Are “Parole Failures” Parolees When They Fail? Incorporating Supervision Terms and the Timing of Arrest Events into Parolee Recidivism Rates

• DOI: 10.3818/JRP.15.2.2013.1
• Published date: 01 December 2013
• Date: 01 December 2013
• Subject Matter: Article



*



Are “Parole Failures” Parolees When They Fail?

Incorporating Supervision Terms and the Timing

of Arrest Events into Parolee Recidivism Rates

Michael Ostermann

Joel Miller

Rutgers University

Jason Matejkowski

University of Kansas

JUSTICE RESEARCH AND POLICY, Vol. 15, No. 2, 2013

© 2013 Justice Research and Statistics Association

DOI: 10.3818/JRP.15.2.2013. 1

* Abstract

Many studies have found that parolees experience high rates of recidivism after they

are released to their communities. However, it is not clear whether these failures occur

during the course of supervision or after the completion of the parole term. This study

uses data from prison releases in a large Northeastern state from 2005 to 2007 (n =

28,869) and a three-year follow-up period to describe the timing of rearrest events as

they relate to the term of active parole supervision. The f‌indings indicate that about

63% of unconditional releases and 54% of parolees are rearrested within three years

of their release. But approximately 55% of arrested parolees were rearrested after

the completion of their supervision term. Parolees who were rearrested after parole

completion differed signif‌icantly from those who were rearrested during the course of

their supervision on several pertinent predictors of recidivism, including age, gender,

ethnicity, instant offense type, and criminal history. Cox regression models indicate

that when only arrests for parolees during the course of their supervision are consid-

ered, the impact of parole is associated with an approximate 22 percentage point de -

crease in the hazard rate when compared to a method that includes arrests that occur

after the culmination of the supervision term. Our discussion presents the implications

of different def‌initions of parolee failure for measuring recidivism rates as well as for

criminal justice policy.



Several studies have explored the impact of parole supervision by analyzing the re-

offense patterns of individuals released to either parole or no supervision. Attempts

to isolate the impact of release status on recidivism typically employ some form of

multivariate modeling to minimize selection biases inherent in discretionary parole

release mechanisms (Gottfredson, Mitchell-Herzfeld, & Flanagan, 1982; Flanagan,

1982; Spivak & Damphousse, 2006; Spivak & Sharp, 2008; Ostermann, 2011a;

Solomon, Kachnowski, & Bhati, 2005). Discretionary parole allows releasing au-

thorities to select inmates who they think are likely to succeed if they were to be

released early and supervised in the community. Studies have generally indicated

that after controlling for differences between supervised and unsupervised groups,

parole’s impact upon recidivism is small and often nonsignif‌icant (Ostermann,

2011a; Petersilia & Turner, 1993; Solomon et al., 2005).

Previous research has typically not incorporated the time of active parole su-

pervision into analytic strategies, however. Within these prior studies, former in-

mates who are released to parole supervision are identif‌ied as parolees when they

have their rap sheets queried for reoffenses after the expiration of the study’s fol-

low-up period. New offenses that are indicated on rap sheets within the follow-up

period are considered parole failures. While this method of identifying recidivists is

convenient for researchers, it does not ref‌lect the way that parole works in practice.

Inmates who are released to parole are supervised until they either fail or meet the

date of their sentence expiration. Without knowing the sentence expiration date,

we cannot know whether those who were released to parole failed while they were

under parole’s watch or failed after they completed their parole term. Considering

a more nuanced approach of measuring recidivism for parolees that incorporates

the time of active supervision may have potentially signif‌icant impacts upon the

measured rates of failure for supervised groups.

This study empirically explores the variability in rates of failure according to

different measurement approaches. To illustrate how conceptualizations of parolee

failure can differ, Figure 1 shows methods for measuring failure across three hy-

pothetical parolees who experience recidivism (or potential recidivism) in different

ways. All three cases are released to parole supervision, have their supervision

terms expire on the same day, and have their rap sheets queried for reoffense in-

formation three years after their release from prison. Case 1 is rearrested after the

parole date and prior to the parole expiration date. This case is considered a fail-

ure. Case 2 is not rearrested, but the terms of parole are revoked during the course

of the supervision term and the individual is sent back to prison. This case can be

considered a failure if parole revocations are considered recidivism events. Case

3 successfully completes the terms and conditions of parole supervision without

incident and is rearrested after the parole term expiration. This case is considered

a failure within the extant literature, and in studies that compare reoffense pat-

terns between parolees and those who are unconditionally released from prison,

this failure is attributed to the parole group (e.g., Flanagan, 1982; Jackson, 1983;

Ostermann, 2011b). This attribution, however, does not coincide with how parole