What Our Friends Can Tell Us About the Elections
Asking people about their friends improves election predictions.
Posted Aug 22, 2018
At a recent girls’ night out, we broke a cardinal rule of polite conversation—we discussed politics.
We “friends of a feather” tend to agree on most things. That night, most agreed that the country needs change, and to move towards this change it is essential to go vote in November. But not everyone fell in line. One girlfriend said she would not vote if her favorite party does not put forward a progressive candidate. This attitude came as a surprise to me. Coming from a Balkan country with turbulent history, I experienced first-hand how quickly political systems can fail. Skip an election and you might find that the whole system has irrevocably changed. Or is no more.
Nevertheless, I relished this information because it helped me to update my own election predictions. If this well-liked and respected friend thinks this way, there could be many others like her in the country. I decided I should dial down my anticipation of the “blue wave” everyone seems to expect.
Similar discussions are happening at dinner parties around the country. The usual reluctance to discuss politics in polite society seems to be vanishing along with diversity in people’s social circles. After all, anyone with access to social media platforms can get plenty of information about their friends’ attitudes even without asking for it. If one so wishes, one can pre-select one’s friends accordingly to avoid discord.
We know better than ever how our friends feel about political issues. Can this knowledge be put to good use? It turns out that the answer is yes, when it comes to predicting elections. In a recently published study, my colleagues and I included a question about social circles in polls preceding the 2016 U.S. and the 2017 French presidential election. In both cases, asking people how they thought their friends would vote improved predictions of the election outcomes, compared to just relying on what they said about themselves .
In the U.S. we had a chance to include these social circle questions in a longitudinal study conducted over several months leading up to the election. We noticed that social circle questions improved not only the predictions of the overall election outcome, but also predictions of people’s own voting intentions and behavior (see figure).
There are several reasons why asking about social circles improves elections predictions. One, people know quite a bit about their friends’ characteristics, including their political orientation . This ‘insider knowledge’ can help the pollsters to overcome the limits of their samples, and include some information about people who otherwise might never participate in the polls in the first place.
The second reason is that it might be easier for people to say that friends will vote for an "embarrassing" candidate than that they themselves will do so . While this might have been the case in the 2016 U.S. election, it seems that people had no problem admitting they would vote for a far right candidate in the 2017 French election. It also doesn’t seem that there is much shame left when it comes to expressing voting preferences in the U.S. of 2018.
The third and perhaps the most interesting reason is that friends influence us over time. So even though we disagree with our friends now, we might come to see the world as they do in a month or two. In a way, reporting about our friends is like looking into our own future.
Of course, we can also choose to disconnect from disagreeing friends. Luckily, at least when it comes to close friendships, political disagreements do not need to be fatal for a relationship. In my case, I’ll certainly continue to like and respect my friend. Who knows, maybe in a few months I’ll come to see things the way she does.
 Galesic, M., Bruine de Bruin, W., Dumas, M., Kapteyn, A., Darling, J., & Meijer, E. (2018). Asking about social circles improves election predictions. Nature Human Behaviour, 2, 187-193.
 Galesic, M., Olsson, H., & Rieskamp, J. (2018). A sampling model of social judgment. Psychological Review, 125, 363-390.
 Barton, A. J. (1958). Asking the embarrassing question. Public Opinion Quarterly, 22, 67-68.