A Practical Measure of Offence Seriousness: Sentence Severity , continued
In most outcome evaluation work, treatment effectiveness is determined by the differential recidivism rates for treated and untreated offenders. Typically, recidivism rates are stated as the proportion of recidivists to the total sample of programme participants over a follow-up period. Recidivism has most often been expressed as a dichotomous variable established by the presence or absence of specific outcome criteria such as further arrests, further convictions or further incarcerations.
Researchers have long recognized the inadequacies of such all-or-none measures and of the recidivism rates derived from them (Maltz, 1984). The usual criticism is that the dichotomous nature of the recidivism variable ignores a substantial amount of relevant information about the recidivistic event. This reduces the ability of researchers to discriminate among groups and lowers the likelihood of being able to assess varying degrees of impact of correctional programmes on individual offender’s post-release performances. For example, the post-release changes stemming from correctional programmes may take the form of a stepwise series of decrements in both quality and quantity of criminal behaviour. It is possible that each successive intervention may have a cumulative effect, with some interventions affecting the types of crimes committed and others affecting the individual’s propensity to commit criminal acts.
A number of attempts have been made to develop methods for the measurement of offence seriousness in order to accommodate this sort of thinking. It has been suggested that improved discrimination might be achieved by examining the distribution of failure times represented by time on the street to recidivism. However, the latter approach measures something equivalent to resistance to recidivism rather than recidivism per se. Thus, while resistance to recidivism is in the form of distributed data, it does not provide a measure of offence seriousness.
Measures of offence seriousness are intended either as tools for evaluating correctional outcomes or else as indices of crime in society. Representative of measures for correctional evaluation are scales where the degree of seriousness is expressed as ordinal weights attached to offence descriptions or offence categories (Gendreau & Leipciger, 1978; Keller & Carlson, 1977;Klein, Newman, Weiss & Bibner, 1983; Moberg & Ericson, 1972; Billingham & Thorpe, 1977; Witherspoon, de Valera & Jenkins, 1973). The best representative of scales developed to measure crime in society is the Offence Seriousness Index (Sellin & Wolfgang, 1964).