Abstract Details
Activity Number:
|
612
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Risk Analysis
|
Abstract #311398
|
View Presentation
|
Title:
|
Construction and External Validation of Risk Models for Violent Offending: Statistical Methodology and Empirical Findings Using Paper Instrument for Violence (PIV)
|
Author(s):
|
Constantinos Kallis*+ and Laura J.W. Bui and Jeremy Weir Coid
|
Companies:
|
Queen Mary University of London and Queen Mary University of London and Queen Mary University of London
|
Keywords:
|
Risk models ;
Violence ;
Predictive accuracy ;
Area under ROC Curve (AUC) ;
Positive and Negative Predictive Values ;
External validation
|
Abstract:
|
Violence is an issue of increasing concern to public health. As a result, a number of prognostic methods are currently implemented in prisons to predict violent offending. We identified methodological problems related to the construction of existing prognostic models for violent offending. In this paper, we introduce the Paper Instrument for Violence (PIV), a novel prognostic model constructed using a new statistical approach that significantly improves predictive accuracy. To create PIV, we used data from the Prisoner Cohort Study (PCS), a sample of serious violent and sexual offenders.
We explore aspects of external validation such as the selection of appropriate measures of predictive accuracy and predictive strength. These measures include Area Under the ROC Curve (AUC), Positive and Negative Predictive Values (PPV and NPV) and z-scores. Our empirical findings on external validation are from a large sample of general offenders.
We conclude that improvement in predictive accuracy can be achieved by selecting a small number of highly predictive variables. Using this approach, shrinkage of predictive accuracy in an external validation sample is at an acceptable low level.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.