On Sun, 3 Jan 1999 13:59:23 -0600, "Paul Barnett" 
wrote:

>Nick Hull wrote in message ...

>However, he does believe that the 8.3% wounding rate (of all defensive gun
>uses) is probably too high, especially in comparison to the 15.6% that
>reporting shooting at their attacker.  That's a 53% hit rate, which is
>triple the 18% of criminals shooting at victims, and even exceeds the 37%
>hit rate of police officers shooting at perpetrators.
>
>While noting that the actual instances of wounding the attacker numbered 17
>out of the entire sample of 222 DGUs (from about 5,000 total samples), Kleck
>appears to think the real issue is that they didn't follow up and ask how
>the gun owner knew they had wounded their attacker. 

The problem of the missing bodies is so obvious a false positive, that
Kleck preemptively subjects this particular subset of bizarre results
to a caveat quite appropriate to his entire bizarre results.  There is
no logical reason not to demand of his entire results a follow-up in
which harder evidence is required of people making boastful claims.

How about the astonishing number of rapes resisted with firearms?  I
would have to look the figure up, but as I recall it Kleck is obliged
to claim an implausibly huge percentage of attempted rapes in this
country were resisted with guns.  There is an independent account of
the number of attempted rapes, you know, provided courtesy of the
NCVS.  The NSPOF survey (which basically used Kleck's methodology),
elicited results in which the number of rapes resisted with guns was
*greater* than the total number of rapes and attempted rapes reported
by the NCVS.  

The reason why many sociologists and criminologists do not control
adequately for the false positive problem, (a problem which is
obviously the source of such outlandish, and in some cases,
impossible, results), is that traditionally sociologists have not
attempted to find hard numbers for very rare events, like DGUs.
Medical scientists have for years recognized this problem, and built
appropriate controls into their research. The NCVS also has built-in
controls which at least reduce this problem, in that it asks only
victims of a crime in which they claim to have had direct contact with
the perp, if they had defended themselves.  

I will add here, that the moronic attack on the NCVS (made by several
people in this thread)  that it ignores defensive gun use before such
events became crimes, is untrue.  An attempted burglary is a crime.
The NCVS does not, however, solicit fantasy speculation about the
intentions of real or imagined perps before the real or imagined perps
make a move that can be independently construed as criminal.  It is
quite true that if I pointed a gun at everyone who came near my house,
that I *might* by doing so prevent a crime.  If I claim that *each
such instance* is a prevention of a crime, however, I am stretching
things a bit.  

Kleck makes the incorrect claim that false negatives are as likely as
false positives.  This is untrue for two reasons, either sufficient.

The first is an arithmetical reason. If 0.5 percent of respondents
actually experience a defensive gun use, only 5 in a thousand
respondents can create a false negative, by failing to report it.
Nine hundred and ninety five out of a thousand, however, have an
opportunity to create a false positive with a false claim of a DGU.
This is obvious to anyone.  
The second is that there is far more of a reason, in a society like
our own, for any one of the 995 to lie about gun use (machismo,
boastfulness, desire to impress, outrage over crime plus
wish-fulfilment, among others), as opposed to the 5 having a reason to
lie about real gun use (paranoia--which will not exist among Kleck's
respondents because the truly paranoid will have deduced that Kleck is
an agent of the BATF and refused to answer any questions.)

And of course none of this addresses the additional question of
whether even those DGUs truthfully reported represent an aggregate
social good.  *That* problem springs from the tendency of people in
general to paint themselves as the good guy.  Ergo, anyone who behaves
in a threatening manner, with a gun, to a neighbor in a barking-dog
dispute, can be relied upon to report this as a DGU if by some chance
he is surveyed by Kleck.  Kleck has no way of knowing who is the good
guy in such a claimed DGU, assuming the supposed instance  is not
simply a fantasy, which in most cases it obviously is.

Best regards,

--Jim McCulloch