General Deterrence Theory

The notion of general deterrence holds that crime rates of populations are held down by the threat of generalized punishments. The probabilities of such punishments include those based on perceptions and reality. However, people's perceptions may not be based on reality but on other sources such as sensational media coverage of crime and punishment. Punishments may include arrest and jail time as well as probation, community service, tethering, and fines.

Three characteristics of a punishment need to be at an optimal level in order for the threat of a punishment to deter crime: certainty, celerity, and severity. First, punishments must be relatively certain. The exact level of certainty that must be reached in order for punishments to deter crime, and the presence of nonlinear floor and threshold effects, are largely unknown. However, some research suggests that a floor of at least 30 percent of offenders may need to be arrested before a deterrent effect of arrest for some crimes is realized. It is plausible that there is also a threshold effect wherein gains in certainty may bring little, if any, deterrent effect. Second, punishments need to be relatively swift. If too much time passes between arrest and actual sentencing/punishment in the criminal justice system, punishment is less apt to deter persons in society from committing crime. For example, half a century ago, less than 2 years elapsed between a death sentence and an execution; now the gap is more than a decade, and in California it has approached 20 years. Third, punishments need to be severe enough so that persons in the general population perceive that the costs of punishments outweigh the potential gains (e.g., financial, emotional satisfaction) from crime. Finally, in order for punishment to have the best deterrent effect on crime rates, all three of these dimensions of punishment need to be at an optimal level.

There is little research on general deterrence theory that incorporates measures of all three principles: certainty, celerity, and severity, both in terms of realities and perceptions, simultaneously. Work often uses just one, or perhaps two, measures of the three constructs, and focuses on either perceptions or reality. Measures have included imprisonment rates and arrest rates for certainty and mean sentence lengths for severity. Additionally, when a deterrent effect is found, it is often unclear how important a predictor it is relative to non-criminal justice system predictors of crime; the latter are typically not fully covered in such research studies. However, a recent meta-analysis sheds some light on this critical issue.

In 2005, Travis Pratt and Francis Cullen conducted a meta-analysis of 214 studies containing 1,894 findings concerning 31 aggregate-level predictors of crime rates. A total of 196/1,894 findings reviewed concern six criminal justice deterrence variables. The incidence of incarceration (46 findings) was the fifth most important predictor [Page 364]of crime rates (with noneconomic institutions, unemployment duration, firearms, and percent nonwhite being stronger predictors). However, five other criminal justice system deterrence factors were weak or marginal predictors of crime rates. These factors with their respective ranks and number of findings upon which the rank was based were arrest ratio (rank 23rd, 77 findings), police expenditures (27th, 13), get-tough policies (28th, 37), police per capita expenditure (30th, 20 findings) and police size (31st, 3 findings).

...