Campaigns and elections

Does how you vote depend on where you vote?

Lee Sigelman May 11 '09

“Where you stand depends on where you sit” — that’s Miles’s Law, the brainchild of a career government administrator who noticed that when his fellow bureaucrats moved from one agency to another, their attitudes seemed to change as well: cost-cutters became big spenders, free-market advocates became regulators, and so on.

Now there’s evidence that something akin to Miles’s law operates in the electoral arena. Call it Berger’s Law, in honor of Jonah Berger, who, with co-authors Marc Meredith and S. Christian Wheeler, has reported that where people vote — that is, the nature of their polling place — affects how they vote.

Berger and his colleagues aren’t just claiming that those who vote out in the ritzy suburbs tend to come down on one side of an issue while those who live on the wrong side of the tracks take the other side. Rather, they’re claiming that characteristics of the polling place itself matter.

In support of that claim, they offer two pieces of evidence. The first comes from a ballot initiative in Arizona in 2000 that proposed raising the state sales tax in order to increase education spending. Some polling places were schools, and 56% of those who voted in schools supported the initiative, as compared to 54% who voted at some other type of location. This difference didn’t seem to stem from uncontrolled differences between the two types of locations, for when numerous characteristics of these locations were taken into account, the observed difference in support for the initiative persisted.

As a follow-up, Berger and his colleagues conducted an experiment. First, they showed participants images of either schools or other buildings. Later, in what participants were told was an unrelated study, they “voted” on the education initiative. Those who had been shown pictures of schools turned out to be more likely (64%) than those who had been shown pictures of other buildings (56%) to support the school funding indicative — results that the authors interpreted as upholding a priming-based account.

As they conclude:

bq. These results illustrate the dramatic and unexpected influence that the environment can have on behavior. Seemingly trivial environmental contexts were found to have significant effects on consequential real-world decision making.

I’m leerier than Berger and his colleagues about calling the difference between 54% and 56% “dramatic,” though I concede that in many real-world elections 2% one way or the other is enough to tip the scales. And I’m not sure about the broader applicability of this idea, because I can’t think of a lot of other polling-place characteristics that could shape outcomes. Berger and his colleagues wonder, for example, whether voting in a church could influence support for gay marriage or stem cell research, but how common is it for churches to serve as polling places? I know that they sometimes do, but is that practice widespread enough to worry about? Perhaps some better-informed-than-I-am “Monkey Cage” reader can answer that question.