cover image: Identifying the Community High Risk Population for Allocation of the Program Integrity Funding /

Premium

20.500.12592/c8m1d0

Identifying the Community High Risk Population for Allocation of the Program Integrity Funding /

26 Feb 2019

The final models were selected based on the strength of their association with reoffending, the feasibility of identifying the factors in the Offender Management System (OMS), and the parsimony of the model. [...] The formula is designed to be dynamic, with the identified FTEs changing with workload demands based on the number and type of offenders in the community and the work completed as a result. [...] Choice of the final model was based on the risk ratio, the overall accuracy of the model (percent true positives and true negatives), the succinctness of the model, and in the reduction of false negatives, which meant a reduction in the Type I error (misclassification of a high-risk offender as a lower risk offender). [...] In order to adjust the level of statistical significance used to assess the identification of branches on the tree and counteract the problem of multiple comparisons, a Bonferroni adjustment was applied to the p-values. [...] The rate of reoffending for the group meeting all the risk criteria was 24% resulting in a risk ratio of 4. That is, offenders in the high-risk group were 4 times more likely to reoffend than the average reoffending rate for non-Indigenous men offenders.
Pages
34
Published in
Ottawa, ON, CA