Predicting Sex Offender Recidivism: Using the Federal Post Conviction Risk Assessment Instrument to Assess the Likelihood of Recidivism Among Federal Sex Offenders

• Published date: 01 September 2018
• Date: 01 September 2018
• Author: Thomas H. Cohen
• DOI: http://doi.org/10.1111/jels.12184

Journal of Empirical Legal Studies

Volume 15, Issue 3, 456–481, September 2018

Predicting Sex Offender Recidivism: Using

the Federal Post Conviction Risk

Assessment Instrument to Assess the

Likelihood of Recidivism Among Federal

Sex Offenders

Thomas H. Cohen*

Sex offenses are among the crimes that provoke serious public concern. The federal

response to the problem of sex offending has resulted in an exponential increase in the

number of sex offenderson federal post-conviction supervision; however, relativelyfew stud-

ies have explored whether and how well the actuarial risk instrument currently used by fed-

eral probation officers—the federal Post Conviction Risk Assessment instrument (PCRA)—

accurately predicts reoffending behavior among the federal sex offender population. This

study uses a uniquedataset of approximately 5,300 convicted sex offenderson federal super-

vision to investigate the PCRA’s capacity to effectively predict subsequent recidivism activity

for convicted federalsex offenders. Results show that the PCRA accuratelypredicts recidivis-

tic behavior involving the commission of any felony or misdemeanor offenses, violent

offenses, and probation revocations for this population. However, the PCRA’s predictive

capacities deteriorate when the instrument is used to assess the likelihood of sexual recidi-

vism. In addition, this study showed that offenders convicted of online child pornography

offenses presented some challenges in terms of predicting their reoffending behavior as

they manifested lower PCRA risk scores and recidivism rates than offenders convicted of

other major federalsexual offenses that involve hands-on behavior.

I. Introduction

Sex offenses are commonly considered one of the most serious and dangerous crimes by

the general public (Hanson & Morton-Bourgon 2005, 2009). The growing exploitation of

children by online sex offenders who use the Internet and related digital technologies to

possess, receive, distribute, or produce online child pornography (hereafter online child

*Administrative Office of the U.S. Courts, One Columbus Circle NE, Washington, DC 20544; email:

thomas_cohen@ao.uscourts.gov. The author thanks Christopher T. Lowenkamp, Michelle Spidell, and the JELS

reviewers for their helpful suggestions and comments. This article benefited from the careful editing of Ellen Fiel-

ding. The views and opinions expressed in this article are solely the author’s and do not represent the views of the

federal probation and pretrial services system or the Administrative Office of the U.S. Courts.

456

pornography offenders) has become a particularly serious concern (Seto et al. 2011). Soci-

etal alarm in the United States over the online sexual abuse of children, combined with

evidence showing that many online sex offenders have self-report histories of contact sex-

ual behavior (Lam et al. 2010; Seto et al. 2011), has resulted in aggressive responses aimed

at targeting sex offenders at the legislative and executive levels (Wolak et al. 2009).

The focus on sex offenders has resulted in substantial increases in the number of

these offenders prosecuted, incarcerated, and placed under post-conviction supervision in

the federal criminal justice system (Faust & Motivans 2015; Motivans & Kyckelhan 2007).

In an examination of key correctional trends involving federal sex offenders, Faust and

Motivans (2015) found that the number of sex offenders placed under federal post-

conviction supervision increased by nearly 1,400 percent—from 321 offenders to 4,714

offenders—between 1994 and 2013, and much of this increase could be attributed to the

growing numbers of online child pornography offenders incarcerated within the federal

Bureau of Prisons (BOP) (USSC 2012). Offenders convicted of federal sex offenses have

also been subjected to increasingly lengthy terms of post-conviction supervision. The

U.S. Sentencing Commission’s (USSC) report to the U.S. Congress on federal child por-

nography offenders stated that the average terms of supervised release sentences imposed

on these offenders during fiscal year 2010 ranged from 220 months for offenders con-

victed of possessing online child pornography to 323 months for offenders convicted of

producing online child pornography (USSC 2012). In comparison, the general federal

offender population received post-conviction supervision terms averaging about43 months

in fiscal year 2010 (USSC 2012).

With the exponential increase in the number of sex offenders (especially those

convicted of online child pornography offenses) under federal post-conviction supervi-

sion, the need to assess and predict their likelihood of recidivism has become ever more

paramount. Although there have been growing efforts to examine the risk characteristics

and recidivism rates of federal sex offenders (see Bourke & Hernandez 2009; DeLisi et al.

2016; Faust et al. 2014; Faust & Motivans 2015; USSC 2012), few empirical studies have

focused on determining whether and how well the actuarial risk instrument currently

used by federal probation officers—the federal Post Conviction Risk Assessment instru-

ment (PCRA)—accurately predicts reoffending behavior among the federal sex offender

population (Cohen & Spidell 2016).

The PCRA is a fourth-generation dynamic risk assessment tool developed to assess

the risk of recidivism among offenders placed on supervised release in the U.S. federal

system (AOUSC 2011; Johnson et al. 2011; Lowenkamp et al. 2013). Although numerous

studies have shown that the PCRA has excellent validity in predicting general and violent

recidivism (Lowenkamp et al., 2013, 2015; Luallen et al. 2016), few efforts have

attempted to assess the PCRA’s capacity to predict sexual recidivism or gauged this instru-

ment’s predictive accuracy for the federal sex offender subpopulation (Cohen & Spidell

2016; Luallen et al. 2016).

The lack of studies examining the PCRA’s capacity to accurately predict recidivistic

outcomes for federal sex offenders represents a striking lacuna that the current study will

attempt to address. It is particularly important to explore the PCRA’s predictive capacities

for this population because, unlike risk instruments designed to predict sexual recidivism

Predicting Sex Offender Recidivism 457