Dan Balz writes in the Washington Post, as many reporters have this week, 

In 2004, Republicans used ballot initiatives barring same-sex marriage to spur turnout among their conservative voters. That strategy helped then-President George W. Bush win reelection.

But did it? I argued in 2006 that it didn’t:

It’s true that states with such initiatives voted for Bush at higher rates than other states, but that’s mostly because the bans were proposed in conservative states. In fact, Bush’s share of the vote rose just slightly less in the marriage-ban states than in the other states: up 2.6 percent in the states with marriage bans on the ballot, up 2.9 percent in the other states.

Political scientist Simon Jackman of Stanford has more here (pdf). He concludes that the marriage referenda tended to increase turnout but not to increase Bush’s share of the vote. And in a county-by-county analysis of Ohio, he found no clear relationship between increased turnout, support for the marriage ban, and increased support for Bush.

Matthew Dowd made the same point yesterday:

Speaking from experience as the chief strategist in 2004 for President Bush, I saw in close detail how little gay marriage could influence turnout of conservatives or evangelicals.   In 2003 and 2004, we did a series of public opinion tests on different messages related to the micro targeting project that would cause voter groups to turn out more in President Bush’s favor.  We tested social issues as well as messages related to the economy, national security, taxes and the size of the federal government.  Not a single social issue (which included gay marriage) fell on the effectiveness scale in the top eight messages.

Further, in analyzing the election returns in the aftermath of the 2004 presidential race an interesting set of data was revealed.   In states that had gay marriage amendments on the ballot including key target states, there was no statistical difference in turnout of conservatives from states that did not have these amendments on the ballot.  Gay marriage had no effect on turnout even among the most conservative potential voters in both the data before Election Day and the returns on Election Day.

Other senior officials from the 2004 Bush campaign confirm: It wasn’t gay marriage that brought social conservatives to the polls, it was national security and the war on terror.

At any rate, as Balz noted, the politics of gay marriage have changed for sure, in Ohio and elsewhere.