6 minute read


Methods Of Research

In spite of the great importance accorded deterrence in lawmaking and sentencing, deterrence remained a neglected field of research until about 1970, in part because of ideology and in part because of great methodological difficulties. In subsequent years research activity has been intense. Most of the research falls under the following categories.

Comparison over time. The most straightforward method of exploring the effects of a change in legislation or enforcement on the rate of crime is before-and-after research. The great difficulty in such research is to identify the impact of the change among all the other factors that have been involved at the same time. Only abrupt and major changes can be expected to give clear statistical evidence of the effects. Changes introduced in the criminal justice system may be accompanied by changes in the tendency of the victims to report the crime or by changes in the practice of crime recording by the police, so that the statistics are not comparable. These difficulties can, to some degree, be overcome by victimization studies undertaken both before and after the reform.

Perhaps the best-known example of before-and-after research was conducted in Great Britain in connection with the Road Safety Act of 1967, which made it an offense to drive with a blood alcohol concentration of 0.08 percent or more. The penalty is normally a fine and loss of driving license for one year on the first offense. From the day the new legislation went into effect, there was a considerable drop in highway casualties as compared with previous years. For the first three months casualties were 16 percent lower than in the preceding year, and deaths were down by 23 percent. For the night hours casualties were reduced by about 40 percent. Unfortunately, it seems that most of the effect has gradually been lost. As time passed it became increasingly difficult to isolate the effects of the law, but H. Laurence Ross's conclusion seems well founded: the benefits produced by the legislation had largely been canceled by the end of 1970 (p. 77).

According to Ross, the explanation of this declining effect lies in a lack of enforcement. The publicity accompanying the law had given the public exaggerated and quite unrealistic ideas about the risk of apprehension and conviction, but little effort was made to enforce the law. The police did not perceive the law as defining an important task, and gradually the public learned that it had overestimated the risk.

The crucial importance of the risk of detection in this area is convincingly demonstrated by the effects of the Finnish reform of drinking-anddriving legislation in 1977 (Andenaes, 1988, pp. 42–63). Before the reform Finland had the most severe sentences for drunken driving among the Scandinavian countries, with prison sentences of several months. After the reform the great majority of offenders got fines or suspended prison sentences. At the same time a fixed limit of 0.05 percent blood-alcohol concentration was established, and the amount of random breath tests of drivers was drastically increased, from about 10,000 in 1977 to about 700,000 in 1984. Roadside surveys of a representative sample of drivers showed that the proportion of motorists driving under the influence of alcohol after the reform had been reduced to about half. The number of alcohol-related accidents also had diminished, although not to the same degree. The main reason for this probably is that many alcohol-related accidents are caused by drivers who have serious alcohol problems and do not react to the threat of punishment in the same way as average drivers.

Comparison between geographic areas. A second method is to compare areas with differences in legislation, in sentencing, or in law enforcement, to see whether these differences are reflected in crime rates. This method was used in research on capital punishment as early as the 1920s, by comparing murder rates in retentionist and abolitionist states. Beginning in the late 1960s the method of geographical comparison has been widely used for different types of crime, by both sociologists and economists, who have employed various statistical techniques in order to discover the effects of differences in certainty and severity of sanction. Most of the American studies use the individual states as units of comparison, are based on official statistics, and are limited to the seven index crimes (homicide, assault, rape, robbery, burglary, larceny, and auto theft, as enumerated by the Federal Bureau of Investigation).

The research has almost invariably found an inverse relationship between certainty of punishment (or rather certainty of imprisonment) and crime rates. Some, but not all, of the researchers have found a similar but mostly lower relationship between severity of punishment (normally measured in length of prison sentences) and crime rates. The findings are, however, difficult to interpret. A few points should be mentioned:

  1. Many of the studies do not try to distinguish between effects of deterrence and effects of incapacitation. The effects they ascribe to deterrence may in fact be a result of the incapacitation of offenders sentenced to prison.
  2. A correlation between crime rates and severity and certainty of sanction does not in itself say anything about the direction of causality. Crime rates may influence severity or certainty of sanction as well as the other way around. The correlation may also be the result of a third factor, for example, the normative climate in a society. Few of the studies tackle these problems in a wholly satisfactory way.
  3. The statistical equations have certain built-in assumptions that are not necessarily true.
  4. If a study does not find a correlation between crime rates and severity or certainty of sanction, this does not prove that the differences in severity or certainty are without effect but only that in the given sample the effect is not of a sufficient magnitude to be statistically demonstrable.
  5. As noted previously, official crime statistics fail to account for variations in rates of victim reporting and police recording of offenses. A low crime-reporting and/or recording rate tends to simultaneously lower the official crime rate, while raising the apparent rate of imprisonment; a high reporting or recording rate has the opposite effect. These variations naturally tend to produce a spurious inverse relationship between official crime rates and imprisonment rates.

For these and other reasons the comparative research should not be accepted uncritically. The highly technical character of such research also constitutes a barrier against practical application until a high degree of agreement among researchers is reached.

Survey research. Survey research can be of interest to the theory of deterrence in many ways. The simplest form of such research consists in collecting data on public knowledge and beliefs about the system of criminal justice. Studies have generally found that such knowledge is low and haphazard. Comparisons over time or between geographical areas of such surveys can be used to explore how perceptions of severity and certainty of punishment vary with actual severity and certainty.

The survey method seems especially suitable for research into the moral effects of criminal law. Attitude surveys in England before and after introduction of the blood-alcohol limit (Sheppard) showed that the new statute and the accompanying publicity did not have any tangible effect on the attitudes to drinking and driving. In contrast, a survey study from Norway, where similar but stricter legislation had been in force for forty years, indicated that the law had been successful in reaching the citizens with its message (Hauge). Thus, the two studies taken together give support to the view that the moral effect of the law depends on a longtime process.

Additional topics

Law Library - American Law and Legal InformationCrime and Criminal LawDeterrence - The Concept, A Historical Perspective, Empirical And Ethical Questions, General Deterrence: Myth Or Reality?