Peter Boucher writes:

Tim wrote that he, at first, agreed with the Kleck DGU estimates, but has
since been convinced by the evidence that they were wrong.

Tim, I’ve known you (well, sort of) for over 5 years, and I’ve never
seen you post anything that indicated that you agreed with Kleck’s DGU
estimates. Did you change your mind more than five years ago? What
was the evidence that forced you to change it?

I first encountered one of Kleck’s estimates in his paper published in
“Social Problems”. This was in 1989, soon after I first started
posting to talk.politics.guns. In this paper he advances an estimate
(derived from the Hart poll) of 650,000 DGUs per year. He also
used NCS data which indicated about 400,000 DGUs to argue that guns
were an effective means of self defence. He explained the discrepency
as being caused by NCS undercounts of commercial robbery and domestic
assaults. This all seemed reasonable at the time, so I accepted
Kleck’s estimate as valid.

Later, I reread Kleck’s paper and noticed that the 400,000 NCS DGUs
were over a seven year period i.e. only about 60,000 per year.
Kleck’s explanation of the discrepency could not explain such a large
gap. The most reasonable explanation was that Hart and the NCS were
measuring two different things — the NCS measuring those occuring in
the context of a well defined crime, while Hart also included those in
more ambiguous situations.

Finally, Kleck’s 1993 survey convinced me that the discrepency was
caused by a small percentage of people making things up.