In this Washington Post article, Robert Barnes writes:
bq. Both sides cite studies that they believe show that the law has not resulted in lower turnouts for minorities and others or, alternately, show that minorities are most likely to be affected. There are numerous media accounts and other reports of fraudulent voting, as well as a corresponding study from the Brennan Center for Justice at New York University law school titled “The Truth About Voter Fraud,” which attempts to knock down many of them.
This is classic “he said, she said” journalism. But is the perspective of both sides based on equally credible research? A few weeks ago, I posted on this topic and cited two studies that showed that voter identification laws are associated with slightly lower turnout and may adversely affect minorities.
Barnes suggests that there are studies that do not show this, or show the opposite. What are they? Can it really be that whites, not minorities, are less likely to have these forms of identification? That more-educated, not less-educated, people are the most likely to be adversely affected by these laws? Even a supporter of these laws, Richard Posner, acknowledges that these laws will adversely affect those least likely to vote. And doesn’t the Brennan Center’s systematic debunking of voter fraud complaints constitute better evidence than “media accounts and other reports of fraudulent voting”?
I am not filing an amicus brief on behalf of the opponents of these laws. I am merely suggesting that the studies cited in my original post constitute the best extant scholarship and shouldn’t be lumped with other studies in a “both sides cite studies” formulation. That the best social science supports one set of partisan views doesn’t mean it deserves to be downplayed so that news reports can seem “objective.”
[Addendum: A reader sends along this paper by Jason Mycoff, Michael W. Wagner, and David C. Wilson. They find:
bq. Examining voting behavior data across four elections (2000, 2002, 2004, and 2006) at both the aggregate and individual levels, our results suggest that voter identification laws do not affect voting at either level.
Their findings about the effect on aggregate turnout conform to those of the Alvarez et al. paper I cited in my original post. Their findings about individual-level turnout are different in that Mycoff et al. find a non-significant effect and Alvarez et al. find a significant effect. There are differences in the details of their statistical analyses that may explain this, but even in the Alvarez et al. analysis, the effect of voter identification laws is substantively small, shifting the likelihood of turnout by 1%. Alvarez et al. also show that the effects of these laws are larger among those with less education and income; Mycoff et al. do not examine these kinds of interactive relationships.
These are working papers, and clearly they are not the final word. But early indications suggest that voter identification laws will disproportionately affect groups that typically turn out at lower rates.]