Law and Policy Quarterly
Issue on Firearms and Firearms Regulation:
Old Premises, New Research
vol. 5, no. 3, 1983: 271.
Posted for Educational use only. The printed edition remains canonical. For citational use please visit the local law library or obtain a back issue.

THE FACTUAL FOUNDATION FOR CERTAIN KEY ASSUMPTIONS OF GUN CONTROL

GARY KLECK
Florida State University

DAVID J. BORDUA
University of Illinois, Urbana

The case for legal restrictions on gun ownership and use as a strategy for reducing criminal violence relies on factual assumptions about the nature of gun ownership and violent behavior. Five of the most crucial ones are identified and subjected to a comparison with the available empirical evidence. All of the following assumptions were found to be substantially at variance with the evidence: (1) Guns are five times deadlier than the weapons most likely to be substituted for them in assaults in which guns are not available. (2) The sight of a gun can elicit aggression, due to the learned association between guns and violence. (3) If guns are made more expensive, more difficult to obtain, or legally risky to own, people will do without them. (4) Guns are useless for self-defense or protection of one’s family, home, or business, and have no deterrent effect on criminals. (5) Homicides are largely "crimes of passion" committed by otherwise law-abiding citizens not distinguishable from other people. Therefore, control must be directed at all gun owners rather than select criminal subgroups.

The gun control issue is complex, involving dozens of interrelated subissues and disputes. Some of the disputes are value disputes or conflicts over fundamental beliefs--a clash of cultures (Wall Street Journal, 1972; Bruce-Briggs, 1976). As such, the issues cannot be [Page 272] resolved solely on the basis of research evidence. However, many of the arguments for gun control depend on certain specific assumptions, sometimes explicitly stated, often left implicit, which can be evaluated on logical grounds, and compared against the available research evidence. We have identified five of the more important assumptions of this sort, and have attempted to subject them to this kind of examination. We take a predominantly critical stance toward these assumptions, because the dominant stance among many members of the academic research community has been one unusually uncritical of gun control policies, in sharp contrast to their ordinarily skeptical view of other governmental policies restricting human behavior in one way or another.

The term "gun control" is very broad, referring to anything from increased penalties for use of guns in a felony to a total ban on ownership of firearms. There are dozens of basic gun control policies and thousands of possible combinations of these policies. Some are directed at ownership, others at illegal use, some at handguns, others at all firearms. We will concentrate in general on policies aimed at restricting or banning ownership of firearms, especially (but not exclusively) policies directed at handguns.

ASSUMPTION ONE

Guns are five times deadlier than the weapons most likely to be substituted for them in assaults in which guns are not available.

This assumption is crucial to gun control arguments because opponents of gun control measures have claimed that where guns are unavailable, other weapons will be substituted for them and homicides will be committed with the alternative weapons at the same rate as would have occurred with guns available. Gun control advocates counter this argument by saying that the substituted weapons will be less technically effective than firearms for inflicting fatal injury, resulting in a lower assault fatality rate and therefore fewer homicides. Two assumptions are involved in the substitution argument of gun control advocates. The first is that knives are the most deadly of the likely substitute weapons and therefore would produce the largest possible substitution effect, and the second is that guns are approximately five times deadlier than knives.

If guns become harder to obtain, or riskier to own, those who feel the greatest need to own guns will be the ones most likely either to retain [Page 273] their guns or to obtain the best available substitute they can afford. If a control policy directed at all guns is under consideration, knives might well be the most common substitute, since they seem to be the next most effective available weapon among those which could be used in the same sort of circumstances as guns. However, many of the policies currently advocated are directed solely at specific types of guns, especially handguns, or even more narrowly, the cheap handguns known as "Saturday Night Specials." If denied one defensive device, a rational, highly motivated person would presumably acquire the next best substitute device. If handguns become harder to get, the next most satisfactory weapon, either for self-defense or for committing crimes, would be a rifle or shotgun, not a knife. These weapons are certainly more expensive than knives, but are also much more effective for the person who desires a weapon because he or she feels unable to physically resist the average robber or rapist, who is most likely to be a strong male. Therefore, restriction of handgun ownership could result in a shift to rifles and shotguns for defensive purposes among those who are highly motivated (on this point see Kates, 1976; Kleck, forthcoming), and also to the use of sawed-off versions of these weapons for criminal purposes. If these weapons are deadlier than handguns (especially at close range, where most assaults occur), such a shift in weapon type would amount to an upgrading of weaponry, and would tend to result in a higher assault fatality rate. In a similar way, effective restrictions on the availability of cheap, small caliber handguns could cause a shift to more expensive handguns of better construction and larger caliber. Since larger caliber guns are deadlier (Zimring, 1972), this policy could also result in a higher assault fatality rate. Whether handgun restrictions would result in a net increase in the assault fatality rate would depend on what proportion of prospective assaulters would substitute knives for handguns, and what proportion would substitute long guns. Benenson and Kates (1979: 111) estimated that even if only 30% switched to long guns and the remaining 70% switched to knives, there would still be a substantial net increase in homicides. (See Kleck, 1983, for an extended analysis of this matter.)

While it may well be that firearms are deadlier in assaults than knives, it is debatable just how much of the greater deadliness is due to the technical characteristics of the weapons and how much is due to differences in the intentions and intensity of motivation of the people who use the weapons. It may be the case that people who are more serious about committing deadly violence for that reason choose more "serious" weapons. However, Zimring (1968) claimed that not only are firearms five times deadlier than knives, but that the difference in gun [Page 274] and knife fatality rates cannot be attributed to differences in motivation or intention of the weapon’s user. As evidence of this latter claim, he purported to show that gun and knife assaulters, described in Chicago police records for 1967, were similar in motive, race, sex, and bodily location of the wounds they inflicted. Despite Zimring’s assertion, by recomputing his figures (1968: Table 5) according to sex, one finds that 87.3% of gun assaulters were male, while only 65.2% of the knife assaulters were male. The relationship between gender and violence is well known, and it is not implausible that male assaulters as a group are more intent on inflicting deadly violence than female assaulters. Weapon preference may be affected by sex-role-structured prior experience with, and attitudes toward, firearms. Sex, in turn, may be related to seriousness of intent.

Zimring’s early work does not allow detailed study of large numbers of gun and nongun assaults which are comparable in presumed degree of intent to kill, since they concern rather heterogeneous samples of assaults. However, a later study (Zimring, 1977) examined only assaults in robberies, presumably a much more homogeneous sample. It indicated that guns were but 1.31 times as deadly as knives in armed robbery assaults (based on police data). Further, a medical study that concerned only abdominal wounds found a 3.1% mortality rate for stab wounds and 9.8% for gunshot wounds, indicating a three-to-one ratio (Wilson and Sherman 1961: 640). Thus, even using fatality rates in the Zimring manner to measure the relative deadliness of different weapons leads to weaker conclusions than Zimring reached. However, the technique is fallacious in any case, since it erroneously assumes comparability of motives and intentions between users of different types of weapons. These considerations suggest that if knives are substituted for guns as a result of an effective gun control program, the savings in lives will be considerably less than would appear if the five-to-one deadliness ratio is believed.

ASSUMPTION TWO

Firearms ownership increases the rate of assaults because the sight of a gun can elicit aggression due to the learned association between guns and violence.

This assumption implies that not only does firearms use in assaults increase the deadliness of those assaults, but that the rate of assaults will [Page 275] also be higher because some assaults that would not otherwise have occurred will be stimulated by the presence of a gun. In two articles in Psychology Today, Leonard Berkowitz made the argument explicit and summarized it with a slogan repeated by others since: "Guns not only permit violence, they can stimulate it as well. The finger pulls the trigger, but the trigger may also be pulling the finger" (Berkowitz, 1968: 22; see also Berkowitz, 1981). Elsewhere, Berkowitz has argued that stimuli commonly associated with aggression, such as guns, can elicit aggression from people ready to aggress (i.e., angry people) when the stimuli are associated with an available target. By a process of classical conditioning, the repeated pairing of guns and aggression in real life and in fiction, creates an association between guns and aggression. As a result, angered persons may respond with aggression when presented with the stimulus, guns. In addition to causing assaults that might not otherwise have occurred, guns may also cause increased intensity of attack, Berkowitz argued. The theoretical rationale for this notion was never made clear and Berkowitz seems to have dropped the idea since the original Berkowitz and Le Page (1967) article on the subject.

The "weapons effect" studies are nearly all experimental studies, usually conducted in laboratories. Typically, confederates of the experimenters in some way anger the subjects, who are then given an opportunity to aggress against the confederates using electrical shocks, supposedly in the context of a "learning experiment." The key experimental condition is the presence of a weapon (usually a gun), toy weapon, or picture of a weapon, which either is or is not associated with the confederate.

Berkowitz and Le Page produced marginal support for the gun effect hypothesis. The weapons effect was observed for strongly angered subjects, but not for weakly angered subjects; significant differences between control and experimental groups were observed for number of shocks given, but not for mean duration of the shocks.

Further research has elaborated, and to some extent undercut, Berkowitz’s original theoretical framework in several important ways. Studies have differentiated between groups that showed the weapons effect and other groups that did not, noted the importance of the differing meanings people attach to guns, and more fully recognized the possibility of guns inhibiting aggression as well as eliciting it.

Turner and his associates (1975) thought that many people may not perceive guns as aggressive stimuli, especially if they have frequently been observed in nonaggressive contexts such as hunting or target shooting. They devised a naturalistic experiment in which a pickup [Page 276] truck driven by a confederate would deliberately fail to move at a traffic light when the light turned green, obstructing traffic behind him. Horn-honking by the drivers of the cars immediately behind the truck was the measure of aggression. The truck sometimes had a rifle in a gun rack, which was clearly visible from behind the truck, and sometimes did not. The rifle was paired with a large bumper sticker on the truck with either an aggressive connotation (the word "vengeance"), or a nonaggressive connotation (the word "friend"). Significantly more honking occurred when the rifle was given the aggressive connotation than when it was not given such a connotation. Further, the rifle paired with the nonaggressive meaning did not produce significantly more aggression than the no-rifle control condition (a fact Berkowitz unaccountably fails to mention in his 1981 Psychology Today discussion of this study). The validity of horn-honking as a measure of aggression or its comparability to physical violence is unknown, and unfortunately the effect of gun meaning on the weapons effect has not been empirically evaluated with any other measure of aggression.

Given that virtually all of the personal experience with guns that most gun owners have is in predominantly nonaggressive recreational activities, these findings suggest that the weapons effect is largely limited to either people who do not own guns or to gun owners whose experience with guns is limited to circumstances of real-life aggression and/or to fictional violence (especially on television or in films).

Four experimental studies have produced findings largely inconsistent with the weapons effect hypothesis--Ellis et al. (1971), Page and Scheidt (1971), Buss et al. (1972), and Turner and Simons (1974)--although the authors of the last study chose not to emphasize the negative weapons effect findings in their Table 1. In other studies, Fischer et al. (1969) found only a small (and statistically insignificant) weapons effect, and that only for people of low emotionality, while Turner and his colleagues (1975) found weapons effects only in a minority of experimental conditions and obtained results indicating inhibition of aggression more often than stimulation of aggression. On the other hand, the findings of Frodi (1973), Leyens and Parke (1975), and Page and O’Neal (1977) largely support the hypothesis. However, Leyens and Parke used pictures of guns rather than actual weapons as stimuli and used as a measure of aggression the number of shocks subjects said they wanted to give to the confederates who had insulted them. The artificiality of these conditions and the dubious validity of the aggression measure makes the findings of questionable generalizability. Further, the [Page 277] Leyens and Parke study was conducted in Belgium and that of Frodi in Sweden. Since Europe has little tradition of widespread participation in gun-related recreational activities such as hunting--see Kennett and Anderson (1975) on the contrast between European and U.S. traditions of gun use--most European subjects are likely to have real-life experience with firearms only in the context of the military, or warfare, if at all. Otherwise, their experience will have been limited to the fictional and largely aggression-laden contexts of television and films. Therefore, these studies may be of limited relevance to an evaluation of the plausibility of the weapons effect hypothesis in the United States.

The social psychologist critics and defenders of the weapons effect hypothesis have clashed with each other primarily over technical issues. These largely inconclusive discussions focus on whether findings were due to demand characteristics of the experiments, such as subjects’ awareness of experimenter’s expectations and subjects’ anxiety at being evaluated on their aggressiveness. However, a more fundamental criticism can be made of almost all of these studies. In nearly all experimental studies of the weapons effect, the weapon is either associated with the potential victim of the aggression (the confederate) or is not associated with anyone in the experimental situation. Weapons were almost never in the possession of or associated with the potential aggressor (the subject). Yet the principal issue of relevance to gun violence is whether the aggressor’s possession of a gun makes his or her physical aggression more likely, not whether it makes the potential victim’s aggression more likely. Thus, the social psychological literature does not address itself directly to the issue of gun owner aggression at all, but rather to the tangential one of aggression directed against gun owners. And in the one study in which guns were linked to the experimental subjects (Buss et al., 1972), no weapons effect was found (subjects in this study fired BB guns before being evaluated for aggression).

In a real-life setting of potential violence, in which one person has a gun and the other does not, it seems highly likely that any potential aggression of the other will be inhibited by the fear of the consequences of assaulting the person with the gun far more than it will be stimulated by the sight of the gun. Consistent with this point, Fischer et al. (1969), Fraczek and Macaulay (1971), and Turner et al. (1975) obtained results indicating significant inhibiting effects of weapons (knives in the Fischer et al. study and guns in the other two studies). Fischer et al. found inhibiting effects for women, while Turner and his associates found inhibiting effects for men and women in a number of experimental [Page 278] conditions. Fraczek and Macaulay found significant inhibiting effects of guns on highly emotional subjects, possibly because such people have learned to fear the possible consequences of their own aggression. These findings, combined with the mixed findings regarding the eliciting of aggression, suggest that, for the population as a whole, guns are as likely to inhibit assaults as to incite them, and that gun ownership therefore has no net effect on the frequency of assaults.

ASSUMPTION THREE

People are only superficially motivated to acquire and own guns. Therefore, if guns are made more expensive, more difficult to obtain, or legally risky to own, people will do without them (i.e., the demand for guns is highly elastic).

The demand for guns is most elastic among those gun owners least motivated to acquire and retain them. If we assume that those motivated by fear of crime are on the average more highly motivated than those motivated by desire for recreation--in hunting or target shooting--then demand for guns is least elastic among those who own guns for self-defense. Consequently, we would expect the resistance to policies restricting firearms to be strongest among the most highly motivated defensive gun owners.

A large proportion of gun owners own guns for the purpose of protection or self-defense. A 1975 national survey found that for 55% of all gun owners, self-defense was at least one of the reasons they owned a gun, although some owners gave other reasons in addition to this one (U.S. Congress 1975: 9). In two national surveys conducted in 1978, 21% and 25% of all gun owners said self-defense was the most important reason they own a gun. Among the handgun owners, 45% owned their guns for this reason (Decision Making Information, 1979: 40). A 1977 survey of Illinois residents indicated that among persons who owned only handguns, 57% owned them exclusively for the purpose of protection, while another 10% indicated protection was their main purpose (Bordua et al., 1979: 231). In a Florida survey, 54.5% of handgun owners said protection was their primary purpose for purchasing the gun. Similar results were obtained in a survey of California handgun owners (California Bureau of Criminal Statistics, 1977).

Given the frequency of defensive ownership of firearms, it would not be surprising if compliance with laws restricting gun ownership would [Page 279] meet with widespread resistance and a low rate of compliance. This expectation is confirmed by survey data regarding anticipated rates of compliance. The Illinois survey asked respondents if they would comply if a law were passed requiring people to turn in their firearms to the federal government; 73% of gun owners stated they would not comply (Bordua et al., 1979). Further, the general public does not believe compliance with such a law would be very great. Fully 95% of a general national sample of adults questioned in 1978 believed that only half or fewer of gun owners would comply with a law requiring a turn-in of handguns to the federal government. The same survey found that 71% of the general public believed that even with a registration of guns, half or fewer of gun owners would comply (DMI, 1979: 66). These data suggest that gun ownership for self-defense, especially handgun ownership, is highly inelastic and that voluntary compliance with restrictions of ownership would be discouragingly low, even in the general, predominantly noncriminal gun-owning population. Presumably voluntary compliance among criminal gun owners would be far worse.

It is a truism that restricting ownership of firearms can have an effect on the homicide rate only to the extent that ownership is reduced or limited among those who are violence prone. This must be true, since everyone who commits a homicide is by definition violence prone, regardless of whether killers could be identified as violence prone in advance of their killings. Therefore, it is crucial to know to what degree gun control laws will limit gun ownership within this group.

It can be hypothesized that it is among the highly motivated defensive gun owners where killers are more likely to be found. This is because, we would argue, killers often perceive themselves as potential victims (see Toch, 1969, for a related view). This should not be surprising since the distinction between the killer and the victim in a homicidal episode often is simply a matter of who strikes the last or hardest blow in a mutual exchange of blows, or of who happens to first introduce a deadly weapon into the exchange (Wolfgang, 1958; Luckenbill, 1978). Many killers actually come close to being victims themselves in exchanges initiated by persons who ultimately become the homicide victim. Further, if killers and victims are often both members of a subculture of violence, as Wolfgang and others have argued, and if they both reside in high crime areas, where risks of victimization are high, it is to be expected that many gun acquisitions by people who eventually become killers are initially made for defensive reasons. Few homicides are premeditated (Wolfgang and Ferracuti, 1967: 141); thus few guns are [Page 280] purchased with the goal in mind of killing a particular individual. Rather, it is reasonable to believe that the weapons were originally acquired for defensive purposes, and only later were used to kill. Indeed, in a sample of Florida prisoners convicted of gun murder, 58.5% had originally acquired their guns for protection (Florida Bureau of Criminal Justice Planning and Assistance, 1977: 4-10).

Therefore, it is among violence-prone people that demand for guns is most inelastic, and it is they who would be the most likely to either violate gun laws or seek effective substitutes for prohibited weapons, whether they be handguns in general, or Saturday Night Specials in particular. This demand would not, for the most part, derive from an intent to use the weapons for criminal purposes; rather, it would be the result of a felt need for protection in an environment accurately perceived to be dangerous. Assuming this analysis is correct, laws aimed at restricting gun ownership will have the least success in the neighborhoods where they most need to succeed in order to produce a reduction in homicide.

ASSUMPTION FOUR

People who buy guns for self-defense are the victims of self-deception and a mistaken belief in the protective efficacy of gun ownership. In fact, guns are useless for self-defense or protection of a home or business, and have no deterrent effect on criminals.

Persons who own guns for defensive purposes, unless they are totally irrational, clearly must at least believe their guns are useful for self-defense, regardless of the actual protective efficacy of guns. Consistent with this point, 83% of a 1978 national sample agreed with the statement, "Most people who have guns feel safer because of it" (DMI, 1979: 43). Gun owners derive at the very least the real, albeit intangible psychological benefit of decreased anxiety regarding criminal victimization. This is a benefit exactly analogous to one that life insurance provides--policy holders do not purchase a policy so they can die and leave their family the insured sum. Rather, the principal benefit is peace of mind. Likewise, the chief benefit of defensive gun ownership is not the actual use of guns for defense against criminals, but rather the peace of mind produced by the knowledge that the gun is available and can be used for defensive purposes if needed. [Page 281]

Gun control advocates argue that, however real these psychological benefits may be, they nevertheless are largely illusory, because gun ownership does not in fact prevent crime victimization. For example, Rushforth et al. (1977: 537) flatly stated without any relevant supporting evidence, that "this goal of security through gun possession is illusory." The argument that guns are ineffective for self-defense has been put forth in its most complete form in the work of Matthew Yeager and his colleagues (1976). They compiled diverse bits and pieces of evidence, some relevant, but much of it bearing little or no relationship to their stated issue of the efficacy of handguns in preventing crime. Because their work is cited by others, it is important to evaluate their arguments and evidence where it is at least marginally relevant to the protection issue.

In connection with burglary, the authors presented evidence indicating that burglary victims seldom have an opportunity to use a gun against a burglar, because there is rarely any confrontation between victim and offender, and they show that almost no burglars are killed by homeowners. Yet, as Bruce-Briggs (1976) has wryly noted, "the measure of the effectiveness of self-defense is not in the number of bodies piled up on doorsteps, but in the property that is protected." A more pertinent question is whether victims’ gun ownership somehow deters burglars from committing burglaries. It is not usually a matter of common knowledge either in the general population, or among burglars, that a particular homeowner owns a gun; therefore, a gun in a given residence is not likely to deter burglars from victimizing that specific residence.[1] Consequently, evidence cited by Yeager et al. regarding individual victim ownership of guns and burglary victimization (1976: 5) is irrelevant to the deterrence issue. However, it may very well be a matter of common knowledge (or belief) that certain neighborhoods are, in the aggregate, heavily armed (e.g., see Hannerz, 1969: 80, regarding a black neighborhood in Washington, D.C.) and it certainly is common knowledge that some regions of the United States, especially the South, are more heavily armed than others. It therefore is conceivable that potential burglars might either refrain from burglary altogether or commit fewer burglaries because of these anxieties regarding the risks of being shot by an armed householder. There is evidence that the legal risk of punishment for burglary, although very low in absolute terms, nevertheless exerts a deterrent effect on burglars. Several researchers have found that burglary rates are lower, ceteris paribus, where the probabilities of legal punishment are higher (e.g., Tittle, 1969; Phillips and Votey, 1972; [Page 282] Ehrlich, 1973). If the nonlegal risk of being shot by a homeowner is taken more seriously by burglars than the risks of legal punishment, which are themselves rather low, this would imply that the perceived risk of being shot by a householder would also exert a deterrent effect.

The clearance and conviction rates for burglaries in the United States in 1976 were 16.8% and 27.8% respectively (Federal Bureau of Investigation, 1977: 162, 217) giving an approximate risk of arrest and conviction for any given burglary reported to the police of 4.7% (0.168 x 0.278 x 100%). There were an estimated 3,252,100 burglaries reported to the police in 1975 (FBI, 1977: 37) and a total of 8,233,000, as indicated by victimization surveys (National Criminal Justice Information and Statistics Service 1977: 17) indicating that only about 40% of burglaries were reported to the police. Thus the overall risk of a burglar being arrested and convicted was only about 1.8% (.40 x .047 x 100%). If half of those burglars convicted received a prison sentence, then the risk of imprisonment was 0.9%. In 1964 (the last year for which relevant national data are available), the median prison term served for burglary was 20.1 months (Federal Bureau of Prisons 1967: 52), a value that is probably lower now. Therefore the legal risk an average burglar in the United States faces is a 0.9% chance of serving a median prison sentence of 20 months. In short, the legal risks of burglary are low in absolute terms, yet they seem to exert a deterrent effect on burglars and potential burglars. On the other hand, the principal nonlegal risk a burglar faces is that of being shot, possibly fatally, by a homeowner armed with a gun. The probability of such an event occurring is also undoubtedly low, but certainly nonzero. A Toronto victimization survey found that in 21% of burglaries the burglar was confronted by a victim (Waller and Okijiro, 1978: 31), indicating that opportunities to use a firearm would not be uncommon among burglary victims, if they owned firearms. Given the seriousness of the possible outcome, even a very slight probability of the event occurring may be taken very seriously by a potential burglar. For example, some professional robbers interviewed by Conklin (1972: 85) stated that they began their careers committing burglaries, but later gave up this type of crime because of "the risk of being trapped in a house by the police or an armed occupant" (emphasis added). Even though burglars may not be deterred from victimizing particular households because of gun ownership, the knowledge that gun ownership is common in general, or in a given area, may very well exert a deterrent effect.

In any case, there is some direct evidence that criminals do take victim gun ownership into consideration in planning crimes and choosing victims. Convicted robbers and burglars interviewed in a California [Page 283] prison stated that they would take into consideration the presence of weapons in a house or business and that they knew of specific cases where robberies were not committed because the prospective victim was known to be armed (Richardson, 1975). Further, the available evidence indicates that criminals perceive gun control measures as facilitating their criminal activities. In an informal survey, Menard (Illinois) Correctional Center convicts in prison for armed robbery unanimously agreed that a recent Morton Grove, Illinois, ordinance banning handgun ownership would be useless in curbing crime and would make things a bit easier for them. (Link, 1982). Another such survey (by a convicted armed robber) of 100 convicts in an Ohio prison indicated that the majority of inmates thought that crimes would increase if civilian guns were confiscated, especially daylight robbery of businesses and homes (Firman, 1975). Preliminary findings from a more rigorous 1982 survey of inmates of a number of prisons confirm these general conclusions (Wright, 1983).

This should not come as a surprise since, if deadly use of firearms is any indication, private citizens seem to use guns (legally) to shoot criminals about as often as the police do. Unpublished FBI data from 44 of the nation’s largest cities indicates that there were 1309 justifiable killings of felons by police and 1108 by private citizens, giving a citizen-to-police ratio of 0.846. In some areas, like Chicago and California, citizens justifiably kill two or three times as many violent felons as do police (Cook, n.d.; Inbau, n.d.; California Bureau of Criminal Statistics and Special Services, 1982). In 1978, the latest year for which national data are available, there were 266 justifiable homicides committed by the police (National Center for Health Statistics, 1981). Assuming 0.846 justifiable citizen slayings (the bulk of them with handguns) for every police case, we can estimate that about 225 felons were justifiably slain by civilians. It has been estimated that there are about 38 nonfatal handgun assaults for every fatal one (Kleck 1983: Table 1), which implies that there were about 8,550 (38 x 225) incidents in which private citizens justifiably assaulted (wounded nonfatally, shot at, or pointed a gun at) criminals. This is likely to be a considerable underestimate, since the 38-to-1 ratio on which it is based was itself based on victimization survey data, which count only nonfatal gun assaults that respondents were willing to admit to government interviewers. Incidents in which the respondent was assaulted while engaged in a crime are not likely to be reported to the interviewers. In any case, it is clear that private citizens are doing quite a bit of shooting at criminals. [Page 284]

Some idea of the magnitude of the risk criminals face from gun ownership among potential victims can be drawn from Cook’s data on robberies. Using Atlanta as an example, he calculated that a robber doubles his chances of dying in a given year by committing just seven robberies, due to the risk of being attacked by victims (Cook, 1979: 755).

Furthermore, the little information available indicates that when using firearms to interrupt or prevent crimes, or to apprehend criminals, private citizens appear to be at least as successful as police. Kates collected and analyzed every story concerning use of firearms for these purposes printed in 42 of the nation’s largest circulation newspapers during periods from January to June, 1975 and May to July, 1976, comparing success rates of police and private citizens. He found that 83% of the instances of firearms use by private citizens were successful in preventing the crime and/or apprehending the criminal, while the figure was only 68% for the police (Kates, n.d.).

Some of the most persuasive evidence regarding the potential deterrent effects of widespread gun ownership among potential crime victims has been generated as a result of what amounted to natural quasi-experiments involving widely publicized programs that provided training in firearms use, as well as a law requiring household gun ownership.

From October 1966 to March 1967 the Orlando Police Department sponsored a program intended to train women in the safe use of firearms. It was introduced in reaction to sharp increases in the number of rapes in 1966 and was given considerable publicity in the local newspaper. Because the program nicely bracketed the division between 1966 and 1967, we can use annual crime figures from the FBI’s Uniform Crime Reports to compare crime rates for 1966 (largely preceding the program) with those for 1967 (largely after the program). Figure 1 and Table 1 indicate crime trends for Orlando and surrounding areas. For 1966 the rape rate was 35.91 in Orlando, while it was only 4.18 for 1967, a one-year drop of 88%. It cannot be claimed that this was merely part of a general downward trend in rape, since the national rate was increasing at the time. No other U.S. city with a population over 100,000 experienced so large a percentage decrease in the number of rapes from 1966 to 1967, and only Philadelphia could boast of even so large a decrease in absolute numbers. (Philadelphia rapes went from 537 to 458, a drop of 79 rapes, but only a 15% decrease.) Further, comparison of Orlando with the surrounding metropolitan area and with Florida as a whole (excluding Orlando) indicates there were no comparable declines; rape was essentially constant or increasing during this period. Nor can it be [Page 285]

 

Figure 1: The Orlando Gun Training Program: Rape Trends, 1958-1972 SOURCE: Table 1 [Page 286]

TABLE 1 Crime Rates in Orlando and Comparison Areas, 1958-1972
Crime rates per 100,000 resident population
	Orlando (city)					Orlando SMSA, minus Orlando					Florida, minus Orlando SMSA
Year	rape	burglary	rbry	murder, nonneg. mnsltr.	auto theft	rape	burglary	rbry	murder, nonneg. mnsltr.	auto theft	rape	burglary	rbry	murder, nonneg. mnsltr.	auto theft
1958	18.52	1033.58	49.39	9.88	242.03	14.74	339.43	28.96	9.15	73.68	6.87	670.67	69.73	11.54	173.89
1959	26.02	1001.70	55.58	9.46	279.10	17.79	480.39	27.16	8.43	102.54	8.03	730.69	65.50	10.69	190.22
1960	11.35	1050.66	62.40	4.54	367.62	18.23	424.57	29.95	7.81	48.62	7.90	845.53	84.05	10.75	202.87
1961	6.72	965.03	102.00	10.09	264.51	16.24	447.51	22.06	4.16	75.35	7.32	742.65	74.95	9.53	175.72
1962	1.11	1001.01	94.12	13.29	324.44	8.40	432.66	18.39	4.40	77.97	5.95	775.07	62.79	7.97	174.72
1963	0.00	1304.21	97.38	9.85	330.43	6.54	603.97	23.08	6.16	90.02	7.40	851.04	75.08	8.50	177.40
1964	27.03	1292.13	161.11	9.73	448.73	6.67	665.65	43.73	7.04	128.24	10.25	982.42	88.10	8.66	192.57
1965	12.82	1079.41	123.97	13.89	421.08	10.01	763.03	40.05	3.93	124.79	13.30	953.42	89.46	8.99	205.90
1966	35.91	1485.35	131.00	17.96	441.59	11.05	822.00	58.02	6.56	131.59	14.20	1041.48	99.49	10.16	240.81
1967	4.18	1161.42	127.42	21.97	515.95	10.02	806.61	63.46	7.68	180.36	15.04	1192.23	129.01	10.06	275.41
1968	9.29	1401.40	208.61	20.65	439.94	10.67	782.82	35.89	9.05	119.96	17.80	1296.01	158.52	11.35	314.32
1969	16.43	2095.61	174.63	17.36	436.07	23.81	1020.89	44.81	7.21	158.87	20.28	1308.83	162.09	10.99	378.07
1970	20.20	2517.02	269.68	10.10	347.45	32.52	1553.21	54.10	11.55	158.36	21.72	1555.21	280.50	12.76	409.73
1971	29.43	2344.34	295.37	20.61	447.48	26.60	1485.29	78.95	8.49	284.10	23.89	1667.45	193.72	13.30	395.11
1972	35.30	2178.33	310.10	18.13	476.12	39.70	1333.52	83.64	10.59	314.72	25.13	1583.28	190.00	12.54	360.87
64-66*	25.25	1285.63	138.69	13.86	437.13	9.24	750.23	47.27	5.84	128.21	12.58	992.44	92.35	9.27	213.09
67-69*	9.93	1552.81	170.22	20.00	463.99	14.83	870.11	48.05	7.98	153.06	17.71	1265.69	149.87	10.80	322.60
% change	-60.7	20.8	22.7	44.3	6.1	60.5	16.0	1.7	36.6	19.4	40.8	27.5	62.3	16.5	51.4
s.d.	12.34	180.68	37.42	3.73	77.79
66- 67	yes	no	no	no	no
>2.s.d.
SOURCES: Crime data--FBI, Uniform Crime Reports, for indicated years. Population data--U.S. Bureau of the Census, Census of Population, 1950; 1960; and 1970. U.S. Bureau of the Census, Statistical Abstract of the U.S., 1981. (Intercensal years estimated by linear interpolation.) [Page 287]
NOTES FOR TABLE 1 NOTE: Orlando SMSA crime figures for 1960, missing in UCR, were estimated as average of 1959 and 1961 figures. Orlando firearms training program ran from October 1966 to April 1967. "S.d." is the standard deviation of crime rates over the period from 1958 to 1966. "% change" is the percentage change in average crime rates from the 1964-1966 period to the 1967-1969 period. The last row refers to whether the change from 1966 to 1967 was greater than two standard deviations. *Average of annual rates.

claimed that Orlando was merely experiencing a continuation of rape trends that it had already been experiencing before the program, since it had experienced an increase the year before, as well as an erratic but generally upward trend over the previous six or seven years.

It might be suggested that Orlando had experienced erratic ups and downs in its rape trends before and that the 1967 experience just happened to reflect one of the brief, sharp downward swings in the rape rate, which it had experienced before, and that the downward swing was therefore unrelated to the gun training program. However, this suggestion too is unlikely, since Orlando had not experienced so large a one-year change in rape rates in its recent past, and the decrease exceeded two standard deviations, a measure of the variability in rape rates over the 1958-1966 period (1958 was the first year the FBI reported rape data for the Orlando SMSA). In other words, the rape decrease was considerably larger than would be expected on the basis of variation in the rate in the recent past.

Further evidence suggesting the program’s effects can be seen by comparing rates for rape (the crime targeted by the program) with rates from crimes not targeted by the program. In Orlando, the surrounding metropolitan area, and Florida, rates for virtually all crimes were increasing or holding steady in the period after the program was carried out. The only major exceptions were the Orlando burglary rate and, to a lesser extent, the burglary rate for the surrounding metro area. Even these exceptions tend to strengthen the case for the deterrent effects of gun ownership, since residential burglaries would seem to be the next most likely crime type to be effected by a program that trained women in firearms use, even though it was not the principal target of the program.

It is highly debatable whether the gun training had its effects because of the training per se. I am inclined to believe it affected the behavior of potential rapists primarily because it served to inform or remind them of widespread gun ownership among women, and thereby increased the [Page 288] perceived riskiness of sexual assaults. If this is true, it would imply that such a training program would not even be necessary in order to produce crime decreases, but rather that any well-publicized reminders of citizen gun ownership could produce a deterrent effect. Newspapers giving prominent coverage to stories of citizens successfully using firearms to prevent crimes or apprehend criminals would be an example.

Orlando’s experience is far from unique. Although almost universally ignored by the mass media, similar crime reductions have been experienced elsewhere in connection with gun training programs, including decreases in armed robberies in Highland Park, Michigan, drug store robberies in New Orleans, and grocery store robberies in Detroit. However, even more remarkable is the recent experience of Kennesaw, Georgia.

In March of 1982, the Kennesaw city council passed a city ordinance requiring householders to keep a firearm in the home, with the exception of households with physically or mentally infirm persons, criminals or persons who conscientiously objected to gun possession. The ordinance was nationally publicized and widely perceived as a reaction to the passage in Morton Grove, Illinois, of an ordinance effectively prohibiting handgun ownership within the city limits. In the seven months immediately following passage of the Kennesaw law (March 15, 1982 to October 31, 1982) there were just five residential burglaries reported, compared to 45 in the same period of the previous year. (Benenson, 1982) An 89% decrease in burglaries in so short a period is hard to explain away; something was clearly happening in Kennesaw that was not happening in the rest of the country.

Again, it is debatable exactly why this ordinance had such an effect. There is no evidence indicating any significant actual increases in household gun ownership; the majority of southern households have guns without being prodded by an ordinance requiring it, and undoubtedly the same was true of Kennesaw. However, once again the publicized passage of the ordinance may have served to remind potential burglars in the area of the fact of widespread gun ownership, thereby heightening their perception of the risks of burglary.

If reminders of widespread citizen gun ownership serve to decrease crime, then publicized decreases in gun ownership or restrictions on their ownership or use, whether naturally occurring or legally mandated through gun control measures, could conceivably have the perverse effect of actually increasing burglary, robbery, and rape rates, by reducing potential offenders’ perception of the risks of committing these [Page 289] crimes. In fact, one could go further. Given the data on private citizens’ use of firearms against criminals and evidence on the slight risks of legal punishment associated with most crimes, it is a perfectly plausible hypothesis that private gun ownership currently exerts as much or more of a deterrent effect on criminals as do the activities of the criminal justice system. The gun-owning citizenry is certainly more omnipresent than the police, and the potential severity of private justice is at least as severe or more severe than more formal legal justice, given the frequency of citizen shootings of criminals and the de facto near-abolition of capital punishment by the federal judiciary. In short, there is the distinct possibility that although gun ownership among the crime-prone may tend to increase crime, gun ownership among the noncriminal majority may tend to depress crime rates below the levels they would otherwise achieve. This possible mixture of effects is something gun control advocates have chosen to ignore.

Yeager and his colleagues (1976) provide some evidence regarding self-protection and robberies, assaults, and rapes. Contrary to their interpretation, the weight of their relatively hard evidence contradicts the claim that guns are ineffective for self-defense. Unpublished data derived from victimization surveys were studied to determine the outcome of crimes during which victims used various forms of self-protection. The surveys did not cover victim use of a gun specifically, but did cover weapon use in general. Regarding robbery, Tables 5, 6, and 7 in Yeager et al. indicate the following: (1) Robberies are less likely to be completed if the victim used a weapon for self-protection, compared both to those who did not use any self-protection methods and to those who used alternative protection methods, such as running away, hitting or kicking, yelling, or reasoning with the criminal; and (2) robbery victims were no more likely to be injured if they resisted with a weapon than those who did not resist at all, and were even less likely to be injured than those who resisted by yelling, hitting or kicking, or by holding onto their property. Presumably those who used guns were even more successful in preventing completion of the crime and in avoiding injury than those using less intimidating weapons.

Regarding assault, the data presented in Tables 11 and 13 of the Yeager et al. study indicate that (1) assaults are less likely to be completed against victims who used weapons, as compared to using no self-protection method; and (2) assault victims who used weapons were less likely to be injured than were those who used no self-protection method. [Page 290]

Regarding rape, the authors had little direct evidence on the success of efforts to resist attackers with a weapon, since too few victims were included in the victimization surveys for meaningful analysis. The authors do point out that most rapists are unarmed (1976: 32), yet fail to draw the obvious inference that this would presumably give an armed victim an even greater chance of successfully resisting the attack (cf. Silver and Kates 1979: 164-165).

One of the few potentially persuasive points made by Yeager et al. is the simple observation that crime victims rarely get the opportunity to use a gun, even if they own one, especially when the crimes are committed away from the victim’s home. Because of laws prohibiting or restricting the carrying of handguns in public places, most potential victims are not likely to get a chance to defend themselves with a gun if victimized away from home, unless they are willing to violate the law by carrying a concealed weapon.

Nevertheless, gun owners do have opportunities to use their guns in self-defense, whether at home or away from home. A 1978 national survey indicated that in 15% of United States gun-owning households, some member of the household had in the past used a gun (even if it wasn’t fired) for self-protection against a person, excluding military service or police work (DMI, 1979: 68).[2] A California survey found that 8.6% of handgun owners responding had used a handgun for self-protection (California Bureau of Criminal Statistics, 1977). A Toronto victimization survey found that 21% of the burglary victims caught burglars in the act, although few of the homeowners had guns, presumably because of generally low Canadian gun ownership (Waller and Okijiro, 1978). Even in connection with robberies, there is some opportunity for victims to use weapons to defend themselves. In 3.5% or robberies reported to victimization surveyors in eight United States cities in 1971-1972, victims admitted using weapons (not necessarily firearms) for self-protection (Yeager et al., 1976). Presumably this is a conservative estimate, since many victims may be doubtful about the legality of their weapon use and, therefore, reluctant to acknowledge it to government interviewers. It is, of course, a matter of personal judgment whether this is a sufficiently large frequency to justify gun ownership for self-defense. Nevertheless, many potential crime victims apparently want to have the option of defending themselves with firearms against criminals should the necessity arise, however rare such a situation may be. Regardless of how one may feel about the desirability of using guns for defensive purposes, it cannot be claimed, on the basis of [Page 291] available evidence, that the belief in the protection efficacy of firearms is just the product of self-delusion.

ASSUMPTION FIVE

Homicides are largely unpredictable "crimes of passion" committed by otherwise law-abiding citizens not distinguishable from other people. Everyone is potentially a killer, and we can not tell in advance who is likely to kill and who is not. Therefore, control must be directed at all gun owners rather than select criminal subgroups.

The position on gun control of the U.S. Conference of Mayors includes the following statement: "those who possess handguns can not be divided into criminals and qualified gun owners" (Yeager et al., 1976: xiii). The assumption is also made in the gun control positions of the AFL-CIO, Common Cause, and the Union of American Hebrew Congregations, among others. This assumption is crucial to gun control proposals directed at all gun owners, rather than just those who misuse their guns. If it is false, then opponents of gun control may argue that such "blanket" measures unnecessarily and unjustly deprive and punish law-abiding people along with criminals.

It is, of course, perfectly true that we cannot identify in advance specific individuals who will kill (or assault), either with or without a gun. The predictive technology simply does not exist and is not likely to exist in the near future (see Wenk et al., 1972). And it is trivially true that everyone is potentially a killer, in the sense that there is at least an infinitesimally small probability that any given person will commit a homicide. However, this does not mean that killers are randomly distributed through the population, or that some aggregates cannot be divided, if we so choose, into two distinct, nonoverlapping groups: those who have been convicted of a violent crime and those who have not. This is, in fact, a distinction already made in existing gun control law. For example, the 1968 Gun Control Act makes it a federal crime punishable by 2 years in prison for any convicted felon to possess a firearm (Internal Revenue Service, 1968: 8).

Opponents of gun control argue that gun control will needlessly deprive ordinary citizens of guns, while leaving criminal misusers of guns armed. Because criminals will disobey gun laws just as they disobey other laws, gun control will fail to affect the small minority of the [Page 292] population that use guns to commit violent crimes. Gun control supporters then reply that even if criminals remain armed, gun control will still help reduce homicides by disarming ordinary citizens who otherwise would have used guns in domestic assaults and other fights between persons who know each other.

This argument has been phrased in a nicely explicit form by gun control advocate Leonard Berkowitz (1981: 11): "Gun control may not be too effective in protecting ordinary citizens against criminals or Presidents against assassins, but it may, nevertheless, save some ordinary citizens from other ordinary citizens like themselves."

There is also an imagery in gun control thinking of domestic homicides as being isolated outbursts of otherwise ordinary, nonviolent, noncriminal persons. Because intrafamily homicides constitute 17% of all United States homicides (FBI, 1980: 11), gun control advocates argue that it therefore is important to restrict firearms among seemingly law-abiding persons as well as among convicted felons.

Because most violent acts are not reported to the police, and many do not result in any kind of officially recorded action (arrest, conviction, or imprisonment), official records of the previous violence of homicide offenders represent only the tip of the iceberg. Records of previous arrests for violent acts seriously underestimate previous violent behavior, records of prior convictions do so even more, and records of prior imprisonment do so still more. Nevertheless, it is useful to look at the evidence regarding officially recorded prior violence of known homicide offenders, in order to establish a minimum, baseline estimate of the prevalence of prior violence among killers. The most representative available samples of known or suspected homicide offenders would be samples of homicide arrestees, since samples drawn at later points in the criminal justice process (e.g., samples of persons convicted or persons imprisoned) would be subject to case loss and various selection biases, including bias associated with prior criminal record (see Wolfgang, 1958: 11-13). There is little evidence concerning prior convictions for such samples. Usually the data either concern prior arrests (and the proportion of arrestees with prior arrests would necessarily be larger than the proportion with prior felony convictions) or refer to samples of incarcerated persons, who would presumably be more recidivists than general samples of arrestees. Wolfgang (1958: 170-172, 183) reviewed earlier studies of the prior records of homicide offenders, most of them done in the 1930s and 1940s. In one sample of persons imprisoned for homicide offenses, 82% had previous criminal convictions, while the [Page 293] figure was 98% and 32% for two other similar samples. In a sample of persons convicted of homicide offenses, 43% had previous convictions. Regarding prior record of arrests, three studies indicated that 54%, 50% and 55%, respectively, of samples of homicide prisoners had previous arrests, while Wolfgang’s own sample of homicide arrestees indicated that 64.4% had a record of prior arrests (1958: 175). More recently the Careers in Crime data of the Uniform Crime Reports indicate that 77.9% of persons arrested for murder or nonnegligent manslaughter in 1970 had previous arrests, and 50.1% had prior convictions (FBI, 1971: 38). Among those homicide offenders arrested in the United States between 1970 and 1975, 67.6% had previous arrest records (FBI, 1976: 43).

The FBI is rather vague about the types of crimes for which offenders were previously arrested or convicted. However, in special computer runs for the 1968 Eisenhower Commission it was determined that 74.7% of persons arrested between 1964 and 1967 for criminal homicide had a record of previous arrests for "a major violent crime or burglary" (Mulvihill et al., 1969: 530, 532). The Careers in Crime data can be questioned regarding sample representativeness, so some independent confirmation of these figures would be helpful. Data for New York City indicate that among those arrested for homicide in 1970, 64.7% had a prior arrest record and 40% had prior arrests for violent offenses (Shinnar and Shinnar, 1975: 596).

Considered as a whole, prior research evidence shows that the majority of homicide arrestees have prior arrest records. As a rough estimate, perhaps half of them have previous convictions of some sort, although the convictions were not necessarily for felonies. A reasonably conservative estimate of the fraction of homicide offenders with prior felony convictions might be about one-quarter.

Further, it should be noted that felony killings account for an increasingly large fraction of United States homicides. In 1964, 17% of murders and nonnegligent manslaughters were known, or suspected to be, the result of other felonious activities. By 1976, the figure was 28%. Among the known felony killings in 1976, 42% resulted from robberies (FBI, 1965: 7, 1977: 10). If robbers are more likely to have criminal records than are persons who commit nonfelony killings, this suggests that the proportion of homicides committed by persons with prior convictions, who fit the popular stereotypes of a "real" criminal, is increasing. Therefore, the potential effect of well-enforced gun control policies aimed specifically at this group has been increasing as well. [Page 294]

Even arrest records grossly underestimate prior violent behavior of assaultive offenders, since for every arrest for a homicide or assault there were at least six such crimes committed.[2] There is, however, another official indicator of prior violence, which produces somewhat less underestimation: police records of patrol car "disturbance calls." A study of Kansas City killings found that 90% of the homicides had been preceded by past disturbances at the same address, disturbances which were serious enough that the police had to be called in, with a median number of five previous disturbance calls per address (Wilt et al., 1977). Thus killings are rarely isolated outbursts of previously nonviolent people, but are usually part of a pattern of violence, engaged in by people who are known to the police, and presumably others, as violence prone.

Assumption 5 is therefore false to the extent that most violent crimes are committed by persons with records of previous violence and criminal behavior, and there is thus significant potential for reducing homicide through measures aimed strictly at convicted criminals. On the other hand, it remains to be seen whether these measures could be effectively enforced, how much substitution of nonprohibited weapons there would be, and how well private transfers (legal and illegal) of firearms from legal gun owners to criminals could be prevented.

CONCLUSION

The research evidence reviewed in this article renders suspect some of the most crucial factual assumptions underlying arguments in favor of policies aimed at restricting the ownership of firearms. These assumptions have gone largely uncriticized and unquestioned by advocates of gun control measures and by social scientists working in the area of crime and violence. In this light, it is suggested that more thorough and rational evaluation of the potential consequences of suggested gun control policies is called for, with greater attention being focused on the issues herein discussed, both by researchers and others who would seek to influence crime control policy.

NOTES

1. However, it is precisely because would-be burglars cannot be sure that a given household is not armed that it is more likely that they would be deterred from burglarizing [Page 295] any residence than would be the case in the absence of widespread gun ownership. Further, private ownership of guns may have the further benefit of minimizing injuries among burglary victims by encouraging burglars to avoid occupied residences and, consequently, confrontations with their victims. In short, because some householders do use guns to confront burglars, many others may thereby be relieved of the obligation to do so.

2. According to the National Crime Survey victimization study, there were an estimated 4,664,000 simple and aggravated assaults in 1977 (NCJISS, 1979: 20), while there were an estimated 718,980 arrests the same year for homicides and simple and aggravated assaults (FBI 1979: 172), giving a ratio of 6.5 crimes to every arrest.

REFERENCES

ALVIANI, J. D. and W. R. DRAKE (1975) Handgun Control--Issues and Alternatives. Washington, DC: U.S. Conference of Mayors.

BENENSON, M. K. (1982) Memorandum reporting telephone conversation with Kennesaw, Georgia, Police Chief Roble.

-------- and D. B. KATES, Jr. (1979) "Handgun prohibition and homicide: a plausible theory meets the intractable facts," pp. 91-118 in D. B. Kates, Jr. (ed.) Restricting Handguns: The Liberal Skeptics Speak Out. Croton-on-Hudson, NY: North River Press.

BERKOWITZ, L. (1981) "How guns control us." Psychology Today 15 (June): 11-12.

-------- (1968) "Impulse, aggression and the gun." Psychology Today 2 (September): 19-22.

-------- and A. Le Page (1967) "Weapons as aggression-eliciting stimuli." J. of Personality and Social Psychology 7 (October): 202-207.

BORDUA, D. J., A. J. LIZOTTE, and G. KLECK, with V. Cagle (1979) Patterns of Firearms Ownership, Use and Regulation in Illinois. Springfield: Illinois Law

Enforcement Commission.

BRUCE-BRIGGS, B. (1976) "The great American gun war." Public Interest 45 (Fall): 37-62.

BUSS, A. J., A. BOOKER, and E. BUSS (1972) "Firing a weapon and aggression." J. of Personality and Social Psychology 22 (June): 296-302.

California Bureau of Criminal Statistics (1977) A Synopsis of a California Poll of Handgun Ownership and Use. Sacramento: California Department of Justice.

-------- and Special Services (1982) Homicide in California, 1981. Sacramento: California Department of Justice.

CONKLIN, J. E. (1972) Robbery and the Criminal System. Philadelphia: Lippincott.

COOK, P. J. (1979) "The effect of gun availability on robbery and robbery murder," pp. 743-781 in R. H. Haveman and B. B. Zellner (eds.) Policy Studies Review Annual. Beverly Hills, CA: Sage.

-------- (n.d.) Unpublished Uniform Crime Reports data supplied by FBI.

Decision Making Information (DMI) (1979) Attitudes of the American Electorate Toward Gun Control, 1978. Santa Ana, CA: Author.

EHRLICH, I. (1973) "Participation in illegitimate activities: a theoretical and empirical investigation." J. of Pol. Economy 81 (May-June): 521-565.

ELLIS, D. P., P. WEINIR, and L. MILLER III (1971) "Does the trigger pull the finger? An experimental test of weapons as aggression-eliciting stimuli." Sociometry 34 (December): 453-465. [Page 296]

Federal Bureau of Investigation [FBI] (1965, 1971, 1976, 1977, 1979, 1980) Uniform Crime Reports, 1964, 1970, 1975, 1976, 1978, 1979. Washington, DC: Government Printing Office.

Federal Bureau of Prisons (1967) State Prisoners: Admissions and Releases, 1964. Washington, DC: Government Printing Office.

FIRMAN, G. R. (1975) "In prison survey the pros are the cons." Amer. Rifleman 23 (November): 13.

FISHER, D., H. KELM, and A. ROSE (1969) "Knives as aggression-eliciting stimuli." Psych. Reports 24 (June): 755-760.

Florida Bureau of Criminal Justice Planning and Assistance (1977) Handgun Regulation Project. Final Report. Tallahassee: Author.

FRACZEK, A. and J. R. MACAULAY (1971) "Some personality factors in reaction to aggressive stimuli." J. of Personality 39 (June): 163-177.

FRODI, A. (1973) "The effects of exposure to aggression-eliciting and aggression-inhibiting stimuli on subsequent aggression." Goteburg Psych. Reports 3, 8.

HANNERZ, U. (1969) Soulside: Inquiries into Ghetto Culture and Community. New York: Columbia Univ. Press.

HARDY, D. T. and J. STOMPOLY (1974) "Of arms and the law." Chicago-Kent Law Rev. 51 (September): 62-114.

INBAU, F. (n.d.) Data from Chicago Police Department, covering 1967-1976 justifiable homicides. (unpublished)

Internal Revenue Service (1968) Gun Control Act of 1968. Washington, DC: Government Printing Office.

KATES, D. B., Jr. [ed.] (1979) Restricting Handguns: The Liberal Skeptics Speak Out. Croton-on-Hudson, NY: North River Press.

-------- (1976) "Reflections on the relevancy of gun control." Law & Liberty 3 (Summer): 1-3.

-------- (n.d.) Unpublished tables.

KENNETT, L. and J. L. ANDERSON (1975) The Gun in America: The Origins of a National Dilemma. Westport, CT: Greenwood.

KLECK, G. (forthcoming) "Handgun-only gun control: a policy disaster in the making," in D. B. Kates, Jr. (ed.) Firearms and Violence: Issues of Regulation. San Francisco, Pacific Institute for Public Policy Research and Ballinger.

LEYENS, J. P. and R. D. PARKE (1975) "Aggressive slides can induce a weapons effect." Europen J. of Social Psychology 5: 220-236.

LINK, M. (1982) "No handguns in Morton Grove--big deal!" Menard Times 33 (January 22): 1.

LUCKENBILL, D. F. (1978) "Criminal homicide as a situated transaction." Social Problems 25 (December): 175-186.

MULVIHILL, D. J. and M. M. TUMIN, with L. A. Curtis (1969) Crimes of Violence. Report of the Task Force on Individual Acts of Violence. Washington, DC: National Commission on the Causes and Prevention of Violence.

National Center for Health Statistics [NCHS] (1981) Vital Statistics of the United States. Mortality, 1978. Washington, DC: Government Printing Office.

National Criminal Justice Information and Statistics Service [NCJISS] (1977, 1979) Criminal Victimization in the United States, 1975, 1977. Washington, DC: Government Printing Office. [Page 297]

NEWTON, G. D. and F. ZIMRING (1969) Firearms and Violence in American Life. Washington, DC: Government Printing Office.

PAGE, D. and E. O’NEAL (1977) "‘Weapons effect’ without demand characteristics." Psych. Reports 41 (August): 29-30.

PAGE, M. M. and R. J. SCHEIDT (1971) "The elusive weapons effect: demand awareness, evaluation apprehension, and slightly sophisticated subjects." J. of Personality and Social Psychology 20 (December): 304-315.

PHILLIPS, L. and H. L. VOTEY, Jr. (1972) "An economic analysis of the deterrent effect of law enforcement on criminal activity." J. of Criminal Law, Criminology and Police Sci. 63 (September): 330-342.

RICHARDSON, H. L. (1975) "Myth #2: criminals won’t have guns." True Magazine (July): 32ff.

RUSHFORTH, N. B., A. B. FORD, C. S. HIRSCH, N. M. RUSHFORTH, and L. ADELSON (1977) "Violent death in a metropolitan county." New England J. of Medicine 297 (September): 531-538.

SHINNAR, R. and S. SHINNAR (1975) "The effects of the criminal justice system on the control of crime: a quantitative approach." Law and Society Rev. 9 (Summer): 581-611.

SILVER, C. R. and D. B. KATES, Jr. (1979) "Self-defense, handgun ownership, and the independence of women in a violent, sexist society," pp. 139-169 in D. B. Kates (ed.) Restricting Handguns: The Liberal Skeptics Speak Out. Croton-on-Hudson, NY: North River Press.

TITTLE, C. R. (1969) "Crime rates and legal sanctions?" Social Problems 16 (Spring): 409-423.

TOCH, H. (1969) Violent Men: An Inquiry into the Psychology of Violence. Chicago: Aldine.

TURNER, C. W. and L. S. SIMONS (1974) "Effects of subject sophistication and evaluation apprehension on aggressive responses to weapons." J. of Personality and Social Psychology 30 (September): 341-348.

TURNER, C. W., J. J. LAYTON, and L. S. SIMONS (1975) "Naturalistic studies of aggressive behavior: aggressive stimuli, victim visibility and horn honking." J. of Personality and Social Psychology 31 (June): 1098-1107.

U.S. Congress, Senate (1975) "Gun control." Congressional Record 121 (December 19): 1-10.

Wall Street Journal (1972) "Understanding gun control." June 7: 14.

WALLER, I. and N. OKIJIRO (1978) Burglary: The Victim and the Public. Toronto: Univ. of Toronto Press.

WENK, E., J. ROBISON, and G. SMITH (1972) "Can violence be predicted?" Crime and Delinquency 18 (October): 393-402.

WILSON, H. and R. SHERMAN (1961) "Civilian penetrating wounds of the abdomen." Annals of Surgery 153 (January): 639-649.

WILT, G. M., J. BANNON, D. K. BREEDLOVE, J. W. KENNISH, D. M. SANDKER, and R. K. SAWTELL (1977) Domestic Violence and the Police: Studies in Detroit and Kansas City. Washington, DC: Police Foundation.

WOLFGANG, M. E. (1958) Patterns in Criminal Homicide. Philadelphia: Univ. of Pennsylvania Press. [Page 298]

-------- and F. FERRACUTI (1967) The Subculture of Violence: Towards an Integrated Theory in Criminology. New York: Tavistock.

WRIGHT, J. D. (1983) Communication to Don B. Kates, Jr.

YEAGER, M.G., J. D. ALVIANI, and N. LOVING (1976) How Well Does the Handgun Protect You and Your Family? Technical Report 2. Washington, DC: U.S. Conference of Mayors.

ZIMRING, F. E. (1977) "Determinants of the death rate from robbery: a Detroit time study." J. of Legal Studies 6 (June): 317-332.

-------- (1972) "The medium is the message: firearm caliber as a determinant of death from assault." J. of Legal Studies 1 (January): 97-123.

-------- (1968) "Is gun control likely to reduce violent killings?" University of Chicago Law Rev. 35 (Summer): 721-737.

GARY KLECK is Assistant Professor in the School of Criminology at the Florida State University. He received his A.B., A.M., and Ph.D. degrees in sociology, all at the University of Illinois. His research interests center around the subject of criminal violence and his publications include articles in the American Sociological Review, the American Journal of Sociology, and Contemporary Sociology. He is currently doing research on the nature and risks of firearms accidents, and the crime-preventive effects of gun control laws.

DAVID J. BORDUA is Professor of Sociology at the University of Illinois in Urbana-Champaign. He received his Ph.D. in sociology from Harvard University in 1957. He taught at the University of Michigan from 1955-1963. He was a Fellow at the Center for Advanced Studies in the Behavioral Sciences in 1962-1963. Professor Bordua has published on crime, juvenile delinquency, and police organization. His recent research has been on private firearms ownership in the United States.

SOURCE: This article is adapted for publication, with permission, from a chapter in the forthcoming book, Firearms and Violence: Issues of Public Policy, edited by Don B. Kates, Jr. (San Francisco: Pacific Institute for Public Policy Research and Ballinger, 1983).