Editorials: Studies Back Up That Few Elections Are Swung by Voter ID Laws | Nate Cohn/NYTimes.com
Last week, I wrote an article arguing that voter ID laws don’t swing many elections. This week, the Brennan Center for Justice says I have it “wrong” on voter ID. Yet, oddly, it’s hard to find a place we disagree. As the Brennan Center puts it in the second sentence of their article: “Yes, it is likely rare for an election to be close enough for voter ID laws to swing the outcome.” The Brennan Center instead disputes my contention that studies tend to “overstate the number of voters who truly lack identification.” My position on the matter, setting aside whether the laws are a cynical attempt to disenfranchise Democratic voters, is based on these facts: Many studies do not use robust matching techniques when comparing state voter registration and licensing databases (and robust matching, even when used, isn’t perfect); and many studies fail to match voter registration files with alternative forms of identification, like United States passports or military identification. The studies with the most sensational and widely publicized findings have generally failed to do these things. The most famous of these was a studyfinding that 758,000 of Pennsylvania’s registered voters lacked identification. It caused liberals to wonder whether voter ID laws could steal elections. The result was publicized by the Brennan Center, but more rigorous studies have since cut that figure nearly in half.