Improving Outcomes and Safely Reducing Revocations from Parole Supervision in Colorado
The majority of the data included in this section were pulled from cohorts of individuals who terminated parole supervision between 2010 and 2019. Given the significant impact of the COVID-19 pandemic on the justice system throughout the country, CJI supplemented these trends with aggregate data from 2020 and 2021 provided by DAP to assess the impact the pandemic had on revocations. Revocation rates have declined consistently over the past decade with the most substantial decline occurring amidst the COVID pandemic CJI examined overall parole outcomes, looking at successful and unsuccessful terminations, and found that revocation rates have been decreasing over the past decade. This decrease speaks to state and agency leadership’s efforts to strengthen community supervision practices and reduce revocations. In 2015, Senate Bill (SB) 124 was enacted, which required parole officers to use swift, certain, and proportional sanctions; authorized the use of short jail terms of no more than five days as an intermediate sanction for individuals on parole who have a pattern of technical violations; redefined drug use on parole as a technical violation rather than a new crime; and repealed the requirement of mandatory arrest if an individual refuses to submit to a urinalysis test. Figure 1 below shows that revocations from parole in Colorado peaked in 2014 (comprising 54 percent of all parole terminations, both successful and unsuccessful) and have been steadily decreasing since, representing 44 percent of all parole terminations in 2019.
Responses to Behavior
To inform its findings related to responses to behavior, CJI examined current policies and practices, analyzed data provided by CDOC on the use of their current graduated sanctions tool and available incentives, and conducted a case file review to understand to what extent sanctions aligned with policy and research. CJI examined administrative and statutory responses to violations, factors that influence a parole officer’s decision-making regarding violation responses, the officer’s level of autonomy, and the use of incentives across the state. Research shows that to effectively change behavior, responses to violations should be proportional to the violation committed, delivered objectively, and focused on the behavior instead of the person. Similarly, incentives should be delivered impartially, focused on the behavior, and used to reinforce continued positive behavior. Incentives should be used four times more often than sanctions to effectively change behavior.
CJI’s review of the CVDMP indicates that many sanction options available to staff are not being used, indicating a disconnect between what officers report and what the data reveal. The CVDMP provides many potential low, medium, and high-level sanctions and interventions for parole officers to use in response to violations. However, the majority of responses available to staff are rarely, if ever, used. An analysis of sanctions used in response to violations from the file review sample indicated that the majority of violation behaviors receive only one or two possible sanctions from each possible category. From all possible low-level sanctions, verbal reprimands and withholding earned time were each used to respond to about 40 percent of violations; the next most frequent low-level sanction, increased therapy, was only used 9 percent of the time. From all possible medium-level sanctions, withholding earned time was used to respond to a third of violations; the next most frequent medium-level sanction, increasing or extending the level of supervision, was used 16 percent of the time. From all possible high-level sanctions, Sure and Swift, a short-term jail sanction, was used to respond to 60 percent of violations; the next most frequent high-level sanction, summons to the Parole Board, was used 17 percent of the time. Table 3 outlines how frequently each CVDMP response was used, by sanction level, in response to violation behavior by individuals in the 2019 file review sample.