Psychologists have long known that situations can shape behavior—for example, when time pressure turns would-be Good Samaritans into callous passersby. More recently, studies have shown that even cues as subtle as a mild scent can trigger changes in participants’ thoughts and actions. But can environmental signals influence important decisions in the real world? A new study of election voting suggests they can.

Investigators at the Stanford Graduate School of Business analyzed data from Arizona’s 2000 general election, looking for a link between where voters cast their ballots and whether they backed a school-funding initiative. Those who voted in a school were more likely than other voters to favor raising the sales tax to fund education. Even after accounting for factors such as where voters lived, the difference was large enough to tip a close vote.

The researchers then confirmed with a controlled experiment that environmental cues were likely to have caused the effect: participants who were primed with images of churches proved less likely to support a stem cell initiative than were subjects who were shown more neutral images, such as office buildings.

This effect occurs only if voters might go either way on an issue, explains study co-author Jonah Berger, now a professor at the University of Pennsylvania. “Environmental cues aren’t going to get people to do things they wouldn’t do otherwise,” Berger says. He suggests that as a first step toward avoiding the induced bias, election officials might try to pick “more innocuous multipurpose rooms” in the polling place to reduce, for example, the religious stimuli in a church setting.