Mr. Kopel is the research director for the Independence Institute
National Review Online. February 27, 2001 10:15 a.m. More by Kopel on public health and gun control.
One of my children's favorite books is The Topsy-Turvies, the story of a likeable but unusual family: They watch TV standing on their heads, eat toast with their toes, wear their pajamas for a stroll outside, comb their hair with a fork, and encourage their children to draw on the walls. And, it seems, the Topsy-Turvies are running the federal government's gun-control policy.
Let's imagine that Mr. Topsy-Turvy were elected president, promising to do something about America's health. Let's also imagine that President Topsy-Turvy, being topsy-turvy, actually tried to keep all his promises. Announcing his health-care reform plan, he might say something like: "For too long, Americans have thought of disease control as a public-health issue. It is time that we all realized that disease is a crime."
"Quite often, disease is spread because people fail to do what they know is right: They don't wash their hands, or they send their child to school when the child has a cold. It's time to crack down on these disease criminals who endanger themselves and other people."
The Topsy-Turvy plan for "treating disease as a crime" would impose criminal penalties, including fines and prison, on people who have communicable diseases. The most severe criminal penalties would be for people whose misconduct inflicts onto others a dangerous disease. Lesser punishments would apply to persons who infect only themselves. A person who caught a cold because he failed to wash his hands often enough would only spend a week in jail. Persons who got sick through no fault of their own would be slapped with administrative fines, but no jail time would be imposed.
To support this "innovative" approach to disease control, the federal government would start doling out millions of dollars for criminal-disease research. Very little of the money would go to people who actually knew anything about disease, such as doctors or epidemiologists. Instead, the federal research money would go to lawyers, police officers, bail bondsmen, and other "researchers" eager to apply their crime-control expertise to disease control.
Now this federally subsidized research would obviously suffer from some quality-control problems. It would not get published very often in serious medical journals, like the New England Journal of Medicine. But it would find its way into traditional crime-control publications, such as police-union magazines, law journals, and the like.
The Topsy-Turvy administration's criminal-disease research would almost never make reference to actual medical research. Indeed, it would recycle half-baked ideas for disease reduction that have already been studied and disproved in the medical journals. Newspapers and the electronic media — which already take a sensational, salacious, and shortsighted approach to the crime issue — would eat up and regurgitate the press releases from the federally funded researchers. So, after a few years, and a few tens of millions of dollars of federal research dollars, most of what the general public hears in a "scientific" way about disease control would come from articles about rubella in Crime Control Digest, and studies of syphilis in the Journal of Penology.
While President Topsy-Turvy was busy helping America understand that "disease is crime," what would the First Lady, Mrs. Topsey-Turvy do? Encourage her children to break all the china in the White House, and then let people sleep in the Lincoln Bedroom without making campaign contributions.
But once the White House was refurbished Topsy-Turvy style, Mrs. Topsy-Turvy might then turn her attention to crime control. "Violence is not just a crime, it's a public-health problem," she might announce, later followed by, "Guns are a virus."
But here, my story changes from fiction to fact. The "violence is…a public-health problem" quote is from former Surgeon General Joceyln Elders. And the second quote comes from Dr. Katherine Christoffel of the American Academy of Pediatrics: "Guns are a virus… They are causing an epidemic of death by gunshot, which should be treated like any epidemic — you get rid of the virus."
And in the topsy-turvy world of modern America, the "guns are a virus" theory of crime control has all the problems that one might expect: shoddy and sometimes fraudulent research, heavily funded by the federal government. (The Centers for Disease Control and Prevention provided funding until Congress forced it to stop; then the National Institute of Justice took over).
Let's take a look at a typical product of the federal factoid factory, the widely publicized claim that owning a gun increases your risk of death by 2.7 times. This claim is based on a 1993 article in the New England Journal of Medicine by Dr. Arthur Kellermann, the king of anti-gun "public-health" research. What Kellermann did was compare people in the areas of Memphis, Seattle, and Cleveland who had been murdered in their homes with people living nearby who had not been murdered. (Murders in the home constitute about one-quarter of all murders.) He reported that the murder victims were more likely to own guns than the other people.
But Kellermann did not show that the gun owners were murdered with their own guns. Only five percent of the victims had even tried to use a gun to resist the murderer.
So even though the gun in the home was rarely implicated in the homicide, we are told that owning a gun greatly increases your chances of being murdered.
One key problem with this study is that it hinges on Kellermann's findings of relatively low gun ownership rates among the "controls" (i.e., people in the same neighborhood who weren't murdered). Yet we know that many gun owners refuse to report their gun ownership to pollsters and inquisitive social scientists. If just 20 per cent (a typical refusal rate) of the controls in the Kellermann article refused to admit to gun ownership, then the whole study falls apart.
Kellermann's answer to this very serious methodological problem is to cite his own research, which ostensibly shows that registered gun owners usually tell the truth when they decide to cooperate with pollsters who ask them if they own guns. But what about the large percentage of gun owners who won't agree to be interviewed by a stranger? And what about unregistered gun owners? People whose gun ownership is private may not feel like talking to a government-paid researcher.
Another problem in the Kellermann study is the issue of cause and effect: Does merely owning a gun actually cause people to be murdered, or are people already at risk of being murdered simply more likely to buy guns?
We gain insight into the cause-and-effect issue by scrutinizing Kellermann's odds ratios. (An odds ratio shows the statistical relationship between two facts; for example, "A person with lung cancer is 6.4 times more likely to be a heavy smoker than a person who does not have lung cancer.") In the Kellermann article, a person who was murdered was said to be 2.7 times more likely to have a gun in the home than a similar person who was not murdered. This means that the gun-risk odds ratio is smaller than the gun-risk odds ratio for renting rather than owning. That is, people who rent are 4.4 times more likely to be murdered than people who own their home. Does this mean that you suddenly become safer when your apartment building goes condo? Of course not. Likewise, are you safer when you get rid of your gun? Not necessarily. Rental status may be an indirect surrogate for a lesser degree of stability in one's life, and that lesser stability might result in a higher risk of being murdered.
The case-control method is a time-tested method of disease research. You compare, for example, 20 people who have lung cancer with 20 similar people who don't. If you find that all 20 of the people with lung cancer smoke at least two packs of cigarettes a day, and none of the people without cancer smoke equally heavily, then the "case-control" study suggests a valuable direction for clinical research about the health effects of smoking.
But when doctors try to apply the case-control method to non-health research, the results can be very unreliable. For instance, let us use the case-control method to compare 200 students, all of whom scored double 800s (a perfect score) on the Scholastic Aptitude Test, with 200 similar students who did not score as high. We find that the high-scoring students are about three times more likely to own a chess set than the lower-scoring students. Accordingly, we might claim that "persons who own chess sets are three times more likely to score an 800 on the SAT than are persons who do not own chess sets."
I hope that the SAT-chess study would not convince high-school students to run out and buy chess sets in the hopes of improving their SAT results. For we don't know if the students with chess sets actually played much chess. Even if they did play chess a lot, we don't know if chess helped build verbal-analogy skills and other skills which lead to SAT success, or if people who were already smart and likely to do well on the SAT were simply more likely to enjoy chess, or to be given chess sets as gifts.
If we just substitute "being murdered" for "scoring an 800 on the SAT," and substitute "gun" for "chess set," we have the specious study that supposedly proves that owning a gun increases the odds that you'll be murdered in your home.
Just as worrisome, a study that concentrates on murder victims will almost certainly fail to turn up evidence of the protective value of firearms. Yet the Kellermann study purports to offer definitive proof that gun ownership does not yield any protective value.
Unfortunately, Kellermann and the "guns-as-viruses" crowd are expertly adept at producing factoids which the anti-gun media eagerly then disseminate as "scientific" proof of the perils of gun ownership. By the time that serious criminological or legal journals have disproved one public-health factoid about guns and public health, another one is being bandied about by the media.
Physicians and public-health researchers tend to be quite knowledgeable about medicine and public health. Unfortunately, some of them get the idea that they are also experts in other areas — like nuclear war, or gun violence. Thus, in the 1980s, the group Physicians for Social Responsibility (PSR) promoted unilateral American nuclear disarmament as the proper scientific response to nuclear danger. Philosophical pacifism was masquerading as science. PSR's Dr. Helen Caldicott predicted in 1984 that nuclear war with the USSR was a mathematical certainty if Ronald Reagan were re-elected.
Dr. Caldicott's scientific conclusions notwithstanding, the world survived Mr. Reagan's re-election. As the threat of nuclear war receded, PSR turned its attention away from disarming the American government, and toward disarming the American people. Like Physicians for Social Responsibility, which found no moral distinction between a nuclear weapon owned by the American government and one owned by a Communist dictatorship, the public-health "experts" consider a firearm owned for protection to be as illegitimate as one owned for the purposes of crime. Wrapping its spurious claims in the cloak of "science," the socialist wing of the medical community cheerily champions the disarmament of law-abiding, mentally healthy Americans.
History showed that the core of the nuclear problem was a political one. When the political system in USSR changed, the threat of nuclear war abated. The way to eliminate nuclear terror was to remove the terrorists who were running the Soviet Union. And the way to eliminate criminal violence is to remove the criminals who are terrorizing the streets and homes of America. But in the topsy-turvy world of anti-gun health research — funded by your tax dollars — law-abiding gun owners are considered the essence of the problem, rather than part of the solution.