Chris Lawrence agrees that there were coding errors but argues that is easy to make such errors. I agree that such errors are easy to make, but, he did it twice, and the errors seem to systematically favour his position. Another thing that strikes me when I read Ayres and Donohue’s paper is that they report many regression results, some of which are favourable to the “more guns, less crime” thesis, but more of them are not. The Lott and Mustard paper and More Guns, Less Crime also report many regressions, all of them supporting Lott’s thesis. It sure looks like that even when he doesn’t have systematic errors in his data, Lott just reports the results that support his position.