RETURN TO OWNERSHIP HOME HAZARDS PAGE
The simple "study" from which these claims arose was performed by Arthur Kellermann and Donald Reay, two medical doctors. They published the report on their study in the June 12, 1986 New England Journal of Medicine (NEJM), which has come now to be always ready to publish anything anti-gun. In the study they used data from medical examiner (Dr. Reay) and police records of the gun deaths over a six year period from 1978 through 1983 in Kings County, Washington (Seattle area). They also interviewed investigating officers in some cases as necessary to obtain details that were left out of the medical examiner and police reports.
According to their NEJM abstract, their objectives were "to obtain specific information about the circumstances, the scene of the incident, the type of firearm involved, and the relationship of the suspect to the victim." They did not describe the extent to which they determined this relationshipwhether they went so far as to find out if the victim and killer were married, family, friends, or just known to each other.
Per the abstract and their NEJM article: A total of 743 firearm-related deaths occurred during this six-year period, 398 of which (54 percent) supposedly occurred in the residence where the firearm was kept. Only 2 of these 398 deaths (0.5 percent) involved an intruder (actually burglar) shot during attempted entry. The police and Dr. Reay were pretty certain that the purposeful killings of seven other persons (1.8 percent) were also self-defense (justifiable homicide), for a total of nine.
There were 333 deaths called suicide, 12 that were called accidental deaths, the nine that were called (by the police and Dr. Reay) justifiable homicide, 41 that were called nonjustifiable homicide (again, by the police and Reay), and three that Dr. Reay could not decide whether they were suicide or accidental. Apparently, there were 11 for which characterization could not be determined (since the other numbers add up to 389, not 398. For each of the nine cases of police-characterized self-defense homicide involving a firearm supposedly "kept in the home," there were (on average) 1.33-1.67 accidental deaths, 4.56 police-characterized "criminal" homicides, and 37.33 suicides "involving" (by?) firearms.
|Simplified: In and around Seattle, over a six year period around 1980, there were 43 times as many "bad" gun deaths than "OK" gun deaths in the homes where the guns were kept. ["bad" deaths being: suicides, accidents & murders]|
Other than the fact that the study turns out to be nearly worthless except as a vehicle for gun control advocates to make false and misleading claims, it appears from the abstract that there was little if anything wrong with the study itself. They didn't say how they determined the gun used in the killing was "kept" in the home, so it is not possible to evaluate how well they made this determination or if they just made an unfounded assumption. But the misuse and mistatements of the findings is where gun control advocates, including Kellermann, have perpetrated a huge lie.
The first thing that should be kept in mind is that the study started with a bunch of gun deaths in homes and, hence, households in which people were killed by someone using a gun. The fact that households where people are killed tend strongly to be nontypical (even aberrant, disfunctional) households, and their inhabitants tend to be abnormal and disfunctional in numerous ways, means that the results are not applicable to people or homes in general or to all households in which there are firearms.
It is not logical to conclude from the study that, on average, households with guns or households in general will experience the kind of ratio the study found.
If one could assume that the results would apply nationwide, which is not a valid assumption, the significance would be only that, in those homes with a gun death, the ratio of gun deaths classified as murders, suicides and accidents by police and one specific medical examiner to those classified as justifiable homicide by those same people would be around 43. National statistics should yield the same ratios, if the police and medical examiner were not biased in the Kellermann-Reay study, which is why the study was worthless.
The second thing to keep in mind is that, because the study involved only one county's gun deaths and only over one time span, the results cannot be expected to be correct everywhereor even a majority of areas of similar population densityand at all times. The study has not been replicated by anyone in other places or times, so there is no telling how general or universal their results are. Science normally involves researchers repeating experiments, possibly with some refinements and corrections, or doing completely different experiments to verify the validity of original experiments.
The third thing to keep in mind is that the study sample size was pretty small (398, or 389), and there were multiple measures (murder, accident, suicide, etc.) so the sampling errors of the study would be somewhat large (especially for the murder, accident, and self-defense categories).
The fourth thing to keep in mind about the study is that the data the doctors were working with were classifications by a police department, not classifications reflecting the true circumstances or even circumstances found in court. Some of the cases probably had not been finally adjudicated since it often takes about five years to do so.
In a very high portion of cases in which a killing is justifiable homicide according to a later court finding, police initially report the case as being one of murder. Some initial adjudications are also overturned. This means that the classifications of killings would over report murders and under report justifiable homicides (self defense).
Because the proportion of justifiable homicides according to the study was quite low, a small increase in that proportion by reducing the proportion of murders would result in a large change in the ratio. Changing 4 (10%) of the murders to justifiable homicide would change the 43:1 ratio to 30:1.
Keeping those four things in mind, let's examine some of the common misrepresentation of the facts about this study.
RETURN TO TOP OF PAGE