Automating Recidivism Risk Assessment

• Author: Kenneth C. Land
• DOI: http://doi.org/10.1111/1745-9133.12271
• Date: 01 February 2017
• Published date: 01 February 2017

EDITORIAL INTRODUCTION

RECIDIVISM RISK ASSESSMENT

Automating Recidivism Risk Assessment

Should We Stay or Should We Go?

Kenneth C. Land

Duke University

An increasingly salient task for practitioners in corrections is recidivism risk assess-

ment of prisoners. Recent decades have experienced a move from more qualitative,

clinical judgment to standardized risk assessment tools of the probability that re-

leased offenders commit new crimes. This trend has been fostered by advances in measure-

ment methodologies and associated scales, as well as by the increasing sizes of correctional

populations. Limited criminal justice system capacities also mean that agencies increasingly

use risk assessment instruments to allocate resources to offenders who are assessed as having

the highest risk to society.

Because the methods of scoring and the instruments themselves vary widely, research is

needed on the approaches that lead to the most reliable and valid outcomes. As a response

to this need, Grant Duwe and Michael Rocque (2017, this issue) study the relationship

between reliability and validity with data from the Minnesota Screening Tool Assessing

Recidivism Risk (MnSTARR), a risk assessment instrument the Minnesota Department

of Corrections (MnDOC) developed and began using in 2013. By using follow-up data

on offenders released in 2014 and manual MnSTARR assessments scored by MnDOC

staff, Duwe and Rocque assess the impact of inter-rater reliability (IRR) on predictive

performance (validity). They also compare the reliability of a manual scoring process with

an automated one.

Duwe and Rocque (2017) find the MnSTARR was scored with a high degree of

consistency by MnDOC staff and that intraclass correlation (ICC) values were at high

levels, which they attribute to the instrument comprising mostly objective rather than

subjective risk factors. Even with high IRR values on the manually scored instruments, they

also report that (a) the automated assessments significantly outperformed those that had

been scored manually and, as might be expected, (b) that the more inter-rater disagreement

Direct correspondence to Kenneth C. Land, Department of Sociology, Duke University, Box 90088, Durham,

NC 27708-0088 (e-mail: kland@soc.duke.edu).

DOI:10.1111/1745-9133.12271 C2017 American Society of Criminology 231

Criminology & Public Policy rVolume 16 rIssue 1