Dozens of mythologies have grown up since November 2004. Among them: that initiatives banning gay marriage cost John Kerry the White House.

Spurred on by the right and by some frightened Democrats, mainstream analysts bought into this canard.

Although it is widely repeated, there is almost no evidence to support the proposition. As is often the case in analysis of elections, facts and evidence take a back seat to opinion, spin and emotion.

The basic facts are clear: Initiatives outlawing gay marriage appeared on the ballot in 11 states. Kerry won two of the states, Oregon and Michigan, but no one could have imagined an ’04 Democratic presidential victory in Georgia, Kentucky, Mississippi, Montana, North Dakota, Oklahoma or Utah. Some may have hoped for a Kerry win in Arkansas, but it was completely unrealistic.

To say George Bush won any of those states because anti-gay-marriage initiatives appeared on the ballot is to misunderstand fundamentally the political dynamics of those places. Bush would have won them whether or not such an initiative was on the ballot.

That leaves Ohio. Since the election was decided there, it is admittedly a pretty important state to leave.

Before returning to Ohio, examine the general arguments. Two are made. One is that the presence of these initiatives brought out pro-Bush voters who would otherwise have stayed home. The second theory is that these initiatives heightened the salience of gay marriage in these states, to Kerry’s detriment.

The logic of the turnout argument has always eluded me. It assumes a significant number of voters who say to themselves, “I don’t care enough about who the next president is to vote, in what is forecast to be a close election, but I do want to make sure the anti-gay-marriage initiative passes with 61 percent, not just 60 percent.” I don’t believe many voters acted on that interior dialogue.
The data bear out my conjecture. Sophisticated new statistical techniques reveal that less than one-half of 1 percent of Ohio voters cast a ballot on the marriage initiative but not in the presidential race. Four times as many voted for president while skipping the initiative.

More generally, statistical analysis of turnout indicates that past participation along with the money spent in each state did significantly affect turnout. Whether there was a competitive Senate or governor’s race may have had some impact as well. But the presence of a gay-marriage initiative had no consistent effect on turnout.

But even if the initiatives did not increase turnout, did they alter the issue terrain? Again the data suggest not.

Voters in Ohio trusted Bush over Kerry to deal with terrorism by an 18-point margin and trusted the president over the Democrat on the economy. In a state with a five-point margin in Republican Party identification and those kinds of advantages on core issues, there isn’t much room for gay marriage to be decisive.

Statistical analysis underlines the point. Vote is directly related to past Democratic performance in each state, the state’s liberal-conservative balance and the condition of the state’s economy but not to the presence or absence of a gay-marriage initiative.

This is not to say that cultural divisions are unimportant. They are central. Nor is it to say that these initiatives did not provide a basis for church-based organizing and exhortations to vote from the pulpit. They may have even been a conduit for additional get-out-the-vote money. And heaven knows, Democrats would have been happier with more news stories about Iraq and the economy and fewer about gay marriage. But the initiatives themselves did not put Bush back in the White House.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982, including Sen. John Kerry (D-Mass.) last year.