On Tuesday, we can vote for whomever we want, but we don't get to pick where we vote. We might get assigned to vote at a school, church, library, community center, or whatever. But where we vote actually makes a difference to how we vote.
A variety of academic research demonstrates this effect. In particular, when people vote in schools, they're more likely to vote for higher taxes that would increase spending on public education. If they're voting in a church, they're more likely to vote against stem-cell research and gay marriage.
Our surroundings matter. The messaging around us matters. The small implicit and explicit environmental factors around us matter. This is a big reason why the entire advertising industry exists — to use messages that shape the choices consumers make. The campaign process is no different, except that the final step between voters and a selection is a totally random, possibly vote-changing environment.
Research from Stanford showed this effect, not just in a hypothetical situation, but in real elections with real differences in results. They looked at the results for several propositions in both Arizona and California, where the topics included stem-cell funding and raising taxes to increase education spending. The research showed that once you controlled for effects like demographics and political views, there was a meaningful impact due to the location of where the votes took place.
For example, people voting at a school were more likely than other voters (55 vs 53 percent) to support an initiative to raise the Arizona state sales tax in order to fund education.