John R. Lott and David B. Mustard conclude that right-to-carry laws de
ter violent crime. Our reanalysis of Lott and Mustard's data provides
no basis for drawing confident conclusions about the impact of right-t
o-carry laws on violent crime. We document that their results are high
ly sensitive to small changes in their model and sample. Without Flori
da in the sample, there is no detectable impact of right-to-carry laws
on the rate of murder and rape, the two crimes that by the calculatio
ns of Lott and Mustard account for 80 percent of the social benefit of
right-to-carry laws. A more general model based on year-to-year diffe
rences yields no evidence of significant impact for any type of violen
t crime, As a result, inference based on the Lott and Mustard model is
inappropriate, and their results cannot be used responsibly to formul
ate public policy.