Gays and the Election

There's been quite a bit of handwringing among Democrats about Kerry's loss and the chief scapegoat seems to be gays. This AM I heard that Bill Clinton had advised Kerry to come out in favor of some of the anti-gay marriage amendments, but that Kerry had refused, and the Clintonistas are citing this as the reason why Kerry lost. And then there's this in today's Washington Post from Tina Brown:

On Wednesday morning, even the gay editors of liberal upscale magazines were prepared to tell you that if there's one person who should get a big bouquet from Karl Rove it's Massachusetts Chief Justice Margaret Marshall, aka Mrs. Anthony Lewis, who forced her state to authorize gay marriage. That was the trade-off for 45 million Americans without health care.

Exactly how much evidence is there that the gay marriage issue hurt Kerry? I did a regression analysis of Bush's vote in 2004, controlling for Bush's vote in 2000 and whether or not there was an anti-gay marriage measure on the ballot. The results below show that, if anything, Bush's vote was a bit lower than expected in states that did have such a measure on the ballot, though this result is not statistically significant.


Number of obs = 51
R-squared = 0.9643
Adj R-squared = 0.9628

	Coef.	Std. Err.	p>|t|
Bush00 %	.98.	.03	.000*
Gay Marriage Measure on the Ballot	-.105	.71	.883
constant	3.85	1.40	.008*

*Statistically Significant at .05 or better.

On the other hand, I did a regression for the counties in Ohio, this time controlling for Bush's 2000 vote and for the percent voting yes on that state's measure to ban gay marriage. These results suggest that Bush's vote increased with support for the marriage measure, even when controlling for Bush's vote four years ago. (These results have been weighted for the number of votes cast in the county).

Number of obs = 88
R-squared = 0.9899
Adj R-squared = 0.9897

	Coef.	Std. Err.	p>|t|
Bush00 %	.97.	.02	.000*
% Voting Yes to Ban Gay Marriage	.23	.02	.000*
constant	-.12	.01	.000*

*Statistically Significant at .05 or better.

Overall, for every 10 percentage point increase in the vote against gay marriage, Bush's vote went up 2.3 percentage points. Some might suggest that that was enough to tip Ohio, and the election, from Kerry. On the other hand, I'm not clear about the causation here. Voters may have turned out to vote against gay marriage and also voted for Bush, but it's also possible that people who were turning out to vote for Bush also were more likely to vote against gay marriage. In other words, I'm not sure if the gay marriage ban helped Bush, or if Bush helped the gay marriage ban.