Year of Award

2012

Document Type

Thesis

Degree Type

Master of Arts (MA)

Degree Name

Sociology

Department or School/College

Department of Sociology

Committee Chair

Dusten Hollist

Commitee Members

Doug Dalenberg, Jame Burfeind

Keywords

RAI, Risk Assessment Instrument

Abstract

The purpose of this research is to examine alternate ways to add meaningful weights to the risk factors on the Montana Risk Assessment Instrument (RAI). An evaluation is made which compares the predictive accuracy of a revised scoring system compared to the one that is currently in use. The data for this analysis is taken from 299 Montana juveniles who were administered the RAI after an offense, between January 01, 2009 to December 31, 2010. The results are based on a Burgess model, a linear probability model, and a logistic regression model. The findings suggest that all three models increased the predictive accuracy of the RAI. The Burgess model and the logistic model showed the greatest improvement. When considering both predictive accuracy and practical usability, the Burgess model for rescoring the RAI was found to be the best approach. The small sample size was a limitation in this research which may have affected the statistical significance of the risk factors found on the RAI when using linear probability and logistic regression. Inconsistencies found between counties when collecting data was another limitation in this research. Finally, the inability to find a continuous outcome variable forced this research to use a linear probability model instead of a linear regression model. Future research to increase the predictive accuracy of the RAI must concentrate on three major topics. First, it must be a priority to find appropriate risk factors for the RAI. Second, continue research that will determine the best approach to add meaningful weight to risk factors. Finally, examine the cut point on the RAI to eliminate the most false positive and false negative predictions.

Download

Share

COinS
 

© Copyright 2012 Patrick David McKay