New paper: "Trust But Verify: Lessons for the Empirical Evaluation of Law and Policy"

I hope that a lot of people download this paper (available here). Here is the abstract of the paper:

In a recent article, Aneja, Donohue and Zhang claim that they are unable to replicate the regressions published by the National Research Council in Chapter 6 of Firearms and Violence. They conclude that the NRC regressions must have been based on bad data supplied by John Lott. The implication is that earlier studies that found that right-to-carry laws reduced crime were flawed because of bad data. However, we can replicate the NRC results with Lott’s original data and with the data set used by the NRC. The earlier studies are not flawed by bad data.