NCJ Number
72826
Date Published
1980
Length
49 pages
Annotation
The Salient Factor Score prediction instrument was tested on an inmate sample from the Massachusetts Department of Correction to examine the feasibility of predicting potential risk of recidivism and of prerelease program non-completion.
Abstract
Predictive instruments of these two outcome situations were developed to be used at different junctures of the incarceration process: the reception/diagnosis stage, the intermediate period of incarceration, and the releasing stage. Salient factor scores for predicting recidivism risk potential and for predicting successful completion of pre-release placement were developed from 1975 construction samples of all releases from Massachusetts correctional facilities and of all inmates in Massachusetts facilities placed in pre-release centers. The resultant scores, when computed for each individual in comparable 1976 validation samples of Massachusetts inmate releases and inmates placed in pre-release centers, showed very low correlations, below the .05 probability level. Therefore, the evidence for validation was quite weak. It is concluded that operational usage of prediction instruments for both outcome situations should proceed with extreme caution. However, use of the scores when approaching the high and low risk extremes appear to be more justifiable. Frequency distributions occurring in the original data sets suggest that a disproportionate number of cases fall in the median category, which points to use of the constructed scores only for experimental purposes. Suggestions for further research into the use of Salient Factor Scores include the need to incorporate more data elements and reduce unknown, inconsistent, and inaccurate data elements to make a stronger predictive instrument. Nine data tables and four appendixes provide samples of score sheets, frequency distributions of salient factor scores for both outcome variables, and salient factor score items.