ABSTRACT
Earlier, in Chapter 4, we described Meehl’s (1954) seminal work distinguishing between clinical and actuarial prediction and his description of actuarial assessment. As we indicated at the end of Chapter 5, Meehl conceived of actuarial assessment as consisting of two stages. In the first stage, which he referred to as classification, individuals are scored using some actuarial instrument, and classified according to their score on the instrument. Virtually all of our discussion in Chapters 4 and 5 pertained to Meehl’s first stage of actuarial assessment. The only allusion we made to Meehl’s Stage 2 was in our Figure 5.1 wherein we plotted absolute recidivism rates for each score value group resulting from our brief actuarial scale (BAS). In discussing Meehl’s Stage 1, we have focused on the development of the actuarial items, the scoring of the instruments, and the resulting accuracy of the actuarial risk assessment instruments (ARAIs) in identifying eventual recidivists. This stage allows evaluators to classify offenders as to their level of relative risk. Now that we have assigned the offender to a level of relative risk (Joe is higher risk than Sam but lower risk than Peter), we turn our attention to establishing the absolute level of risk by determining a numerical probability of recidivism, often expressed as a percentage. In other words, if we release 100 sex offenders into the community for a period of five years, how many will be charged or convicted of a sexual re-offense? The reader will remember that this last step in actuarial prediction was called for in Monahan and Steadman’s (1994a) model of statistical prediction.
