45

This paper presents some methodological, epistemological, and ethical considerations relevant to an « offbeat » use of statistics on the criminal justice system. The use is « offbeat » because it does not seek identifying predictors related to the probability that an offender may recidivate but at the identifying predictors related to the criminal justice decision makers’ activities (recommendations and decisions). Producing this type of statistical data can be analysed as counter-conduct (on the part of the researcher) seeking to produce a similar phenomenon in the minds of offenders and decision makers. The paper illustrates this offbeat use using research on parole decision making in Belgium.