Myth: Gun ownership is not the cause of America's high murder rate.

Fact: Gun availability is correlated with murder; gun controls laws see the murder rate fall.


Murder can be broken down into three components: desire, ability and feasibility. A society's aggregate desire to commit murder depends on social factors, but its ability and feasibility to commit murder are heightened by widespread gun availability. Indeed, most studies find a clear correlation between the murder rate and gun ownership, especially handgun ownership. Most gun owners claim they buy guns for protection, but it appears that the problem (murder) and the solution (gun deterrence) are one and the same: 70 percent of all murders are committed with guns. We should also consider a failure any "deterrence" that is correlated to the very crime it is supposed to deter. But most damagingly, the FBI deems only 1 percent of all murders to be "justifiable homicides" using a firearm. Statistics from the nation's largest crime survey also show that a gun in 19 times more likely to be used in nonfatal crime than in nonfatal self-defense. Pro-gun advocates respond by trying to refute these statistics, citing a study that shows that defensive uses of guns outnumber their criminal uses. However, the survey they cite is not credible. Even if these dubious statistics were true, one cannot praise a social pathology (i.e., gun violence) that only partially cures itself (i.e., through gun deterrence).


This essay is divided into three parts: 1) a pro-gun control philosophy on the relationship between gun ownership and the murder rate; 2) a review of the most relevant statistics; and 3) an analysis of the statistics to see if they support this philosophy.


Every act of murder can be broken down into three components: desire, ability and feasibility. Without any one of these components, an individual cannot commit murder. Let's look at each one:


To be sure, not every person in society has a desire to commit murder. Murderous impulses occur only to a very small percentage of the population, to those who are sufficiently jealous, angry, drunk, drugged, insane, irrational, pathological, self-destructive or deceived enough to seriously contemplate killing someone. Some people may be only temporarily afflicted; others much more permanently. The more permanently ones we see as repeat offenders in our criminal justice system. How many people entertain an urge to kill in society varies; perhaps one country sees only 0.5 percent of its population in this state of mind over the course of a year, while another country sees only .001 percent. The difference can be attributed to social factors, like the availability of mental health treatment, substance-abuse programs, family counseling, poverty, media violence, racial tensions and hostilities, or any of countless other imaginable factors that contribute to the murderous impulse.

Some social factors appear to have enormous impact on violent crime. Two social factors in particular have been getting increased attention from researchers lately; the first is media violence. Many sociologists do not consider it an accident that the crime wave that hit America in the 60s and 70s coincided with the first television generation coming of age. Dr. Brandon Centerwall has produced one of the most famous studies, which found that the mere introduction of television into a region causes its crime rate to double as soon as the first television generation comes of age. (1) In a 22-year study of 800 children from grade 2 to early adulthood, Leonard Eron and Rowell Huesmann found that the best predictor of later aggression was a heavy childhood diet of TV violence -- more so than poverty, grades, a single parent in the home or exposure to real violence. (2)

The second is income inequality. Although absolute poverty levels do not correlate too significantly with the crime rate, income inequality does (oddly enough). Two separate studies, one from Harvard, the other from Berkeley, compared state crime rates to their income inequality rates, and found that the states with the most inequality had the highest rates of homicide, violent crime and incarceration. This correlation holds internationally as well; Europe has much lower levels of inequality than the U.S., and much lower violent crime rates as well. In the U.S., the rising murder rate has accompanied a rising level of income inequality. In 1968, the Gini index of income inequality was a record low .348; by 1994, it had risen to .426, the highest level since the Great Depression. (3)

Unfortunately, the social factors that contribute to a society's overall desire to commit murder do throw a wild card into this debate. Still, there is enough data to reach certain conclusions about the link between guns and murder.


The second component of murder is ability. Killing a human being is a surprisingly difficult task, and that means that weapons with higher ability will kill in a greater percentage of attempts. Among common weapons, guns are unmatched for their killing ability, and this efficiency can be seen in attempted suicide statistics. On a national level, the statistics on suicide attempts and methods are sketchy, but there have been a number of more reliable smaller studies. The following one from Dallas yields a typical result. It showed that suicide attempts with a gun are fatal 76 percent of the time, compared to 4 percent when other methods are used. (4) The American Association of Suicidology gives an even higher estimate: 92 percent of all suicide attempts with a gun are fatal.

Gun researcher Josh Sugarmann writes:

Pro-gun criminologist Gary Kleck has argued that the higher fatality rate of gun suicide attempts "could be due to greater seriousness of intent among gun users. There is evidence that suicide attempters who use more lethal methods are more intent on killing themselves, rather than merely making an attempt as a 'cry for help' to those around them." (6) But this observation is irrelevant, because those with more serious intent are presumably attracted to guns for their greater killing ability in the first place.

The heightened ability of guns has important implications for murder. When people experience a murderous impulse, they may attack no matter what the situation, and with whatever weapon is handy. Although they may attack with the same degree of ferocity and blind passion, a knife attack will probably result in injury, a gun attack in death. Thus, enhanced ability alone will drive up the murder rate. It follows that if a gun is lying around the house waiting for the next violent family argument to happen, the chances for tragedy are greater.


The third component of murder is feasibility. A person might have both the desire and the weapon needed to kill, but if the circumstances don't offer a feasible opportunity to carry it out, then a person will probably decide against it. With weapons like knives or clubs, a would-be murderer faces enormous personal risk. Here's why:

A gun, however, dramatically reduces all these risks. In order of the above, guns allow people to:

What guns do, then, is make it more feasible for a would-be killer to act out his murderous impulses. Gun possession thus allows a crime to occur that wouldn't have otherwise. A good analogy is robbery. In medieval times, wealth was usually stolen only when it was in transit, by highway brigands who outnumbered the drivers. But in modern times, a lone individual with a gun can walk into a bank and rob the entire establishment. Bank robbery became a widespread phenomenon only after the invention of guns.

Certain types of guns enhance feasibility more than others. Long-barreled guns like rifles and shotguns are difficult to hide and bulkier to carry; therefore, easily concealed handguns are the weapon of choice among murderers. One exception to this rule is the long-barreled gun which is super-efficient, namely, machine guns or assault weapons. These enhance feasibility in other ways that might be preferred by a would-be killer. For these reasons, gun-control advocates especially focus on handguns and assault weapons when it comes to regulation.

Let's imagine now two countries, both of which have populations of 250 million. Suppose 50,000 people in each of these countries are going to experience murderous impulses over the course of a year. One nation has a complete ban on guns. The other has universal gun ownership. Which nation will see the higher murder rate?

Common sense would dictate that the nation with guns will realize its enhanced ability and feasibility to commit murder. However, we should never underestimate the gun advocates' powers of rationalization! A common counter-argument is Robert Heinlein's: "An armed society is a polite society." This is hardly true, as the statistics below demonstrate; you could not get a more polite and murder-free society than Japan, which bans virtually all guns, or a more violent society than America, which owns the most guns in the world. But let's treat this counter-argument on a philosophical level.

If we were to arm everyone in society, then the ability to commit murder would become universal. This is a serious step in the wrong direction; to rescue their point, gun advocates must rely even more heavily on arguments of defeasibility. The fear of getting shot back, they argue, will deter most murderers. And there is a degree of truth behind this argument -- police, for example, wear sidearms precisely for their deterrence effect and protective benefits. Surveys of criminals show that they tend to avoid targets they feel might be armed.

But ultimately this argument fails, even in principle. A central tenet of game theory is that attackers have the advantage over defenders. A defender must defend against all possibilities of attack, and in doing so defends none of them very well. An attacker has to choose only one line of attack, and therefore can do it extremely well. Attackers have the advantage of surprise, planning and initiative. An example is a careful, well-considered plan to shoot someone in the back, even if the person is openly carrying a sidearm. Another example is bank robbery. The fact that banks are extremely well-protected hasn't stopped their robbery even today -- criminals simply arm themselves more heavily and take advantage of the fact that they are the attackers. Or, in the face of heavily armed targets, attackers may simply alter their line of attack, selecting weaker targets: the old, the disabled, or children. A useful analogy here is war. The fact that the entire world out-armed Hitler did not stop him from attacking it. And he nearly succeeded -- because, as the attacker, he had the initiative.

Furthermore, even the certain threat of retaliatory force will not stop someone whose senses are impaired by drugs, alcohol, jealousy, anger, insanity, pathology, self-destruction or deception. Although we can identify some groups at risk for these behaviors, we can hardly predict them all -- jealous husbands, for instance. And the "failure to defease" is a tremendously costly one, now that the ability to commit murder is universal.

Thus, many people suggest gun control as a solution to high murder rates. We should note there is a spectrum of gun control, ranging from licensing laws to the total banning of all guns. Comparing different societies for gun availability alone is insufficient, since we must also consider their different gun control laws. We should also note that gun control only reduces the ability and feasibility to commit murder -- it does not limit desire. That's something for other social policies to address. The interplay between these three components is where the debate becomes complex. Desire may sometimes counter ability and feasibility, which only confounds the issue. Suppose two nations have the same population size, although nation A sees 50,000 people a year with the murderous impulse, but only 25,000 in nation B. If nation A has gun control, and nation B has high gun ownership, then nation A might still see a higher murder rate. Gun advocates might then claim that this "proves" their case -- although gun control advocates would still claim that reducing gun ownership would reduce murder rates in both nations. Many people debating these issues often fail to take these considerations into account.

The interplay between desire, ability and feasibility makes for some unique case studies. One example is Israel, whose entire population is armed, and yet has a low murder rate. However, their desire to commit murder is low, because Israel is usually either at war or the threat of war, and criminologists have long known that the crime rate drops during wartime. (One could say that the desire to kill is externalized in the case of war.)

Another example is Switzerland, which also has universal gun ownership by military-aged males, and a low murder rate compared to the U.S. Contrary to the suggestions of pro-gun advocates, however, gun ownership in Switzerland is not universal; only 32 percent of the general population own guns. By comparison, this figure is 49 percent in the U.S. And Switzerland also has much stricter gun control laws. All military weapons (which are long-barreled) must be kept locked up, with their ammunition sealed, stored in a separate place, and strictly accounted for. Hence it is almost impossible to use these weapons for crime without detection. Handguns are also highly regulated. Even then Switzerland has both the highest handgun ownership and highest handgun murder rate in Europe.

Now let's review the statistics, to see how the correlation between gun ownership and murder rates is borne out according to the above philosophy. Although few correlations are ever exact in sociology, the ones below are generally clear (that is, you can see them without mathematically measuring them). To the extent that they vary, the differences can be attributed to other social factors, like gun control laws, income inequality, etc.


According to a 1992 review of the scientific literature, most studies find that gun density is positively associated with the murder rate. (7). The National Institute of Justice, for example, reports a study of U.S. cities which found a positive correlation between gun ownership levels and felony gun use and felony murder. (8)

How about other violent crimes, like rape and assault? The NIJ report says: "Greater gun availability increases the rates of murder and felony gun use, but does not appear to affect general violence levels." In other words, we generally have a constant level of violence in our society, but guns allow a greater portion of that violence to become deadly. "The fact that the United States is a violent society does not have much to do with guns," writes researcher Philip Cook. "The fact that our violent crime is so deadly has much to do with guns." (9) This coheres with the above philosophy that only a certain percentage of the population experiences the impulse to commit murder, and is prevented only by its lack of ability and feasibility.

Here's a closer look at the numbers:

In 1991, there were 211 million privately-owned firearms in the U.S., which then had a population of 252 million people. Of these firearms, about 71 million were handguns.(10) The long-term trend in both handgun production and criminal use has been away from manual revolvers and towards rapid-firing, semi-automatic pistols. (10) The domestic production of pistols has doubled since 1980, while domestic production of rifles has fallen 40 percent, and shotguns 14 percent. In 1980, pistols made up less than 15 percent of total firearm production in the U.S.; by 1993, they had climbed to 40 percent.(12)

The following chart shows the general climb of both the murder rate and firearm sales in the U.S.:

Murder rate (per 100,000) and firearm sales (millions of constant 
dollars, CPI-U)(13)

       Murder  Firearm
Year   Rate    Sales
1985    7.9   $1,548
1986    8.6    1,647
1987    8.3    1,667
1988    8.4    1,810
1989    8.7    1,777
1990    9.4    1,602
1991    9.8    1,859
1992    9.3    1,829
1993    9.5    2,095

Since 1989, manufacturers and importers introduced an average of 3.5 million new guns into the U.S. market each year. By contrast, the U.S. resident population has grown an average of 2.7 million a year. That's roughly 800,000 extra guns a year. (14)

In 1993, about 1.3 million Americans faced an assailant armed with a firearm. Of those, 86 percent (or 1.1 million) of the incidents involved a handgun. (15)

Here is the breakdown for all the weapons or methods used to commit murder from 1980 to 1993. Notice the trend for guns and handguns:

Murder method or weapons used; 1980-1993 (16)

Weapon or Method     1980    1993
Guns (all types)     62.4%   69.6%
     Handguns        45.8    56.9
Cutting or stabbing  19.3    12.7
Blunt objects         5.0     4.4
Personal weapons      5.9     5.0
Asphyxiations         2.3     1.9
Fire                  1.3     0.9
All others            3.8     5.5

And here are the circumstances surrounding murders for 1993. ("Arguments" include those over money, property, romantic triangles, etc. "Felonies" include robbery, narcotics, rape etc.)

Circumstances surrounding murder, 1993 (17)

Circumstances 1993
Argument       30.8%
Unknown        27.7
Other Motives  21.7
Felonies       19.1

For murders in 1994, almost half of the victims were either related to (12 percent) or acquainted with (35 percent) their killers. Only 13 percent were killed by total strangers. Of female victims, 28 percent were killed by their husbands or boyfriends. (18)

Types of Firearm deaths, 1993 (19)

Suicide                  18,940
Firearm homicide         18,571
    Handgun homicide     13,980
    Justifiable homicide    251
Accidental                1,521
Undetermined                563
Total                    39,595

In 1993, the FBI counted 24,526 murders (13,980 by handguns), yet only 251 of these were justifiable homicides by civilians using handguns.(20) This is only one percent of all murders! However, "justifiable homicide" is a narrowly-defined legal term, meaning the killing of an assailant in self-defense, and as a last resort. For example, shooting someone for stealing your car is not considered justifiable homicide (unless your life is in danger). More on this below.

And then there are the international statistics, which also show a clear correlation between handgun ownership and murder rates. (Note: the first two statistics are for handguns, not guns in general.)

Percent of households with a handgun, 1991 (21)

United States  29%
Switzerland    14
Finland         7
Germany         7
Belgium         6
France          6
Canada          5
Norway          4
Europe          4
Australia       2
Netherlands     2
United Kingdom  1

Handgun murders (1992) (22)

                Handgun    1992          Handgun Murder
Country         Murders    Population    Rate (per 100,000)
United States   13,429    254,521,000    5.28
Switzerland         97      6,828,023    1.42
Canada             128     27,351,509    0.47
Sweden              36      8,602,157    0.42
Australia           13     17,576,354    0.07
United Kingdom      33     57,797,514    0.06
Japan               60    124,460,481    0.05

As for overall firearm possession, the U.S. again comes in first, with half of all households owning a firearm. Canada is also near the top of the possession list, with a 29 percent ownership rate. Not surprisingly, they lead the list in murder rates:

Murders per 100,000 of population, 1991 (21)

United States   8.40
Canada          5.45
Denmark         5.17
France          4.60
Portugal        4.50
Australia       4.48
Germany         4.20
Belgium         2.80
Spain           2.28
Switzerland     2.25
Italy           2.18
Norway          1.99
United Kingdom  1.97
Austria         1.80
Greece          1.76
Sweden          1.73
Turkey          1.45
Japan           1.20
Ireland         0.96
Finland         0.70


The general correlation between the murder rate and the ownership of guns, especially handguns, is clear. Some might try to muddy this correlation by appealing to differences in gun control laws, but that doesn't help much. Europeans have far stricter gun regulation than the U.S., so their lower murder rates are actually an argument in favor of gun control.

The correlation between gun availability and murder begs the question: which causes which?

Before delving into this argument, we should note that the correlation itself is embarrassing to the gun lobby. They would love nothing more than to see the U.S. with both the highest gun ownership and lowest murder rate in the world. But this is not the case, and gun lobbyists are reduced to esoteric, "what-if" types of arguments. For example, what if the U.S. had even fewer guns than it has now? Then the murder rate would be even higher, they claim. (!) It's only because the murder rate is soaring that people are defending themselves by buying more guns.

There are several weaknesses to this argument. One might ask what kind of a "deterrence" is correlated to the very crime it is supposed to deter. The gun advocate might respond, "Well, firefighters are correlated to forest fires." But in the latter stages of a fire there is a negative correlation, as firefighters increase and fires diminish. A similar negative correlation between guns and murder has yet to be observed, anyplace, anywhere.

Furthermore, when guns are involved in the vast majority of murders -- 70 percent and growing -- it is clear that the "solution" and the "problem" are one and the same. One might also ask how a nation achieves a high murder rate in the first place without guns. After all, it's not easy to kill by clubbing, stabbing or hanging; these methods lack the super-ability and feasibility that guns provide. This is borne out by the fact that the murder rate is significantly lower in places where these are the primary murder methods. An even stronger rebuttal is the effect of gun control laws. If the above pro-gun argument were true, we should expect to see the murder rate climb, not fall, after the passage of gun control laws. But the introduction of gun control in Washington D.C., Kansas City, Canada, the Massachusetts 1974 Bartley-Fox Amendment, and the Brady Law shows that the murder rate indeed falls. (More on this in the next essay.)

But perhaps the greatest weakness of the pro-gun argument is that only 1 percent of all murders are considered by the FBI to be justifiable homicide by firearm. Self-defense might be the intention of people who buy guns, but when these weapons actually get used, it's almost always for murder. The implications of this are fatal to the pro-gun argument, because people's intentions are irrelevant -- the only thing that matters is how these guns are actually used. If they are used mostly for murder, with little deterrence effect, then the arrow of causality runs from gun availability to murder. Even then, causality wouldn't be the central issue here; guns could be banned simply on the grounds that they are used mostly for murder.

Not surprisingly, the gun lobby has mounted a furious assault on this statistic. They have done this in two ways:

1) Question the "justifiable homicide" figures. Some criticize the FBI for reading police reports, not trial outcomes, to determine the number of justifiable homicides. Possibly, a different truth might emerge in trial. But if a criminal homicide may turn out to be justified in trial, there is no reason to believe that the opposite wouldn't occur as well: a similar number of justified homicides turning out to be criminal. The likelihood for the latter is quite great, since, after all, self-defense is a common excuse for murder. Pro-gun advocates have yet to provide evidence that this phenomenon alters the statistics in any significant way.

Others claim that the legal term "justifiable homicide" is too narrow. For example, if an intruder enters your house, you are legally required to run out the back door (if there is one); shooting him is not considered justifiable homicide. It is only considered justified if you have no back door, and his advances are such that you believe your life to be in danger. Gun owners scoff at this law, but there is actually a good reason for it. The intruder may be drunk, drugged, mentally ill, poisoned, and not at all predisposed to robbing your house under normal conditions. The proper place for him could be a treatment center, a hospital, or even prison. But to kill him is a miscarriage of justice -- especially in the case of a mentally ill person. Better to escape and let your insurance cover any damage to your property than to have a homicide on your conscience the rest of your life.

But on the rationale that "justifiable homicide" is too narrow a term, pro-gun researchers have attempted to include other worthy categories, like "sudden combat" or "excusable homicides." An example is when someone pushes you down, and in the suddenness and confusion of combat you stop thinking and react instinctively, shooting him, even though he meant no further harm and your life was not in danger. But the FBI reports that only 1.4 percent of all homicides are "excusable." Pro-gun criminologist Gary Kleck (whose figures, as we shall see, are not deemed credible by the rest of the scientific community) therefore uses the term "Civilian Legal Defensive Homicides" (or CLDH's), a category which includes justifiable and excusable homicide. Even by his more liberal definition, however, only 7.1 to 12.9 percent of all murders are firearm CLDH's. (23) This hardly proves that guns are used to kill more in self-defense than in the commission of a crime.

2) Adding nonfatal gun defenses. Fatally shooting an attacker isn't the only way a gun can prevent crime; wounding him, firing warning shots, or simply waving the gun may do the trick. How often does this happen? Therein lies the controversy.

The most reliable figure comes from the National Crime Victimization Survey (NCVS), one of the nation's two main methods of measuring crime. In recent years, incidents where victims have used a gun in one way or another against their robbers or assailants have numbered 70,000 a year. (24) This is almost three times more than the murder rate of 24,000, which suggests guns are more beneficial than they are detrimental. However, that's not really the right way to look at it. If it's nonfatal gun defenses we are considering, then nonfatal crimes must be considered as well. And, again, those 70,000 nonfatal gun defenses formed only 1 percent of the total robberies and assaults performed each year. (25) In 1993, 1.3 million of these crimes involved guns. In other words, guns are 19 times more likely to be used in a nonfatal crime than a nonfatal defense.

Kleck and others have criticized the NCVS for undercounting the number of times victims use guns against their attackers. Kleck himself surveyed 4,979 households, and his results project that there were 2.4 million gun defenses in 1992, 1.9 million of them with handguns. About 72% of these gun defenses occurred in or near the home. (26) If his results are credible, then guns protect far more than they are used in crime, and arguably have social utility.

Should Kleck's figures be regarded as more accurate than the NCVS? To those familiar with both, the answer is a resounding no. The NCVS is the nation's second largest on-going survey. It questions 59,000 households twice a year, and has been in operation for over 20 years. It employs state-of-the-art methodology, with some of the nation's finest statisticians constantly refining and testing the validity of its results. Most of the surveys are conducted over the phone, and it has a 97 percent participation rate. A respondent's anonymity is also guaranteed by law. Unfortunately, its survey results often describe a world quite different from some people's political beliefs, so the NCVS is regularly blasted as "untrustworthy," "inept," "ideologically driven," etc.

By comparison, Kleck's survey was 12 times smaller, and not conducted by any nationally known survey organization. His sample appears to have concentrated on urban men from the South and West, populations which identify most closely with America's gun culture. His projection of 2.4 million gun defenses was based on a mere 54 responses describing incidents of self-defenses with a gun. The exact nature of these defenses, and how often they occurred per respondent, is unknown. Why? Kleck did not write a paper for more than a year after his survey, and as of 1995 has still not written a technical article for peer review. Instead, he has hit the publicity circuit, promoting his findings in newspapers, magazines and talk shows. Ducking peer review is a common method of pseudo-scientists and cranks, one for which there is no valid reason or excuse.

Gun researcher David Hemenway writes of Kleck's survey:

Kleck has a history of producing analysis that is roundly rejected by academia. His earlier estimates of successful gun defenses have differed substantially not only from academic consensus, but from each other -- 340,000 in 1986, 645,000 in 1988, and 2.4 million today. The earlier estimates were based on eight small private surveys that asked a single, vague question about using a gun for protection or self-defense. These studies failed to question a cross-section of the nation, or determine the nature of the self-defense, or the time period involved. They failed to distinguish from police and military uses, or uses against humans and animals, or the "self-protection" of a guard who merely wears a sidearm, or even two fighting gangsters who draw their weapons in self-defense. There is also a question of perception -- in almost all arguments, both parties perceive their behavior as self-defensive. Even criminals frequently see themselves as the victims of aggression. A National Institute of Justice report states: "Among a sample of prisoners, 48 percent of those who fired their guns while committing crimes claimed they did so in self-defense." (28) Really, now!

The University of Maryland conducted an academic review of Kleck's earlier work and found that "Kleck's conclusions rest on limited data. Small changes in the procedures would produce large differences in the findings. The estimates are questionable, and it appears unwise to place much weight on them." (29)

Until Kleck submits formal research that can be positively appraised by peer review, there is no reason to trust his alleged and highly contradictory findings. The NCVS remains the most authoritative source on the criminal and defensive uses of guns, and it shows that these weapons are overwhelmingly used more for crime than self-defense.

But much of the controversy over how guns are used overlooks an even more basic issue. And that is that you cannot credit a disease for its own partial cure. Even if Kleck could prove that guns were used in 100 million cases of self-defense each year, that still would not prove that guns have social utility, as long as they still drive up the murder rate. Suppose that our nation, in the name of personal security, started selling everyone vials of the Ebola virus, the horrific and contagious plague that causes agonizing and certain death in a matter of days. A few unbalanced criminals may use this virus to create plagues that kill 20,000 people a year before containment. But that's no problem: supporters have 100 million proven cases, with rock solid documentation, that the counter-threat of Ebola poisoning stopped criminals from further such murder. Needless to say, we would certainly find it odd if a "National Ebola Association" were to argue that Ebola vials provide social utility and greater security based on these numbers. Any rational society would simply choose not to hand out vials of Ebola to its population in the first place.

Return to Overview


1. Brandon S. Centerwall, "Exposure to Television as a Risk Factor for Violence", American Journal of Epidemiology, (Vol. 129, 1989), pp. 643-652.

2. Huesmann, Rowell L. and Leonard D. Eron., eds. Television and the Aggressive Child: A Cross-National Comparison. Hillsdale, NJ: Lawrence Erlbaum, 1986.

3. U.S. Bureau of the Census, Current Population Reports, Series P60.

4. P.J. Cook, "The Technology of Personal Violence," in Michael Tonry, ed., Crime and Justice: An Annual Review of Research, vol. 14, (Chicago: University of Chicago Press, 1991).

5. Josh Sugarmann, "Reverse Fire," Mother Jones, September 13, 1995.

6. Gary Kleck, "Guns and Violence: A Summary of the Field." Prepared for delivery at the 1991 Annual Meeting of the American Political Science Association, The Washington Hilton, August 29 through September 1, 1991. Address was a summary of his book, Point Blank: Guns and Violence in America (Hawthorne: Aldine de Gruyter, 1991).

7. R.L. Ohsfeldt and M.A. Morrisey, "Firearms, firearm injury and gun control: a critical survey of the literature," Advances in Health Economics and Health Services Research 13, pp. 65-82, 1992.

8. Jeffrey A. Roth, "Firearms and Violence," NIJ Research in Brief, February 1994, National Institute of Justice.

9. P.J. Cook and M.H. Moore, "Gun Control," in J.Q. Wilson and J. Petersillia, eds., Crime (San Francisco: ICS Press, 1995), pp. 269-94; P.J. Cook, "The technology of personal violence," in Michael Tonry, ed., Crime and Justice: An Annual Review of Research, vol. 14 (Chicago: University of Chicago Press, 1991).

10. Bureau of Alcohol, Tobacco and Firearms (ATF), How Many Guns? ATF News Release FY-91-36, G.P.O., Washington, 1991.

11. Marianne W. Zawitz (BJS statistician), "Guns Used In Crime" (NCJ-148201), Bureau of Justice Statistics.

12. U.S. Firearms Production 1993, American Firearms Industry, March 1994, p. 41; and U.S. Firearms Production, American Firearms Industry, January 1995, p. 74.

13. Murder rate: FBI, Crime in the United States, annual. Firearm sales: National Sporting Goods Association, The Sporting Goods Market in 1994 (and prior issues), based on a sample survey of 80,000 households, excluding Alaska and Hawaii. Current dollars converted to constant dollars by CPI-U.

14. Firearms: ATF, June 29, 1994; U.S. Firearms Production, American Firearms Industry, March 1994, p. 41; ATF, February 23, 1993; and U.S. Firearms Production 1993, American Firearms Industry, January 1995, p. 74. Population: U.S. Bureau of the Census, Current Population Reports, P25-1045 and P25-1126.

15. Zawitz.

16. FBI, Crime in the United States, 1980 and 1993.

17. FBI, Crime in the United States, 1993.

18. Violence Policy Center, "Firearm Facts."

19. Advance data from vital and health statistics, no. 242, National Center for Health Statistics (NCHS).

20. FBI, Crime in the United States, 1994, 1995.

21. Where We Stand, Michael Wolff, Peter Rutten & Albert F. Bayers III and the World Rank Research Team (New York: Bantam Books, 1992), pp. 297,289.

22. Handgun murders: Handgun Control, Inc. Population Figures: July 1992 count for each country as reported by CIA World Factbook, 1992.

23. Kleck, Point Blank, pp. 111-114.

24. National Crime Victimization Survey, annual.

25. Roth.

26. Gary Kleck, National Self-Defense Survey, Spring, 1993.

27. David Hemenway, "Guns, Public Health and Public Safety," in Dennis Henigan et al., eds., Guns and the Constitution (Northampton, Mass.: Alethia Press, 1995), pp. 63-64.

28. Roth.

29. D. McDowall and B. Wiersema, "The incidence of civilian defensive firearm use," Violence Research Group Discussion, Paper 10, Institute of Criminal Justice and Criminology, University of Maryland.