Attachment C_Edit Document.docx

Juvenile Recidivism in Washington State:

A 2013 court cohort and 2014 JRA release cohort

Introduction In 2016 the National Coalition for Juvenile Justice (NCJJ) in conjunction with the Pew Charitable Trusts (Pew) published a report on juvenile recidivism in Washington State. They used data prepared by the Washington State Center for Court Research (WSCCR) and the Juvenile Justice Rehabilitation Administration (JRA) to examine a cohort of youth disposed and another cohort of youth admitted to a JRA facility during the calendar year 2012. This represented the first, large-scale juvenile recidivism study for Washington State to include court and JRA records. In an effort to continue and improve upon this important first step, we have completed a new juvenile recidivism study using a cohort of juveniles disposed in 2013 and a cohort of juveniles released from JRA in 2014.

The initial NCJJ and Pew study was a significant step towards improving juvenile recidivism reporting in Washington State. However, we took some additional steps to improve the quality of the information reported by accounting for time incarcerated during the follow-up period and using a JRA release cohort rather than an admission cohort. In addition to improving upon the definitional and methodological approaches, we also added new variables related to their criminal history and previous incarcerations to help explore potential associations between the offenders and recidivism.

Data & Methods The qualifying event for the study was the first criminal justice cycle[footnoteRef:1] for which an individual received a disposition during 2013 for the court cohort and the individual’s first release from JRA during 2014 for the JRA cohort.[footnoteRef:2] Only the most serious disposed charge in that criminal justice cycle was counted. For those individuals with more than one court disposition during 2013 or more than one JRA release during 2014, we counted the first disposition or release as the qualifying event for inclusion in the study and all follow-up periods are based upon that date. [1: A criminal justice cycle attempts to define the period around an offense with a disposition. As such, we filter out multiple charges with the same case number, offense date, and adjudication date, to not overcount the number of offenses committed by an individual.

] [2: The first disposition of the calendar was taken for each disposition cohort. There were a number of individuals who had both an adjudication and diversion in the same year, so those categories are not exclusive.

We used WSCCR court records to identify qualifying dispositions, as well as track criminal history and recidivism events. JRA provided admission and release records to identify qualifying events for the JRA release cohort, as well as track previous JRA admissions for study subjects. We also incorporated WSCCR juvenile detention facility records to allow us to track whether individuals met the minimum amount of follow up period.

With access to the offender’s complete criminal history, we attempted to categorize criminal history in a number of different ways to examine the relationship between past and future offending, as well as analysis of the relationship between the age of first conviction and recidivism. While WSCCR detention records are not complete[footnoteRef:3], they are sufficient records to establish the detention history for the majority of juveniles detained in Washington State. We included measures of detention prior to the qualifying offense, pretrial detention related to the qualifying offense, and prior incarcerations at a JRA facility. We recognize that many of these measures are proxies for the seriousness of past and current offending, but there are a number of instances where the relationship may not be so direct. [3: Currently, WSCCR receives records from all juvenile detention facilities across Washington State, except for Martin Hall. In addition, our King County detention records incorporated in this study are limited to 2013-2016. Mason County detention data may be incomplete prior to 2015, and Pacific/Wahkiakum detention data may be incomplete prior to 2014.

In an attempt to improve upon the previous work from NCJJ and Pew, we attempted to account for time incarcerated during the recidivism follow-up period. We included records of time spent in JRA and detention and discounted that time from the period between the juvenile’s adjudication date and most recent WSCCR data available. Those individuals without the minimum amount of follow-up period (18 months)[footnoteRef:4] as “street time”[footnoteRef:5], were excluded from the study. The second large methodological change from the NCJJ and Pew study was use of a JRA release cohort rather than an admission cohort. Using a release cohort rather eliminates the need to account for the JRA sentence as part of the follow-up period allowing a more contemporary cohort to be analyzed and fewer subjects to excluded because they did not have enough street time. [4: For the court cohorts, the recidivism clock began on their disposition date and for the JRA release cohort it began on the date of their release. Recidivism was defined as an offense that occurred within eighteen months after their program eligibility date or program start date, and that resulted in a conviction within twelve months of the offense date.

] [5: “Street time” refers to time not spent in incarceration. Loftstrom, M., Raphael, S., & Grattet, R. (2014). Is Public Safety Realignment Reducing Recidivism in California? Public Policy Institute of California. ]

We divided our analyses into three different categories, dispositions, adjudications, and diversions. These groups are distinguished by the qualifying case outcome. Dispositions include convictions, deferrals, and diversions. Adjudications include convictions and deferrals, but do not include diversions. Diversions only include cases where the outcome was a diversion. In addition, there were some analyses that included the JRA release cohort. Due to multiple factors, including a lack of a consistent case-related identifier between JRA and Administrative Office of the Courts (AOC) data, we did not connect the JRA release cohort to their original disposition.[footnoteRef:6] [6: From past attempts at analysis, we know that the majority come from convictions for felony charges, along with revocations for disposition alternatives or juveniles with extensive criminal histories.]

Study Population In the tables below, differences in gender, race, age, and qualifying offense are evident between the disposition, adjudication, diversion, and JRA release cohorts.

The gender gap varies greatly across the various cohorts from 66.3 percent male among dispositions, 74.7 percent among adjudications, 60.5 percent among diversions, and 89.8 percent among the JRA release cohort. The mean age only moves 0.2 years among the court cohorts, it jumps almost a full year with the JRA release cohort. This is not suprising as the JRA release cohort has served a sentence since their court adjudication. There are large changes among racial/ethnic groups across the cohorts, as the highest percentage of Whites are found in the diversion cohort (61.0%) and the lowest percentage is found in the JRA release cohort (51.8%). Blacks also saw a large change from the diversion cohort (9.4%) to the JRA release cohort (18.7%). Other racial/ethnic groups saw much smaller variations across the different cohorts.

Qualifying offense also varied across the cohorts. Adjudications consisted of 66.7 percent misdemeanors and 33.3 percent felonies, while diversions consisted of 98.3 percent misdemeanors and 1.7 percent felonies.

Finally, age at first conviction showed variation across the various cohorts, as well. The most frequent age of first conviction for the JRA release cohort was 13 years old, while the most frequent age of first conviction was 14 years old for adjudications and all dispositions, and most frequent age of conviction for the diversion cohort was 16 years old.

Results The analyses include a variety of demographic, offense, and court factors related to the study population and the outcomes analysis.[footnoteRef:7] There are several, easily identifiable trends that can be picked out from the charts and tables in the subsequent subsections.[footnoteRef:8] Consistently, the highest rates of recidivism are found with the JRA release cohort. The adjudicated cohort has the highest recidivism rates among the court cohort, followed by all dispositions and then the diversion cohort. [7: Recidivism analyses with fewer than 50 subjects were not reported to avoid misleading statistics.

] [8: Results with less than 50 subjects were removed from all tables to limit inferences from small samples. ]

Due to difficulties linking JRA admissions to qualifying offenses, those analyses based upon a qualifying disposition, criminal history, or previous incarcerations do not include the JRA release cohort.

Recidivism The most consistent trend was that recidivism rates increased along with the severity of the disposition - diversions to adjudications to all dispositions to the JRA release cohort. This was consistent regardless of the demographic, criminal history, or qualifying offense variable used to filter the results or the type of recidivism measured (overall, misdemeanor, felony, or violent felony).

The most inclusive category, all dispositions, had an overall recidivism rate of 28.1% with a felony recidivism rate of 9.7%. The adjudication cohort saw increases in both of those rates (43.5% and 18.3%), while the diversion cohort saw decreases in recidivism (20.0% and 4.6%, respectively).

Gender Gender is commonly associated with recidivism as recidivism studies consistently find that males recidivate more often than do females. In this case, males recidivated more than females for all disposition types and all types of recidivism.

For felony recidivism, the rates among males were almost two to three times higher than those of females (12.2% to 4.9%, 20.5% to 11.9%, and 6.2% to 2.3%) across the different disposition cohorts. The exception was the JRA release cohort, where females had higher rates of misdemeanor and overall recidivism rates (39.9% to 33.3% and 59.0% to 53.8%, respectively).

Race The relationship between race and recidivism rates is fairly consistent across different types of recidivism and different dispositions.

Black juveniles had the highest felony and violent felony recidivism rates for all cohorts and the highest overall recidivism rates for adjudications and the JRA release cohort. Asians and Pacific Islanders had the lowest all and misdemeanor recidivism rates across all dispositions and the JRA release cohort, while Whites had the lowest felony and violent recidivism rates across all dispositions.

Age at Qualifying Offense All court dispositions cohorts show a consistent trend across overall recidivism and misdemeanor recidivism.

For adjudications and all dispositions, the recidivism rates increase until age 14 and then decrease to their lowest point at age 17. These trends change in the felony and violent felony recidivism categories. For felony recidivism, the recidivism rates for the adjudication and all disposition cohorts increase steadily until age 16 (10.7%, 19.8%, and 4.5%) and then decrease (8.8%, 17.2%, and 2.6%). The diversion cohort’s recidivism rates, across each recidivism categories, appear to peak around age 12 (25.9%, 22.5%, 7.3%, and 2.6%) and then fall consistently until their lowest points at age 17 (11.4%, 9.7%, 2.6%, and 0.7%).

Age at First Disposition All court disposition cohorts show either a consistent decrease in recidivism as the age at first conviction increases, or show an increase in recidivism from the first recorded age group to the second group and then a consistent decrease, across all types of recidivism.

The JRA release cohort showed less consistent recidivism results and their patterns, across different types of recidivism. The only consistent among the JRA release cohort is higher recidivism rates than any other cohort and that exists across all intersections of court cohorts and recidivism types.

Criminal History There appears to be a consistent relationship between the subject’s criminal history and their likelihood of recidivism and the likelihood of recidivism increases with the quantity and severity of the criminal history. This was the only demographic were all disposition cohorts demonstrated the same pattern across all recidivism groups.

Criminal history is a consistent and prominent factor in risk assessment instruments designed to predict recidivism (ASRA, Arnold Foundation, COMPAS, LSI-R) and the reason for that is demonstrated in the above table.

Previous Incarcerations There appears to be a clear relationship between the type of previous incarceration and the recidivism rate.

Those with any kind of prior incarceration (qualifying offense pretrial detention, previous offense detentions, or previous JRA admissions) had higher recidivism rates than those without any incarcerations and those with JRA admissions had higher recidivism rates than those that only had been in detention across all types of recidivism.

County Level County level analysis of recidivism is difficult, because several counties have a low number of juveniles represented in the study. As a result, it is often difficult to draw county level conclusions. For example, 13 of the 35 juvenile jurisdictions in Washington State had fewer than 100 juveniles with dispositions in 2013. The number of counties with more than 100 subjects decreases further, as the dispositions are disaggregated. There were fewer than 100 subjects in 18 jurisdictions for those with diversions, 19 jurisdictions for those with adjudications, and 34 of the jurisdictions for the JRA release cohort.

One of the most noticeable results regarding the counties comes from the population demographics. There were four counties that had much higher rates of diversions to adjudications (Benton/Franklin, Pierce, Snohomish, and Spokane) and four counties with much higher rates of adjudications to diversions (Clark, Cowlitz, Kitsap, and Thurston). Despite the discrepancies between the types of qualifying disposition, counties with a larger relative share of adjudications did not demonstrate higher recidivism than the counties with a larger relative share of dispositions. In fact, recidivism rates of counties in each group were relatively similar to the state average.

Conclusion As discussed, the highest rates of recidivism are found with the JRA release cohort, followed by the adjudicated cohort, all dispositions and then, the diversion cohort. These results were anticipated, because dispositions do not occur in a vacuum, nor are they meted out randomly. From this study, the more severe dispositions are associated with longer and more severe criminal histories and more severe qualifying offenses. So, those who have committed crimes are more likely to receive the more severe case dispositions and then are more likely to commit future offenses.

Gender and criminal history were most often associated with higher recidivism. Males are overrepresented in the criminal justice system at all stages and offender ages. A third factor related to recidivism was the individual’s age at first disposition. The earlier a person had their first disposition, the more likely they were to recidivate after the qualifying offense.

While we believe this study provides an accurate and important picture of the state of recidivism among juvenile offenders in Washington State, we know there is room for improvement. In the next report we hope to incorporate data from Washington’s Department of Corrections data to better track how many subjects truly had the minimum follow up period outside of incarceration. We would also hope to incorporate older King County detention facility records, to provide a more comprehensive record of each subjects’ previous detentions. In addition, we will attempt to incorporate additional explanatory variables into the analysis to provide deeper insight into the court and JRA population, their recidivism, and possible explanations for these outcomes.

image4.png image5.png image6.png image7.png image8.png image9.png image10.png image11.png image12.png image1.png image2.png image3.png