Jim McCulloch wrote:

> In article , "John M. Wildenthal"
>  wrote:
>
> > Jim McCulloch wrote:
> > >
> > > I figured it would sound about right to you. Inasmuch as Kleck's figures
> > > are wrong by a factor of 30, if we take the NCVS statistics as our guide,
> > > we can assign you the full NRA credulity ranking of 30.  Most of the
> > > readers of tx.guns, at least, will probably score higher than that.
> >
> > But Jim, the NCVS wasn't designed to properly estimate defensive firearm use.
>
> Really? That would be news to the NCVS's statisticians. They need to know
> this. Perhaps you should write and tell them.

I'll do you one better.  I'll list some of the NCVS shortcomings when used as a
tool to estimate DGUs.  Feel free to forward them to whomever you feel
appropriate.  Of course, I am cribbing these shortcoming from Kleck & Gertz.  I'll
also ask again now, and throughout this response - what do you feel are the
methodological deficiencies of the Kleck & Gertz study?


> >The NCVS is a useful tool for several things, but
> > it wasn't designed to estimate the number of defensive firearm uses.
>
> Again, this would be news to the NCVS.

Maybe you should read what the BJS says about the NCVS (from the BJS website):

National Crime Victimization Survey (NCVS) is the Nation's primary source of
information on criminal victimization. Each year, data are obtained from a
nationally representative sample of roughly 45,000 households comprising more than
94,000 persons on the frequency, characteristics and consequences of criminal
victimization in the United States. The survey fully reports the likelihood of
victimization by rape, sexual assault, robbery, assault, theft, household burglary,
and motor vehicle theft for the population as a whole as well as for segments of
the population such as women, the elderly, members of various racial groups, city
dwellers, or other groups. The NCVS provides the largest national forum for victims
to describe the impact of crime and characteristics of violent offenders.

We could also look at the NACJD description:

The National Crime Victimization Survey (NCVS) series was designed to achieve three
primary objectives: to develop detailed information about the victims and
consequences of crime, to estimate the number and types of crimes not reported to
police, and to provide uniform measures of selected types of crime. All persons in
the United States 12 years of age and older were interviewed in each household
sampled. Each respondent was asked a series of screen questions to determine if he
or she was victimized during the six-month period preceding the first day of the
month of the interview. Screen questions cover the following types of crimes,
including attempts: rape, robbery, assault, burglary, larceny, and motor vehicle
theft. The data include type of crime; severity of the crime; injuries or losses;
time and place of occurrence; medical expenses incurred; number, age, race, and sex
of offender(s); and relationship of offender(s) to the victim (stranger, casual
acquaintance, relative, etc.). Demographic information on household members
includes age, sex, race, education, employment, median family income, marital
status, and military history.  A stratified multistage cluster sample technique was
employed, with the person-level files consisting of a full sample of victims and a
10 percent sample of nonvictims for up to four incidents. The NCVS data are
organized by collection quarter, and six quarters comprise an annual file. For
example, for a 1979 file, the four quarters of 1979 are included as well as the
first two quarters of 1980.

Jim, where do you see DGUs in the above descriptions?  They are not the primary
focus of the NCVS.  Decisions improving the accuracy of the primary focus could
reduce the accuracy of other estimates.

The most common complaint against the NCVS as a measure of DGUs is that if an
incident is not a crime by the definition of the NCVS survey, a DGU for that
incident cannot be listed.  If you warned or fended off an attacker without any
loss or injury, would you consider yourself a victim?  I wouldn't always,
particularly if no explicit crime occurred.  The NCVS only surveys DGU by
individuals who answer "yes" to having an enumerated crime against them.  This
allows for standardization and increased accuracy in classifying the various
criminal activities.  But if an individual is overwhelmingly successful in
defending themselves, possibly preemptively, they might not consider themselves a
victim of any of the enumerated crimes.

Let's think about a few other instances where the NCVS isn't going to even ask
whether a DGU occurred:

Some spousal abuse - if the abused spouse does not consider the abuse a "crime,"
question 42 notwithstanding.

DGU by a felon - since the NCVS does not allow anonymity, individuals might not
report a DGU since they might fear that accurately reporting their use of a firearm
could be illegal and lead to subsequent prosecution.  Or the incident where they
had a DGU might have been illegal - maybe a drug buy turned heist - so they don't
report the crime in the first place.  Since criminals frequently target other
criminals, this undercount could be significant.

These are problems in the NCVS methodology *if* you are interested in the actual
number of DGUs.  These problems result in an undercount of DGUs.


> Not only do Kleck's figures diverge wildly from the NCVS's figures, but
> they diverge wildly from each other.  Kleck has actually done a number of
> studies of this problem, and his methodology, which Dr. Wolfgang (who I
> would guess is getting along in years) admires so much, gives him vastly
> different results each time.

Actually, it isn't only Dr. Wolfgang.  Kleck publishes in peer-reviewed journals.
In my field, that means that three other professionals with experience in that area
would have looked over the paper, including the original data set, and tried to
find errors.  They would first look for whether the paper had some new insights to
bring to this area of study.  Then they would look for methodological errors.  I
presume that Criminology has similar standards.

I haven't seen Kleck's numbers widely diverge from each other because usually I
look at the ranges he estimates, not the point estimates.  The ranges of his and
most other comparable studies overlap in the 1.5-2 million DGU range.

Again, what are the methodological deficiencies in the Kleck & Gertz paper?  If you
can't point them out and don't like the results, you have nothing constructive to
add to accurately estimating DGUs.


> One of the more amusing improbabilities of Kleck's recent figures, is that
> the logically required body count of criminals killed or wounded by the
> defensive uses he reports exceeds the known number of persons wounded by
> guns from all causes every year.

Ah, the first possible criticism - consistency.  Your appear to suggest that since
some people probably overestimate the effects of their shooting, are they also
lying about whether they used a gun or not?  I can imagine an interviewee making
errors on the effects - they may not have direct knowledge of whether their shots
hit.  But they would have direct knowledge about whether they used a gun or not.


> Leaving aside the question of where the bodies of the killed have been
> hidden, the question of where wounded got treatment is an interesting one
> indeed.  Kleck, I have read, believes they got treatment secretly.

Why do you believe they all went for professional treatment?  A wound through the
arm or leg won't necessarily cause any problems requiring professional medical
attention.

I've asked you to describe the methodological faults in the Kleck & Gertz DGU
study.  You haven't named any, though you do question the consistency of some
estimates.  You only disagree with the results, not the methodology.  Would you
therefore agree that Kleck & Gertz have a superior methodology for determining
DGUs, compared to the NCVS?  Let's discuss the methodologies.  I've already pointed
out two reasons why the NCVS results might be biased downward.  Kleck & Gertz
specifically addresses the NCVS shortcomings.  Or do you feel that biased estimates
are better than unbiased estimates?


John M. Wildenthal
mailto:j-wildenthal@tamu.edu
--
"I don't like the idea that the police department seems bent on keeping a pool of
unarmed victims available for the predations of
the criminal class." - David Mohler, orthopedic surgeon, on being denied a permit
to carry a handgun by the New York City police, _Manahattan, Inc._ magazine, April,
1989