help solve problems that political scientists see when they predict voter choice using the current models. “Behavioral political science faces many of the same challenges as behavioral economics,” says Huettel. “We think we need to start identifying cases where the traditional model might fail and use those cases to guide more research.”