Note that even if you are in love with Kleck’s estimate for DGUs, you
can’t honestly compare it with the NCVS estimate for gun crimes, since
it is not possible for both Kleck’s estimate for DGUs and the NCVS
estimate for gun crimes to be correct.

Dr. Paul H. Blackman writes:

Why not? NCVS’s purpose is to measure crimes, not defensive gun
uses. Why can’t one think the NCVS does a pretty good job measuring
what it’s trying to measure, and Kleck-Gertz did a pretty good job
measuring what they were trying to measure? Kleck did, of course;

Kleck does not seem to have noticed that his “generous” NCVS derived
estimate of 550,000 gun crimes is inconsistent with his survey’s
finding of 450,000 DGUs against gun crimes. It is absurd to suppose
that more than 90% of gun criminals encounter armed victims.

It is possible that the NCVS measures crimes accurately. It is
possible that Kleck’s survey measures DGUs accurately. It is not
possible for both to be accurate.

with the caveat that NCVS gun-related crimes will include things
perceived as gun-related crimes which were not. (For example,
Charlie Orasin, who founded what is now Handgun Control, Inc., did
so following his being the victim of a robbery where the robber made
reference to a gun but Charlie never saw it, and doesn’t know if
there really was one. If asked by NCVS, that would have been a
gun-related crime, as would others at zucchini- or banana-point,
etc.)

If you want to exclude those, then you shouldn’t count DGUs where the
criminal didn’t see the gun. (25% of the total in Kleck’s survey.)