Contagious: Why Things Catch On by Jonah Berger Page B

question with the name of their city or state. Evanston. Birmingham. Florida. Nevada. If asked to clarify, they might add “near my office” or “across from the supermarket.” Few will be more specific. And why should they be? Although geography clearly matters in voting—the East Coast leans Democratic while the South skews Republican—few people would think that the exact venue in which they vote matters.
    But it does.
    Political scientists usually assume that voting is based on rational and stable preferences:people possess core beliefs and weigh costs and benefits when deciding how to vote. If we care about the environment, we vote for candidates who promise to protect natural resources. If we’re concerned about health care, we support initiatives to make it more affordable and available to greater numbers of people. In this calculating, cognitive model ofvoting behavior, the particular kind of building people happen to cast their ballot in shouldn’t affect behavior.
    But in light of what we were learning about triggers, we weren’t so sure. Most people in the United States are assigned to vote at a particular polling location. They are typically public buildings—firehouses, courthouses, or schools—but can also be churches, private office buildings, or other venues.
    Different locations contain different triggers. Churches are filled with religious imagery, which might remind people of church doctrine. Schools are filled with lockers, desks, and chalkboards, which might remind people of children or early educational experiences. And once these thoughts are triggered, they might change behavior.
    Could voting in a church lead people to think more negatively about abortion or gay marriage? Could voting in a school lead people to support education funding?
    To test this idea, Marc Meredith, Christian Wheeler, and I acquired data from each polling place inArizona’s 2000 general election. We used the name and address of each polling location to determine whether it was a church, a school, or some other type of building. Forty percent of people were assigned to vote in churches, 26 percent in schools, 10 percent in community centers, and the rest in a mix of apartment buildings, golf courses, and even RV parks.
    Then we examined whether people voted differently at different types of polling places. In particular, we focused on a ballot initiative that proposed raising the sales tax from 5.0 percent to 5.6 percent to support public schools. This initiative had been hotly debated, with good arguments on both sides. Most people support education but few people enjoy paying more taxes. It was a tough decision.
    If where people voted didn’t matter, then the percent supporting the initiative should be the same at schools and other polling locations.
    But it wasn’t. More than ten thousand more people voted in favor of the school funding initiative when the polling place was a school. Polling location had a dramatic impact on voting behavior.
    And the initiative passed.
    This difference persisted even after we controlled for things like regional differences in political preferences and demographics. We even compared two similar groups of voters to double-check our findings. People who lived near schools and were assigned to vote at one versus people who lived near schools but were assigned to vote at a different type of polling place (such as a firehouse). A significantly higher percentage of the people who voted in schools were in favor of increasing funding for schools. The fact that they were in a school when they voted triggered more school-friendly behavior.
    A ten-thousand-vote difference in a statewide election might not seem like much. But it was more than enough to shift a close election. In the 2000 presidential election the difference between George Bush and Al Gore came down to less than 1,000 votes. If 1,000 votes is enough to shift an election, 10,000 certainly could. Triggers matter.
    So how do