Saturday, April 02, 2005
Terrorism v. Gay Marriage
I ran a model of the Bush-Kerry vote controlling for party identification, ideology, sex, age, race, income, and education. I also controlled for whether someone thought invading Iraq was worth it and whether they approved of gay marriage. Both the Iraq and gay marriage variables were statistically significant, but the Iraq variable had much more impact on vote choice. The probablity of a Bush vote was 87 percent among those who approved of the invasion, but only 31 percent among those who did not, a difference of 56 points. In contrast, the probability of a Bush vote among those approving of gay marriage was 42 percent versus 67 percent among those opposed, a difference of only 25 points. In other words, attitudes about Iraq and more than twice impact on the vote than did opinions about gay marriage. I then re-ran the model controlling for 2000 vote. With this variable included, the variable for gay marriage was no longer significant, suggesting that the gay marriage issue did nothing to change votes from 2000 to 2004.