Mike Allen’s morning newsletter, Playbook, features this item today:
MOST SURPRISING STORY OF DAY — Atlanta Journal-Constitution p. A1, ‘Voter turnout surges amid five-year ID law: ‘Rhetoric on both sides’ overblown, AJC review finds,’ by Shannon McCaffrey (who moved to AJC from AP last year) : ‘Few things stir as much political passion as voter ID laws. Since Georgia lawmakers passed one of the nation’s first and strictest laws in 2005, each side has charged the other with trying to undermine the electoral process. When Georgia became one of the first states in the nation to demand a photo ID at the ballot box, both sides served up dire predictions. Opponents labeled it a Jim Crow-era tactic that would suppress the minority vote. Supporters insisted it was needed to combat fraud that imperiled the integrity of the elections process. But both claims were overblown, according to a review by The Atlanta Journal-Constitution of statewide voting patterns [in general elections] in the five years since the law took effect.
As many NRO readers might immediately realize, this story isn’t surprising at all, actually – anyone who’s consulted the facts on voter ID, rather than inhaling the D.C. conventional wisdom, knows that there’s no empirical evidence new laws have depressed minority turnout especially, and pretty weak evidence that they reduce turnout overall. In fact, the most prominent evidence that such measures don’t hurt minority turnout comes from . . . Georgia, as readers of the Atlanta Journal-Constitution, if not Politico, already know. The state’s voter-ID law was passed in 2006. From the AJC, 2010: “Voter turnout boosted in metro Atlanta.” 2012: “Despite voter ID law, minority turnout up in Georgia.”
Obviously, such assessments are unscientific, but the burden of proof lies on voter-ID critics. Look at the social-science research compiled by the Brennan Center for Justice, which opposes voter-ID requirements: The studies ostensibly demonstrating the threat to minority turnout rely on a priori assessments of who has an ID (not even, often, which registered voters have IDs), rather than research conducted on actual turnout, generally or among minority groups. Nate Silver examined the question last summer and summarized the research on turnout, which shows that it’s not affected or reduced in a statistically insignificant way. He argues that statistical significance is a problematic concept here because it relies on ”the null hypothesis,” which assumes the treatment (here, requiring voter ID) will not have an effect. Silver thinks that’s problematic here (how convenient) because by definition requiring an ID prevents some people from voting, but if the metric we’re measuring is just turnout, I’m not sure it is: The passage of a voter-ID law also is likely to drive attention toward voter-turnout and voter-registration efforts (as, indeed, it has in Georgia), which means, ceteris paribus, turnout is just as likely to go up — balancing out whatever suppressive effect the law may have. This doesn’t prove the law can’t be racially discriminatory in some way, as the research on the question notes, but it does belie the argument that it’s clearly depressing minority turnout, for which the evidence is scant or nonexistent.