An assessment of the reliability and validity of the Security Rating and Inmate Risk Assessments used in correctional institutions in Newfoundland and Labrador

• Main Author: Burden, Kelly L.
• Format: Thesis
• Language: English
• Published: Memorial University of Newfoundland 1996
• Subjects: Newfoundland
• Online Access: https://research.library.mun.ca/5878/; https://research.library.mun.ca/5878/1/Burden_KellyL.pdf; https://research.library.mun.ca/5878/3/Burden_KellyL.pdf

The Security Rating and Inmate Risk Assessments, two instruments used in correctional institutions in Newfoundland and Labrador, were examined for their ability to predict recidivism. The Security Rating Assessment is primarily composed of items related to an offender's past and present criminal behaviour. In contrast, the Inmate Risk Assessment is composed of "need" items: items that deal with the type and severity of social, emotional, and economic problems experienced by an offender. Recidivism was defined as reincarceration and both a dichotomous criterion, recidivist vs. non-recidivist, and a continuous criterion, number of violation-free days, were used. Three institutions with different male inmate populations were separately examined to determine the reliability and validity of the two instruments. The correlations obtained at the three institutions were coverted to z-scores and tests of differences were conducted. In cases where significant differences were not found, a single correlation was calculated using the combined samples. The recidivism rate for all the institutions combined was 18.57% with a low of 11.74% and a high of 29.21%. Both assessments were found to be reliable as shown by the intercorrelations and Cronbach's alpha. In general, the assessments were also found to be valid as shown by the criterion-to-total score correlations. Also, the recidivism rates correlated positively with the different security and risk levels as measured by the Security Rating and Inmate Risk Assessments. Specifically, more medium-security inmates than minimum-security inmates were recidivists. As well, low-risk inmates were less likely to recidivate than medium or high-risk inmates. The Inmate Risk Assessment could not discriminate between the medium and high-risk categories. In conclusion, more accurate predictions of recidivism can be made by using the Security Rating and Inmate Risk Assessments.