I posted a couple of items here on the dubious methodological assumptions lurking in Richman, Chattha, and Earnest (2014), a paper that purported to show widespread evidence of non-citizens illegally voting in elections. I didn’t want to be particularly hard on Richman; anyone could make mistakes and write a bad paper, there’s professional pressure and get out there and promote your work to a broad audience when it might be ‘relevant’ in some fashion, and findings shouldn’t be ignored or buried simply because of their political implications.
Recent events suggest any sympathy or assumptions of good faith were almost certainly misplaced:
Jesse Richman endured a blistering critique Tuesday of his estimate of 18,000 noncitizen voters in Kansas and said he couldn’t support claims by Kansas Secretary of State Kris Kobach that Hillary Clinton won the popular vote because more than 3 million illegal ballots were cast in the 2016 presidential election.
Richman, who teaches political science at Old Dominion University, testified as an expert witness for Kobach in a trial over the state’s voter registration law. Kobach, who is seeking the GOP nomination in this year’s governor’s race, has referred to the 18,000 figure as the best available estimate for showing proof of citizenship is needed to address widespread voter fraud.
American Civil Liberties Union attorneys took aim at shortcomings in Richman’s methods and presented two experts to refute his conclusions.
Varying estimates from Richman are based on small-sample surveys, including one in which six of 37 noncitizens said they tried to register to vote. Under questioning by ACLU attorney Dale Ho, Richman acknowledged he had no way of knowing if those six were successful in their efforts.
In another survey, Richman looked at suspended voters — those who didn’t provide proof of citizenship — and flagged any names that appeared to be foreign. Ho asked if, for example, a name like “Carlos Murguia” would be flagged. When Richman said yes, Ho informed him Murguia is a federal judge in the same courthouse where the trial is taking place.
The good news is this really is probably the best Kobach can do. The bad news is it may be good enough.