Report: Examining Parole Revocation Patterns 2018 | Department of Criminal Justice

Summary:

In the state of Utah, a revocation to prison from parole can occur in two instances: (1) if the offender failed to comply with their supervision conditions, and (2) if the offender was convicted of a new crime while on parole. It should be emphasized that changes in revocation rates over time may reflect changes in offender behavior rather than the nature and enforcement of criminal justice policies and practices. Additionally, it may not reflect general health trends or other dynamic social phenomena. With this in mind, this study examines revocation patterns through the lens of Utah’s current criminal justice policies, using a 1-year follow-up period.1 Its main findings may be summarized as follows:

While ceilings on parole revocations are reducing parole violators’ prison length of stay, previous parole violators are continuing to cycle through the system. Such findings hint at an increased need of effective interventions for this on average, high risk population.
Increases in new convictions while on parole are being driven by non-violent offenses. Such non-violent offenses are primarily comprised of drug possession only crimes.

The above findings warrant a closer look at the causes of revocations and their interplay with revocation length of stay, implementation fidelity, organizational capacity, and current enforcement practices. Overall, the nuances of these findings highlight the ongoing need to collect and monitor data, as well as the importance of conducting further analyses to elucidate current criminal justice trends.

Data & Method

A total of 4,541 unique parole starts were obtained from the Utah Department of Corrections’ (UDOC) database O-Track.5 In order to delineate the pre- and post-reform time frames, these data were divided into four distinct cohorts. The first cohort had a parole start date between October 1, 2013, and March 31, 2014; the second cohort started parole between October 1, 2014, and March 31, 2015; the third cohort had a parole start date between October 1, 2015, and March 31, 2016; the fourth cohort started parole between October 1, 2016, and March 31, 2017. Each cohort was followed for one year from their respective parole start date. The outcome variables of interest were dichotomized into two variables: the first denotes whether the parole was returned on a technical violation only, and the second denotes a revocation based on a new conviction.

To account for changes in the parole population and as informed by prior literature, demographics in the form of age and gender serve as important covariates. Additionally, criminal justice variables comprising of offense type, offense severity, and the parolee’s risk to re-offend (available through the Level of Service: Risk-Need-Responsivity [LS:RNR] instrument) are included.

Image of bar graphs displaying Figure 1 percent expected to be revoked on a technical violation

Results

This section explores prison revocations due to one or more technical violations. Figure 1 illustrates the expected revocation rate by cohort. As shown in the figure, the two pre-reform cohorts have a similar expected return rate of 36 percent, while cohort three shows a higher expected return rate of 45 percent. The variable denoting whether it was the offender’s first parole start was among the most important variables, as determined by the BMA process, and showed a relatively large negative average coefficient (see Appendix A). Because this part of the parole population is decreasing over time, this finding warrants a closer look at this subset.

Figure 2 illustrates the expected revocation rates by cohort, restricting the sample to those on their first parole start. As seen in the figure, all three cohorts have a similar expected rate of just below 30 percent. Hence, the BMA process does not highlight a difference in the likelihood of revocation for a technical violation across groups when restricting the sample to this subset of the population.

The limitations of comparing recidivism rates through time are emphasized in this section. An important nuance of the current study is its inability to distinguish changes in enforcement policies and practices from changes in offender behavior. Because a revocation to prison from parole may be a direct result of enhanced or reduced enforcement efforts, both in relation to violations and new convictions, study findings may not reflect actual changes in behavior among the parole population.

The importance of implementation fidelity should also be emphasized. Per implementation science standards, full implementation, particularly with fidelity, takes time and requires adequate resources. Hence, the impact of any policy change may not come to full fruition until such resources are available and the subsequent requirements of each policy recommendation are met. As an example, specific recommendations in Utah’s broad reform package included improving and expanding reentry and treatment services. It further emphasized the need to ensure quality of treatment by establishing statewide standards and certification processes. Regarding the design and tailoring of community treatment services to meet the needs of those involved in the criminal justice system and to ensure such services adhere to evidence-based practices, these processes may take longer than two years post-reform to actualize (Morris, Woodling & Grant, 2011).

Indeed, the complexity of examining the effectiveness of criminal justice policies is highlighted by the underlying assumption of fidelity of such policies, which takes time to realize. Future research may explore processes related to fidelity through the lens of implementation science. Furthermore, accounting for changes in broader social and public health trends, including the current opioid epidemic, may assist forthcoming research in contextualizing and identifying pertinent factors for inclusion in future analyses.

Read the full report:

Examining Parole Revocation Patterns