Lott's new book has been published. This means that I can disclose the results of his 2002 survey. In that survey, 7 people said that they had used a gun for self defence. Of those 7 people, only one reported firing the gun (in fact, that person reported wounding the attacker). This means that a 95% confidence interval for the percentage firing is 3%–50% (calculated using Wilson's method). This confirms what I suggested earlier---the number of defensive gun users in Lott's 2002 survey is far too small for this survey to give any useful information about the percentage of defenders who fire their weapons. This is what Lott wrote about his new survey:

"let me note the most important bottom line: the survey that was done last fall produced very similar results. The earlier results were replicated."

Given the size of the sample of defensive gun users, this is an extremely misleading statement. Either Lott does not understand the basic concept of statistical significance, or he knowingly made a false statement about replicating the earlier results. It is extremely misleading for Lott to say that this new survey "replicates" the survey Lott claims to have conducted in 1997.

I estimated the firing percentage above using the unweighted sample. Weighting the sample by demographics, household size and/or frequency of defensive gun use would make the effective sample size smaller and the confidence interval even larger. For example, of the 7 defensive gun users, one reported three uses, and four reported two uses. If these frequencies are used to weight the results, then the effective sample size is reduced to 6.3 and the 95% confidence interval for the percentage firing is 1%–46%. I haven't worked out the confidence interval if demographic weighting is used, but it should be similar. It isn't possible to weight by household size, since Lott failed to collect that information.