the data this study is based on provide few opportunities for examining sanction effects other than those derived from incapacitation. Nonexperimental studies require a high number of individual-level variables to estimate the effects of sanctions, as it must be ascertained that offenders who are sentenced to different sanctions are actually comparable in regard to those factors that influence recidivism risk. The techniques used in such ex post facto studies are difficult, and even the best research often gives rise to debate and doubt over the validity of the results (see Börjeson, 1966; Elmhorn, 1966; Quensel, 1966; Bondeson, 1977, 1986, and 1987; Kühlhorn, 1987). For instance, one can never be certain whether the researcher has taken into account precisely those conditions that have caused comparable offenders to receive different sanctions – conditions that can also influence the risk of recidivism.

As mentioned in Chapter 6, legislation aimed at lowering the level of sanctions for certain crimes took place within the study period. It is therefore possible to examine the effects of this change. Opportunities such as this are referred to as natural experiments, since, if no significant social changes occur at the same time, it can be assumed that offenders sentenced after the change are comparable to those sentenced before.

This chapter is titled “The Deterrent Effect of Sanctions.” In principle, however, it could also be “The Rehabilitative Effect of Sanctions.”