This article is from the talk.politics.guns Official Pro-Gun FAQ, by Ken Barnes (kebarnes@cc.memphis.edu) with numerous contributions by others.
McDowall, David; Wiersema, Brian; and Loftin, Colin; "Easing
Concealed Firearms Laws: Effects on Homicide in Three States,"
J. of Criminal Law and Criminology v.86, n.1, pp.195-204 (1995)
Polsby, Daniel; "Firearms Costs, Firearms Benefits, and the
Limits of Knowledge," J. of Criminal Law and Criminology v.86,
n.1, pp.207-220 (1995)
In summary: A recently published study by the University of
Maryland's Violence Research Group, claimed that average monthly
homicides by gun increased in four of five urban areas studied,
after the adoption of liberalized concealed carry reform laws.
This statistical "factoid" was first offered up March 13, 1995
by the Associated "Black Rhino" Press, and was widely reported
(and misreported) elsewhere. However, the researchers do not
attempt to argue that_any_of these gun-related homicides were
committed by concealed carry permit holders, nor do they consider
the relative proportion of concealed carry permit holders in each
of the studied localities, which if one were to assume that less
restrictive concealed carry is associated with increased homicide,
would be a particularly important factor to consider. Information
that is available from other sources, such as the state government
in Florida, suggests that criminality by concealed carry permit
holders is virtually unknown. (See 3.8, and Appendix II.) If
increased carry by citizens with CCW permits is responsible for
any increased homicide, the researchers do not establish a plausible
explanation for the change in rate, if such a change actually exists.
Also the researchers selected the cities they did on the basis of
their status as "large urban areas," rather than looking at the
homicide rates for each state as a whole, when the concealed-carry
laws were enacted statewide.
Three of the five localities studied were in Florida (where
a statewide CCW reform law was enacted in 1987, and where "gun
control" advocates predicted increased violent crime would make
it into the "Gunshine State"). Both in Jacksonville, where the study
claims that the monthly average number of gun-related homicides
increased by 74 percent, and in Tampa, where the study claims the
monthly average number of gun-related homicides jumped 22 percent,
the study notes that the rates of homicide_without_firearms also
increased. In Miami, a city of intermediate size relative to
Tampa and Jacksonville, the study claims gun-related homicides
per month increased 3 percent, but this analysis is based upon a
different time span than that for Tampa and Jacksonville. The
researchers claim that using a different baseline was necessary
"because of an unusually sharp increase in homicide rates [in Miami]
in May 1980 after an influx of Cuban refugees." Results of time
series analyses such as this, and the similar study of the effects
of Washington D.C.'s gun registration law by the same researchers
(see 3.0.b., and Appendix IV), are highly dependent upon the time
spans chosen for study. In those Florida cities for which the
most dramatic results were obtained, the baseline used stretched
all the way back to 1973, but for Miami, the baseline began in 1983.
Presumably, given the higher initial Miami homicide rates that the
researchers attribute to the Cuban refugees, starting off with
same baseline year for Miami as for the other Florida cities would
have produced a resulting increase_less_than the 3 percent observed.
Complicating any analysis of Florida's crime rates is the fact that
almost immediately following the passage of Florida's 1987 CCW reforms,
the state changed the manner in which it collects crime statistics,
so comparisons before and after the implementation of the law_can_be
invalid on that basis. However, the number of homicides ought not
to be affected by that, since to have a homicide, you need to have
a body. Statistical changes ought not to affect the ability to count
bodies, which is essentially what the Maryland study does (albeit
very crudely, and without distinguishing between justifiable homicides
and murders). The researchers cite FBI's Uniform Crime Reports data
in asserting that justifiable homicides are uncommon, but the FBI
numbers actually undercount the number of justifiable homicides by
civilians, since FBI statistics rely upon the initial determination
made about an incident (see 3.8). A statewide analysis of the Florida
data (see Appendix II.), reveals that the U. Maryland researchers
missed the overall downward trend in murder/manslaughter rates since
the Florida concealed-carry law was enacted.
The question of the direction of causality's arrow is critically
important to consider. Does increased homicide lead to more people
obtaining permits to carry, or does increasing the availability of
permits to carry increase the homicide rate? The Maryland study's
researchers seem to want to argue the latter, but they have thus far
offered no evidence that CCW permit holders are doing the killing!
The sampling of these particular localities, since nonrandom, can
also be used to introduce bias into such a study, and sociological
differences between localities also need to be controlled for.
The wide disparities in the with-gun homicide rates given in
the study seem very unusual at first glance, and this is the result
of the fact that the study is calculating its percent increases in
absolute terms (4 going up to 7) rather than per 100,000 population,
as is necessary for any realistic assessment. The attempt by the
researchers to control for population changes by lumping together
data from all five studied cities is a less than adequate solution
if they intend to consider the homicide rates seperately for each
city elsewhere in their report (and in their quotes to the press).
The researchers themselves admit that "[t]he population of all five
areas grew over the study period, especially in the Florida cities.
Homicide counts thus may have changed after the laws in part because
of increases in the populations at risk."
The two other localities examined were Jackson, Mississippi,
where the average monthly gun-related homicide rate increased by 43
percent; and Portland, Oregon, where the average monthly gun-related
homicide rate_fell_by 12 percent. The Portland data, the researchers
note, did not provide enough homicides, so data from two adjoining
counties were included as well. (Again, one presumes that if this
had not been done, the Portland rate would have been seen to decline
even more than the 12 percent reported.) Calculating average numbers
of homicides per month serves to make the numbers smaller, and if
there are few homicides in a locality to begin with, as was apparently
the case in Portland, the degree of "noise" in the data can produce
some "startling" percentage changes, if considered in absolute terms.
(If one average has 6 homicides, and the next, 3, that's a "50% DROP!")
It's also widely known in criminology that more violent crime occurs
in summer months than in winter months, so averaging over the year,
as a "moving average" technique does, can also serve to produce
smaller numbers, more suceptible to "startling" percentage changes.
This study design (and the similar study of Washington D.C.'s gun
registration law) seems geared towards maximizing "noise," rather
than discerning a longer-term trend.
The researchers' conclusion is tentative: "While advocates of
these relaxed [carry] laws argue that they will prevent crime, and
suggest that they have reduced homicides in areas that adopted them,
we strongly suggest caution. When states weaken limits on concealed
weapons, they may be giving up a simple and effective method of
preventing firearm deaths." This quotation also points up another
aspect of bias in the study. What exactly is significant about_gun"
related homicides, versus total homicides? Does the fact that
a homicide is committed with a firearm make the slain any_more
dead than if the homicide was committed with a knife, or with hands
and feet? The researchers claim that homicides by other means
remained mostly steady, yet in the two Florida cities showing the
greatest increases, both gun and non-gun related homicides went up.
If the increased presence of firearms in the hands of the law-abiding
serves to deter murders with other weapons, this effect would be
discounted when only homicides involving firearms are considered.
In the case of Florida, since the murder rates have_decreased_when
considered on a statewide basis, these supposed increases in urban
areas must have been offset by drastic drops elsewhere.
The researchers do not consider Florida data later than 1991,
since they say other laws, such as the 1991 passage of waiting
periods and background checks on new gun purchases in Florida, would
make it difficult to disentangle the results of the new laws from
the concealed carry law. Yet presuming that the declines in
murder rates since 1991 are attributable to background checks and
waiting periods runs counter to the assumption that gun availability
is a crucial factor driving murder rates. Neither background checks
nor waiting periods would have any appreciable effect on the existing
supply of guns available, while concealed carry opponents claim that
CCW laws make guns more available in the heat of an argument. To
compare the potential effects of these laws on the availability of
guns is to compare apples and oranges. Concealed carry can be the
ultimate test of the "gun availability" hypothesis. So far, the
awful consequences that are habitually predicted by those opposed to
concealed carry reform laws haven't been shown to occur. If anything,
the available data is consistent with the view that concealed carry
has a deterrent effect, or at the least has no direct negative effect.
The researchers_attempt to_speculate that increased availability
of legal concealed weapons is leading to an "arms race" between
criminals and their potential victims, and more criminals are
using firearms than previously. They write "[g]reater tolerance
for legal carrying may increase levels of illegal carrying as well.
For example, criminals have more reason to carry firearms --and to
use them-- when their victims might be armed." As Daniel Polsby
points out in the same journal, this "arms race" hypothesis would
presumably also operate when greater numbers of armed police officers
are introduced into a locality. But the legitimacy of availability
of concealed carry permits to the law abiding is not predicated on
the frequency of misbehavior of criminals-- except in the minds
of "gun control" advocates who wish to prohibit any item that could
potentially be misused by criminals (chemical defense sprays and
stun guns included), regardless of its effectiveness in protecting
the weak from the predations of the strong. (See 1.1.a) Theirs
is a policy which demands that victims "lie back and enjoy it,"
rather than_fight_back, and reduce their risk of injury or death.
 
Continue to: