Guns and Violence: A Summary of the Field

This article is copyrighted. It was provided by the author, criminologist Gary Kleck, and is distributed with the permission of the author. It can be uploaded to other BBSs as long as it is not altered, and it may be cited as long as credit is given. As per Don B. Kates Jr.'s note:

"I have obtained Gary's permission for you to use this summary of his book. It should indicate the information noted at the bottom as to where it was presented and should indicate that it is a summary of his book POINT BLANK: GUNS AND VIOLENCE IN AMERICA (Aldine de Gruyter, 1991)".

This book can be ordered directly from Aldine de Gruyter at 200 Saw Mill River Road, Hawthorne, NY 10532.


Guns and Violence: A Summary of the Field

Gary Kleck
School of Criminology and Criminal Justice
Florida State University
Tallahassee, Florida 32312

Prepared for delivery at the 1991 Annual Meeting of the American Political Science Association, The Washington Hilton, August 29 through September 1, 1991. Copyright by the American Political Science Association.


Introduction

This paper of a summary of my book, Point Blank: Guns and Violence in America, which in turn summarizes the literature on guns, violence and gun control, as well as reporting new research. The purpose of the paper is to outline the main findings and conclusions, without systematically establishing the empirical basis for each conclusion. These can be found in the book itself. And since the book has about 570 references, it is not practical to cite supporting materials for each assertion. Only studies summarized in the tables are contained in the References. Instead, I have simply indicated the chapter of the book where interested readers may find the full set of supporting citations, empirical evidence, and detailed argumentation.

Why the Issue Matters (Chapter 2)

In 1985, about 31,600 persons were killed with guns, and perhaps another 130,000 people suffered nonfatal gunshot wounds. The majority of the deaths, 55%, were suicides, rather than criminal homicides. Only 37% were homicides, 5% were fatal gun accidents, and 1.5% each were due to legal intervention (police officers killing suspects in the line of duty) and to death where it was undetermined whether injury was intentionally or accidentally inflicted. Among all deaths due to "external cause," i.e. accident, suicide or homicide, guns were involved in 22% of them, handguns in about 13% of them. The majority of all gun deaths involve handguns, mainly because 79% of the gun homicide deaths involved handguns. Guns were involved in 1.5% of all deaths, from all causes, in 1985. They were involved in 59% of suicides, 60% of homicides, and 1.8% of accidental deaths in 1985.

There were also over 650,000 violent crimes involving guns in some way in 1985, over 540,000 of them (82%) involving handguns. Guns were involved in about 12% of all violent crime, and handguns in about 10%. The majority of the gun crimes were assaults, mostly threats without any injury or any element of theft or rape.

Gun Ownership (Chapter 2)

The prospects for reducing violence by restricting guns depends to a great extent on how many guns there are, how people get them, why they own them, and how strongly they would resist or evade gun controls in order to hold onto them. Also, one's interpretation of a positive relationship between violence rates and gun ownership rates depends on the degree to which one believes that violence can drive up gun ownership, by motivating people to get guns for protection, as well as gun ownership increasing violence.

There were probably over 200 million guns in private hands in the U.S. by 1990, about a third of them handguns. One straightforward policy implication is that policies which seek to reduce gun violence by reducing the overall supply of guns, as distinct from reducing the number possessed just by high-risk subsets of the population, face an enormous obstacle in this huge existing stock. Even if further additions to the stock could somehow be totally and immediately stopped, the size of the stock and durability of guns imply that, in the absence of mass confiscations or unlikely voluntary surrenders of guns, it might be decades before any perceptible impact of a supply-reduction strategy became apparent.

Gun ownership increased from the 1960's through the 1980's, especially handgun ownership. Some of the increase was due to the formation of new households and to growing affluence enabling gun owners to acquire still more guns; however, a substantial share of the increase was also a response to rising crime rates among people who previously did not own guns. Most handguns are owned for defensive reasons, and many people get guns in response to high or rising crime rates. Therefore, part of the positive association sometimes observed between gun ownership levels and crime rates is due to the effect of the latter on the former, rather than the reverse. Nevertheless, most guns, especially long guns, are owned primarily for recreational reasons unconnected with crime.

From the mid-1960's to the mid-1980's, scattered evidence strongly suggests that, while gun ownership increased in general, it did so even more among criminals and violence-prone people than it did among the nonviolent majority of the population. Because these "high-risk" groups are largely unrepresented in national surveys, this would partially account for the fact that household gun prevalence in national surveys remained fairly constant during this period, despite huge additions to the total stock of privately owned guns.

Gun owners are not, as a group, psychologically abnormal, nor are they more racist, sexist, or pro-violent than nonowners. Most gun ownership is culturally patterned and linked with a rural hunting subculture. The culture is transmitted across generations, with recreation-related gun owners being socialized by their parents into gun ownership and use from childhood. Defensive handgun owners, on the other hand, are more likely to be discon- nected from any gun subcultural roots, and their gun ownership is usually not accompanied by association with other gun owners or by training in the safe handling of guns. Defensive ownership is more likely to be an individualistic response to life circumstances perceived as dangerous. Defensive ownership is also a response to the perception that the police cannot provide adequate protection. This response to dangers, however, is not necessarily mediated by the emotion of fear, but rather may be part of a less emotional preparation for the possibility of future victimization.

The strongest and most consistent predictors of gun ownership are hunting, being male, being older, higher income, residence in rural areas or small towns, having been reared in such small places, having been reared in the South, and being Protestant. The social origins of Rs consistently predict having firearms, supporting the view that early socialization into gun owning subcultures is important in explaining gun ownership. However, traits like racial prejudice and punitiveness towards criminals are not important. Most gun ownership in the general public is related to outdoor recreation like hunting and its correlates, rather than crime. On the other hand, ownership of handguns may well be linked with fear of crime and prior burglary victimization, though find- ings are necessarily ambiguous due to questions of causal order - fear could motivate gun acquisition, but having a gun could also reduce the owner's fear.

The pattern of results as a whole is compatible with the thesis that gun ownership is a product of socialization into a rural hunting culture. The findings support a simple explanation of the high level of gun ownership in the United States, an explanation which rejects the notion that weak gun laws are somehow responsible. Unlike European nations with a feudal past, the U.S. has had both widespread ownership of farmland and millions of acres of public lands available for hunting. Rather than hunting being limited to a small land-owning aristocracy, it has been accessible to the majority of ordinary Americans. Having the income and leisure to take advantage of these resources, millions of Americans have hunted for recreation, long after it was no longer essential to survival for any but an impoverished few. Hunting in turn encouraged other recreational uses of guns, including target and other sport shooting, and collecting, of both handgun and long guns. Rather than high gun ownership being the result of a lack of strict gun control laws, it is more likely that causation ran in the other direction, i.e. that high gun ownership discouraged the enactment of restrictive gun laws, and that the prevalence of guns was mostly a product of the prevalence of recreational hunting. Only since the mid-1960s has a large share of gun ownership been attributable to concerns about crime.

Probably fewer than 2% of handguns and well under 1% of all guns will ever be involved in a violent crime. Thus, the problem of criminal gun violence is concentrated within a very small subset of gun owners, indicating that gun control aimed at the general population faces a serious needle-in-the-haystack problem.

Criminal gun users most commonly get their guns by buying them from friends and other nonretail sources, or by theft. Therefore, gun regulation would be more likely to succeed in controlling gun violence if it could effectively restrict nondealer acquisitions and possession of guns by this small high-risk subset of gun owners.

Focussing on Special Gun Types (Chapter 3)

Since about half of U.S. households have a gun, broadly directed restrictions on the acquisition, possession, and use of guns impinge on the lives of millions of Americans, not just a small, politically powerless subset of them. This is the essential political obstacle which faces advocates of stricter gun control - legislators who vote for strong gun laws must face the prospect of offending large numbers of gun-owning voters. Perhaps in response to this simple fact, many advocates of more restrictive controls have directed their focus away from measures which regulate all types of guns and toward those which regulate special subtypes of firearms, i.e. types of guns which are owned by smaller numbers of voters and which are consequently more vulnerable to regulation.

Pro-control groups have increasingly stressed the need to control various special weapon categories such as machineguns, "assault rifles," plastic guns, "Saturday Night Special" handguns, and "cop-killer" bullets, or sometimes all handguns. For each weapon or ammunition type, it is argued that the object is espe- cially dangerous or particularly useful for criminal purposes, while having little or no counterbalancing utility for lawful purposes. A common slogan is "This type of gun is good for only one purpose - killing people."

The specific weapon type so described shifts from one year to the next, in response to shifts in the political winds rather than actual criminologically significant shifts in criminal use of guns. For example, the so-called "cop killer bullets" which were restricted in 1986, as far as anyone can tell, have never killed a cop. Likewise, the all-plastic guns which would have been undetectable by airport security equipment were never actually manufactured, and thus had never been involved in a single act of violence.

"Assault rifles" and "assault weapons" became important objects of gun control efforts in the 1980s. Contrary to widespread claims, these semi-automatic "military-style" weapons are rarely used by criminals in general or by drug dealers or juvenile gang members in particular, are almost never used to kill police officers, are generally less lethal than ordinary hunting rifles, and are not easily converted to fully automatic fire. They do offer a rate of fire somewhat higher than other gun types and can be used with magazines holding large numbers of cartridges, but there is at present little reason to believe either attribute is relevant to the outcome of any significant number of gun crimes. While the involvement of commonplace semiautomatic pistols has been common in U.S. violence since the 1920's, probably fewer than 2% of gun homicides involve the military-style semiautomatic weapons which are commonly labelled "assault weapons.".

Saturday Night Specials (SNSs) are small, cheap handguns. They have been the target of special control efforts in the past because it was claimed that they were the preferred weapon of criminals, and were especially useful for criminal purposes, based on the twin notions that they are especially concealable because of their small size, and that their low price makes them especially affordable for predominantly low-income criminals. The best available information indicates the following about SNSs. Only about 10-27% of crime handguns (in the 1970's) fit the U.S. Bureau of Alcohol, Tobacco and Firearms (BATF) definition of SNSs (barrel length under three inches, .32 caliber or less, and price under $50 in mid-1970's dollars). Thus, most crime handguns were not SNSs, nor did they claim a share even approaching a majority. Because only about 10% of violent crimes involve a handgun, SNSs are involved in only about 2-7% of all violent crimes. Further, the SNS share of crime guns appears to be no larger than the SNS share of the general civilian handgun stock - at least 20% of all handguns introduced into the general civilian stock were SNSs. Thus, there is no strong reason to believe that criminals are any more likely to use SNSs than noncriminal members of the general public are. More specifically, criminals are no more likely to use cheap or small caliber handguns than noncriminal gun owners. Therefore, there is no meaningful sense in which criminals can be said to "prefer" SNSs. On the other hand, there is some mixed sup- port for the idea that some criminals prefer short-barrelled handguns over longer-barrelled ones, though the weapons tend to be middle or large caliber and of good quality. At most, perhaps 7%, and more realistically 1-2%, of SNSs will ever be involved in even one violent crime. In sum, most handgun criminals do not use SNSs, and most SNSs are not owned or used for criminal purposes. In- stead, most are probably owned by poor people for protection.

One policy implication of the last conclusion is that gun control efforts directed specifically at SNSs, such as the Ken- nedy-Rodino bill, would have their greatest impact in reducing the availability of defensive handguns among low income people. The identical observation was made by liberal critics about the ban on importation of SNSs contained in the Gun Control Act (GCA) of 1968. Effective SNS-specific measures would disproportionately affect the law-abiding poor, since it is they who are most likely to own SNSs and obey the laws, and who are least likely to have the money to buy better quality, and therefore higher-priced, weapons.

Considering the obvious flaws of a policy focussing solely on SNSs, why would anyone advocate it? One answer is that SNSs may not be the real target of the policies, but rather that all handguns are. Given the somewhat obscure and technical definitions that are actually used in legislation and administrative regulations, it would be easy to manipulate such a definition in a politically low-profile way such that most handguns fell within the SNS category. Another possible motivation is that prohibiting those types of firearms which poor people can best afford is the next best thing to an overtly discriminatory policy of banning gun ownership by poor people, a policy which would be politically, and perhaps constitutionally, impossible to implement in any but a covert form.

A SNS-specific control policy could be worse than merely ineffectual. If it actually did deprive any criminals of SNSs, some would adapt by substituting larger and/or marginally more expensive guns, which would imply the substitution of larger cali- ber, longer barrelled handguns. Wounds inflicted with larger caliber handguns are more like to result in a death; longer barrelled guns fire bullets with greater accuracy and a higher muzzle velocity, thereby increasing their deadliness. Conse- quently, among those persons who previously would have used SNSs but who, as a result of the control policy, substituted larger handguns, the attack fatality rate would almost certainly increase.

Most U.S. gun laws are aimed largely or solely at handguns. This focus has the same flaw as the focus only on SNSs, but on a larger scale. While some potentially violent people denied handguns would do without guns of any kind, others would substitute shotguns and rifles, which are generally more lethal. Under any but the most optimistic circumstances, this would result in a net increase in the number of homicide deaths.

One of the political temptations of handgun-only control is that it appears to be a satisfactory compromise between doing nothing about gun violence, which would alienate pro-control vot- ers, and restricting all gun types, which would alienate many long gun owners. It is tempting to assume that the results of this apparent compromise policy would correspondingly lie somewhere between the results of a policy of doing nothing and the results of one restricting all guns. This assumption is false - the "middle" course of restricting only handguns is worse than either of the other two alternatives.

A clear policy recommendation follows from what should be the first principle of weapons regulations: Never place restrictions on a subcategory of weapons without also placing restrictions at least as stringent on more deadly, easily substituted alternative weapons.

Focusing on specialized weapon categories will be an unpro- ductive, but unfortunately increasingly popular gun control stra- tegy in the foreseeable future. The very features that make the piecemeal approach ineffective also make it politically attractive. Thus, policies focusing on machine guns, "assault rifles," plastic guns, and armor-piercing bullets are inoffensive to most voters and have little cost, but they also address weapons that are only very rarely used by criminals.

So far, this is merely a special case of a political universal applying to any policy area - weak approaches carry less risk to policymakers, while also having less impact on the target problem. However, many special-weapon gun control measures are worse than this, since they have serious potential for making the violence problem worse. Policies targeting only less lethal weaponry, such as handguns generally or "Saturday Night Specials" specifically, can increase the gun death total by inadvertently encouraging the substitution of more lethal types of guns.

Defensive Use of Guns by Crime Victims (Chapter 4)

Policy analysts seeking to assess the relative costs and benefits of gun control sometimes simplify their task by assuming that gun ownership has no significant benefits, beyond the relatively minor ones of recreational enjoyment of shooting sports like hunting. Under this assumption, it is unnecessary to show that a given law produces a large reduction in violence, since even one life saved would surely outweigh the supposedly negligible benefits of gun ownership. This simplification, however, is unrealistic, because it erroneously assumes that gun ownership and use has no defensive or deterrent value, and thus no potential for preventing deaths or injuries.

Each year about 1500-2800 criminals are lawfully killed by gun-wielding American civilians in justifiable or excusable homicides, far more than are killed by police officers. There are perhaps 600,000-1 million defensive uses of guns each year, about the same as the number of crimes committed with guns. These astounding totals may be less surprising in light of the following facts. About a third of U.S. households keep a gun at least partially for defensive reasons; at any one time nearly a third of gun owners have a firearm in their home (usually a handgun) which is loaded; about a quarter of retail businesses have a gun on the premises; and perhaps 5% of U.S. adults regularly carry a gun for self-defense.

Keeping a gun for home defense makes most defensive gun owners feel safer, and most also believe they are safer because they have a gun. The belief is not necessarily a delusion. People who use guns for self-protection in robberies and assaults are less likely to have the crime completed against them (in a robbery, this means losing their property), and, contrary to widespread belief, are less likely to be injured, compared to either victims who use other forms of resistance or to victims who do nothing to resist. (Criminals take the gun away from the victim in less than 1% of these incidents.) The evidence does not support the idea that nonresistance is safer than resisting with a gun.

Defensive uses of guns most often occur in circumstances where the victims are likely to have access to their guns, mostly in their homes or places of business. Thus, defensive gun uses are most commonly linked with assaults in the home (presumably mostly domestic violence), commercial robberies, and residential burglaries.

The fact that armed victims can effectively disrupt crimes suggests that widespread civilian gun ownership might also deter some criminals from attempting crimes in the first place. There probably will never be definitive evidence on this deterrence question, since it revolves around the issue of how many crimes do not occur because of victim gun ownership. However, scattered evidence is consistent with a deterrence hypothesis. In prison surveys criminals report that they have refrained from committing crimes because they thought a victim might have a gun. "Natural experiments" indicate that rates of "gun deterrable" crimes have declined after various highly publicized incidents related to victim gun use, including gun training programs, incidents of defensive gun use, and passage of a law which required household gun ownership. Widespread gun ownership may also deter burglars from entering occupied homes, reducing confrontations with residents, and thereby reducing deaths and injuries. U.S. burglars are far less likely to enter occupied premises than burglars in nations with lower gun ownership.

Gun use by private citizens against violent criminals and burglars is common and about as frequent as legal actions like arrests, is a more prompt negative consequence of crime than legal punishment, and is more severe, at its most serious, than legal system punishments. On the other hand, only a small percentage of criminal victimizations transpire in a way that results in defensive gun use; guns certainly are not usable in all crime situations. Victim gun use is associated with lower rates of assault or robbery victim injury and lower rates of robbery completion than any other defensive action or doing nothing to resist. Serious predatory criminals perceive a risk from victim gun use which is roughly comparable to that of criminal justice system actions, and this perception may influence their criminal behavior in socially desirable ways.

The most parsimonious way of linking these previously uncon- nected and unknown or obscure facts is to tentatively conclude that civilian ownership and defensive use of guns deters violent crime and reduces burglar-linked injuries.

Rates of commercial robbery, residential burglary injury, and rape might be still higher than their already high levels were it not for the dangerousness of the prospective victim population. Gun ownership among prospective victims may well have as large a crime-inhibiting effect as any crime-generating effects of gun possession among prospective criminals. This could account for the failure of researchers to find a significant net relationship between rates of crime like homicide and robbery, and measures of general gun ownership - the two effects may roughly cancel each other out. Guns are potentially lethal weapons whether wielded by criminals or victims. They are frightening and intimidating to those they are pointed at, whether these be predators or the preyed upon. Guns thereby empower both those who would use them to victimize and those who would use them to prevent their victimi- zation. Consequently, they are a source of both social order and disorder, depending on who uses them, just as is true of the use of force in general.

The failure to fully acknowledge this reality can lead to grave errors in devising public policy to minimize violence through gun control. While some gun laws are intended to reduce gun possession only among relatively limited "high-risk" groups such as convicted felons, through such measures as laws licensing gun owners or requiring permits to purchase guns, other laws are aimed at reducing gun possession in all segments of the civilian population, both criminal and noncriminal. Examples would be the Morton Grove, Illinois handgun possession ban, near approximations of such bans (as in New York City and Washington, D.C.), prohibitions of handgun sales (such as those in Chicago), and restrictive variants of laws regulating the carrying of concealed weapons. By definition, laws are most likely to be obeyed by the law-abiding, and gun laws are no different. Therefore, measures applying equally to criminals and noncriminals are almost certain to reduce gun possession more among the latter than the former. Because very little serious violent crime is committed by persons without previous records of serious violence (Chapter 5), there are at best only modest direct crime control benefits to be gained by reductions in gun possession among noncriminals, although even marginal reductions in gun possession among criminals might have crime-inhibiting effects. Consequently, one has to take seriously the possibility that "across-the-board" gun control measures could decrease the crime-control effects of noncriminal gun ownership more than they would decrease the crime-causing effects of criminal gun ownership. For this reason, more narrowly targeted gun control measures like gun owner licensing and permit-topurchase systems seem preferable.

People skeptical about the value of gun control sometimes argue that while a world in which there were no guns would be desirable, it is also unachievable. The evidence summarized here raises a more radical possibility - that a world in which no one had guns might actually be less safe than one in which nonaggressors had guns and aggressors somehow did not. As a practical matter, the latter world is no more achievable than the former, but the point is worth raising as a way of clarifying what the goals of rational gun control policy should be. If gun possession among prospective victims tends to reduce violence, then reducing such gun possession is not, in and of itself, a social good. Instead, the best policy goal to pursue may be to shift the distribution of gun possession as far as practical in the direction of likely aggressors being disarmed and likely nonaggressors being armed. To disarm noncriminals in the hope this might indirectly help reduce access to guns among criminals is not a cost-free policy.

Effects of Guns on Assaultive Violence (Chapter 5)

Guns in the hands of prospective victims of violence can deter criminal attempts or disrupt crimes once they are attempted, thereby exerting a violence-reducing effect. Oddly enough, guns in the hands of aggressors also have certain violence-reducing effects, along with the more obvious violence-increasing effects. The power which weaponry confers has conventionally been treated as exclusively violence-enhancing - it has commonly been assumed that weapon possession and use serves only to increase the likelihood of the victim's injury and death (e.g. Newton and Zimring 1969). This is an unduly restrictive conceptualization of the significance of weaponry. A broader perspective starts with a recognition of weaponry as a source of power, frequently used instrumentally to achieve goals by inducing compliance with the user's demands. The ultimate goal behind an act of violence is not necessarily the victim's death or injury, but rather may be money, sexual gratification, respect, attention, or the terrorizing, humiliation, or domination of the victim. Power can be, and usually is, wielded so as to obtain these things without inflicting physical injury. Threats, implied or overt, usually suffice and are often preferred to physical attack.

The effects of guns in the hands of aggressors can be better understood if we view violent events as being composed of an ordered series of stages, with the occurrence and outcome of each stage being contingent on previous stages. Figure 1 lists the stages, along with the likely effects which gun possession by aggressor or victim is likely to have on the outcome.


Figure 1.  Effects of Possession and Use of Guns
               on Assaultive Violence


Stage in                      Guns in the Hands of the:
Hostile Encounters            Aggressor           Victim

Confrontation                      (+)            (+/-)
Threat, given confrontation        (+)            (-)
Attack, given threat                -              -
Injury, given attack                -              0
Death, given injury                 +              0

Notes:
+ means gun possession of use increases the probability that the
encounter will proceed from the previous stage to the current one,
- means a gun decreases this probability, and 0 means no effect.
Parentheses around symbols indicates there is insufficient
information to do more than state a hypothesized direction of
effect.

(1) Confrontation. First, the prospective aggressor and victim coincide in time and space, entering into a potentially conflictual encounter with each other. Possession of a gun can embolden both victims and aggressors to go where they like, including dangerous places where they might adventitiously encounter a stranger who, in the course of the interaction, becomes an adversary, or it may even encourage them to stop avoiding, or even deliberately seek out, contact with persons with whom they already had a hostile relationship. Thus, gun ownership could increase the rate of assaultive violence by giving people freedom of movement without regard to the risks of entering into dangerous circumstances, thereby increasing the rate of hostile encounters. There is, however, no systematic evidence on these possible effects.

(2) Threat. Once aggressor and victim find themselves confronting one another in a hostile encounter, a gun in the possession of the aggressor could encourage him to threaten the victim, with words or a gesture, possibly alluding to the gun. On the other hand, the prospective victim's possession of a gun could, if it was known to the would-be aggressor, discourage the aggressor from expressing a threat. Again, there is no systematic evidence bearing directly on this effect.

(3) Attack. Some hostile encounters go beyond verbal or gestural threats, escalating to an attempt to physically injure the victim, i.e. proceeding to an attack. An aggressor's possession of a gun can either increase or decrease the probability that he will attack his victim. At least four categories of effects on attack can be conceptualized, and they can be labelled facilitation, triggering, inhibition and redundancy.

Facilitation. A gun could make possible or easier an attack which would otherwise be physically or emotionally impossible, dangerous, or difficult to carry out. It has often been remarked that a gun serves as an "equalizer," that it is a way of making power relations more equal than they otherwise would be. Just as a prospective victim's possession of a gun can give him power greater than or equal to his adversary and discourage an attack, the aggressor's possession of a gun could encourage it. The gun might assure the aggressor that his attack will so effectively hurt his victim that counterattack will be impossible, or at least that his victim will be afraid to strike back, even if physically capable of doing so. Guns can thereby encourage weaker adversaries to attack stronger ones. Thus guns are more commonly used when women attack men than when women attack other women, are more common when an individual attacks a group than when the situation is reversed, and so forth. Guns also facilitate attack from a distance. As someone once observed noted, "a gun may not be absolutely necessary to kill, but at fifty yards it's certainly a help." Further, a gun may facilitate an attack by a person who is unwilling to attack in a way which involves physical contact with his victim, or by a person too squeamish to use a messier weapon like a knife or club.

Triggering. This is the effect which experimental psychologists label the "weapons effect." Since it is but one of many effects of weaponry, this term is unsuitable, so I have relabelled it the triggering effect. Psychologists have argued that a person who is already angered may attack when they see a weapon, due to the learned association between weapons and aggressive behavior. The experimental research literature on this hypothesis is almost exactly divided between studies supporting it and studies failing to support it. Generally, the more realistic the study's conditions and the more relevant to real-world aggression, the less supportive the results were. There may be triggering effects, but they appear to be very contingent effects, which depend on settings and conditions not yet very well- specified.

Inhibition. Some of the "weapons effect" studies found evidence that weapons could inhibit aggression as well as trigger it. While the reasons for these experimental findings are not clear, in real world violence, one reason for such an effect might be that a gun provides an aggressor with a more lethal weapon than he wants. Most aggressors do not want to kill, but this could easily happen if they attacked with a gun. Therefore, an aggressor may refrain from attacking altogether, for fear that he might do end up inflicting more harm than he wanted to.

Redundancy. This inelegant term alludes to the possibility that possession of a gun could make a physical attack unnecessary, by making it possible for an aggressor to get what he wants without attacking.

Weapons are an important source of power frequently wielded to achieve some emotional or material goal - to obtain sexual gratification in a rape or money in a robbery, or, more frequently, to frighten and dominate victims in some other assault. All of these things can be gained without an attack, and indeed the possession of a gun can serve as a substitute for attack, rather than its vehicle. In robberies, offenders without guns often feel they must attack their victim in order to insure that the victim will not resist, while robbers with guns are confident they can gain the victim's compliance merely by pointing their gun at them. In assaults, a gun can enable an aggressor to terrify his victim or emotionally hurt him, making a physical attack unnecessary.

It is not yet possible to separately assess the relative importance of each of these possible causal effects. However, the total effect of all them considered together is fairly clear. The net effect of aggressor gun possession on whether the aggressor attacks is negative. In at least 17 prior studies, mostly of robbery, but also of assault, aggressors with guns were less likely to attack and/or injure their victim.

(4) Injury. Once an aggressor makes an attack, it may or may not result in injury. That is, only some attempts to injure are successful. The rate at which attacks result in physical injury to the victim is lower when the attacker fires a gun than when he throws a punch, attempts to cut or stab his victim, or tries to strike the victim with a blunt instrument of some kind. This presumably is because it is difficult to shoot a gun (usually a handgun) accurately, especially under the emotionally stressful conditions which prevail in most violent encounters. Only about 19% of incidents where an aggressor shot at a victim result in the victim suffering a gunshot wound, while the comparable attack completion rate is about 55% for knife attacks. Since guns facilitate attacks at a distance and attacks against more difficult targets, they may thereby also reduce the attack completion rate. (5) Death. Finally, if the aggressor does inflict a physical injury on the victim, it may or may not result in death. Less than 1% of all criminal assaults result in death, and the measured fatality rate is under 15% even if we limit attention just to gunshot woundings. Further, because nonfatal attacks are substantially undercounted, while fatal attacks are fairly completely counted, the true fatality rate in gunshot woundings is actually still lower, probably under 10%.

Nevertheless, the measured wounding fatality rate for guns is about four times higher than that of woundings with knives, the next most lethal weapon, among those which could be used in the same circumstances as guns. This might seem to indicate that if guns became scarce and attackers used guns rather than knives, only one fourth as many victims would die. This reasoning, however, is invalid because it implicitly attributes all of the difference in fatality rates to the weapon itself, and assumes that all else, including the intentions and motives of the aggressors, is equal in gun and knife attacks. This assumption is unrealistic. Evidence indicates that aggressors who use guns choose them over other available weapons - a gun is not used just because "it was there;" weapon choice is not random. Rather, more serious aggressors use more serious weaponry. For example, aggressors with longer records of violence in their past are more likely to use guns. Thus, some of the 4-to-1 difference in fatality rates between guns and knives is due to differences in the people who used the weapons, rather than just the technical differences between the weapons themselves. Since weapon scarcity would presumably not alter the intentions and aggressive drive of aggressors, this implies that the fatality rate would drop by a factor of less than four if knives were substituted for guns. It is impossible to say how much less, since it is impossible to measure and control for the intentions and intensity of an aggressor's anger and willingness to hurt his victim at the moment of the attack. Nevertheless, studies that have imperfectly controlled for aggressor traits thought to be correlated with these factors indicate that guns still appear to be more lethal than knives.

To summarize, an aggressor's possession and use of a gun apparently reduces the probability that he will attack, reduces the probability that the attack will result in an injury, and increases the probability that the injury will be fatal. Therefore, it is not at all obvious that threatening situations with a gun-armed aggressor are more likely to result in the victim's death, since it is not obvious what the relative balance of these three countervailing effects is. The best empirical evidence on real- life violent incidents indicates that the net effect is essentially zero. That is, the overall probability of a threatening situation ending in the victim's death is about the same when the aggressor is armed with a gun as it is when the aggressor is unarmed. In short, guns have many strong effects on violent encounters, but they work in both violence-increasing and violence-decreasing directions, and these effects apparently more or less cancel each other out.

Note that this conclusion takes no account of gun effects on confrontations and threats. It is still possible that gun availability in a population could affect the rates of assault and murder, despite the foregoing conclusions, if it significantly encouraged people to more frequently enter into dangerous confrontations and to issue threats or otherwise initiate hostile interactions. Also, an analysis focussing solely on individual violent incidents cannot take account of possible deterrent effects of victims having guns, which would tend to discourage aggressors from seeking contact with victims or threatening them. Consequently, the net impact of widespread gun ownership must be assessed using data on aggregates like cities or states, where the combined impact of all of these separate effects can be estimated. These kinds of studies will be summarized later.

Effects of Guns on Robbery (Chapter 5)

A robber's goal is to get his victim's property. Injury to the victim appears to be more of unintended by-product of the crime than an important goal, in contrast to homicides and assaults. Consequently, guns have some additional effects peculiar to robberies, as well the effects observable in assaultive crimes. They may have a facilitative effect similar to that connected with assaultive crimes, since they may encourage some people to rob who would not be willing to do so without a gun. They also appear to encourage robbers to tackle more difficult, better guarded (and more lucrative) targets, such as stores or groups of people on the street, rather than lone individuals. While this might seem to imply that gun availability should increase the robbery rate, the best available evidence indicates that the former has no apparent net effect on the latter. This may be due partly to deterrent effects of victim gun ownership, especially the impact of defensive gun ownership and use by store owners on commercial robberies. However, gun possession by robbers also may have its own negative effect. Because the average "take" in gun robberies is higher than in nongun robberies, a robber can acquire a given amount of money (e.g. that needed to support a drug habit) with fewer robberies.

Concerning the attacks, injuries, and deaths linked with robberies, the effects of robber gun use parallel those observed in assaults, with some additional elements also apparent. Robber gun use appears to inhibit victim resistance, thereby reducing the robber's need to attack and injure the victim. And indeed, studies have invariably indicated that gun robbers are less like to attack or injure their victims than are unarmed robbers. On the other hand, if the victim is injured, he is more likely to die if shot with a gun than if injured in some other way. As with assaultive crimes, it is unclear how much of this greater fatality rate is attributable to the weapons and how much to robber differences.

Impact of Gun Ownership Levels on Violent Crime Rates (Ch. 5)

The findings of aggregate studies are summarized in Table 1. Their findings are almost exactly evenly split between 12 findings that support the idea that higher gun levels increase crime rates and 11 findings that do not. All but a handful of the studies are technically very weak. They rely on small samples, sometimes including as few as nine, or even four cases; only Bordua (1986) had more than 50 cases. In combination with the multicollinearity that typically characterizes aggregate data, this implies very unstable results. Most use measures of gun ownership which are either known to be invalid or whose validity is unknown. Eight of the studies did not control for any other factors that might be as- sociated with gun ownership and could affect crime rates, making it impossible to check whether any observed association between gun and violence levels were spurious; 11 studies controlled for no more than two other variables.

The most critical flaw in the aggregate-level studies is the failure to model the two-way relationship between crime rates and gun levels. Higher crime rates can cause more people to acquire guns for self-defense. Consequently, any significant positive associations generated in studies failing to model the possible two-way relationship will at least partially reflect the effect of crime rates on gun rates, rather than the reverse. Whether there is also any effect of guns on violence is impossible to detect from these findings. Of eighteen studies, the problem was statistically addressed in only four of them. These studies generally found no impact of gun ownership levels on violent crime rates.


Table 1.  Studies of the Effect of Gun Ownership Levels
                    on Violent Crime Rates


                              2-way   Measure of  Crime
Study              Sample    Relat.?  Gun Level   Rates  Results

Brearley (1932)  42 states     No       PGH        THR     Yes
Krug (1967)      50 states     No       HLR        ICR      No
Newton and       4 years,      No       NPP      THR,TRR   Yes
 Zimring (1969)   Detroit                        AAR,GHR
Seitz (1972)     50 states     No     GHR,FGA      THR     Yes
                                       AAR
Murray (1975)    50 states     No     SGR,SHR    GHR,AAR    No
                                                   TRR
Fisher (1976)    9 years,      No     NPP,GRR      THR     Yes
                  Detroit               PGH
Phillips et al.  18 years,     No      PROD        THR     Yes
  (1976)           U.S.
Brill (1977)     11 cities     No       PGC        ICR      No
                                                   THR     Yes
                                                   TRR      No
Kleck (1979)     27 years,    Yes      PROD        THR     Yes
                   U.S.
Cook (1979)      50 cities     No     PGH,PGS      TRR      No
                                                   RMR     Yes
Kleck (1984a)    32 years,    Yes      PROD        THR      No
                   U.S.        No                  TRR     Yes
Maggadino and    31 years,     Nob     PROD        THR     Yes
  Medoff (1984)    U.S.
Lester (1985)    37 cities     No       PCS        VCR      No
Bordua (1986)   102 counties   Noc    GLR,SIR    HAR,THR,   No
                  9 regions                        GHR      No
McDowall (1986)  48 cities,   Yes     PGH,PGS      TRR      No
                  2 yearsd
Lester (1988)    9 regions     No       SGR        THR     Yes
McDowall and     36 years,     Noe    PGR,FGA      THR     Yes
 Loftin (1988)    Detroit
Linsky et al.    50 states     No       GMR        GHR     Yesf
 (1988)
Kleck and        170 cities   Yes        g          g       No
 Patterson (1991)

Results: Yes=Study found significant positive association between
 gun levels and violence; No=Study did not find such a link.

Measures of Gun Level:
FGA = Fatal gun accident rate
GLR = Gun owners license rate
GMR = Gun magazine subscription rates
GRR = Gun registrations rate
HLR = Hunting license rate
NPP = Number of handgun purchase permits
PGA = % aggravated assaults committed with guns
PGC = % homicides, aggravated assaults and robberies (combined
      together) committed with guns
PCS = same as PGC, but with suicides lumped in as well
PGH = % homicides committed with guns
PGR = % robberies committed with guns
PGS = % suicides committed with guns
PROD= Guns produced minus exports plus imports, U.S.
SGR = Survey measure, % households with gun(s)
SHR = Survey measure, % households with handgun(s)
SIR = Survey measure, % individuals with gun(s)

Crime Rates:
AAR = Aggravated assault rate
GHR = Gun homicide rate
HAR = Homicide, assault and robbery index (factor score)
ICR = Index crime rate
RMR = Robbery murder rate
THR = Total homicide rate
TRR = Total robbery rate
VCR = Violent crime rate

Notes:
a. Table covers only studies and findings where the dependent
   variable was a crime rate, as opposed to the fraction of
 crimes committed with guns.
b. Authors modelled two-way relationship, but only report gun
   impact results for a model where this was not done.
c. A few gun-violence associations were positive and significant,
    but almost all involved female gun ownership or male longgun
   ownership. Author interpreted the pattern to indicate the
  effect of violence on gun ownership.
d. Panel design, two waves.
e. Attempt to model two-way relationship probably failed due to
   an implausible identification restriction.  See text.
f. Only established an association with gun homicide rate.  No
   result for total homicide rate reported.
g. Gun ownership treated as a latent construct, measured with five
   indicators: PGH, PGS, PGR, PGA and the % of the value of stolen
   property due to stolen guns.  Crime rates modelled were total
 rate, gun rate, and nongun rates of homicide, rape, aggravated
assault, and robbery (e.g. the rates of total homicides, gun
homicides and nongun homicides).

Effects of Guns on Suicide (Chapter 6)

In a suicide, victim and offender are the same person, so there is no victim resistance to overcome. This radically changes the nature of the technology needed to carry the act out. The gun's capacity to facilitate attacks against strong victims or attacks at a distance is irrelevant. On the other hand, its lethality, and the quickness with which it can be used, may be significant for suicides.

Gun availability might increase suicide rates by giving suicide attempters a more lethal method. It could be argued that, in the absence of a gun, while some attempters would still persist after a nonfatal suicide attempt, others would not and lives would therefore be saved. This argument differs, however, from, the parallel argument made for gun effects in assaultive crimes. Unlike in the latter case, there are many common methods of committing suicide which are nearly as lethal, and in other ways even more satisfactory, than guns. The fatality rate in gun suicide attempts is about 85%, but it is about 80% in hanging attempts, 77% with carbon monoxide, and 75% with drowning. These are only slight differences, and some or all of them 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.

Other ways of committing suicide are in many ways as satisfactory or even superior to using a gun. For example, using carbon monoxide in the form of exhaust fumes does not disfigure the victim as much as shooting, is not as messy, is less painful, is nearly as lethal, and is quieter and therefore less likely to summon people who might intervene to save the attempter's life. Consequently, there is more reason with suicide than with homicide to expect that nongun methods could be substituted for guns with equally frequent fatal results.


Table 2.  Studies of the Effects of Gun Ownership Levels
                         on Suicide Rates
                                                                    Kleck
          Markush and                                               & Pat-
           Bartolucci Lester  Lester   Lester   Clarke and   Lester terson
Study:        (1984)  (1987)  (1988a)  (1988b)  Jones (1989) (1989) (1991)

Sample        9 U.S.    50   6 Austl.  9 U.S.    26(13)       50     170
             regions  states  states  regions    yearsa     states  cities
# Control
  Variables    0        0       0        3         0          0       13
Measure of Gun
 Ownershipb    S        O       S        S         S          O        Oc
Impact on Gun
 Suicide Rate? Yes      Yesd    Yes      Yesd    Yes/Noe      Yes     No
Impact on Total
 Suicide Rate? Nof      No      No       No      Yes/Noe      Yes     No

Notes:
a. Time series dataset included 26 years total, but only thirteen had real
     data on gun ownership levels; the rest were interpolations.
b. S=survey measure - % of Households with guns; O=other measures
c. See note i, Table 5.
d. Only bivariate association reported.
e. Handgun prevalence related to suicide rates, total gun prevalence
      unrelated.
f. Significant positive correlation was only obtained if eccentric
     weighting scheme was applied.  Conventional unweighted results
    indicated no significant association.

Consistent with this assessment, previous research has indicated that while gun ownership levels are consistently related to the rate of gun suicides, they are unrelated to total suicide rate (see Table 2). That is, where guns are common, people will more frequently use them to kill themselves, but this does not affect the total number of people who die. Apparently, gun availability affects only method choice, not the frequency of fatal outcomes.

Gun Accidents (Chapter 7)

While gun accidents contribute only about 5% of the deaths linked with guns, they play an important rhetorical role in the gun control debate. They are used in attempts to persuade people that keeping guns in their homes for protection is foolish because the risks of a gun accident exceed any defensive benefits. Gun accidents play a different rhetorical role in the debate from homicides or suicides because most people can accurately tell themselves that there is no one is their household like to assault another person or attempt suicide, but it is harder to confidently state that no one will be involved in an accident. Since anybody can have an accident, every household with a gun is at risk of suffering a gun accident.

There are several problems with this argument. First, gun accidents are quite rare relative to the numbers of people exposed to them. The rate of accidental death per 100,000 guns or per 100,000 gun-owning households is less than 4-6% of the corresponding rates for automobiles, and has also been sharply declining for over 20 years, despite rapid increases in the size of the gun stock. Second, the risk of a gun accident is not randomly distributed across the gun-owning population and is not a significant risk for more than a small fraction of owners. Gun accidents are apparently largely confined to an unusually reckless subset of the population, with gun accidents disproportionately occurring to people with long records of motor vehicle accidents, traffic tickets, drunk driving arrests, and arrests for violent offenses. Accidents are most common among alcoholics and people with personality traits related to recklessness, impulsiveness, impatience, and emotional immaturity. The circumstances of gun accidents commonly involve acts of unusual recklessness, such as "playing" with loaded guns, pulling the trigger to see if a gun is loaded, and playing Russian roulette with a revolver. Gun accidents are largely confined to defensive gun owners - less than one sixth of accidental deaths are connected with hunting. Consequently, gun accidents are quite rare for ordinary gun owners, especially when compared with the frequency of defensive uses.

Contrary to impressions left by the news media, gun accidents rarely involve small children. There are probably fewer than 100 fatal handgun accidents involving preadolescent children in the entire nation each year. Instead, gun accidents are largely concentrated in the same age groups where assaultive violence is concentrated, among adolescent and young adult males.

Most gun safety training is aimed at hunters, rather than the defensive gun owners who make up the bulk of people involved in gun accidents. Because of this narrow focus, and because the training does not treat alcoholism or modify the shooter's personality, it probably has little impact outside of the hunting community. On the other hand, it might be possible to reduce gun accidents through gun laws (mainly aimed at reducing crime) which prohibit gun acquisition or possession by high-risk groups like felons or alcoholics.

Types of Gun Controls (Chapter 8)

"Gun control" encompasses many different forms of laws intended to regulate human behavior in some way related to firearms. Some controls regulate gun acquisition, restricting the purchasing, trading, or receiving of guns. Gun owner license laws require that people have a license in order to lawfully possess a gun, even in the home, and in order to acquire the gun in the first place. This license is not issued until the applicant has passed through a check of official records to see if the person has a prior criminal conviction, and possibly to see if they have some other disqualifying traits, such as alcoholism or mental illness. Purchase permit laws require a person to get a permit before buying a gun, and applicants must first pass through a records check. "Application-to-purchase" systems are similar to purchase permit systems, except that the records check is typically optional, and the system usually requires a minimum waiting period between initial purchase attempt and final delivery of the gun. Registration systems merely record the acquisition or possession of a gun, linking each gun with a particular owner. They do not screen for unqualified gun buyers.

Other laws regulate gun transactions from the other end, licensing and regulating the selling of guns, or regulating their manufacture or importation. Still others regulate various kinds of gun use. Some laws forbid the carrying of guns in public places, while others require licenses to do so. Restrictions are generally stronger regarding concealed carrying than open carrying, and stronger with respect to carrying on the person than carrying in a motor vehicle. Some attach mandatory penalties to unlawful carrying. Other laws attempt to discourage gun use in crimes by attaching additional penalties (some discretionary, others mandatory) if various dangerous felonies are committed with a gun.

Almost all states prohibit possession of guns by high-risk subgroups of the population, most commonly convicted criminals, mentally ill people, drug addicts, alcoholics, and minors. These laws do not directly restrict the original acquisition of guns, but instead make it somewhat more legally risky to be in possession of guns at any one time.

The strongest gun laws of all impose bans on the possession, sale, and/or manufacture of various categories of guns. While no U.S. jurisdiction forbids gun ownership altogether, New York City and Washington, D.C. have de facto bans on the private possession of handguns, and some small towns have formal handgun bans. Some cities, such as Chicago, forbid the sale of handguns within city limits, without banning their possession. Finally, a number of states have banned the sale and manufacture of "Saturday Night Specials," usually defined in practice as guns made of cheap metal with a low melting point.

Public Opinion and Support for Gun Laws (Chapter 9)

Levels of support for gun control have shown no clear long- term trends in the past decades. There is short-term volatility in reported levels of support for some measures, consistent with evidence that opinion is easily changed and that gun control is not a salient issue for many Americans, despite the emotional intensity of debates among activist minorities. The intensity of support for gun control appears to be weaker than opposition, in the sense that opponents report that they are much more likely to actually do something based on their beliefs, such as contributing money to an organization connected the issue or writing a letter to a public official. Much of the support for gun control is not utilitarian or instrumentalist in character: that is, many people support gun control even though they do not believe it is an effective tool for reducing violence. Instead, positions on gun control seem symptomatic of culture conflict, with gun law used as a way of declaring gun ownership and gun owners to be morally inferior, parallel to the way alcohol prohibition was used as a way for older Anglo-Saxon Protestants to condemn the culture of supposedly free- drinking Catholics from Irish or Southern and Eastern European backgrounds.


    Table 3.  What Kinds of Gun Control Do Americans Favor?
          (generally in increasing order of popularity)
                                                            % in
Control Measure                                Date  Survey Favor

All-Guns Ban
  Private citizens surrender all guns to govt. 1976  NORC    <17a
  Ban sales of all guns                        1985  Roper    22
  Making it illegal for civilians to own guns  1989a Time/CNN 29
Handgun Bans
  Buy back, destroy handguns, mandatory basis  1978  Caddell  26
  Buy back, destroy handguns, voluntary basis  1978  Caddell  33
  Ban further manufacture, sale of handguns    1978  Caddell  32
  Ban private possession of handguns           1990  CSUR     36
  Ban sales of handguns                        1989  CBS      40
  Ban handgun possession in high crime areas   1975  Harris  <47
  Local ban on sale, possession of handguns    1986  Gallup   47
     in R's own community
  Ban further manufacture, sale of             1978  Caddell  48
     nonsporting handguns
  Federal ban on interstate sales of handguns  1986  Gallup   67
  Ban further manufacture, sale of             1978  Caddell  70
     small, cheap handguns
  Federal ban on manufacture, sale, possession 1989  Gallup   71
     of "Saturday Night Specials"
Other Gun Bans
  Federal ban (as above) on semiautomatic      1990  CSUR     69
     assault guns, such as the AK-47
  Federal ban (as above) on plastic guns       1989  Gallup   75
Ban on Keeping Loaded Guns
  Illegal to have loaded weapons in home       1965  Gallup  <47a
Ammunition Purchase Permit
  Police permit required to buy ammunition     1965  Gallup  <56a
Mandatory penalty, carrying
  Mandatory minimum 1 year jail term,          1981  Gallup   62
    carrying gun without a license
Ban on Use by Minors
  Completely forbid use of guns by those <18   1967  Gallup  <68a
Purchase Permit
  Require permit to purchase a rifle           1975  Harris  <69a
  Require police permit to purchase a gun      1990  CSUR     68
Owner's License
  Require license to own a handgun             1978  Caddell  74 Registration
  Register all guns owned                      1989  Time/CNN 73
  Register all rifles owned                    1989b Time/CNN 68b
  Register all shotguns owned                  1989b Time/CNN 65b
  Register all handguns owned                  1990  Harris   73
  Register all gun owners                      1940  Gallup   74
  Register all semiautomatic weapons owned     1989b Time/CNN 77b
  Register all gun purchases                   1990  Harris   79
  Register all handgun owners                  1938  Gallup   79
  Register all handgun purchases               1978  Caddell  84
Safety Training
  Require mandatory safety training to buy gun 1989b Time/CNN 82b
Mandatory Prison Sentence, Crime with Gun
  Require mandatory prison sentence for        1978  Caddell  83
     persons using a gun in a crime
Carry Permit
  Require a permit to carry a gun outside home 1988  Gallup   84
Waiting Periods
  21 day waiting period to allow criminal      1981  Gallup   91
     records check, handgun purchases
  14 day waiting period, any gun purchase      1989a Time/CNN 89
   7 day waiting period, handgun purchases     1988  Gallup   91

Sources:
1989b Time/CNN - Quinley (1990); 1981, 1986, 1987 Gallup - Gallup (1987);
1988 Gallup - Gallup (1989); 1965, 1967 Gallup, 1975 Harris, 1976 NORC -
Smith (1980); 1990 CSUR (Center for Social and Urban Research) - Mauser and
Margolis (1990); 1985 Roper, 1989a Time/CNN, 1989 CBS, 1990 Harris - computer
search of DIALOG database, POLL file; all others - Crocker (1982).  See
original sources for exact question wordings.
Notes:
a. Source only reported % opposing measure; 100 minus % opposing is maximum
possible % in favor.
b. Computed as simple average of separate percentages for gun owners and
nonowners.
NORC = National Opinion Research Center, producer of General
Social Surveys


Table 3 shows the level of public support for many different specific gun control proposals. There are a large number of weak or moderate controls which a majority of Americans will endorse if asked, though few will volunteer "gun control" as an answer if asked an open-ended question soliciting their opinion about how crime might be reduced. Bans on gun possession do not have majority support, but many moderate regulatory measures do. Controls on handguns enjoy more support than controls on the more widely owned rifles and shotguns. There is more support for "getting tough on criminals" than for controls likely to restrict or impose costs on ordinary gun owners. In short, Americans support controls unlikely to have any direct impact on themselves, while opposing those which might impose some costs on them or interfere with their own gun ownership.

The Impact of Gun Control Laws on Violence Rates (Chapter 10)


Table 4.  Studies of the Effect of Gun Control Laws
                    on Violent Crime Rates
                                                          Gun
                                             Control
Study                                   Weakness       Effective?
                                   1  2  3  4  5  6  7
Wisconsin (1960)                   X  X  X  X  X          No
Krug (1967)                        X  X  X  X  X  X       No
Geisel et al. (1969)                    (X) X  X  X       No
Olin Mathieson (1969?)             X  X  X  X  ?  X       No
Seitz (1972)                       X  X  X (X) X  X       Yes
Murray (1975)                      X  X  X (X) X          No
Zimring (1975)                     X  -  -     -  -  X    Mixed
Beha (1977)                        X     X (X) -  -  X    Mixed*
Deutsch and Alt (1977)             -     -  X  -  -  X    Mixed*
Cook (1979)                              X     ?          No
Hay and McCleary (1979)            -     -  X  -  -  X    No*
Nicholson and Garner (1980)        X     -  X  -  -  X    Mixed
Sommers (1980)                     X  X  X  X  X     X    Mixed
Jones (1981)                       X  -  -  X  -  -  X    Mixed
Lester and Murrell (1981)          X     X  X  X  X       No
Pierce and Bowers (1981)           X     -  X  -  -  X    Mixed*
Lester and Murrell (1982)          X  X  X  X  X  X       Mixed
Magaddino and Medoff (1982)        X  X  X  X  X          No
DeZee (1983)                          X  X  X  X          No
Loftin et al. (1983)               -  -  -  X  -  -  X    No
Loftin and McDowell (1984)         -  -  -  X  -     X    No
Magaddino and Medoff I (1984)         X  X  X  X          No
Magaddino and Medoff II (1984)        -  -     -  -  X    No
McPheters et al. (1984)            -  -  -  X  -  -  X    Yes
Lester and Murrell (1986)          X  X  X  X  X  X       No
Lester (1987)                      X  X  X  X  X  X       No
Lester (1988)                     (X) X  X     X  X       Yes
Jung and Jason (1988)              -     -  X  -     X    No
Kleck and Patterson (1991)                                No

Summary:  3 Yes, 8 Mixed, 18 No

"Gun Control Effective?" means "Did gun laws appear to significantly reduce
total rates of violence or crime?"

Weakness Codes:  X indicates problem existed, blank indicates no problem,
dash (-) indicates problem is an inherent property of time series studies,
and (X) indicates partial presence of problem, or problem inadequately dealt
with.

Weaknesses:
 1.  Included no, or very few, control variables.
 2.  State level of analysis used, rather than city.
 3.  No measure of local gun control laws.
 4.  No measure of gun ownership included.
 5.  Only one source of information on gun control laws was used.
 6.  Lumped heterogenous mixture of gun laws together, without separate
     measures of impact of different types of gun laws.
 7.  Studied just one specific law; little generalizability.

*These four studies are not independent since they are all evaluations of the
same law (Mass. Bartley-Fox law) in the same time period, using the same
general methods.  They contributed 3 of the 8 studies classified as "Mixed."
Their findings are classified this way because, taken as a whole, they
indicate that the law had no effect on homicide, may have reduced robbery
(two studies indicated this, one did not), and reduced gun assaults by a
moderate amount while increasing nongun assaults by a larger amount.


Table 5.  Studies of the Effect of Gun Control Laws on Suicide Rates
                                                          Gun Controls
                                                         Significantlya
                                        Gun      # Gun   Reduce Rate of:
                            # Control Ownership Controls  Gun     Total
Study               Sample  Variables Measured? Assessed Suicide Suicide
Geisel et al.(1969) 50 states,  7       No       1(8)b   Yes/Noc   No
                      1960
                    50 states,  8       No       1(8)b   Yes/Noc   No
                      1965
                    129 cities, 8       No       1(8)b      -      No
                      1960
Murray (1975)       50 states,  9       No         7        No     -
                      1970
Lester and Murrell  48 states,  0       No         1b      Yes    Yes
  (1980)            1960, 1970
Nicholson and       Time series 0       No         1       Yes    Yes
  Garner (1980)     Wash., D.C.
Lester and Murrell  48 states,  0       No        3(8)d    Yes     -
  (1982)            1960, 1970
Medoff and          50 states,  5       No        1(2)e     -     Yes
  Maggadino (1983)    1970
DeZee (1983)        50 states,  7       No          7       No     -
                      1978
Sommers (1984)      50 states,  2       No          9       Nof    -
                      1978
Lester (1987)       48 states,  0       No          1b       -    Yes
                      1970
Lester (1988)       9 regions,  2      Yes          1b     Yesg    No
                      1970
Boor and Bair       50 states,  9       No        2(8)h      -    Yes
  (1990)             D.C., 1985
Rich et al. (1990)  time series 0       No          1      Yes     No
                     2 cities
Kleck and Patterson 170 cities 10      Yes         13       Noi    Noi
 (1991)
Notes:
a. Significant at .05 level.
b. Measured "strictness" of gun control - all control types lumped
    together.
c. Overall "strictness" index was significantly and negatively related, but
     separate gun law dummies yielded no significant results.
d. Used 3 factor scores grouping 8 gun control types together; individual
    controls not separately assessed.
e. Lumped two gun law types together into a single dummy variable.
f. Only one of 9 gun law coefficients significant at .05 level.
g. Only bivariate association reported.
h. Grouped eight types of gun control into two summary indexes.
i. Of 13 types of controls assessed, 11 showed no negative relationship with
     either gun or total suicide.  Bans on gun possession by mentally ill
    people were negatively related to gun suicide but not total suicide,
   while licensing of gun dealers was negatively related to both.


Table 4 summarizes prior research on the impact of gun laws on violent crime rates, while Table 5 summarizes research on their impact on suicide rates. Given the previously noted lack of support for the notion that guns have a net violence-increasing impact on either violence rates or the outcomes of individual violent incidents, it is not surprising that research has failed to indicate consistent support for the view that gun laws reduce violence. Most studies do not support this idea, and the few that do are extremely weak methodologically. The more common technical weaknesses are listed in Table 4.

Kleck and Patterson (1991) sought to avoid all of these technical problems. Their analysis covered all forms of violence which involves guns, encompassed every large (over 100,000 population) city in the nation, and assessed all major forms of existing gun control in the U.S. Their findings are summarized in Table 6. They indicate that gun ownership levels have no net positive effect on the total rate of any major form of violence, and that, with few exceptions, existing gun control laws have no net negative effect on violence rates.


   Table 6. The Effect of Gun Control Laws and Gun Ownership Levels
               on Violence Rates

                                                  Model
                                                                   Fatal Gun
                                  Murder Aslt Robbery Rape  Suicide Accidents
Significant
positive effect
of gun ownership
on violence?                       NO    NO    NO     NO     NO       NO

Significant
negative effect
of gun laws on
violence?

License to possess gun in home     NO    NO    NO     NO      NO     YES
Permit to purchase                MAYBEa NO    NO     NO      NO      NO
Application to purchase            NO    NO    NO     NO      NO      NO
Waiting period to receive gun      NO    NO    NO     NO      NO      NO
Ban on possession by criminals     NO   MAYBEaMAYBEa  NO      NO      NO
Ban on possession by mentally ill MAYBEa NO    NO     NO     MAYBEa   NO
Ban on possession by addicts       NO    NO    NO     NO      NO      NO
Ban on possession by alcoholics    NO    NO    NO     NO      NO      NO
Ban on purchase by minors          NO    NO    NO     NO      NO      NO
Registration of guns               NO    NO    NO     NO      NO      NO
State or local dealer license      NO   YESa   NO     NO     MAYBEa   NO
Concealed handgun carrying         NO    NO    NO     NO
  forbidden or permit hard to get
Open handgun carrying forbidden    NO    NO    NO     NO
  or permit hard to get
Mandatory penalty, unlawful carry MAYBEa NO   YESa    NO
Discretionary add-on penalty for   NO    NO   YESa    NO
  crimes committed with a gun
Mandatory add-on penalty for       NO    NO    NO     NO
  crimes committed with a gun
State Constitutional guarantee of  NO    NO    NO     NO
  individual right to bear arms
De facto ban on handgun possession NO    NO    NO     NO      NO      NO
Ban on sale of Sat. Night Specials NO    NO    NO     NO      NO      NO

Summary: 4 YES, 7 MAYBE, 91 NO

Notes:
a. Gun law appeared to reduce gun use in this category of violence.
Source: Kleck and Patterson (1991)


The only clear exceptions were owner licensing, which seems to reduce fatal gun accidents, add-on penalties for committing crimes with a gun, which appear to reduce robbery, mandatory penalties for unlawful gun carrying, which also seem to reduce robbery, and state or local licensing of gun dealers, which (surprisingly) appears to reduce suicides and assaults.

Policy Conclusions (Chapter 11)

Despite substantial variation in gun control severity and gun ownership levels across U.S. cities, there is no evidence that these have any measurable impact on violence levels, although they do affect the frequency with which guns are used in some kinds of violence. On the other hand, the frequency with which guns are carried may have an impact on robbery which gun ownership levels do not, and gun ownership within special high-risk subsets of the population may have an impact on violence rates which general gun ownership levels do not.

Therefore, the significance of the few gun control measures found to be effective should not be overlooked. There is empirical support for some moderate gun controls. I favor a national "instant records check," which would screen for high-risk gun buyers similar to owner license and purchase permit systems, but without the delays and arbitrary administration which sometimes characterizes those controls. The system should cover nondealer transactions as well as dealer sales, and apply to rifles and shotguns, as well as handguns. Also, tighter licensing of gun dealers and increased enforcement of carry laws may be useful.

Gun control is a very minor, though not entirely irrelevant, part of the solution to the violence problem, just as guns are of only very minor significance as a cause of the problem. The U.S. has more violence than other nations for reasons unrelated to its extraordinarily high gun ownership. Fixating on guns seems to be, for many people, a fetish which allows them to ignore the more intransigent causes of American violence, including its dying cities, inequality, deteriorating family structure, and the all- pervasive economic and social consequences of a history of slavery and racism. And just as gun control serves this purpose for liberals, equally useless "get tough" proposals, like longer prison terms, mandatory sentencing, and more use of the death penalty serve the purpose for conservatives. All parties to the crime debate would do well to give more concentrated attention to more difficult, but far more relevant, issues like how to generate more good-paying jobs for the underclass which is at the heart of the violence problem.


REFERENCES