Gay marriage: three key factors behind voting for bans

http://www.guardian.co.uk/commentisfree/cifamerica/2012/may/10/gay-marriage-three-key-factors-voting-bans

• wanted to examine whether the percentage of religious adherents, taken from the Association of Religion Data Archives (ARDA), within a state could explain county-to-county variance in support of marriage bans. I also added variables for partisanship (as measured by Obama's 2008 vote percentage) and for educational attainment (as measured by the percentage of voters within each county with at least a bachelor's degree, according to census figures).
• What I found many would call predictable: 89-93.5% of the variation in vote for same-sex marriage amendments could simply be accounted for these three variables. One of the more interesting pieces of data is that the variable with the least impact is religion. For every extra percentage point of a county's population that is religiously adherent, there is only a 0.1-point change we would see in support for marriage bans.
• Education, it turns out, is far more important at accounting for the differences between county votes in favor of same sex marriage bans. When holding religious adherence and Obama support constant, we would expect that the percentage of support for same-sex marriage bans to drop by 0.8-0.9 of a point for every extra 1% of a county's population that has at least a bachelor's degree.