Here is Kleck on his study....
Klecks defense of his study
http://www.rkba.org/research/kleck/md-rebuttal.3sep95
Vernick refers to "a relatively small sample size" used in
my research, noting that "about 5,000 respondents" were
interviewed. This was substantially correct (it was 4,977), but
this is in fact an unusually large sample for survey research.
Most national surveys have samples in the 600-1600 range. The
number of persons who reported a DGU is not "the sample size."
Rather, the sample size is the number of persons who were asked
the DGU question, i.e. 4,977. It is this number which influences
the precision of the estimates, not the number who answer "Yes"
to the DGU question. In any case, Vernick's guess that only 50
people reported a DGU is incorrect. A total of 194 persons
(weighted; 213 unweighted cases) reported a DGU involving either
themselves or someone else in their household, 165 reported a DGU
in which they had personally participated in the previous five
years, and 66 reported a personal DGU in the past one year
preceding the survey (see Table 2, p. 54 of the report).
Vernick speculates that some substantial number of survey
respondents who reported a defensive gun use (DGU) were actually
describing "distant-in-time events" and that this resulted in
enormous overstatement of the frequency of DGUs. This problem,
known as "telescoping," does occur but in surveys of this type
its effects are cancelled out by problem~ in the opposite
direction (i.e. problems tending to make estimates of DGU
frequency too small) of respondents forgetting DGU events which
really did occur in the period that was asked about. In any case,
effects of telescoping are far too weak to account for the
results we obtained. These issues are discussed on pp. 34-35 of
the report.
Vernick speculates that respondents "may have not understood
what would qualify as a 'defensive use' of a firearm - perhaps
including events where the gun was carried for 'self- defense'
but never actually displayed in response to a specific threat"
(my emphasis). In addition to the highly conjectural nature of
these remarks, they are also wrong. Contrary to Vernick's rather
elitist assumption that members of the general public are too
stupid to know the simple distinction between merely carrying a
gun for protection and actually using it for self-defense, none
of the respondents who initially answered "yes" to our DGU
question were describing instances of merely carrying guns for
protection.
In any case, our estimates of DGU frequency were based
solely on cases that qualified as bona fide DGUs. Two of the
conditions needed for incidents to qualify as genuine DGUs were
that (1) there had to have been an actual confrontation between
the defender and an adversary, and (2) the defender had to have
actually used the gun in some way, some as pointing it at their
adversary in a threatening manner, or using it in a verbal threat
(e.g. 'Stop, I've got a gun.") None of the cases that went into
our estimation of 2.5 million annual DGUs involved person who
merely owned or carried a gun for protection.
Vernick hints that this estimate somehow must be unreliable
because "prior work by Kleck using similar methodology" yielded
the very different estimate of 1 million. I have not done any
"prior work" using "similar methodology." In past publications I
have merely noted the number of annual DGUs that are implied by
the results of surveys previously done by other people, including
the 1 million estimate. The Spring, 1993 National Self-Defense
Survey is the only survey I have conducted on this topic.
Indeed this is the only survey ever designed by anyone
specifically to estimate the frequency of DGU. Given the
technical flaws of prior surveys yielding DGU estimates, there is
no reason why my survey should have yielded the same, presumably
erroneous, estimates as previous surveys. Indeed, there would be
something seriously wrong if, despite my considerable efforts to
improve the methodology, I just got the same results as the
previous, seriously flawed surveys yielded.
In this connection, Vernick misleads by omission, failing to
inform the Commission just how common surveys yielding large DGU
estimates are. To date, there have been at least 14 surveys
implying anywhere from 700,000 to 3.6 million DGUs per year (see
Table 1 of enclosed report). For Vernick to hint that my estimate
was an isolated fluke rather than a common result is more than a
little deceptive. That there are many other surveys implying
frequency DGUs is common knowledge among scholars who study this
subject, as it has been reported in both previous published
articles (e.g. Social Problems, volume 35, p. 3, February, 1988)
and in my book, Point Blank: Guns and Violence in America (p.
146), winner of the 1993 Hindelang Award, granted by the American
Society of Criminology to the most outstanding book of the
preceding several years. These are hardly obscure information
sources to serious scholars, and no competent student of the
subject could claim to be unaware of these numerous surveys.
