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.
Deterrence In Cn~runology arid Social Policy H. LAURENCE ROSS and GARY D. LAFREE INTRODUCTION Social policy and research on the deterrence of crime have often been unrelated in the United States. While politicians have periodically called for harsher punishments to deter crime, most criminologists prior to the 1970s either ignored the deterrence issue or voiced strong skepticism toward it (e.g., Sutherland, 1924:360; Reckless, 1967:5041. This gap between policy and research is unfortunate, manifesting itself in policy initiatives unrefined by empirical evaluation and empirical research with little policy · · ~ slgnl~lcance. In recent years the estrangement between criminology and social policy on deterrence has shown signs of abating. This chapter examines recent social research on two important categories of modern human misconduct- street crime and drunk driving to explore the implications of recent crim- inological studies for policy on the deterrence of crime. TWO FUNDAMENTAL PERSPECTIVES ON CRIMINAL CONDUCT Two broad perspectives on human behavior have long competed for preeminence in efforts to control crime in America. Both have historical roots as well as present-day champions, and both have been evident in the operation of our legal system since its inception. The first asserts that human behavior may be usefully represented as the product of rational individual calculation; the second asserts that behavior is largely guided by nonrational biological, psychological, or social forces. 129
130 H. LAURENCE ROSS and GARY D. LAFREE The "rational actor" perspective assumes that human beings behave to maximize personal pleasure and minimize pain. As elaborated by social reformers like Bentham and Beccaria, or jurists like Blackstone, Romilly, or Feuerbach, this view argues that crime can be deterred by increasing the costs of criminal behavior or increasing the rewards of noncriminal behav- ior. Contemporary criminologists generally refer to this perspective as the "Classical School" (Jeffery, 1972; Void, 19791. By contrast, the second perspective assumes that human behavior, in- cluding crime, is governed by forces over which the individual has relatively little conscious, rational awareness or control. Starting with Cesare Lom- broso and his students in the lSOOs, criminologists have labored to discover, describe, and understand these forces. Contemporary criminologists refer to this perspective as the "Positive School." For those who assume that crime is caused by factors outside the of- fender's control, the proper role of criminology is not to investigate the deterrent effects of variations in the law and its enforcement but, rather, to help ameliorate the problem of crime by identifying and taking steps to alter the biological, psychological, or social conditions that produce it. Whereas the classical perspective suggests the possibility of deterring crime through manipulating actual or expected rewards and punishments, the positive school recommends changing nonrational elements of the offender's psyche or environment. The most common policy approaches for bringing about these changes include a variety of intervention strategies that re- gardless of their actual performance, have generally been justified as "re- habilitative" (e.g., probation, parole, indeterminate sentencing, and institutional treatment). Public policy on crime in the twentieth century has drawn from bow the positive and classical schools. The belief that credible Greats of punishment deter criminal behavior is probably as old as criminal law itself and has broad appeal to policymakers and the public. From an intuitive point of view it seems reasonable. Surely Chinese citizens were less likely to exceed the speed limit in Peking early in this century when authorities exhibited the heads of drivers executed for sewing alongside speed limit signs (Zimring and Hawk- ins, 1973:11~. The adoption of harsh laws against crime in the United States, as well as the mobilization of ainunal penalties to deal with specific behavioral problems such as drunkenness and other drug abuse show continued fain in He efficacy of deterrence. At He same time, twentieth century policymakers have created vast programs to rehabilitate criminals, including probation, pa- role, and specialized correctional facilities for juveniles and for specific cat- egories of convicted offenders. While some of these efforts at rehabilitation have been called half-hearted, no one can seriously deny that substantial resources have been devoted to the rehabilitative ideal.
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 131 In contrast to the dual approach of policymakers, balancing (perhaps vacillating) between deterrence and rehabilitation, criminologists have by and large rejected deterrence principles. For over a century a host of dis- tinguished scholars with viewpoints as different as Enrico Fermi and Edwin Sutherland were able to agree on one point: deterrence does not work. ~ One major reason for this long-term rejection of deterrence is that criminologists, especially in the United States, traditionally were humanists and reformers (Gibbons, 1979; Wilson, 1983a). Since the early years of this century American criminology has had a strong social reform component, based on the belief that government is not merely a device to facilitate the pursuit by individuals of their private ends, but also a device to shape and improve the character of its citizenry. The reformers held Hat if only the right institutions were built, the right people properly trained to staff them, and the right classification procedures used to fill them, then surely rehabilitation would occur. However, in the mid-1970s a profound retreat from these assumptions became evident among criminologists and policymakers. For the first time in 150 years American criminologists seriously questioned whether reha- bilitation was a reasonable goal. State governments across the United States were moving away from indeterminate sentencing (long associated with the rehabilitative ideal), curtailing the use of probation and parole, and speaking against "correctional" programs slanted toward rehabilitation. The reasons for the recent decline in the popularity of rehabilitation, both among policymakers and criminologists, undoubtedly warrant a separate, detailed account. Here we can merely summarize the more common ex- planations for the shift. First, renewed interest in deterrence appears to reflect the perceived failure of rehabilitation policies. This failure was typically "proven" by the precipitous increase in crime rates in the last two decades (Wilson, 1975, 1983a, 1983b); by widely publicized prison disasters, such as those in Attica in 1969, and Santa Fe in 1980; and by mounting research evidence that many programs aimed at preventing re- cidivism through rehabilitation programs have been relatively ineffective (Martinson, 1974; Lipton et al., 19751. Second, traditional methods of rehabilitation have come under increas- ingly strong attack in the last two decades on the basis of their intrusiveness. In a series of decisions in the 1960s, most notably Escobedo v. Illinois (378 U.S. 478, 1964) and in re Gault (387 U.S. 1, 1967), the Supreme Court revolutionized the meaning of due process rights in American law tin fact, this belief was fully articulated by Edwin Sutherland in the chapter he wrote on "Crime and Punishment" for Recent Social Trends in the United States, the Ogburn Report.
132 H. LAURENCE ROSS and GARY D. UNFREE with important consequences for the idea of rehabilitation, with its emphasis on individualized intervention to bring about psychological changes in of- fenders. Disdain among both politicians and criminologists for this type of intervention became increasingly evident in the late 1960s. For example, Struggle for Justice ( 1971), the influential report prepared for the American Friends Service Committee, asserted (p. 85) that rehabilitation rests "largely on speculation or on assumptions unrelated to criminality," and that de- cisions made about offenders are routinely made "in the absence of credible scientific data on the causation or treatment of crime." This critique and others like it attacked the fundamental assumptions of rehabilitation: that crime is caused by forces over which the individual has little control and that the criminal justice system is able to identify and correct these forces. Finally, a less theoretical but eminently plausible explanation for the decline in support for rehabilitation programs is their cost. Rehabilitation, especially the kind envisioned by much of the criminological literature, is expensive. Many states find themselves spending increasing amounts of scarce revenues on correctional programs that are often difficult-to justify to taxpayers. Andrew Scull (1977) and others have argued that these purely economic forces, rather than humanitarian or scientific concerns, have led to declining support for rehabilitation. With diminishing support for rehabilitation, the justification for punishing criminals has shifted toward deterrence, and the forms of scholarship have likewise changed. It is difficult to find more than a half-dozen professional articles or books written on the subject of deterrence from 1900 to 1965. But starting in the 1960s, the study of deterrence has become a crimino- logical growth industry. Within criminology, the deterrence proposition has generated new interest and a large and rapidly expanding research literature (see Zimring and Hawkins, 1973; Andenaes, 1974; Gibbs, 1975; Cook, 1977;Blumsteinet al., 1978;Tittle, 1980;Archeretal., 1983;forreviews). Stated simply, this proposition asserts that proscribed behavior is deterred by perceptions that legal punishments are swift, sure, and severe. Our main purpose in this chapter is to appraise and interpret research on the deterrence proposition. Because researchers know little about the con- sequences of swift punishment in the context of laws there is too little swift punishment available in our legal system to study our exclusive focus is on variations in the certainty and severity of punishment. Moreover, evaluations of the effects of certainty and severity of punishment must distinguish between the legal existence (de jure) of punishment, and its actual use (de facto). Despite de jure changes in punishment, meaningful de facto changes have been rare. Thus, evaluations of the deterrence prop- osition are confounded by the fact that real changes in the imposition of punishment seldom accompany legal changes. We review prior research
DETERRENCE IN CRlMlNOLOGY AND SOCIAL POLICY 133 for two important types of criminal behavior, street crime and drunk driving, present generalizations that have been established in these fields, evaluate the strength of the evidence, and interpret their meaning for social policy. STREET CRIMES Perhaps no area of deterrence research has generated as much public interest in recent years as attempts to reduce street crime, which generally means robbery, rape, assault, or murder that occurs in public places between people previously unacquainted. The public is concerned and fearful: the President's Commission on Law Enforcement (1967) found that one-third of all Americans were afraid to walk alone at night in their own neigh- borhoods, and many reported that they stayed off the streets altogether because of their fear of crime. Subsequent victimization surveys conducted by the Department of Justice (see, e.g., Hindelang, 1976) have confirmed the enormous impact that fear of street crime has on the behavior of citi- zens especially the poor, members of minority groups, and urban resi- dents. "Crime in the streets" has been a recurring national and political issue since the mid-1960s, and strategies for dealing with street crime are the subject of debates, media programming, political campaigns, and gov- ernment commissions. In light of this interest social science knowledge concerning the effect of deterrence-based legal interventions bears important policy implications. This knowledge is summarized here. Certainty of Punishment The deterrence proposition predicts that proscribed behavior will be re- duced to the extent that the relevant public perceives a high likelihood of punishment for violations.2 As other reviewers (e.g., Blumstein et al., 1978) have noted, relatively little effort has been made to measure this perception directly; the bulb of what we know simply relates aggregate measures of street crime to policy or program innovations that are intended to increase the actual chance of punishment, implicitly assuming that increases in the (intended or actual) likelihood of punishment will in tuna lead to increases in its perceived certainty. This chain of assumptions may reasonably hold where the innovations are accompanied by official publicity and mass media 2Deterrence research distinguishes between "general" and "special" types. General deterrence is the inhibiting effect of sanctioning an offender on other potential offenders' criminal behavior. Special deterrence is the inhibiting effect of sanctioning an offender on his or her own future criminal behavior.
134 H. LAURENCE ROSS and GARY D. LAFREE attention, but its unproved assertion is a point of weakness in much of the existing evidence. The best-studied legal innovation to increase the certainty of punishment has been intensive policing efforts designed to raise the risk of apprehension and charging. Police crackdowns are fairly common responses to public concern about street crimes, and a number of these have been submitted to evaluation. Of particular interest are two studies of patrol efforts in New York City, "Operation 25" and the "20th Precinct" studies; the San Diego Field Interrogation project; a study of robberies in the New York City transit system; the LEAA High-Impact Anti-Crime project; and the Kansas City Preventive Patrol project. One of the earliest evaluations of increased patrol's effect in reducing street crimes was Operation 25 in New York City (see Zimring and Hawkins, 1973:348-3491. The police department selected the twenty-fifth precinct, a small district with a high crime rate, for greatly increased patrol during a four-month experimental period. The number of foot-patrol officers within this district was quadrupled for the experiment, and crime rates declined in all categories during those four months. However, no data were available to investigate the possibility that Operation 25 may have shifted the location of crimes from the experimental precinct to adjacent areas. This weakness in Operation 25 was avoided in a subsequent, similar study in the twentieth precinct, which received a 40 percent increase in police manpower and also noted decreases in the rates of major crimes (Press, 19711. The evaluation controlled the experimental data with findings from adjacent districts to test for displacement effects and from distant districts to test for the possibility that a general decline in crime rates could have explained the decline observed in the experimental district. The control data supported the conclusion that the patrol was effective in reducing street crimes (i.e., those visible from the street) and that it did not merely displace the criminal activity into adjacent districts. Of course, these results can also be questioned. The period of the experiment was only eight months and only changes in crimes reported to the police were measured. Moreover, the official records were maintained by police who were aware of the experiment and/or the previous findings from Operation 25. One of the best-designed experiments on the deterrence of street crime, the San Diego project (Boydstun, 1975), concerned the practice of stopping, questioning, and frisking persons who aroused police suspicions (i.e., con- ducting "field interrogations". In one area of the city, field interrogations were eliminated, whereupon the number of "suppressible" crimes (robbery, burglary, theft, auto theft, assault, sex crimes, malicious mischief, and disturbances) increased by about a third; when field interrogations were
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 135 resumed, the number of such crimes dropped back to preexperimental levels. There was no change in the frequency of suppressible crimes in two control areas where either field interrogation practices remained unchanged or po- lice officers were specially trained to conduct them in light of legal pro- cedures and human relations principles. Because the presence or absence of field interrogations did not affect the number of arrests in either control or experimental areas, Boydstun concludes that the visibility of police activity was responsible for the apparent deterrence of crime. In response to a large increase in subway robberies (especially of toll- booth stations) in 1965, the New York Transit Authority introduced special patrols on the subways during nighttime hours. Evaluators (Chaiken et al., 1974) found that crime rates dropped substantially during the patrol hours, but not during the balance of the day, for up to six years following the crackdown. An interesting sidelight in the study that has important impli- cations for deterrence research was the discovery of a "phantom effect." For eight months, while there were stepped-up patrols only at specific times and places, serious crime rates declined throughout the subway system. The evaluators assert that uncertainty as to the deployment of the police had a deterrent effect on potential offenders in nontarget areas and times. However, in the long run this phantom effect disappeared, possibly because potential felons became familiar with actual deployment practices. One difficulty with this study was that the evaluation was based on crime reports made by the participating transit police officers. After the evaluation was completed, researchers alleged that police officials miscoded the times of some offenses, apparently to exaggerate the reduction in frequency of of ~ ~ HA tot ~ ~ ~ T ~ tenses during peak patrol hours Wagner, Aim: m). In a reexam~nauon of the evidence, Chaiken (1976) concludes that despite the falsification, there was a significant deterrent effect, although of lesser magnitude than the original evaluation suggested. Perhaps the most ambitious experiment in deterring street crime ever attempted in the United States was the High-Impact Anti-Crime Program, funded for $160 million by the Law Enforcement Assistance Administration in 1972 (Chelimsky, 19761. Eight cities with high crime rates were targeted for crime-reduction programs with the goal of reducing stranger-to-stranger personal crime by 5 percent in two years and 20 percent in five years. Each city had complete discretion in designing individual programs and evalu- ating the consequences. Most of the money was allocated to increased enforcement, although some projects also aimed at streamlining court op- erations. Unfortunately, the variability in programs and evaluations strongly compromised the scientific utility of the program. A summary evaluation (Chelimsky, 1976) found the individual project descriptions to be flawed, but the summary was forced to rely on Uniform Crime Report (UCR) data
136 H. LAURENCE ROSS and GARY D. LAFREE that omitted several theoretically crucial variables, including conviction rates, sentencing patterns, and arrest-to-crime rates. As Franklin Zimring points out (1978:144-149), the UCR data do not allow specific measures of stranger-to-stranger offenses, those targeted for reduction by the project. Even so, the final evaluation offered no explicit statistical comparison for crimes other than burglary. The report found that five of the eight Impact cities had 1974 levels of burglary lower than would have been predicted on the basis of extrapolations from a set of comparison cities, but no such differences were apparent in the remaining three cities. Thus, modest sup- port was obtained for the deterrence proposition from this very costly ex- periment. The Kansas City Preventive Patrol project was designed to test the relative effect on crime rates of three policing strategies: "proactive" patrol, with patrol car levels between two and three times the normal level; "normal" patrol; and no routine patrol, police entering the area only in response to calls for assistance (Kelling and Pate, 19741. These strategies were assigned randomly to 15 contiguous districts within the city and were maintained for 12 months. The major dependent variables were official police statistics and victim reports. No deterrent results could be found for patrol at either "normal" or"proactive" levels. Critics of the Kansas City study (cited in Zimring, 1978:142-143) have pointed out that the districts were small, and that residents not subjected to patrol could still see police patrolling peripheral areas and responding to calls. There were, in fact, no significant differences in police response time or arrest rates between the three kinds of patrol districts. The com- parison between no-patrol and routine patrol areas might have been con- taminated by this proximity effect. It is notable that the Kansas City patrols were by car whereas prior studies reporting positive effects used foot patrols. However, the study was widely interpreted as failing to show that doubling or tripling of police patrol could measurably affect crime rates. Severity of Punishment The deterrence proposition also predicts that proscribed behavior will be reduced to the extent that the relevant public perceives great severity of punishment for violations. Most research on this hypothesis compares ju- risdictions that have death penalty provisions with those having (presumably less severe) prison sentences; compares jurisdictions with longer and shorter prescribed or actual prison sentences for various offenses; or examines longitudinal effects of changes in punishment seventy on official crime rates (for reviews, see Andenaes, 1974; Zeisel, 1976; Blumstein et al., 19781. Regardless of methods used, there is little evidence directly bearing
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 137 on perceptions of severity; the tacit assumption is that perceptions of severity generally accord with the severity of formal legal prescriptions. This as- sumption may be false in specific circumstances. The Death Penalty Capital punishment obviously cannot be experi- mentally manipulated, and examinations of its deterrent effect have been limited to comparisons of homicide rates in contiguous states with and without the death penalty (Campion, 1955; Sellin, 19591; to examinations of time-series data on homicide rates within one or more jurisdictions that change capital punishment laws (Sellin, 1959; Walker, 19691; and to com- parisons of homicide rates within a jurisdiction before and after the im- position of a death sentence or execution (Graves, 1956; Savitz, 1958~. Although these studies have generally failed to find evidence for a deterrent effect of capital punishment, they have serious methodological problems that compromise their probative value. The main problems lie in their inability to control for demographic, cultural, and socioeconomic factors other than the death penalty that could affect rates of serious criminality, and their failure to distinguish between the formal prescription of the death penalty and its actual application. Recent support for the deterrence proposition in the matter of capital punishment has been reported in a well-known study by Isaac Ehrlich (1975), who examined aggregate U.S. data on homicide rates and capital punishment for the years 1932-1970. After performing an elaborate set of statistical analyses, Ehrlich concluded that capital punishment does deter homicide, offering a specific estimate of the magnitude of the effect (p. 3981: "On the average the tradeoff between the execution of an offender and the lives of potential victims it might have saved was of the order of magnitude of 1 for 8 for the period 1933-1967 in the United States." Ehrlich's study, introduced to the Supreme Court by the Solicitor General in Fowler v. North Carolina (95 Sup. Court 223, 1975), has been widely cited in support of capital punishment legislation and its imposition in individual cases. Because of the importance of the findings and the fact that it is one of the few studies to report a deterrent effect for capital punishment, its supporting data have been reanalyzed several times (e.g., Bowers and Pierce, 1975; Klein et al., 1978~. These analyses show that Ehrlich's findings are sensitive to minor changes in the form of the analysis. Among the most striking is the consequence of changing the time period over which the analysis is made: the negative relationship between homicide rates and executions is present only when the years 1962-1969 are included in the analysis, and these were unusual years for the United States in that both homicide and all other street crimes increased dramatically while the frequency of executions declined steeply (and ceased entirely in 1968; see
138 H. LAURENCE ROSS and GARY D. LAFREE Bowers and Pierce, 1975: 197-202~.3 Thus, the probative value of the Ehr- lich study is doubtful. The bulk of the existing literature on whether capital punishment deters crime more than other forms of punishment (e.g., life imprisonment) is based on data gathered in situations in which few convicted offenders actually receive capital punishment (and the method for selecting which offenders receive it appears highly capricious, despite recent Supreme Court decisions aimed at clarifying standards), so the generally negative findings must be understood as limited to situations in which actual likelihood of punishment is low. It is possible that if executions were applied with higher likelihood they might have an effect on homicide rates, but the wisdom of such a policy is primarily a moral and ethical matter, not a scientific question. In any event, policy decisions about capital punishment are more likely to be affected by moral and ethical considerations than by the crude estimates of deterrent effects that present social research has produced. Other Punishments The quantity of studies of the deterrent effects of noncapital punishments on street crime is somewhat more impressive (for a review, see Nagin, 1978~. However, there is only one study we know of where the design rises to the level of a quasi experiment. Schwartz (1968) studied the effect of increased statutory penalties for rape and attempted rape on the frequency of these crimes in Philadelphia. Following a brutal rape case that received a great deal of media attention, the state of Penn- sylvania raised the maximum penalty for rape by a factor of two or three, depending on the severity of the case. Schwartz examined reported rape rates for the period surrounding the change and concluded that neither the frequency nor the seriousness of rape changed significantly after the new law was passed. The study has obvious weaknesses, most notably the use of reported cases of a crime that is notoriously underreported. However, it avoids the general problem that affects the balance of the known studies: the impossibility of adequately controlling statistically for all the important social, economic, and demographic variables other than deterrence-based laws that can affect crime rates. The general picture painted by the bulk of studies relating severity of punishment for street crimes as measured by length of sentences (prescribed or actual) to crime rates is less favorable to the deterrence proposition (Chiricos and Waldo, 1970, Forst, 1976~. The predominant finding is of no significant relationship. In an exhaustive review of the evidence for a 3A detailed methodological critique of Ehrlich's analysis by Klein et al. (1978:343-351) cites several other reasons for questioning his results, including omitted variables.
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 139 deterrent effect of sentence severity on crime, Nagin (1978:110) concludes that at best the results are "equivocal." As with the death penalty studies, though perhaps to a lesser degree, these studies also take place against the background of a relatively low risk of any punishment. Hence, results more favorable to the deterrence proposition might be found if severe punishments were perceived to be likely in the event of violations by the population to which the threat is addressed. Summary On the matter of certainty of punishment, we find some support for the deterrence proposition in the literature on street crime. With few exceptions, crime rates are found to decline when measures are adopted to increase the certainty of punishment. However, there is much weaker conflation for the deterrence proposition in the matter of severity of punishment, with a few studies claiming an effect contradicted by numerous studies finding no effect. Because all empirical research in the area of severity of penalties takes place in situations where the objective likelihood of punishment is very low, scientific generalizations and policy decisions based on this lit- erature should be appropriately qualified. DRUNK DRIVING Unlike the case for street crimes, active public interest in the issue of drunk driving is relatively recent, and its depth and persistence have not yet been tested. The acute current concern about drunk driving has resulted in a flood of deterrence-based legal interventions that promise to increase our understanding of deterrence both in the specific case and in general. Laws have been passed and enforcement campaigns undertaken with the aim of increasing the perceived certainty and severity of punishments for drunk driving, forming a pool of natural experiments that can be subjected to analysis. Furthermore, the results can be ascertained using indexes such as weekend night fatalities, which are often both validly and reliably mea- sured by official statistics agencies, and which correlate strongly with al- cohol-impaired driving. There now exists a relatively large body of knowledge in this area and additional experience is rapidly accumulating. Certainty of Punishment Two types of legal interventions regarding drunk driving are directed primarily at increasing the objective (and hence, presumably perceived) certainty of punishment. The first of these is the replacement of laws that
140 H. LAURENCE ROSS and GARY D. LAFREE define the offense in behavioral terms with "Scandinavian-type" or "per se" laws, which define the offense in terms of blood-alcohol concentrations, measurable by instruments. Typically these laws require only that a driver exceed the tolerated level in order to justify arrest and to secure conviction; there is no need to demonstrate drunken behavior or to produce other evidence of impairment, a matter Hat under previous law was a severe handicap in detecting and prosecuting drivers whose chances of experienc- ing a crash were substantially increased by the consumption of alcohol. The second type of intervention is the enforcement crackdown, in which police resources devoted to drunk-driving patrols are abruptly increased. We will review the literature accumulated to date on the deterrent effects of these interventions. Scandinavian-type Laws These laws originated in Norway in 1936 and Sweden in 1941, where they formed part of general accumulations of legal restrictions on drunk driving and on overall alcohol use. In the original countries these laws were not very much noticed, being relatively small incremental steps in the accretion of policy. However, over the years, the Scandinavian countries won a reputation (not completely earned, in our opinion) for having dealt successfully with drunk driving, and the laws were copied in other jurisdictions where they represented a sharp break with tradition and therefore were much more noticeable. The first important adoption of this model outside Scandinavia came in Great Britain in 1967. The Road Safety Act of that year prohibited driving or attempting to drive with a blood-alcohol concentration greater than .08 percent (the level that a man of medium build might reach after drinking four or five drinks on an empty stomach in the period of an hour). It was proposed that police be empowered to stop any driver and administer a screening breath test for blood alcohol (using a new device imported in mass quantities from West Germany, with much fanfare). Although this provision was rejected by Parliament, the final legislation permitted a test on anyone involved in an accident (regardless of fault) or committing a serious traffic law violation. In addition to receiving official publicity, the British Road Safety Act was strengthened by media attention that continued for years as the complex law was challenged on numerous grounds by defendants seeking to escape the mandatory penalty of a year's license suspension. Evaluations of the law initially showed substantial deterrent effects: weekend night fatalities and serious injuries declined by more than half immediately following the imposition of the new rule, and attribution of the decline to the law was supported by the failure of comparable casualties to decline during non- drinking hours, as well as by behavioral data reported in polls and other sources (Ross, 1973; Saunders, 19751. However, the effect of the law was
DETERRENCE IN CRIMINOLOGY AND SOCIAL POl ICY 141 not permanent. Matters started to return to the status quo ante shortly after the effect was evident, and within about two years the effect could no longer be demonstrated. Very much the same pattern has been noted in other countries. In 1978, a Scandinavian-type law was adopted in France, a country with one of the highest per capita levels of alcohol consumption in the world. Police were given permission to conduct road blocks in which all drivers passing through could be tested for alcohol, with predictable controversy and publicity. Again, weekend-night serious casualties declined significantly with the inception of the law, and again a reversion was immediately evident, bring- ing matters back to the status quo ante within a year (Ross et al., 19821. In He Netherlands, a Scandinavian-type law introduced in 1974 was evaluated by the superior method of roadside surveys of blood-alcohol concentrations among random samples of drivers, which found an important decline coinciding with the law's inception, along with a trend toward prior levels (Noordzij, 19801. The coincidence of the Dutch law with the fuel crisis of the 1970s unfortunately compromises the probative value of this example, as does failure to find confirmation of the effect in data on total crash-related fatalities. A Canadian law of 1969 was found in two independent studies (Carr et al., 1975; Chambers et al., 1976) to be effective in reducing crash-related fatalities, despite the fact that it was much less threatening than the others mentioned (police needed to have reasonable suspicion of an offense before requiring the test); and there appears to have been much less resistance to and hence notoriety for the Canadian law. As in the other instances, the effect of the Canadian law was not found to be long-lasting. Two negative reports have been made concerning Scandinavian-type laws in Australia (Birrell, 1975) and New Zealand (Hurst, 1978), but both were methodologically deficient studies of what appear to have been relatively unpublicized innovations. Enforcement Crackdowns Police activity to enforce drunk-driving laws is relevant to the deterrence prediction that threatened behavior will be reduced when punishment appears to be more certain. Perhaps the classic case occurred in 1975 when the chief constable of the county of Cheshire, England, experimentally required his men to demand breath tests for alcohol in all situations where the law permitted the request. This experiment was "discovered" by the local press, which elevated the police activity to the status of a campaign. The consequence was significant decreases in serious injuries and fatalities during the month in which the increased enforcement was maintained (Ross, 19771. Similar results were found in both Australia (Cameron et al., 1980) and New Zealand (Hurst and Wright, 1980), where
142 H. LAURENCE ROSS awl GARY D. LAFREE despite the initial negative results reported for their Scandinavian-type laws, campaigns of enforcement for these laws were accompanied by impressive declines in casualties in the affected jurisdictions. Among the most ambitious enforcement efforts were the U.S. Alcohol Safety Action Projects (ASAPs) funded by the United States Department of Transportation in 35 sites in the mid-1970s. It is estimated that more than $200 million in public funds was spent on these projects, which cen- tered on increasing patrol as well as on streamlining the processing of the accused in the criminal justice system. Unfortunately, very much like the High-Impact Anti-Crime program for street crimes, the structuring of the ASAPs varied by city, and the local evaluations were on the whole incom- petent. However, a final evaluation by competent if perhaps not disinterested U.S. Department of Transportation staff (1979) did find evidence for a deterrent effect, as measured by greater reductions in nighttime than in daytime fatal crashes, in 12 of the 25 sites, and in 8 of the 13 sites where the absolute level of nighttime crashes and a moderate population growth rate rendered evaluation less problematic. Severity of Punishment Although efforts to increase the severity of punishment for drunk drivers have probably been much more frequent than those directed at certainty, there are few published evaluations. The efforts have taken the form of increasing statutory penalties and of judicial crackdowns increasing the actual penalties for drunk driving in various jurisdictions. Statutory Changes Many statutory changes in the penalty for drunk driv- ing have been accomplished as part of broad packages of countermeasures, some of which also relate to increasing certainty. One example of a law that appears to have been directed only toward increasing the perceived severity of penalties took place in Finland in 1950, when the maximum sentence for drunk driving was doubled from two to four years, with the provision for six years in He event of serious bodily injury resulting from He offense, and seven years for causing death. Although there was a decline in crash-related fatalities in subsequent years, it proved not to be possible to attribute this decline to the law because it was greater for less serious crashes than for fatal crashes, whereas the latter are more likely to involve drunk driving. Further- more, the drop was greater for multiple-vehicle crashes than for single-vehicle crashes, the latter again more likely to involve alcohol (Ross, 19751. Judicial Crackdowns In Chicago in 1970 the supervising judge of the traffic court decreed that all defendants judged guilty of drunk driving during
DETERRENCE lN CRIMINOLOGY AND SOCIAL POLICY the Christmas holidays should receive a seven-day 143 1' ail sentence. Although subsequently the number of crashes declined, and officials made broad claims of success for this campaign, careful analysis of the data found that the decline could not be distinguished from chance variation. Furthermore, data from Milwaukee, chosen as a comparison jurisdiction, showed an even greater proportional decline, although as in Chicago it was not statistically significant (Robertson et al., 1973~. Similar findings are reported from a city in New South Wales, Australia, where a local magistrate declared his intention to increase greatly the pen- alties for drunk drivers. Research showed that serious crashes did not decline perceptibly even though, unlike in Chicago, the threatened penalties were actually put into effect in most cases (Misner and Ward, 1975~. There is evidence in these studies that the criminal justice system reacts in a way that vitiates the declared severity of actual punishments. For example, in Chicago, it was found that convictions declined where drivers were accused in the absence of chemical tests in evidence. In a more recent study of jurisdictions adopting mandatory jail sentences for first offender DUI (driving under the influence) defendants, Gropper et al. (1983) report that there were important increases in not-guilty pleas, in jury trials, and in failures to appear for trial as well as dismissals and not-guilty findings. Moreover, the eventual punishment for those nonetheless convicted was slowed by considerable increases in delay between arrest and conviction. Summary The research to date on attempts to deter drunk drivers suggests that measures directed at increasing the perceived certainty of punishment can have a sharp, immediate deterrent effect on the proscribed behavior. Rates of crashes likely to involve alcohol decline sharply at the inception of well- publicized laws that simplify apprehension and prosecution, and during well-publicized campaigns of police enforcement. The extent of the ob- served declines in crashes is impressive in light of the fact that crashes involve many causal factors other than alcohol. Deterrent effects have been found in virtually all well-designed studies of significant interventions, in many countries throughout the world. However, these effects universally disappear over time, a matter of several months or a few years at most. One possible explanation for this fact is that the very low levels of actual likelihood of punishment are insufficient to continue an initial impression of reasonable certainty of punishment for the violator. No deterrent effect is evident for legal interventions that are directed solely at increasing the severity of punishment. Although there are only a few reported studies supporting this generalization, Mere are no negative find
144 H. LAURENCE ROSS arid GARY D. LAFREE ings. It seems plausible to attribute this disconf~ation of deterrent expec- tations to the very low risk of punishment of any kind, which permits the violator to regard the Great as negligible, and to possible public perception of the fact that Me criminal justice system does not necessarily deal any more severely after those interventions than before. Severity-based interventions are found to produce undesired and unanticipated side effects Trough the discretion of legal actors, behavior that may reflect the sense that actual offenses detected are a haphazard selection from a much larger population of undetected offenses, so that Rose charged are in a meaningful sense unlucky. RESEARCH AND POLICY IMPLICATIONS We must conclude that, as tests of the scientific validity of the deterrence proposition, existing research is inadequate. For many years social science research simply ignored the issue of deterrence. More recently the size and scope of the deterrence literature has increased dramatically, but it is still incapable of resolving basic theoretical questions, for several reasons. First, evaluations of the deterrent effects of punishment are often based on changes in formal laws rather than changes in actual enforcement be- haviors. This issue is particularly important with regard to sanction severity. Proclaimed increases in the severity of penalties repeatedly have been found to be vitiated by the reluctance or incapacity of legal agents to actually implement the new penalties. As for certainty, most interventions may be described as having increased the objective probability of punishment from "negligible" to "trivial" levels. The sheer resistance of the criminal justice system to piecemeal implementation of new penal sanctions may be the major finding of studies ostensibly testing the deterrence proposition. Second, most prior research on deterrence relies on the unsupported as- sumption that changes in objective levels of certainty and severity of punish- ment are reflected in the perceptions that are the subject of the theoretical proposition being tested. Where the risk levels are extremely low, as is common in Me situations being studied here, and where actual punishments are not increased despite policymakers' intentions, it is hazardous to assume that perceived certainty and severity of punishment have been changed. Third, and more serious for studies of street crime than of drunk driving, is the general inadequacy of research design. Much of the research on street crime is based on correlational and econometric analyses, the defects of which have been well exposed (e.g., Greenberg, 1977; Blumstein et al., 19781. In the few classical experiments, control groups are often contam- inated by proximity to experimental groups. Reliability and validity of measurement are serious problems for most of the studies of street crime interventions, and many of Me time-series quasi experiments fail to control
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 145 for the possibility that events other than the deterrence-based legal inter- ventions may have caused declines in violations. For example, interventions against drunk driving in recent years have frequently coincided with eco- nomic transitions or crises in fuel availability. But despite the admitted weaknesses of evidence in individual studies, the accumulated literature supports a number of tentative conclusions. The deterrence proposition is generally supported in evaluations of legal inter- ventions bearing on the certainty of punishment for the offender. In the short run, at least, there is clear evidence that offense rates decline. In the long run, however, matters are not so clear, very likely because the deterrent effects depend on an overestimation of the chances of apprehension by the relevant public due to publicity and media attention surrounding the inter- ventions. This impression may be difficult to maintain in the face of daily experience that fails to support it. In contrast, there is very little evidence favoring deterrence in the matter of severity of punishment, even in the short run. One explanation is that the relevant public may readily learn or come to expect that the declared severity of penalties is compromised by resistance to change on the part of legal actors. An even more appealing explanation lies in a possible inter- action between perceived severity and certainty: where the likelihood of any punishment is very low, as it very often seems to be, the prospective offender discounts even severe penalties as negligible. Accepting this statement of the evidence, several questions can be raised for policy considerations. First, why does the deterrence approach, espe- cially in the matter of severe punishment, remain so popular as a basis for countermeasures against these and other social problems? Second, what are the prospects for obtaining long-term deterrent results through increasing the probabilities of punishment? Third, what alternatives might be proposed as a basis for more rational countermeasures? The Continued Reliance on Deterrence One reason for the continued tendency to invoke deterrence-related coun- termeasures may be the intuitive appeal of the deterrence proposition. In- trospection informs us that we often refrain from prohibited acts because of threatened punishment. Moreover, our daily experience in the market- place confirms an economic counterpart of the deterrence proposition, which is that when the price of any good is raised, everything else remaining the same, less is consumed. If this intuitive confirmation fails us in nonmarket circumstances, it may be because much criminal and other socially prob- lematic behavior faces threats with unusually small probabilities of fulfill- ment, in which rational calculation and behavior are noted to be uncertain.
146 H. LAURENCE ROSS awl GARY D. LAFREE The utilities involved in deciding whether to drink and drive, or to steal from despised or helpless victims, may be more like those associated with gambling, where hordes of people enjoy participation in mathematically unfair games, than like ordinary market behavior. Second and, we think, very important, the cost of deterrence-based coun- termeasures tends to be delayed and obscure, while the costs of alternatives may be daunting. Increased severity, in particular, can seemingly be invoked with the stroke of a pen. In fact, as the grossly overcrowded conditions in many of the nation's prisons now attest, increasing the severity of punish- ment may be more costly in practice than it appears when considered as a basis for policy. Third, deterrence-based measures frequently relieve established institu- tions and vested interests that might have much to lose from other coun- termeasures. The appeal of deterrence-based approaches to drunk driving, for instance, rests heavily on the assumption that the problem lies with a small minority of irresponsible deviants. Neither the alcoholic beverage nor the automobile industry bears responsibility in this conception of the prob- lem. Likewise, massive public expenditures aimed at deterring street crime, though ineffective, focus attention on individual deviants and away from possible defects in the social structure, including gross inequalities in life chances among different perpetrator groups. Fourth, deterrence-based reasoning may provide a cover for retribution- based motives fueling popular movements. The person injured by a mugger or drunk driver understandably may feel that muggers and drunk drivers deserve punishment, but demands for action based on this feeling may be more successful when legitimated by the promise of reductions in future damage from these causes. Perhaps in part for this reason, the anti-drunk- driving movement has been more insistent on enacting severe penalties than on requiring occupant-protection devices that would be much more effective in reducing fatalities. The Prospect for Increased Certainty Deterrence-based policy seems to founder on the low actual rates of apprehension and punishment for offenders. It is possible that greatly in- creased investments in criminal justice might in the end be effective; nothing in the research literature refutes this possibility. But we doubt that this is a profitable line of endeavor. There is little in accumulated experience with street crimes or drunk driving to suggest where, in the scale of probability of punishment, a threshold of appreciably greater effectiveness may lie, but it seems likely to involve levels of police intrusiveness and expense otherwise unknown in democratic societies. The American public is deeply
DETERRENCE IN CRIMINOLOGY AND SOCIAL POl.[CY 147 ambivalent about this approach and fundamentally negative when presented with unbiased estimates of the costs. For example, although the San Diego study showed that field interrogation by police might be an effective crime deterrent, many commentators have rejected this kind of police activity. Charles Reich (1979:113) has put it thus: "I fully recognize that safety is important and that safety requires measures. But other qualities also require measures: I mean independence, boldness, creativity, high spirits." Many Americans envy the tranquility of Japanese urban society, but most would find the Japanese system of policing, in which neighborhood police provide their headquarters with updated information on the daily lives of residents of their jurisdiction, too intrusive. The research literature also indicates that legal actors resist substantive changes in their job-related behavior. This may be caused by feelings of injustice when the prescribed penalty fails to fit the crime, or it may simply be rationalized human laziness or procedural resistance to change. As Zim- ring (1978:171) states the case, "The resiliency of courts and police when policy changes are induced by outside money investments is formidable." Thus, in a recent study of the response of police to the mandate to "do something about rape" in a large midwestern city (LaFree, 1981), it was clear that the police changed those things easiest to change primarily recordkeeping while doing little about the things that the deterrence prop- osition suggests would be most important for actually reducing rape, such as increasing arrests and filing more felony complaints. Alternatives to Deterrence The social science literature on both street crime and drunk driving suggests then it may be more fruitful to consider offenders as rational reactors to their social environments rather than as irrational, maladapted, or pathological deviants. The case may be easier to make for the general population in such matters as white-collar crime (e.g., filing deceptive tax returns) than among street cnminals, but the literature suggests that when the world is viewed from the deviant's position in society much of the problematical behavior appears normal and predictable (cf. Lempert, 1981- 1982, for a discussion of this point among fathers ordered to pay child support). Surely the case is easily made for the drunk driver, who exists in a society that institutionalizes the use of alcohol as a social lubricant and mandates dependence on the private automobile for most of its members. Because the probability of the most severe consequence is in reality min- uscule (one fatality for about one third of a million miles of drunk driving), it seems easy for the individual driver departing a tavern late at night to answer the question, "How do I get home?"
148 H. LAURENCE ROSS and GARY D. LAFREE For both street crime and drunk driving, this line of thought has led to suggestions for modifying the situation, rather than the individual, to reduce the problematic behavior. Street crimes may perhaps be better reduced by "hardening the target" than by deterring or, for that matter, reforming potential criminals. For example, Oscar Newman (1972) has written ex- tensively on how urban design can reduce crime by making living areas more "defensible." Similarly, drunk driving has been reduced in the 18- to 21-year-old age group as a consequence of raising the drinking age in various states (Wagenaar, 19821. Effects are also being found for the re- s~iction of young drivers to daylight (nondrinking-hour) driving only (Preusser et al., 1984~. There is evidence for the view that raisin" the price of alcoholic beverages through taxation would reduce drunk driving along with a host of other alcohol-related problems (Moore and Gerstein, 19811. Much better public transportation, including subsidized taxi-like service, is another promising countermeasure. A somewhat different approach to these problems is based on accepting the difficulty of fundamental changes in social institutions as well as be- havior, and striving to make the results of the problematic behavior less damaging to the victims. For example, drunk driving would be of relatively less consequence if it did not entail deaths and injuries. By "padding" the car with passive restraints and the highway with soft shoulders and yielding barriers around fixed obstacles, the inevitable crashes caused by alcohol (and a host of other factors) could be better absorbed by society. Similarly, the cost of such activities as burglary could be lowered, though not elim- inated, by social insurance schemes to repay the victims' financial losses. Perhaps the main reason these types of countermeasures are less attractive than those based on deterrence is that their costs are out front, and Hey are not trivial. We further admit that they are unlikely to be "solutions" to He problems they address; most are merely mitigating. However, it spikes us as a more rational policy to experiment along these lines in hopes of finding economically sound mitigants than to follow the chimera of deterrence-based solutions that experience repeatedly shows to be inade- quate in the context for which they are proposed. ,, ,6 We would like to acknowledge the helpful comments of Gwynn Nettler and Jack Gibbs. REFERENCES American Friends Service Committee 1971 Struggle for Justice: A Report on Crime and Punishment in America. New York: Hill and Wang.
DETERRENCE IN CRIMINOLOGY AND SOCIAL POLICY 149 Andenaes, J. 1974 Punishment and Deterrence. Ann Arbor: University of Michigan Press. Archer, D., Gartner, R., and Beittel, M. 1983 Homicide and the death penalty: a cross-national test of a deterrence hypothesis. Journal of Criminal Law and Criminology 74:991-1013. Birrell, J. 1975 The compulsory breathaliser .05 percent legislation in Victoria. Pp. 775-785 in S. Israelstam and S. Lambert, eds., Alcohol, Drugs and Tragic Safety. Toronto: Addiction Research Foundation of Ontario. Blumstein, A., Cohen, J., and Nagin, D., eds. 1978 Deterrence and Incapacitation: Estimating the Elects of Criminal Sanctions on Crime Rates. Panel on Research on Deterrent and Incapacitative Effects, National Research Council. Washington, D.C.: National Academy of Sciences. Bowers, W., and Pierce, G. 1975 The illusion of deterrence in Isaac Ehrlich's research on capital punishment. Yale Law Journal 85:187-208. Boydstun, J. 1975 San Diego Field Interrogation: Final Report. Washington, D.C.: The Police Foun- dation. Cameron, M., Strang, P., and Vulcan, A. 1980 Evaluation of Random Breath Testing in Victoria, Australia. Paper presented at the Eighth International Conference on Alcohol, Drugs and Traffic Safety, Stockholm. Campion, D.R. 1955 Does the death penalty protect state police? Appendix F. Part I, Minutes of Proceedings and Evidence, No. 20, Joint Committee of the Senate and House of Commons on Capital Punishment and Corporal Punishment and Lotteries, Canadian Parliament. Reported in part in H. Bedau, ea., The Death Penalty in America, revised ed. Garden City: Anchor Books. Carr, B., Goldberg, H., and Farbar, C. 1975 The Canadian breathaliser legislation: an inferential evaluation. Pp. 679-687 in S. Israelstam and S. Lambert, eds., Alcohol, Drugs and Traffic Safety. Toronto: Addiction Research Foundation of Ontario. Chaiken, J. 1976 What's known about deterrent effects of police activities. The Rand Paper Series. Santa Monica: Rand Corporation. Chaiken, J., Lawless, M., and Stevenson, R. 1974 The Impact of Police Activity on Crime: Robberies on the New York Subway System. Report number R-1424-NYC. Santa Monica: Rand Corporation. Chambers, L., Roberts, R., and Voeller, C. 1976 The epidemiology of traffic accidents and the effect of the 1969 breathaliser amendment in Canada. Accident Analysis and Prevention 8:201-206. Chelimsky, E. 1976 HighImpact Anti-Crime Program. National Institute of Law Enforcement and Criminal Justice, Law Enforcement Assistance Administration: U.S. Department of Justice. Chiricos, T., and Waldo, G. 1970 Punishment and crime: an examination of some empirical evidence. Social Problems 18:200-217. Cook, P. 1977 Punishment and crime: a critique of current findings concerning the preventive effects of punishment. Law and Contemporary Problems 41:164-204.
150 H. LAURENCE ROSS and GARY D. LAFREE Ehrlich, I. 1975 The deterrent effect of capital punishment: a question of life and death. The American Economic Review 65:397-417. Forst, B. 1976 Participation in illegitimate activities: furler empirical findings. Policy Analysis 2:477 492. Gallagher, F. 1978 Appendix: an annotated bibliography of deterrence evaluations, 1970-1975. Pp. 174- 186 in A. Blumstein, J. Cohen, and D. Nagin, eds., Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. Washington, D.C.: National Academy of Sciences. Gibbons, D. 1979 The Criminological Enterprise: Theories and Perspectives. Englewood Cliffs, N.J.: Prentice-Hall. Gibbs, J. 1975 Crime, Punishment and Deterrence. New York: Elsevier. Graves, W. 1956 The deterrent effect of capital punishment in California. Reprinted in part in H. Bedau, ea., The Death Penalty in America, revised ed. Garden City: Anchor Books. Greenberg, D. 1977 Deterrence research and social policy. Pp. 281-295 in S. Nagel, ea., Modeling the Criminal Justice System. Beverly Hills: Sage. Gropper, B., Martorama, C., Mock, L., O'Connor, M., and Travers, W. 1983 The Impacts of Mandatory Confinement for Drunk Driving on Criminal Justice Op- erations (summary report). Technical report. Washington, D.C.: U.S. Department of Justice, National Institute of Justice. Hindelang, M. 1976 Criminal Victimization in Eight American Cities: A Descriptive Analysis of Common Theft and Assault. Cambridge: Ballinger. Hurst, P. 1978 Blood test legislation in New Zealand. Accident Analysis and Prevention 10:287-296. Hurst, P., and Wright, P. 1980 Deterrence at Last: The Ministry of Transport's Alcohol Blitzes. Paper presented at the Eighth International Conference on Alcohol, Drugs and Traffic Safety, Stockholm. Jeffery, C. 1972 The historical development of criminology. Pp. 458-498 in H. Mannheim, ea., Pio neers in Criminology, 2nd ed. Montclair, N.J.: Patterson Smith. Kelling, G., and Pate, A. 1974 The Kansas City Preventive Patrol Experiment. Washington, D.C.: The Police Foun- dation. Klein, L., Forst, B., and Filotov, V. 1978 The deterrent effect of capital punishment: an assessment of the estimates. Pp. 336 360 in A. Blumstein, J. Cohen, and D. Nagin, eds., Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. Washington, D.C.: National Academy of Sciences. LaFree, G. 1981 Official reactions to social problems: police decisions in sexual assault cases. Social Problems 28:582-594. Lempert, R. 1981- Organizing for deterrence: lessons from a study of child support. Law and Society 1982 Review 16:514-568.
DETERRENCE IN CRIMINOLOGY AND SOCIAL policy 151 Lipton, D., Martinson, R., and Wilkes, J. 1975 The Effectiveness of Correctional Treatment: A Survey of Treatment Evaluation Stud- ies. New York: Holt, Rinehart, and Winston. Martinson, R. 1974 What works? Questions and answers about prison reform. The Public Interest 35:22- 34. Misner, R., and Ward, P. 1975 Severe penalties for driving offenses: deterrence analysis. Arizona State Law Journal 1975:677-713. Moore, M., and Gerstein, D. 1981 Alcohol and Public Policy: Beyond the Shadow of Prohibition. Washington, D.C.: National Academy Press. Nagin, D. 1978 Generaldeterrence: a review of the empiricalevidence. Pp. 95-139inA. Blumstein, J. Cohen, and D. Nagin, eds., Deterrence and Incapacitation: Estimating and Effects of Criminal Sanctions on Crime Rates. Washington, D.C.: National Academy of Sciences. Newman, O. 1972 Defensible Space. New York: Macmillan. Noordzij, P. 1980 Recent Trends in Countermeasures and Research on Drinking and Driving in the Netherlands. Paper presented at the Eighth International Conference on Alcohol, Drugs and Traffic Safety, Stockholm. President's Commission on Law Enforcement and Administration of Justice 1967 The Challenge of Crime in a Free Society. Washington, D.C.: U.S. Government Printing Office. Press, S. 1971 Some Effects of an increase in Police Manpower in the 20th Precinct of New York City. Santa Monica: Rand Corporation. Preusser, D., Williams, A., Zador, P., and Blomberg, R. 1984 The effect of curfew laws on motor vehicle crashes. Law and Policy. Reckless, W. 1967 The Crime Problem, 4th ed. Englewood Cliffs, N.J.: Prentice-Hall. Reich, C. 1979 Police questioning of law abiding citizens. Pp. 107-113 in J. Bonsignore, E. Katsh, P. d'Errico, R. Pipkin, S. Arons, and J. Rifkin, eds., Before the Law. Boston: Houghton Mifflin. Robertson, L., Rich, R., and Ross, H. 1973 Jail sentences for driving while intoxicated: a judicial policy that failed. Law and Society Review 8:55-67. Ross, H. 1973 Law, science and accidents: the British road safety act of i967. Journal of Legal Studies 2:1-78. 1975 1977 The Scandinavian myth: the effectiveness of drinking-driving legislation in Sweden and Norway. Journal of Legal Studies 4:285-310. Deterrence regained: the Cheshire constabulary's "breathaliser blitz. " Journal of Legal Studies 6:241-249. Ross, H., McCleary, R., and Epperlein, T. 1982 Deterrence of drinking and driving in France: an evaluation of the law of July 12, 1978. Law and Society Review 16:345-374.
152 H. LAURENCE ROSS and GARY D. LAFREE Saunders, A. 1975 Seven years experience of blood-alcohol limits in Britain. Pp. 845-853 in S. Israelstam and S. Lambert, eds., Alcohol, Drugs and Trawl c Safety. Toronto: Addiction Research Foundation of Ontario. Savitz, L. 1958 A study in capital punishment. Journal of Criminal Law, Criminology and Police Science 49:338-341. Schwartz, B. 1968 The effect in Philadelphia of Pennsylvania's increased penalties for rape and attempted rape. Journal of Criminal Law, Criminology and Police Science 59:509-515. Scull, A. 1977 Decarceration. Englewood Cliffs, N.J.: Prentice-Hall. Sellin, T. 1959 Sutherland, E. The Death Penalty. Philadelphia: American Law Institute. 1924 Criminology. Philadelphia: Lippincott. Tittle, C. 1980 Sanctions and Social Deviance. New York: Praeger. U.S. Department of Transportation 1979 Alcohol Safety Action Projects: Evaluation of Operations, Data, Tables of Results and Formulation. Washington, D.C.: National Highway Traffic Safety Administration. Void, G. 1979 Theoretical Criminology, 2nd ed. New York: Oxford University Press. Wagenaar, A. 1982 Effects of the minimum drinking age on automotive crashes involving young drivers. UMTRI Research Review 13:3-12. WaLker, N. 1969 Sentencing in Rational Society. London: Penguin Press. Wilson, J. 1975 Thinking About Crime. New York: Basic Books. 1983a Thinking about crime: the debate over deterrence. The Atlantic Monthly. September, Pp. 72-88. 1983b Crime and American culture. The Public Interest 1983:22-48. Zeisel, H. 1976 The deterrent effect of the death penalty: facts vs. faiths. In P. Kurland, ea., The Supreme Court Review: 1976. Chicago: University of Chicago Press. Zimring, F. 1978 Policy experiments in general deterrence: 1970-1975. Pp. 140-174 in A. Blumstein, J. Cohen, and D. Nagin, eds., Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. Washington, D.C.: National Academy of Sciences. Zimring, F., and Hawkins, G. 1973 Deterrence. Chicago: University of Chicago Press.
