Location, location, location
From an article at The Blaze : Polling places often include schools, churches and some other public buildings. A growing body of research states that the location of where you cast you vote could be influencing those decisions to an extent, a factor called “priming.” A new study from Baylor University found that of those in England and the Netherlands who stopped to take a survey near a Christian church were likely to report themselves as more politically conservative and had a more negative attitude toward non-Christians, compared to those answering survey questions near a government building. I see just a few flaws in the assumptions here: Correlation, causation. They have demonstrated that a set of people who answer a survey in one location tend to be "conservative", and a different set of people answering the same survey tend to be more "liberal", using their own definitions of conservative and liberal. "Church-going" does not always mean "...