The Deterrent Effect of Perceived Severity: A Reexamination

In a recent study, Grasmick and Bryjak argue that the failure of previous perceptual deterrence researchers to find an inverse relationship between perceived severity of punishment and criminal involvement is due to the fact that they used an invalid measure of perceived severity and tested an additive rather than an interactive model of the deterrence process. Using cross-sectional data, Grasmick and Bryjak found a moderate inverse relationship between their "refined" measure of perceived severity and self-reported past criminal conduct for those who also perceived the certainty of punishment to be high. This paper attempts to replicate and extend Grasmick and Bryjak's research using panel data. With a correct temporal ordering of variables and an identical severity measure, however, we find that perceived severity has no deterrent effect on later deviant behavior. Our data reveal that the "refined" measure of Grasmick and Bryjak confounds the threat of legal sanctions with the fear of informal penalties, and that the greatest effects on delinquent involvement are from those informal sources of social control. In the empirical literature on social control and criminal behavior much has been written about the deterrent effect of the threat of legal sanctions. Although research on the deterrence question was long dormant, an explosion of research occurred after publications by Gibbs (a) and Tittle (a). After some early work on the punishment properties of statutes and aggregate crime rates by Antunes and Hunt, Bailey et al., Bean and Cushing, Chiricos and Waldo, Gray and Martin, Logan, and others,' the central deterrence proposition became recognized as one relating perceived properties of punishment to involvement in crime. The bulk of this literature has shown that the perceived certainty of legal punishment does have a mod