Understanding and Preventing Violence: Volume 1 (1993)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

FIREARMS AND VIOLENCE 255 6 Firearms and Violence An analysis of violence in the United States would be incomplete without a discussion of firearms, which are involved in about 60 percent of all homicides. The most widely debated methods of preventing gun violence involve legal restrictions on firearm owners or licensed dealers. While evaluations have found several regulations that have reduced gun homicides when they are enforced, regulations arouse controversy among law-abiding dealers and owners—some of whom acquired guns precisely because they feared gun violence. Our review of available evidence revealed several promising preventive strategies, none requiring new laws, which are ripe for rigorous evaluation. We also concluded that like "drug-related violence," "gun violence" may be best understood in terms of illegal markets and reduced through tactics that police already apply in illegal drug markets. We begin by discussing overall patterns and trends in firearms ownership and use in violent crimes. Then, proceeding from best understood to least understood, we examine how firearms alter a series of conditional probabilities in violent events: the chance that an injury causes death, the chance that an encounter (e.g., robbery, assault) produces an injury, and the relationship of firearms availability to the frequencies of such encounters. We then discuss what little is known about the channels through which firearms reach violent crime scenes and conclude by examining 

FIREARMS AND VIOLENCE 256 the empirical basis for alternative strategies for reducing the incidence and consequences of violent firearm use. PATTERNS AND TRENDS IN FIREARM OWNERSHIP AND VIOLENT USE Firearms are widely owned and widely available in the United States. No precise count is available since no enumeration system exists, but the best estimates are that the national gun population was in the range of 60-100 million in 1968 (Newton and Zimring, 1969), 100-140 million in 1978 (Wright et al., 1983), 130-170 million in 1988 (Cook, 1991), and 200 million in 1990 (Bureau of Alcohol, Tobacco, and Firearms, 1991). For at least three decades, the fraction of all households owning any type of gun has remained stable at about 50 percent; however, the fraction owning a handgun rose from 13 percent in 1959 to 24 percent in 1978, where it has remained, more or less, since then (Bureau of Justice Statistics, 1989, cited in Cook, 1991). Surveys generally conclude that the fraction of households owning firearms is greatest in rural areas and small towns, higher for whites than for blacks, highest in the South, lowest in the Northeast, and higher for high-income households than for low-income households (Cook, 1991; Wright et al., 1983). Violent uses of guns impose both human and monetary costs. In 1989 gun attacks resulted in about 12,000 homicides—about 60 percent of all homicides (Federal Bureau of Investigation, 1989). In addition, Cook (1991) estimates that 5.7 nonfatal gunshot injuries occur for every homicide—a projection on the order of 70,000 for 1989. For 1985, Rice and associates (1989) estimate the total cost of intentional and unintentional gun injuries at over $14 billion, including both the direct costs of hospital and other medical care and the indirect costs of long- term disability and premature death. The risk of death from homicide by a gun is not uniformly distributed throughout the population; rather, it is highly elevated for adolescents, particularly black males. In 1988 the gun homicide rate per 100,000 people was 8 for teenagers ages 15-19, but less than 6 for the population as a whole. Among teenagers ages 15-19, the 1989 rate was 83.4 for black males compared with 7.5 for white males (unpublished data from the National Vital Statistics system, 1992)—a ratio of more than 11:1. The elevated risk for teenagers reflects both a concentration of homicides in ages 15-34 and a peak in gun use by killers of young 

FIREARMS AND VIOLENCE 257 people. Figure 6-1, which displays 1989 age distributions for homicide victims in five-year intervals, shows that the largest fraction of nonfirearm homicide victims was in the 25-29 age range; the largest fraction of firearm homicide victims was slightly younger, in the 20-24 age range. The fraction of homicides committed with guns peaks even earlier, at 81 percent for victims ages 15 to 19, then declines gradually but monotonically for victims age 20 and older (Figure 6-2). Figure 6-1 Age distribution for homicide victims at five-year intervals. SOURCE: Adapted from Federal Bureau of Investigation (1989:11). Although the 1989 risk of gun homicide for black teenage males ages 15-19 was 11 times the risk for their white counterparts, black victims outnumbered white victims in that age range by 

FIREARMS AND VIOLENCE 258 only 1,163 to 547, because whites are so much more numerous in the population (unpublished data from the National Vital Statistics system, 1992). Despite the small numbers involved, the firearm homicide risk for young black males warrants special attention. Not only is the recent increase in this risk a significant social inequity, but also such a huge increase is unique to this particular intersection of sex, race, age, weapon type, intent, and time period. Figure 6-2 Fraction of homicides committed with firearm by age of victim. SOURCE: Adapted from Federal Bureau of Investigation (1989:11). Trends between 1979 and 1988 in teenage deaths were reported by Fingerhut et al. (1991). During that period, no increases of similar proportions occurred in firearm homicides of black females 

FIREARMS AND VIOLENCE 259 or white males of the same age, or among older black males ages 25-29 (Figure 6-3).1 And between 1987 and 1988 nonfirearm homicides actually declined for black male teenagers, from 10.8 per 100,000 to 9.5, essentially the 1984 level. The rate for unintentional firearm death, a measure more closely related to firearm availability, rose only about 50 percent (from 2.3 to 3.4 per 100,000) among black male teenagers between 1984 and 1988. And every rate displayed in the figure declined during the immediately preceding period, 1979-1984. The most comparable trend during the period 1984-1988 is in the other form of intentional injury: a 100 percent increase (from 3.4 to 6.8 per 100,000) in the firearm suicide rate for black male teenagers—a higher relative increase than for other race and sex categories in this age group or for older black males. Figure 6-3 Death rates for selected demographic groups and manners of death: United States, 1979-1988. F/H = firearm homicide; N/H = nonfirearm homicide; F/S = firearm suicide; F/U = unintentional firearm death. SOURCE: Fingerhut et al. (1991: Table 1). In short, the increase during the period 1984-1988 in firearm homicides cannot be explained solely in terms of either demographics or firearm availability; rather, one must look at these factors in conjunction with other events that affected young black males during those specific years. 

FIREARMS AND VIOLENCE 260 The risks of violent gun use are not evenly distributed across type of gun. Of all guns in the United States, approximately one-third are handguns and two- thirds are long guns—a heterogeneous category that includes, for example, both small-caliber and large-caliber rifles, shotguns, and semiautomatic and automatic rifles. Handguns are disproportionately likely to be used in homicides, by a factor of more than 3 to 1. In gun homicides for which the type of weapon was known, handguns accounted for nearly 80 percent in 1989, compared with 8 percent for rifles and 12 percent for shotguns (Federal Bureau of Investigation, 1989). Despite substantial recent public attention to what are called assault weapons, we were unable to find national statistics on the availability, use, or lethality of this rather ill-defined category of firearm (Zimring, 1989; Kleck, 1991). Even for handguns, the weapon of choice in violent crimes in the United States, the risk of violent use is not evenly distributed. Based on estimates by Cook (1981), it appears that, of each cohort of new handguns sold, about one- third are involved in crime (i.e., displayed or fired) at least once during their usable lifetimes.2 New handguns, those acquired most recently, are especially likely to be involved in violent crimes (Zimring and Hawkins, 1987). But there is little information available about the kinds of places, encounters, handgun attributes, ownership patterns, and distribution channels that heighten the risk that a given handgun will be used in a violent manner. Research on this question is complicated by the lack of data systems to track unique identifiers of firearms and by their frequent movement, through burglary, unregulated sales, and simple carrying, from one situation to another. (On request from a law enforcement agency, the Bureau of Alcohol, Tobacco, and Firearms attempts to trace the ownership paths of guns used in crime, but these do not provide a basis for comparing their ownership histories to those of other guns.) There are no accurate profiles for distinguishing “dangerous" handguns or locations from "safe" ones. THE LETHALITY OF FIREARM INJURIES A focus on violent injuries produces one clear correlation. The case fatality rate—the fraction of injuries that lead to death—is much greater for guns than for knives, the next most lethal weapon. Zimring (1968) reports fatality rates of 12.2 percent for gun attacks and 2.4 percent for knife attacks, a lethality ratio of about 5 to 1. Others report somewhat smaller lethality ratios, but no 

FIREARMS AND VIOLENCE 261 estimate is less than 2 to 1 (Curtis, 1974, reported in Wright et al., 1983; Vinson, 1974). From scientific and policy perspectives, the difficulty lies in determining how much of the difference in lethality among weapons is due to instrumental effects (i.e., properties of the weapon or ammunition) and how much to the perpetrator's intent. For more than 30 years, people have argued that the perpetrators most intent on killing not only select firearms more frequently but also aim more carefully, fire more bullets, and take other steps that further increase the lethal effects (Wolfgang, 1958). In this view, intent contributes most of the lethality and, if guns became unavailable, such intent killers would succeed in killing with knives, blunt instruments, fists, or feet. This argument is plausible for some shootings, but there is no basis for estimating how often it applies. Firearms are rarely used by serial killers, for example, who are considered among the most intent of all killers. Often guns are used to achieve some objective other than killing, such as to threaten. The use of guns to threaten that has been the best measured is robbery, in which the purpose is to get persons, usually strangers, to hand over property. Guns are also used to threaten or to inflict harm against family members or casual acquaintances in many different situations—disputes, revenge for some perceived slight or chronic harassment, and the like. Another use of threat is self-defense, in which, for example, a surprised homeowner may brandish or fire a gun primarily to frighten a burglar away. The choice of a firearm to communicate threats has two potentially offsetting effects on lethality. If use of a gun instead enables the initiator to achieve his or her goals without firing, physical harm is prevented (Kleck, 1991). But having a gun may encourage some people to initiate robberies or altercations against others who would otherwise appear too invulnerable to challenge; occasionally the gun will be discharged in those encounters, causing physical harm. Guns may also be fired in circumstances that involve no strategy. Such incidents include gratuitous murders at the successful conclusion of a robbery— perhaps as a manifestation of hate or of what has been termed recreational violence. Also, guns may be fired in conflicts between intimates or more casual acquaintances involving intense anger, fear, or both. Between 1976 and 1987, more than twice as many women were shot and killed by their husband or intimate partner as were killed by strangers using any means. In violent domestic disputes, the choice of a weapon may 

FIREARMS AND VIOLENCE 262 well be the nearest available object that can project force, and it seems likely that instrumentality rather than intent contributes most of the firearm's lethal effect. Police and court records make clear that neither these scenarios nor countless others should be ignored in debates over firearms policy. The difficulty, as any trial judge knows, is that establishing the motive for an instance of gun use is a formidable, time-consuming, judgmental task, often with an uncertain outcome. Available information is therefore indirect and imprecise. In an analysis of gun assaults in Chicago (Zimring, 1968), the victim received a single wound in over 80 percent of both fatal and nonfatal gun assaults. Since most guns have the capability to inflict multiple wounds, this suggests that at least some assaulters with guns lacked sufficient intent to exhaust all options for killing the persons they attacked.3 An analysis of survey data for victims of 12,000 robberies (Cook, 1980) is consistent with Kleck's conclusion that guns are sometimes used to threaten victims who would otherwise be relatively invulnerable: 55 percent of commercial robberies but only 13 percent of other robberies involve guns; the few gun robberies of persons tend to target relatively invulnerable groups and young adult males; and nongun street robberies are most likely to target women, children, and elderly victims. Similarly, Sampson and Lauritsen (Volume 3) report that, between 1973 and 1982, approximately 41 percent of male victims of violent crimes other than homicide, but only 29 percent of female victims, were attacked by persons with weapons—a pattern consistent with the use of weapons to overcome victim invulnerability rather than necessarily to kill. A survey of 1,874 incarcerated convicted felons interviewed during 1982 provides a more precise but perhaps unrepresentative distribution of the motivations of criminal gun users (Wright and Rossi, 1985). Of the 184 who fired a gun while committing the offense for which they were incarcerated, only 36 percent reported doing so "to kill the victim." More commonly reported motivations were "to protect myself" (48%) and "to scare the victim" (45%). Incarcerated offenders are not representative of all offenders and may be reluctant to report an intent to kill. Nevertheless, the data suggest strongly that many criminals who fire guns do so for reasons other than a single-minded attempt to kill. In fact, as the authors conclude, "most of the men [76%] who actually fired guns in criminal situations claimed to have had no prior intention of so doing" (Wright and Rossi, 1985:15; emphasis in original). 

FIREARMS AND VIOLENCE 263 These findings indicate that, even in violent felonies, firearms are sometimes fired without a premeditated intent to kill. Consequently, a problem for future research is to measure what fraction of the difference in lethality between firearms and other weapons is due to instrumentality rather than intent. Since this fraction of lethality would be reduced if guns became less available, the research findings should be integral to debates over firearms policy. FIREARMS, ENCOUNTERS, AND INJURY RATES To this point, our discussion has treated the gun victim as a passive target. Usually, of course, violent gun use occurs during an encounter between two or more people and depends on the nature of their interaction. As we indicate in Chapter 3, far too little is known about these interactions generally, in part because records of representative samples of assaults are difficult to construct, compile, and analyze. However, the choice of weapon for attacks and for self- defense is important in determining the probability and severity of injury when a gun is involved. Weapon Choice By Perpetrators Most of the evidence about perpetrators' weapon choice and lethality comes from studies of robberies and assaults. Studies of robbery reviewed by Cook (1991) indicate that, compared with other robbers, those who carry a gun are more likely to complete their robberies without experiencing victim resistance and without injuring the victims. However, because gun injuries are so much more likely to be lethal, the fatality rate for gun robberies, 4 per 1,000, is about triple the rate in knife robberies and 10 times the rate in robberies with other weapons (Cook, 1980). Zimring and Zuehl (1986) report a similar differential for nonresidential robberies known to the police. But because nonlethal residential robberies are more likely to be reported if a gun was used, including residential robberies in the analysis attenuates the lethality differential. Kleck and McElrath (1989) report a somewhat different pattern for assaults. The presence of a lethal weapon decreases the chance of actual physical attack. If an attack occurs, injury is less likely with a gun than with a knife. When an injury occurs, the fatality rate is higher in gun incidents than in knife incidents. The latter effect approximately offsets the first two, so that fatalities are about equally likely in gun assaults and knife assaults. 

FIREARMS AND VIOLENCE 264 The Bureau of Justice Statistics (1986: Table 12) reported, by weapon type, the distribution of injury outcomes for violent victimizations of all types, based on National Crime Survey data for the period 1973-1982. Escalation from threat to attack, and especially to an attack causing injury, was far less probable when the offender was armed with a gun (probability = .14) than when he or she was unarmed (.30), armed with a knife (.25), or armed with any other weapon (.45). Given that they were injured, just over half of both gun victims and knife victims required medical attention, but gun victims were somewhat more likely to be hospitalized (.83) than were knife victims (.74) if they required medical help. Nevertheless, because of the low probabilities of attack and injury when the offender displayed a gun, the overall probability of being hospitalized following victimization by a gun-wielding offender was only .06—the same as when the offender was unarmed, and less than when the offender was armed with a knife (.10) or with another weapon (.16).4 The data do not of course explain the low probabilities of attack and injury in events in which guns are involved. The attacker's display of the gun may intimidate the victims into submission rather than resistance; the attacker may fire the gun but miss (an event that has no analog in statistics on attacks by unarmed offenders); or the intended victims of gun wielders may themselves be more likely to be carrying guns, which they display or use in self-defense. Firearm Use in Self-Defense The use of guns for self-defense is one of the most controversial topics in firearms policy, and it promises to become more so if residents of high-crime areas grow more skeptical that police can or will effectively protect them and their property. In a 1979 survey, 20 percent of all gun owners and 40 percent of all handgun owners cited self-defense at home as the most important reason for their gun ownership; for handgun owners, self-defense was the most commonly cited reason (Decision-Making Information, Inc., 1979, cited in Wright et al., 1983:96). But self-described "self-defense" is ambiguous; it may refer, for example, to homeowners with a generalized fear of burglary, to single women who have received specific threats of rape, to belligerent regular patrons of "fighting bars" who anticipate confrontations, to gang members preparing for neighborhood intrusions by violent rival gangs, to criminals fearing retaliation for their previous crimes, and even to 

FIREARMS AND VIOLENCE 265 criminals in the course of their crimes—recall that in Wright and Rossi's sample of incarcerated felons, 48 percent of those who fired guns cited self-defense as a motive. Police report anecdotally hearing "It was him or me" as an increasingly common excuse offered by alleged youthful killers with guns. Counts of Gun Use for Self-Defense Because self-defense is such an ambiguous term, it is not surprising that there is great dispute over how often firearms are actually used for self-defense. Using National Crime Survey data, Cook (1991) estimates incidents of self- defense firearm use at 78,000 per year, just below the number of people killed and wounded by firearms. Drawing on a number of surveys that ask about self- defense uses of guns without tying them to specific incidents, Kleck puts the annual number of self-defense uses about 10 times higher—between 700,000 (G. Kleck, personal communication to Jeffrey A. Roth, National Research Council, 1990; Kleck, 1991) and 1 million (Kleck, 1988) per year, about equal to the number of violent crimes involving guns. Part of the discrepancy is no doubt accounted for by methodology. The National Crime Survey excludes commercial robberies, and it undercounts attempted and completed rapes and sexual assaults, particularly by family members; self-defense gun uses in those incidents would be missed as well. The surveys reviewed by Kleck (1988, 1991) have smaller sample sizes and longer reference periods, and they leave definitions of self-defense and use to respondents. Besides the ambiguity in defining self-defense, the counts may be inflated by respondents' inadvertent telescoping of incidents of self-defense into the reference period from previous periods. Also, some reported instances of self-protection are not comparable to specific criminal victimizations: respondents in some surveys could well have been concerned about animal rather than human attackers; others may have simply brought the gun nearby in anticipation of an encounter that never occurred.5 To correct for these and various other methodological artifacts, Kleck uses a number of adjustment factors, some of which rely on untested assumptions. He presents no sensitivity analyses of alternative adjustment procedures, and some of them are incompletely documented. Our calculations indicate that making alternative plausible assumptions would substantially decrease Kleck's estimates of annual self- defense uses.6 Because of the likely undercounts in the National Crime Survey 

FIREARMS AND VIOLENCE 266 and the uncertainties surrounding Kleck's sources and adjustments, the discrepancy about how and how often guns are used for self-defense remains approximately 80,000 to 700,000. Some of it can be accounted for by Kleck's adjustment factors. But an additional cause is almost certainly that some of what respondents designate as their own self-defense would be construed as aggression by others. As recommended by Cook (1991), there is a need for estimates that are free of the threshold and reporting problems associated with the National Crime Survey and the FBI's Supplementary Homicide Reports but that differentiate among self-defense uses and that rest on surveys of sufficiently high quality to eliminate the need for dubious adjustment procedures. The Probability of Gun Use for Self-Defense In making the decision whether to purchase a gun for protection, it is less relevant to know how often guns are used for self-defense than to know either the probability of using the gun if victimized or the consequences of using it. According to National Crime Survey tabulations, victims rarely defend themselves with firearms—1.2 percent of robberies, 1.4 percent of assaults, 3.1 percent of all residential burglaries, and 4.6 percent of residential burglaries by strangers. 7 Self-defense gun use, however, is associated with a reduced risk of physical attack and injury. Of intended robbery victims who defended themselves with guns, only 17 percent reported being injured, compared with 33 percent overall and 25 percent of those who made no self-protection effort. For assaults, the injury rates were 12 percent for targets who used guns, compared with 30 percent overall and 27 percent who took no self-protective action (Kleck, 1988: Table 4). Statistical associations such as these have been interpreted as demonstrations of the self-defense value of guns. However, since such self-defense uses are so rare in victimizations, they may well represent anomalies—scenarios in which, for example, the initiator failed to surprise the intended victim. In losing the advantage of surprise, the initiator may simultaneously become more vulnerable to any form of self-defense and give a gun-owning intended victim time to arm himself. Without detailed comparisons of situational dynamics in events in which gun-owning victims did and did not use their guns in self-defense, claims that average victims can successfully defend themselves with guns against robbery, burglary, and assault remain unproven. Firearm ownership can carry substantial risks. In confrontations 

FIREARMS AND VIOLENCE 267 with assailants, even experienced owners of firearms can find their weapons turned against them. Between 1984 and 1988, 93 percent of all law enforcement officers killed in the line of duty died of gunshot wounds. Sixty-four of these officers (19%) lost their lives when their service weapons were turned against them (unpublished data, Federal Bureau of Investigation). Guns in the home may also increase a family's risk of serious injury or death through mishandling, family violence, or suicide. Kellermann and Reay (1988) studied all gunshot deaths occurring in King County, Washington, between 1978 and 1983 and found that 52 percent occurred in the home where the firearm involved was kept. For every time a gun in the home was involved in a self- protection homicide, they noted 1.3 unintentional deaths, 4.5 criminal homicides, and 37 firearm suicides. Wintemute et al. (1987) studied 88 incidents in which young California children fatally shot a playmate or themselves and noted that 75 percent occurred while children were playing with a gun or demonstrating its use. In at least 48 percent of fatal residential shootings, children gained access to unlocked firearms that were stored loaded. Below the age of 8, few children can reliably distinguish a toy gun from a real one. Firearm Availability and Violence How is the availability of guns related to overall levels of violence? The discussion above has established that greater gun availability is associated with more robberies, home burglaries, assaults, and homicides using guns—but this is only one of several countervailing relationships to be considered. By reducing robbery and home burglary completion rates, self-defense gun use theoretically decreases the rewards for these crimes and increases the perceived risks to offenders. Yet guns in the homes of law-abiding citizens are themselves tempting targets for burglars. Moreover, by analogy from theoretical models of international arms races (Downs, 1991), the fear that potential victims or adversaries may be armed may simply encourage someone planning an attack to acquire superior firepower and carry out the attack regardless. Because of these conflicting potential incentives, it is not clear a priori whether greater firearms availability should be expected to increase or decrease aggregate violence levels. Indirect empirical evidence on this question is available from four kinds of studies: multivariate statistical analyses of nonfatal violent crime levels, using the prevalence of gun use in homicides 

FIREARMS AND VIOLENCE 268 and suicides as a proxy measure for gun availability (Cook, 1979, 1985); cross- national comparisons of gun availability measures and gun crimes (van Dijk et al., 1990); time-series analyses of gun availability measures and homicide rates (Phillips et al., 1976; Kleck, 1984; McDowall, 1991); and comparisons between jurisdictions in which substantial availability differences are generally acknowledged without being precisely measured (Sloane et al., 1988). Studies of all these types are discussed by Cook (1991). They find no relationship between gun availability and the number of nonfatal violent crimes. For crimes that end in death, they generally find that greater gun availability is associated with somewhat greater use of firearms and somewhat greater rates of felony murder, but do not account for a large fraction of the variation.8 Cook (1979) finds no discernible effect of gun ownership prevalence on robbery rates. Comparing experience in Seattle and Vancouver, two neighboring jurisdictions that are demographically and socioeconomically similar but have different gun laws, Sloane et al. (1988) found no differences in event measures—burglary and simple and aggravated assault rates—but Seattle, which has more permissive gun laws, had an overall homicide rate more than 60 percent higher and a firearm homicide rate 400 percent higher. Even where positive correlations are found between measures of gun availability and nonfatal violent crime, the direction of the causal chain is unclear. Some households may be arming themselves in response to increases in violent crime rates (G. Kleck, personal communication to Jeffrey A. Roth, National Research Council, 1990). Differences across nations, states, and communities may reflect separate local traditions about guns and about violence, rather than any direct connection between guns and violence (Cook, 1991). To the extent that alienation from public institutions exists in black communities (discussed in Chapter 3), it could account for higher levels of gun ownership and of violent crime; mistrust of the police and courts could trigger a "vigilante" mentality in which citizens arm themselves in order to be prepared to settle disputes without recourse to the civil or criminal justice system. The strongest empirical evidence on how variations in firearms availability affect levels of violent crime and felony homicides can be obtained from carefully controlled evaluations of changes in laws and especially in enforcement efforts, that reduce firearms availability. As background for a such a discussion, we summarize what little is known about the sources of firearms used in violent events. 

FIREARMS AND VIOLENCE 269 Sources of Firearms According to recent estimates, only about one firearm of every six used in crimes was legally obtained. The high rate of handgun murders in cities such as New York and Washington, which have highly restricted legal access to handguns, is further evidence that guns are often obtained illegally, as is the recent increase in firearm homicides among black males ages 15-19, since minors are legally prohibited nearly everywhere from owning handguns. Most published research on illegal and unregulated sources of firearms is several years old and is, as one might expect, sketchy and unrepresentative. But at least during the 1970s and early 1980s, the path from manufacturer to gun felon involved theft and unregulated private transfers far more often than licensed dealers. The small volume of illegal sales by dealers is usually a small-scale, decentralized, off-the-books activity involving used guns rather than a trade in new guns dominated by large organizations. Theft—from licensed dealers, from residences, and from other criminals—is an important source of firearms used in felonies. Moore (1981) reports the results of manufacturer-to-user traces by the Bureau of Alcohol, Tobacco and Firearms of 113 handguns used in Boston felonies during 1975-1976. Of the guns whose histories could be traced, 40 percent were stolen at some point: 12 percent of those used in assaults, and 56 percent of those used in other crimes, which presumably were more likely to involve advance planning. Another 29 percent were purchased directly from licensed retailers (27% by legally eligible purchasers, and 2% by proscribed persons), and 20 percent involved a chain of private transfers, which are not federally regulated. In a survey of 1,874 incarcerated felons, 32 percent of those who used guns in the instant offense reported stealing them, 16 percent purchasing them from dealers, and 52 percent buying or borrowing them from private sources (Wright and Rossi, 1985). Of special interest are the 52 percent of respondents who rated "need gun to do crime" as a "somewhat" or "very" important motive for obtaining their guns. Of these, 45 percent reported stealing them and 30 percent reported buying them on black markets—compared with only 24 percent and 22 percent, respectively, among respondents for whom use in crime was rated "not important" as a motive for obtaining their guns (calculated from Wright and Rossi, 1985: Table 21). To explore the nature of the illegal market in firearms, Moore (1981) tabulated the investigative files for 131 cases of "dealing 

FIREARMS AND VIOLENCE 270 without a license” that were closed by the Bureau of Alcohol, Tobacco, and Firearms during 1974-1976; these were virtually all the cases closed in seven regional offices. According to these files, the trade is of small scale: the majority of traders had no inventories at the time of arrest, only 10 percent had more than 20 guns on hand, and the majority sold fewer than 5 guns per month. Despite the small scale of operations, the sources of illegally sold guns were diverse: thefts from residences and other dealers, legal purchases from other dealers, and unregulated purchases from private owners were all represented in substantial numbers. While this information must be considered dated and imprecise, it establishes four points about the acquisition of guns used in crimes during the 1970s and early 1980s. First, all distribution channels—legal and illegal purchases from licensed dealers, unregulated private transactions, and theft—are potentially important sources of guns used in violent crime. Second, illegal sources were most important for those acquiring guns in explicit preparation for later crimes. Third, illegal sellers of firearms tend to be small-scale independent operators rather than members of large organizations. Fourth, most people who buy guns from licensed dealers that they later use in crimes were legally entitled to make their purchases. If these findings still hold today, they suggest that illegal firearms transactions more closely resemble today's decentralized crack trade than the earlier highly organized distribution of heroin and powdered cocaine (see Chapter 4). In turn, this suggests that local-level interdiction efforts in illegal firearms traffic may reduce firearm violence more effectively than expanded federal regulation of licensed dealers and interstate commerce. This inference was drawn by Moore more than a decade ago, and it is echoed (for different reasons) in the call by participants in the December 1990 Forum on Youth Violence in Minority Communities for greater community involvement in the formulation of local firearm policy (Centers for Disease Control, 1991). However, there is as yet no strong evidence bearing on which centralized or decentralized approaches, if any, are most effective in enforcing existing firearms regulations, and on the importance of community involvement in increasing the effectiveness of enforcement. INTERVENTIONS TO REDUCE FIREARM VIOLENCE The strongest evidence on how firearms availability affects levels of violent crime and felony homicides can be obtained from 

FIREARMS AND VIOLENCE 271 carefully controlled evaluations of interventions that reduce firearms availability. This principle is true not only of interventions to modify the availability of firearms, but also of interventions to modify their uses, their allocation across owner categories, and their lethality. (This classification of interventions appears in Moore, no date; other useful classifications appear in Zimring, 1991; and Kellermann et al., 1991). The need for evaluations encompasses legal interventions; technological interventions that modify firearms, ammunition, and shields; and interventions involving public education about uses and misuses of firearms. Although substantial experience has been gained with several of these intervention strategies, there is a dearth of evaluations sound enough to permit us to learn from that experience. Table 6-1 lists legal, technological, and public education interventions classified by their objective: to alter the uses, allocation, lethality, or availability of firearms. These objectives are not mutually exclusive, and some interventions are intended to achieve several of them simultaneously. For example, the Federal Gun Control Act of 1968 contained provisions to reduce gun availability (by restricting imports) and allocation (by prohibiting sales to convicted felons, minors, and certain others). Statutory waiting periods for purchases from licensed gun dealers, which are intended primarily to affect firearms allocation (by improving enforcement of the 1968 act), may affect uses (e.g., if a prospective purchaser's passionate anger subsides during the waiting period). Theoretically, some of these intervention strategies may reinforce each other if introduced as a package. For example, mandatory waiting periods may be more effective if accompanied by interventions to disrupt illegal gun distribution channels. Or public education on safe use and storage of guns may be a useful requirement as part of an owner licensing program; however, experience with driver's education in reducing automobile deaths is not encouraging (Robertson, 1983). Registration of guns to particular owners, like automobile registration, might facilitate other interventions intended to achieve one of the listed objectives (Moore, no date). Unfortunately, there is little evidence that bears on these speculations about control strategies. The technological strategies—many of which arise from the public health approach to injury control—are relatively untested. And there are few evaluations of the effects of legal strategies embodied in existing federal, state, and local laws (Wright et al., 1983:244). The opportunities to learn from enactment and enforcement of restrictions are sometimes 

FIREARMS AND VIOLENCE 272 TABLE 6-1 Evaluation Status of Strategies and Interventions for Reducing Gun Violence Strategy and Intervention Evaluated? Effective? Strategy 1: Alter gun uses or storage Place and manner laws Restrict carrying Bartley-Fox Amendment Yes Yes Enhance sentences for felony gun use Michigan Yes Partiala Pennsylvania Yes Partiala Increase probability of sentences for felony gun use Operation Triggerlock No ? Civil/administrative laws Owner liability for damage by gun No ? Technological Enhance/maintain firearm detectability No ? Metal detectors in dangerous places No ? Enhance visibility of dangerous illegal uses No ? Shields for vulnerable employees No ? Public education Safe use and storage No ? Role in self-defense Yes ? Strategy 2: Change gun allocation Civil/administrative laws Permissive licensing of owners (e.g., all but felons, drug No ? users, minors, etc.) Waiting periods for gun purchases No ? Restrict sales to high-risk purchasers Gun Control Act of Yes No 1968 Law enforcement Disrupt illegal gun markets No ? Mandatory minimum sentences for gun theft No ? Technological Combination locks on guns No ? Strategy 3: Reduce lethality of guns Protective clothing in dangerous encounter No ? Reduce barrel length and bore No ? Reduce magazine size No ? Ban dangerous ammunition No ? 

FIREARMS AND VIOLENCE 273 Strategy and Intervention Evaluated? Effective? Strategy 4: Reduce number of guns Restrictive licensing systems D.C. Firearms Control Act Yes Yes of 1968 Restrict imports No ? Prohibit ownership No ? a Reduced gun homicides, no consistent effect on robberies, assaults, or nongun homicides. overlooked by both their advocates and their opponents in public debates. The Legal Environment As background, a brief overview follows of laws that attempt to implement the strategies in Table 6-1. The reader is referred to Wright et al. (1983) for a comprehensive overview as of a decade ago, and to Bureau of Alcohol, Tobacco and Firearms (1988, 1989) for more detailed compilations of federal, state, and local laws. Briefly, the Federal Gun Control Act of 1968 attempts to restrict imports of military weapons and to interdict interstate retail trade in all firearms by limiting legal firearms shipments to federally licensed dealers. Licensed dealers are prohibited from selling guns through the mail or across state lines except to other licensed dealers. The law also prohibits gun sales to certain categories of persons: minors, illegal aliens, felons convicted or under indictment, drug users, and former involuntarily committed mental patients. The 1968 act is enforced by the Bureau of Alcohol, Tobacco and Firearms. Bureau staffing decreased by about 24 percent between 1981 and 1983 and did not return to 1981 levels until 1989, despite its assumption of new responsibilities in arson and drug law enforcement (General Accounting Office, 1991:13-15). There have been no rigorous evaluations of the effects of these changes in enforcement resource levels. However, the General Accounting Office (1991:39) noted a recent doubling of the police waiting time for routine traces of firearms used in crimes. Also, newspaper accounts provide anecdotal reports of convicted felons and active drug dealers obtaining licenses as federal fire-arms dealers; the licenses enable them to obtain large numbers of 

FIREARMS AND VIOLENCE 274 firearms for personal use or for off-the-books sales to others (Isikoff, 1991). Federally licensed dealers are also required to comply with state and local requirements, most of which are intended to keep guns away from designated high-risk categories of persons. The most common such restriction is a requirement that gun purchasers obtain a permit or license before taking possession. In most of the 21 states with license requirements as of 1989, eligibility for a license was "permissive" (i.e., open to most categories of persons), but a few jurisdictions have adopted "restrictive" licensing requirements, with eligibility limited to a few select groups. Among the nation's most restrictive laws is the District of Columbia's 1977 Firearms Control Act, which prohibits handgun possession by virtually anyone except police officers, security guards, and those who owned their handguns before the law went into effect. Nevertheless, Isikoff (1991) reported that there were 41 federally licensed dealers in Washington, D.C. A fundamental political and ethical issue in firearms regulation is the trade- off between the effects on deaths and injuries and the costs imposed on members of law-abiding society who use guns for recreation or other legal purposes—a policy evaluation criterion noted by Zimring (1991), Wright et al. (1983), and others. Consequently, some firearm regulations apply a strategy to a limited domain: only "high-risk" firearms, firearm users, or firearm uses. Examples include bans on Saturday night specials or assault weapons (considered high-risk weapons, respectively, because of concealability and rapid-fire capability), computer systems to verify that prospective gun buyers are not members of a prohibited category, and "place and manner" laws that prohibit designated high- risk firearm uses such as carrying concealed weapons or firing them in densely populated areas. In short, there is a variety of legal, public education, and technological strategies for interventions with licensed firearm dealers, firearm owners, and the firearms themselves. For several reasons, it is impossible to predict a priori their effects on frequencies of violent events and violent deaths. The effects of strategies to alter firearm uses depend on individuals' behavioral responses to public education and to the disincentives of legal penalties; these cannot be known in advance. Interventions intended to reduce the lethality of firearms, or to reduce the availability of firearms in certain categories, may be thwarted by the substitution of other, more lethal categories; one cannot know the extent of substitution in advance. And interventions to alter 

FIREARMS AND VIOLENCE 275 the allocation or availability of firearms may be circumvented by illegal markets, to an extent that cannot be known in advance. Therefore, carefully controlled evaluations represent the only way of ascertaining intervention effects. Findings from Previous Evaluations As previously mentioned, there have been few evaluations of the legal, technological, and public education intervention strategies displayed in Table 6-1. We summarize findings from those that are available. Strategy 1: Alter Gun Uses or Storage Of the tactics listed in the table to alter gun uses or storage, three have been more or less rigorously evaluated: restrictions on carrying firearms, sentence enhancements for using guns to commit felonies, and public education concerning gun use for self-defense. Evaluations provide strong evidence that carrying restrictions and sentence enhancements reduced the harm from gun violence in the urban jurisdictions in which they were evaluated. The evaluations of gun use promotion for self-defense were of poorer quality, and so their violence control effectiveness remains to be demonstrated. Although we could find no evaluations of actual experience with technological protections against dangerous uses of firearms or with owner liability laws for gun injuries or damage, their logic seems sufficiently persuasive to warrant systematic evaluation programs. The most thoroughly evaluated restriction on gun uses is the 1974 Bartley- Fox Amendment, which expanded Massachusetts licensing procedures and mandated a one-year sentence for unlicensed carrying of firearms in public. Process evaluations suggest that the law was vigorously enforced immediately following passage, but that over the following two years police confusion about the law and growing judicial discretion in its application partially undercut its intent (Beha, 1977; Rossman et al., 1979). Early evaluations of the law's effects in Boston produced conflicting findings, but a more extensive evaluation that compared statewide trends with trends in neighboring states demonstrated rather clearly that the law decreased gun use in assaults and robberies and also decreased gun homicides during the two-year evaluation period (Pierce and Bowers, 1979). Another approach to restricting high-risk gun use is a mandatory 

FIREARMS AND VIOLENCE 276 sentence increase for felonies in which a gun is used. This approach has been evaluated in six jurisdictions, and a meta-analysis of the findings concludes that the sentence enhancements decreased gun homicides, left nongun homicide levels unchanged, and produced no consistent effect on gun robberies or assaults (McDowall et al., 1992a, b). The homicide effects are consistent with a strong instrumental effect for guns and very little substitution of other weapons in homicides. The mixed results for robberies and assaults are puzzling because the one-year sentence enhancement is a smaller relative increment to the average homicide sentence than to the average assault or robbery sentence. It may well be that a clearer picture would have emerged if the available data had permitted a separate evaluation for commercial robberies (in which guns are most likely to be used) and provided more systematic recording and classification of assaults. Currently, under Operation Triggerlock, special investigation and prosecution efforts are being employed to increase the probability of punishment for felony firearm use. Evaluation results are not available at this writing. There have been at least two evaluations of interventions to promote the use of firearms for self-defense. Both involved simple comparisons of violent crime levels before and after the interventions, with no measures taken of gunshot injuries, fatal homicides, or firearm accidents. A highly publicized antirape police initiative that trained some 2,500 women residents of Orlando, Florida, in the safe use of firearms was followed by an 88 percent decrease in the recorded rape rate the following year (Kleck and Bordua, 1983). But the recorded rape rate in one year not long before the intervention had actually dropped to zero, suggesting problems with recording accuracy. Green (1987, cited in Cook, 1991) questioned the reliability of the police department's records in evaluating its own antirape program, and McDowall et al. (1991) found the decrease well within the bounds of normal year-to-year variation at that time. Similarly, a local ordinance requiring each household to own a firearm was passed in Kenesaw, Georgia, in 1982 and was followed by a drop in burglary rates (Kleck, 1988). However, a subsequent analysis of long-term trends by McDowall et al. (1991) demonstrated that the drop was well within the bounds of year-to-year variation in Kenesaw's burglary rates. Further evaluations, preferably involving randomized experiments and multiple outcome measures, are needed to yield a strong conclusion on the effectiveness of this approach. 

FIREARMS AND VIOLENCE 277 Strategy 2: Change the Allocation of Firearms Of the tactics for changing the allocation of firearms, we could find a rigorous evaluation of only one: the provision of the Federal Gun Control Act of 1968 that prohibits gun dealers from selling to certain categories of persons designated "dangerous." This provision seems to have had no significant effect on firearms injuries or deaths. Because this law focused on interstate transactions, its effects on handgun use in assaults and homicides should have been evident in New York and Boston, where the incentives for out-of-state purchases were high because of unusually restrictive local license requirements. In fact, Zimring (1975) found that, before and after the law was enacted, trends in those jurisdictions were not significantly different from trends in other large cities. The lack of effect may have been due to a lack of enforcement effort. A 1970 crackdown to enforce the federal gun law in the District of Columbia, involving a seven- to tenfold increase in enforcement agents on the street, did accompany a six-month decrease in gun homicides, while other homicides remained constant. But evaluation results for the 1968 act may also reflect the fact that purchases by ineligible persons from licensed dealers make only a small contribution to gun use in violent crimes. We consider interventions in gun markets involving unlicensed dealers in the conclusion to this chapter. Strategy 3: Reduce the Lethality of Available Firearms The third strategy is to designate certain firearms dangerous because of their concealability, firepower, or other characteristic, then either to restrict access to them by law or to make them less dangerous through technological means. Recently, the most conspicuous proposals to implement this strategy have focused on assault weapons. The term has no generally accepted precise definition but seems to apply to rifles and pistols capable of automatic or semiautomatic fire and having a military appearance (Kleck, 1991). Except for military cosmetics (e.g., plastic rather than wood stocks, nonreflective rather than reflective surfaces) and a shift from American to foreign brands, the semiautomatic rifles that have recently become more visibly associated with urban homicides differ little from semiautomatic weapons that have been available for decades. This fact, which receives little attention from either advocates or opponents of 

FIREARMS AND VIOLENCE 278 special restrictions on these weapons, has complicated the problem of drafting legislation. There is controversy over how much more lethal assault weapons are than other rifles and pistols (Kleck, 1991). But this controversy has been carried out entirely in terms of laboratory observations of the firepower of particular weapons. Resolution of that controversy in laboratory settings would provide little useful information about actual use of such weapons in violent crimes. Given the imprecision of the term assault weapon, it should not be surprising that there are no generally accepted estimates of the number of such weapons in the United States, of ownership patterns, of their lethality compared with other weapons, or of their uses in crimes (Zimring, 1991). Therefore, legislative efforts to restrict their availability are proceeding with very little basis in knowledge. Even if workable restrictions on assault weapons can be specified and implemented, past experience suggests that continuing enforcement efforts would be required to avoid circumvention through illegal markets. Therefore, the needs for evaluation cited in connection with Strategy 2 apply here as well. Strategy 4: Reduce the Availability of Firearms Perhaps the most ambitious effort to reduce the number of firearms in a community was the 1977 District of Columbia Firearms Control Act, which prohibited handgun ownership by virtually everyone except police officers, security guards, and previous gun owners. Three evaluations of this intervention are available: U.S. Conference of Mayors (1980), Jones (1981), and Loftin et al. (1991). As discussed by Cook (1991), the first two evaluations indicate that, during periods of vigorous enforcement, the District of Columbia law did reduce the rates of gun robbery, assault, and homicide during the three years following implementation. The effect was especially strong for homicides arising from disputes among family members and acquaintances. The Loftin et al. (1991) evaluation found decreases of about one-fourth in D.C. gun homicides and suicides immediately after passage of the law. The effect persisted until 1988, when gun homicides associated with crack markets increased, and it was not mirrored by trends in D.C. nongun homicides or suicides, or in gun homicides or suicides in nearby suburban areas that were not subject to the law. While none of these findings should be considered conclusive or universally applicable, their convergence—despite different approaches—suggests 

FIREARMS AND VIOLENCE 279 that local restrictive licensing laws, when enforced, may reduce firearm homicides and warrant evaluations in other communities. RESEARCH AND EVALUATION NEEDS Violence involving firearms exacts a large toll in terms of deaths, injuries, and monetary costs. The risk of firearm violence is high and rising, especially among young black males. Research is needed to determine answers to a number of questions about the multiple complex relationships between firearms and violence. Ignorance is especially profound concerning so-called assault weapons, a poorly defined category. One or more surveys are needed to develop accurate estimates of ownership by gun type, of motives and sources for obtaining guns, and especially of gun acquisition patterns among juveniles and criminals. Case control, ethnographic survey, and other studies are needed of robberies and assaults involving guns to discover the risk factors associated with actual shootings and the role of guns in self-defense. Existing data are inadequate for measuring the annual incidence of self- defense gun use. The National Crime Survey is subject to event threshold and recall problems, and other estimates are subject to definitional ambiguities and other methodological problems that cannot be solved without resorting to dubious adjustment factors. For informing individuals' choices about gun ownership, however, the more important questions are how often and under what circumstances intended victims deploy their guns in self-defense and how the deployment affects the outcome of the event. These questions bear further analysis not only through victimization surveys but also through analyses of event records. Analyses of natural variation have established that increased firearm availability is associated with increased firearm use in violent crime. But the methods used to date have been insufficient to answer a more basic question: Do changes in gun availability cause changes in violence levels overall? To answer this question, evaluations are needed of the effects of gun supply reduction initiatives on violence levels in particular geographic areas. Educational, technological, and legal strategies can be devised with the objectives of changing how handguns are used and stored, changing their allocation from higher-risk to lower-risk segments of the population, reducing their lethality, or reducing their numbers. For any of these policies to reduce homicides, they must 

FIREARMS AND VIOLENCE 280 reduce violent uses of at least some types of guns to an extent that is not offset by substitution of other more lethal weapons. It is not clear in advance whether this condition would be met for any policy focused on handguns. While intentional firearm injuries are two to five times as lethal as injuries from knives, the next most lethal weapon, it is not known how much of this difference in lethality is due to weapon characteristics and how much to more lethal intentions by those who attack with guns. We also do not know the extent to which long guns or assault weapons would be substituted if handguns became less available, or indeed whether those weapons are more lethal than handguns when used in violent crimes. Without a priori knowledge, an assessment must depend on evaluations of experience, which are conditionally encouraging. Examples exist of reductions in violent gun uses caused by legal restrictions on carrying guns, by enhanced sentences for felony gun use, and by restrictive licensing of firearms in particular jurisdictions. However, the success of legal strategies depends critically on the nature and intensity of enforcement efforts, since illegal or unregulated gun transfers supply most of the guns used in crimes. Enforcement efforts directed against illegal gun markets combine elements of the first three strategies shown in Table 6-1. Also, unlike interventions involving licensed dealers and gun owners, most of whom are law abiding, disruption of illegal markets focuses on persons who are violating laws. Interventions against illegal drug markets should be more effective if they are informed by more specific research on the nature of those markets—their size, their distribution channels, and their retail-level marketing techniques. As a starting point, however, efforts to disrupt local illegal gun markets might reasonably borrow tactics that are currently being used against illegal drug markets. These could include buy-bust operations,9 high-priority investigation and prosecution of alleged unregulated gun dealers, the development of minors arrested in possession of guns as informants against gun sources, phony fencing operations for stolen guns, high-priority investigations and prosecutions of burglaries and robberies in which guns are stolen, and high mandatory minimum sentences for those who steal or illegally sell guns. In Chapter 4 we suggest that the success of street-level disruption of drug markets may depend on the extent of local community support. Community support was also cited as a necessary ingredient of firearm policy development at the December 1990 

FIREARMS AND VIOLENCE 281 Forum on Youth Violence in Minority Communities (Centers for Disease Control, 1991). This suggests that community-oriented or neighborhood-oriented police work involving close coordination with community-based organizations may be useful in reducing residents' motivations to acquire guns for self-defense, in encouraging community members to notify police anonymously about illegal gun transactions and persons illegally carrying guns, and in supporting market disruption tactics when necessary. In view of the lack of available evidence, a high priority should be placed on evaluations of the violence-reduction effects—on gun and nongun violence and on illegal and unregulated gun transfers—of interventions to reduce illegal sales of firearms. Specifically, high priority should be placed on evaluating the effects of three intervention strategies: • disrupting illegal gun markets using the centralized and street-level tactics currently in use for disrupting illegal drug markets; • enforcing existing bans on juvenile possession of handguns; and • community-oriented or neighborhood-oriented police work involving close coordination with residents and community-based organizations. At this writing, the Department of Justice is expanding its pilot testing of “Operation Weed and Seed"—a program under which these and other initiatives involving police-community cooperation could be systematically evaluated. Rigorous evaluations should also be made of the effects of strategies other than law enforcement interventions: public education interventions to change how guns are stored and used and technological interventions to reduce the lethality of firearms. To the extent that consideration is given to reducing the availability of assault weapons through new statutes, there is a need to reach consensus on a definition of that term, and then to ascertain their level of violent use, their lethality in actual use, and the effects of alternative enforcement strategies in reducing the violence associated with them. NOTES 1 For legibility, trends in the 1984-1988 firearm homicide rate for black males are not displayed in Figure 6-3 for all available age groups. The rate for ages 15-19 increased from 30 to 68 per 100,000, a 126 percent increase. The corresponding increases were from 2.1 to 4.5 (114%) for 

FIREARMS AND VIOLENCE 282 ages 10-14; from 63 to 104 (65%) for ages 20-24; from 70 to 90 (29%) for ages 25-29; and from 62 to 69 (11%) for ages 30-34 (Fingerhut et al., 1991). 2Cook has estimated that for each new cohort of 100 guns, 33 uses in crime are reported. But because the frequency of repeat uses is unknown, the fraction that is ever used in this way is unknown. 3 Wright et al. (1983) point out that in some of the 80 percent, the shooter may have shot a second time but missed, and in others the killer may have failed to shoot a second time because the victim appeared to be dead already. It seems likely that both explanations apply to some incidents but that even together they are insufficient to account for all one-shot incidents. 4Because National Crime Survey data include only nonfatal victimizations, the publication does not report conditional probabilities of being hospitalized or killed. Because deaths are so rare, however, those probabilities would not be substantially altered. 5 Respondents to one of the surveys of the uses of self-defense (Decision-Making Information, Inc., 1979) stated that only 31 percent of the incidents involving gun use for self-defense were "important enough to report to the police." 6 As one example, in a survey by Decision-Making Information (1979, cited in Wright et al., 1983:96), 15 percent of respondents reported "ever" using a gun for self-protection. To convert this lifetime prevalence to an annual prevalence, Kleck multiplies it by the ratio 1.4/8.6, the ratio of annual to lifetime prevalence rates for self-defense uses obtained in another survey and obtains an estimate of 897,000 annual self-defense uses. Yet elsewhere in the same publication, he reports that the mean duration of gun ownership in a representative sample of gun owners is 23.4 years. If one accepted the 1.4 percent rate as accurate but assumed that self-defense uses were distributed uniformly over the life of each gun, the implied adjustment ratio would be 1.4/23.4, which would reduce Kleck's annualized estimate to 329,000. As another example, Kleck (1988:5 and Table 2) adjusts justifiable civilian homicides with guns for assumed underreporting in the Uniform Crime Reports Supplementary Homicide Report. The adjustment involves 57 "other civilian legal defensive homicides" (OCLDHs) in Detroit and leads to a national estimate of 2,819 total firearm OCLDHs for 1980. Yet the estimation method for the Detroit OCLDHs is not explained, their cited source does not appear in the reference list, and no more than one OCLDH was estimated for any of the other five large jurisdictions included in Kleck's Table 2. Omitting the Detroit OCLDHs from the adjustment would reduce the national estimate from 2,819 to 1,711. 7 Because some incidents of self-defense with a firearm in events outside the home entail illegal firearms carrying, the rates for assaults and robberies may be underreported to interviewers. However, given the difficulty of gun deployment during a surprise robbery or assault, it seems 

FIREARMS AND VIOLENCE 283 unlikely that the rates for assaults and robberies are as high as those for residential burglaries, in which self-defense with a gun is generally legal. 8 Based on a multivariate analysis, Cook (1979) finds that a 10 percent reduction in the prevalence of gun ownership in a city is associated with about a 5 percent reduction in the gun robbery rate and a 4 percent reduction in the robbery murder rate. Killias (1990, cited in Cook, 1991) reports a correlation of .72 between the gun fraction in homicide and the prevalence of gun ownership for 11 countries. 9In a buy-bust operation, a dealer is arrested following one or more illegal sales to undercover officers, which are usually recorded on videotape for use as evidence. References Beha, J.A., III 1977 "And nobody can get you out": The impact of a mandatory prison sentence for the illegal carrying of a firearm on the use of firearms and the administration of criminal justice in Boston. Boston University Law Review 57:96-146, 289-233. Bureau of Alcohol, Tobacco and Firearms 1988 (Your Guide to) Federal Firearms Regulation, 1988-89. Bureau of Alcohol, Tobacco and Firearms publication No. 5300.4. Washington, D.C.: U.S. Government Printing Office. 1989 State Laws and Published Ordinances—Firearms. Washington, D.C.: U.S. Government Printing Office. 1991 How many guns? ATF News (Press release). May 22. Bureau of Justice Statistics 1986 The Use of Weapons in Committing Crimes. Washington, D.C.: U.S. Government Printing Office. 1989 Criminal Victimization in the United States, 1987. Washington, D.C.: U.S. Bureau of Justice Statistics. Centers for Disease Control 1991 Forum on Youth Violence in Minority Communities: Setting the Agenda for Prevention, December 10-12, 1990. Summary of the proceedings. Public Health Reports 106(3):225-277. Cook, P.J. 1979 The effect of gun availability on robbery and robbery murder: A cross section study of fifty cities. Policy Studies Review Annual 3. 1980 Reducing injury and death rates in robbery. Policy Analysis 6(1):21-45. 1981 Guns and crime: The perils of long division. Journal of Policy Analysis and Management 1:120-125. 1985 Report on a City-Specific Gun Prevalence Index. Mimeographed. Durham, N.C.: Duke University, Institute of Policy Sciences. 1991 The technology of personal violence. Pp. 1-72 in M. Tonry, ed., 

FIREARMS AND VIOLENCE 284 Crime and Justice: A Review of Research. Vol. 14. Chicago: University of Chicago Press. Curtis, L.A. 1974 Criminal Violence: National Patterns and Behavior. Lexington, Mass.: D.C. Heath. Decision-Making Information, Inc. 1979 Attitudes of the American Electorate Toward Gun Control. Santa Ana, Calif.: Decision-Making Information, Inc. Downs, G.W. 1991 Arms race and war. Pp. 73-109 in P.E. Tetlock, J.L. Husbands, R. Jervis, P.C. Stern, and C. Tilly, eds. Behavior, Society and Nuclear War. New York: Oxford University Press. Federal Bureau of Investigation 1989 Crime in the United States, 1988: Uniform Crime Reports. Washington, D.C.: U.S. Government Printing Office. Fingerhut, L.A., J.C. Kleinman, E. Godfrey, and H. Rosenberg 1991 Firearm mortality among children, youth, and young adults 1-34 years of age, trends and current status: United States, 1979-88. Monthly Vital Statistics Report 39(11, Supplement). Hyattsville, Md.: National Center for Health Statistics. General Accounting Office 1991 BATF: Management Improvements Needed to Handle Increasing Responsibilities. Report to the Chairman, Subcommittee on Oversight, Committee on Ways and Means, House of Representatives. GAO publication No. GGD-91-67. Washington, D.C.: U.S. General Accounting Office. Green, G.S. 1987 Citizen gun ownership and criminal deterrence: Theory, research, and policy. Criminology 25:63-82. Isikoff, Michael 1991 Gun dealers' "great scam": U.S. licenses grow popular with criminals. Washington Post May 8, pp. A1, A8. Jones, E.D., III 1981 The District of Columbia's Firearms Control Regulations Act of 1975: The toughest handgun control law in the United States—or is it? Annals of the American Academy of Political and Social Science 455:138-149. Kellermann, A.L., and D.T. Reay 1988 Protection or peril? An analysis of firearm related deaths in the home. New England Journal of Medicine 319:1256-1262. Kellermann, A.L., R.K. Lee, J.A. Mercy, and J. Banton 1991 The epidemiologic basis for the prevention of firearm injuries. Annual Review of Public Health 12:17-40. Killias, Martin 1990 Gun ownership and violent crime: The Swiss experience in international perspective. Security Journal 1(3):169-174. 

FIREARMS AND VIOLENCE 285 Kleck, G. 1984 The relationship between gun ownership levels and rates of violence in the United States. In D.B. Kates, Jr., ed., Firearms and Violence: Issues of Public Policy. Cambridge, Mass.: Ballinger. 1988 Crime control through the private use of armed force. Social Problems 35:1-22. 1991 Point Blank: Guns and Violence in America. New York: Aldine de Gruyter. Kleck, G., and D.J. Bordua 1983 The factual foundation for certain key assumptions of gun control. Law and Policy Quarterly 5:271-298. Kleck, G., and K. McElrath 1989 The Effects of Weaponry on Human Violence. Mimeographed. Tallahassee, Fla.: Florida State University, School of Criminology. Loftin, C., d. McDowall, B. Wiersema, and T.J. Cottey 1991 Effects of restrictive licensing of handguns on homicide and suicide in the District of Columbia. New England Journal of Medicine 325(December 5):1615-1620. McDowall, D. 1991 Firearm availability and homicide rates in Detroit, 1951-1986. Social Forces 69 (4):1085-1101. McDowall, D., A.J. Lizotte, and B. Wiersema 1991 General deterrence through civilian gun ownership: An evaluation of the quasi-experimental evidence. Criminology 29(4):541-559. McDowall, David, Colin Loftin, and Brian Wiersema 1992a A comparative study of the preventive effects of mandatory sentencing laws for gun crimes. Journal of Criminal Law and Criminology , forthcoming. 1992b Preventive effects of mandatory sentencing laws for gun crimes. Pp. 87-94 in Proceedings of the Social Statistics Section. Annual meeting of the American Statistical Association, Atlanta, Ga., August 18-22, 1991. Alexandria, Va.: American Statistical Association. Moore, M.H. 1981 Keeping handguns from criminal offenders. The Annals of the American Academy of Political and Social Science 455:92-109. No date Alternative Policy Approaches to the Control of Handgun Abuse. Unpublished manuscript available from the Kennedy School of Government , Harvard University. Newton, G.D., Jr., and F.E. Zimring 1969 Firearms and Violence in American Life. Washington, D.C.: U.S. Government Printing Office. Phillips, L., H.L. Votey, Jr., and J. Howell 1976 Handguns and homicide: Minimizing losses and the costs of control. Journal of Legal Studies 5:463-478. 

FIREARMS AND VIOLENCE 286 Pierce, G.L., and W.J. Bowers 1979 The Impact of the Bartley-Fox Gun Law on Crime in Massachusetts. Unpublished manuscript. Boston: Northeastern University, Center for Applied Social Research. Rice, Dorothy P., Ellen J. MacKenzie, and Associates 1989 Cost of Injury in the United States: A Report to Congress. San Francisco, Calif.: Institute for Health and Aging, University of California and Injury Prevention Center, The Johns Hopkins University. Robertson, L.S. 1983 Injuries: Causes, Control Strategies, and Public Policy. Lexington, Mass.: Lexington Books. Rossman, D., P. Froyd, G.L. Pierce, J. McDevitt, and W. Bowers 1979 The Impact of the Mandatory Gun Law in Massachusetts. Boston University School of Law: Center for Criminal Justice. Sloane, J.H., A.L. Kellermann, D.T. Rey, J.A. Ferris, T. Koepsell, F.P. Rivara, C. Rice, L. Gray, and J. LoGerfo 1988 Handgun regulations, crimes, assaults, and homicide: A tale of two cities. New England Journal of Medicine 319:1256-1262. U.S. Conference of Mayors 1980 The Analysis of the Firearms Control Act of 1975: Handgun Control in the District of Columbia. Mimeographed. Washington, D.C. van Dijk, Jan J.M., Pat Mayhew, and Martin Killias 1990 Experiences of Crime Across the World: Key Findings From the 1989 International Crime Survey. Deventer, The Netherlands: Kluwer Law and Taxation Publishers. Vinson, T. 1974 Gun and knife attacks. Australian Journal of Forensic Science 7(2):76-83. Wintemute, G.J., S.P. Teret, J.F. Kraus, and M.A. Wright 1987 When children shoot children. 88 unintentional deaths in California. Journal of the American Medical Association 257:3107-3109. Wolfgang, M. 1958 Patterns in Criminal Homicide. New York: Wiley. Wright, J.D., and P.H. Rossi 1985 The Armed Criminal in America: A Survey of Incarcerated Felons. National Institute of Justice Research Report (July). Wright, J.D., P.H. Rossi, and K. Daly 1983 Under the Gun: Weapons, Crime and Violence in America. Hawthorne, N.Y.: Aldine Publishing Company. Zimring, F.E. 1968 Is gun control likely to reduce violent killings? The University of Chicago Law Review 35:721-737. 

FIREARMS AND VIOLENCE 287 1975 Firearms and federal law: The gun control act of 1968. Journal of Legal Studies 4:133-198. 1989 The problem of assault firearms. Crime & Delinquency 35(4):538-545. 1991 Firearms, violence, and public policy. Scientific American (November):48-54. Zimring, F.E., and G. Hawkins 1987 The Citizen's Guide to Gun Control. New York: Macmillan Publishing Company. Zimring, F.E., and J. Zuehl 1986 Victim injury and death in urban robbery: A Chicago study. Journal of Legal Studies XV(1):1-40. 

FIREARMS AND VIOLENCE 288