Validation and examination of the Ohio Youth Assessment System with juvenile sex offenders

• Author: William T. Miller,Christina A. Campbell,Jordan Papp
• Published date: 01 May 2020
• Date: 01 May 2020
• DOI: http://doi.org/10.1111/1745-9133.12464

DOI: 10.1111/1745-9133.12464

ORIGINAL ARTICLE

RISK ASSESSMENT

Validation and examination of the Ohio Youth

Assessment System with juvenile sex offenders

Jordan Papp1Christina A. Campbell2William T. Miller2

1University of Michigan

2University of Cincinnati

Correspondence

JordanPapp, Institute for Social Research,

Universityof Michigan, 426 Thompson

Avenue,Ann Arbor, MI 48104.

Email:pappjr@mail.uc.edu

Research Summary: In this study, we examined the use

of an actuarial risk assessment tool—the Ohio Youth

Assessment System-Disposition Tool (OYAS-Disposition

Tool)—with juvenile sex offenders. Specifically, the main

goals of the study were to (a) examine the predictivevalidity

of the tool with sex offenders and (b) explore the nature of

the use of professional discretion used to override the tool.

The sample consisted of 3,235 youth from a large juvenile

county court in the Midwest. The results indicated that the

OYAS-Disposition Tool was a significantly better option

for predicting general recidivism for sex offenders than it

was for non–sex offenders. The tool was also an effective

method for predicting sexual recidivism. Most importantly,

however, the use of professional overrides significantly

reduced the ability of researchers to apply the tool to pre-

dict new court petitions and adjudications to nonsignificant

levels. Finally, several justifications were commonly used

for overrides: treatment needs, offense seriousness, and

use of an alternative sex-offender–specific assessment.

Policy Implications: The findings in this study highlight

several important policy implications that would improve

the assessment process for juvenile sex offenders. First,

agencies using specialized risk assessments designed for

sex offenders may consider applying a general risk assess-

ment tool to identify a broader set of criminogenic needs

Criminology & Public Policy. 2020;19:433–450. wileyonlinelibrary.com/journal/capp © 2019 American Society of Criminology 433

434 PAPP ET AL.

and to predict risk of general recidivism. Second, there is a

need to evaluate policies and practices that allow fort he use

of professional discretion with sex offendersgiven that they

reduce the predictive validity of the risk tool evaluated.

The high rate of overrides for juvenile sex offenders and

justification for their use go against best practices in correc-

tions. For instance, overrides were often justified based on

offense seriousness; however, focusing on evidence-based

criminogenic risk factors provides the best accuracy in

predicting future offending. In this study, we call into

question court policies that allow for overrides based on

crime type or based on a practitioner’s professional judg-

ment concerning a juvenile’s level of service needs. Last,

agencies should consider validation research within their

agency before full adoption of a general risk assessment

tool to quell concerns about the use and accuracy of a tool

for special populations like juvenile sex offenders.

KEYWORDS

overrides, professional discretion, risk assessment, sex offenders

The successful supervision and treatment of juvenile sex offenders depends on the ability to predict

with accuracy the risk that they pose for reoffending (Worling & Langstrom, 2008) as it is estimated

that approximately 20% of sex offenses are committed by persons younger than 18 years of age in any

given year (Barbaree & Marshall, 2008). Significant improvements in risk prediction have been made

in the past several decades. In fact, actuarial risk assessment has been demonstrated to be one of the

best methods of assessing criminogenic risk (Desmarais & Singh, 2013). Debate continues, however,

about the most effective wayto design actuar ial tools used byresearchers to predict recidivism among

juvenile sex offenders. Moreover, the amount of research pertaining to juvenile sex offenders is sparse

relative to the amount pertaining to their adult counterparts.

The findings from the research available have shown that juvenile sex offenders possess divergent

criminogenic risk factors that are much less common among nonsex offenders (Prentky & Righthand,

2003; Worling & Langstrom, 2008). For example, deviant sexual interests are a risk factor among

juvenile sex offenders, which is not the case fornonsex offenders (Worling & Curwen, 2000). For this

reason, some scholars have questioned the effectiveness of risk assessments that were not designed

with sex offenders in mind (Tully, Chou, & Browne, 2013). Contrary to this argument, some research

findings also indicate that risk assessment tools designed to aid in predicting recidivism for general

types of offending can aid in predicting recidivism for sex offenders as well (Olver, Stockdale, &

Wormith, 2009; Ragusa-Salerno, Ostermann, & Thomas, 2013). These inconsistencies demonstrate

the importance of continued research into the effectiveness of generally designed risk assessment tools

with sex offenders (Baldwin, 2014).