ON POSITIVE METHODOLOGY, ETHICS, AND POLEMICS IN DETERRENCE RESEARCH

Abstract

and the role of positive research, and point out what I consider the basic deficiencies in Beyleveld's analysis. My approach to criminal behaviour rests in essence on the following general premises: (a) human beings respond to incentives; (b) offenders are human beings; therefore (c) offenders respond to incentive. I consider the first generalisation to be a good approximation for the study of human behaviour in general, and the second to be a good starting-point for studying criminal behaviour. I do not see, however, the basic strength of my approach to criminal activities as deriving from the plausibility of these two general isations or from their logical derivative as summarised in my third statement, but from the fact that it can be tested against, and rejected by, empirical evidence. This, I submit, is the basic merit of any positive analysis whether applied to natural or human phenomena. The hypothesis that offenders respond to incentives has not been automatically treated in my empirical research as an integrating assumption of general applicability, but has first been developed systematically in order to derive its various logical ramifi cations or behaviourial propositions, and then tested empirically to see if it could be accepted as a plausible mechanism to explain observed criminal behaviour. It is this general analytical issue, not capital punishment, which has been the focus of my reaserch, as is best summarised in articles by me which receive no mention in Beyleveld's critique.1 What I term the economic approach to crime has been tested by me and by others in connection with a wide variety of offences ranging from auto-theft and violations of anti-trust laws to skyjacking and murder. My own findings have been replicated and corroborated by other researchers using independent data sets from the United States, Great Britain and Canada. The evidence examined has included aggregate time series data as well as cross-sectional statistics at the state, S MSA, and city levels from different time periods and different geographical regions.2 The evidence from all of these studies is not inconsistent with a detailed set of implications derived