by Kathleen A. Tomberg and Gina Moreno
John Jay Research and Evaluation Center
March 14, 2024
Overview
In 2016, the Osborne Association began a new reentry program called “Prepare” for fathers and father figures returning home from prison. With funding from the Office of Family Assistance within the Administration for Children and Families in the U.S. Department of Health and Human Services, the Prepare program aims to improve relationships between formerly incarcerated fathers and their children using a family-centered approach focusing on parenting and workforce skills with one year of follow-up support. The Osborne Association engaged JohnJayREC to evaluate Prepare. The 2021 evaluation report answered several research questions:
- Is the program successful in helping individuals reengage with their children and families?
- Is the program successful in helping individuals obtain and maintain employment?
- Is there a relationship between obtaining and maintaining employment and improved relationships between children and co-parents?
- Do interactions between staff and clients, as well as between clients, influence client outcomes?
As part of their 2021 report, the research team intended to measure participant recidivism, comparing recidivism rates for participants who successfully completed the program against those who did not. However, at the end of the initial evaluation, the necessary criminal history data was not yet available from the New York State Division of Criminal Justice Services (DCJS).
Obtaining DCJS Data
In 2023, JohnJayREC obtained de-identified DCJS data for Prepare participants through an existing data-sharing agreement.1 The agreement provided JohnJayREC with unsealed2, fingerprintable3 criminal history data to help New York organizations working with justice-involved individuals understand the justice-related outcomes of their clients. Researchers inspected and standardized Osborne program participant information files (including name, sex, race, date of birth, Social Security number, and New York State ID number) before sending the list to DCJS to request a criminal history match for each name. Survival analyses estimated how many participants were arrested for a fingerprintable offense within five years of exiting the Prepare program.
1. New York’s Division of Criminal Justice Services (DCJS) maintains, analyzes, and publishes justice system data, including incidents of crime, arrests, and dispositions. Police, sheriffs, probation departments, and courts report the data to provide the public and policymakers with information detailing how the justice system operates in their communities.
2. When JohnJayREC entered into the data acquisition agreement with DCJS, sharing of sealed records was not permitted due to requirements of other DCJS legal agreements. Not being able to match client data with a full set of an individual’s recorded justice system contacts could create missing observations and bias results.
3. JohnJayREC received arrest disposition data from DCJS related to fingerprintable arrests only. These arrests include all Penal Law felony and misdemeanor charges as well as other felonies and misdemeanors that can be elevated to felonies for repeat offenders. The JohnJayREC team did not receive disposition data from any arrests involving an infraction or violation as the top charge.
Time-to-Rearrest Analysis
Survival analyses estimate the predicted time to particular events, in this case, the average time until program participants were re-arrested. The research team used the Kaplan-Meier estimate to calculate the likelihood of a participant remaining arrest-free during the study follow-up period. Values on the survival curve charts are estimates generated by the model and do not always match counts presented in the tables. Each chart includes a 95 percent confidence interval, indicating the upper and lower bound of values around a given estimate.
Since JohnJayREC sampled a subset of all participants, the research team cannot be certain that the average time-to-rearrest or survival rate observed in the sample is the same as the population as a whole. To account for this uncertainty, researchers used a confidence interval that suggests the degree of uncertainty around the percentage of participants remaining arrest-free in a given period post-program exit. In other words, if researchers conducted many survival analyses using different participant samples, there would be a 95 percent chance the interval would contain the true survival rate — i.e., the true percentage of participants not arrested after leaving the program.
The study includes different participant cohorts with varying entry and exit dates. Time zero refers to when individuals exited the program regardless of the date. Results obtained from the survival analyses estimated the likelihood of participants remaining arrest-free during a determined follow-up period after program completion.
Sample Description
Osborne submitted the names of 671 former Prepare clients for analysis, all of whom left the program before 2020. Names were matched to the criminal history records maintained by DCJS. The matching process accurately identified 542 Prepare clients (81% match rate). By the end of the first post-program year, 92 percent of those participants remained arrest-free. The number dropped to 74 percent by the end of the third year and 64 percent by the end of the fifth year.
The sample of Prepare clients (n=542) was primarily male (96%) and either non-Hispanic Black (66%) or Hispanic (27%) (Table 1). One-fifth (20%) of sampled participants were between ages 18 and 30, while 36 percent were between 31 and 40, and 44 percent were age 41 or older.
The age range is unsurprising, given that Prepare is a program geared towards fathers and father figures. Most participants completed 8 to 12 years of education (70%), most exited the Prepare program successfully (88%), and more than half were placed in a job during the program (54%).
More than one-quarter of the 542 Prepare clients in the analytic file (26%, n=139) had at least one arrest following their exit from the program (Table 2). Most of the 139 participants who experienced an arrest were arrested one to two times (85%), and almost half were convicted (47%, n=66). Convictions were most often for misdemeanor charges (41%), while the rest were for felonies (26%), violations (16%), and infractions (13%).
Survival Analysis
Entire Sample
Researchers conducted a survival analysis using arrest data for the entire Prepare sample (n=542) after participants exited the program. This type of modeling generalizes time-to-arrest and generates estimated probabilities/percentages of staying arrest-free for the larger population of individuals participating in Prepare. There were 476 (88%) successful exits (participants who completed the program), while others were considered either unsuccessful exits (n=63) or exits due to administrative reasons (n=3).
According to estimates generated by the survival model used, by the end of the first post-program year, 92 percent of Prepare participants remained arrest-free. This number dropped to 84 percent by the end of the second year and to 64 percent by the end of the fifth year (Figure 1).
Sub-sample: Successful Program Exits
JohnJayREC then analyzed the sample of 476 participants who completed the program successfully. The sample was largely male, non-Hispanic Black (65%) or Hispanic (27%). Almost half of the sample was age 41 or older (47%). Another large group was between ages 31 and 40 (35%), and the rest were 18 to 30 (18%). Half of the sample completed 12 years of education, while a quarter had finished 8 to 11 years.
Around a fifth of the participants in this sample (22%, n=105) were arrested within five years of exiting the program (Table 3). Most (86%) were arrested between one and two times, and slightly fewer than half (43%) of those arrested were convicted. More than a third of convictions (38%, n=17) were for misdemeanors, and another third (31%, n=14) were for felonies.
Researchers conducted a survival analysis to examine arrest data for the sample of participants who successfully exited the program (n=476). By the end of the first post-program year, 93 percent of participants remained arrest-free. This number dropped to 87 percent by the end of the second year and 67 percent by the end of the fifth year (Figure 2).
Sub-sample: Unsuccessful Program Exits
JohnJayREC analyzed the subsample of 66 participants who did not successfully exit the program (including three administrative exits). Almost half (41%, n=27) left the program due to non-compliance. The Osborne Association lost contact with another quarter (24%, n=16), and a fifth (18%, n=12) voluntarily withdrew.
The demographics of this sample were similar to the group with successful exits. Participants were primarily male, non-Hispanic Black or Hispanic, and most were older than 30. More than a third of these participants completed 12 years of education, and a third had 8 to 11 years.
More than half of the participants in the unsuccessful sample (34 of 66) were arrested within five years of exiting the program. More than eight in ten were arrested once or twice, with more than half resulting in conviction. Approximately half of those convictions were for misdemeanors, a quarter were for violations, and another fifth were for felonies.
Researchers conducted an additional survival analysis to examine the arrest data for participants who did not successfully exit the program. By the end of the first post-program year, 82 percent remained arrest-free. This number dropped to 65 percent by the end of the second year and 44 percent by the end of the fifth year.
Differences by Job Placement
Researchers compared the 294 participants receiving job placements during Prepare and the 248 not receiving job placements. One-fifth of participants with job placements (20%, n=58) were arrested within five years (Table 4). Just 10 percent were arrested more than two times, and fewer than half (45%) of those arrested were later convicted. Of the participants not receiving job placements, 87 (33%) were arrested within five years and 66 (or 81% of those arrested) were arrested three or more times.
Researchers examined all participants receiving job placements after leaving the program (Figure 3). By the end of the first year, 95 percent of participants remained arrest-free. The number dropped to 89 percent by the end of the second year and 75 percent by the end of the fifth year.
Another survival analysis examined arrest data for participants according to whether or not they received job placements during the program. By the end of the first post-program year, 95 percent of participants with job placements remained arrest-free compared with 88 percent among those not receiving job placements. The numbers dropped to 75 and 56 percent five years after program exit, reflecting a 19 percent difference overall.
Finally, researchers conducted a survival analysis of participants (n=289) who completed successfully and received job placements (Figure 4). By the end of the first post-program year, 95 percent remained arrest-free. The number dropped to 88 percent by the end of the second year and remained at 74 percent by the fifth year.
Discussion
JohnJayREC’s 2021 evaluation report included a discussion of participant attitudes measured with surveys and interviews. However, the project was not able to obtain job retention data from the New York State Department of Labor (despite attempts to do so), to obtain criminal history data, or include a comparison sample. This analysis of criminal history data supplements the previous report, and this study sample includes additional Prepare participants.
Nearly three-quarters (74%) of Prepare participants remained arrest-free five years after leaving the program. The percentage was higher (78%) among participants who successfully graduated from the program and higher still (80%) among those who successfully graduated from the program after being placed in a job during the program. Of all participants receiving job placements during the program — whether or not they completed the program successfully — 80 percent remained arrest-free five years after leaving the program.
The survival analysis models suggest that, on average, nearly two-thirds (64%) of the larger population of participants in Prepare are likely to remain arrest-free five years after graduating from the program. This percentage goes up if participants successfully complete the program (67%) or are placed in a job (75%), regardless of program completion status. The models estimate that 74 percent of individuals who graduate from the program with a job will remain arrest-free after five years.
These findings are promising but are not sufficient to establish Prepare’s effectiveness. Job placements appear to be related to positive program outcomes, and this warrants closer examination, but other factors likely affect someone’s likelihood to recidivate. Familial connections, housing stability, health issues, and age influence recidivism. This analysis is unable to account for all such factors.
A more comprehensive evaluation of the Prepare program would need to measure a range of outcomes (shifts in participant attitudes, reconnection with children, job attainment and retention, avoiding future arrest, etc.) while including one or more comparison groups of individuals with similar characteristics and backgrounds who do not participate in programs similar to Prepare upon returning home from prison.
Research Note
The research team acknowledges that New York City’s experience of the COVID-19 pandemic complicates the interpretation of this analysis. Research that relies on crime data to assess the outcomes of an intervention in the justice sector must account for recent changes in key metrics. Violent crime rates in New York surged from 2020 through 2021 before declining in the latter period of the pandemic. During that time, arrests for less serious offenses and the court system’s ability to handle those offenses plummeted at first and then rebounded. The apparent outcomes of any justice-sector program would be affected by these patterns. Indeed, the proportion of Prepare participants arrested after completing the program fell from 20 percent to 11 percent between 2019 and 2020. By 2022, however, the figure had increased to 27 percent. The results of this analysis are strengthened by analyzing outcomes over a period of several years.
About the Authors
Kathleen A. Tomberg is the Deputy Director for Strategy and the IRB Officer (Institutional Review Board) for the Research and Evaluation Center. She earned a Masters in Forensic Mental Health Counseling from John Jay College of Criminal Justice.
Gina Moreno is a senior research analyst within the Research and Evaluation Center. She earned bachelors and masters degrees from John Jay College of Criminal Justice and an additional masters degree in statistics from Baruch College.
Acknowledgments
Funding for this report was provided by the Osborne Association. The authors are grateful to the staff and leadership of the Osborne Association for their guidance and support during the development of the project. They are also grateful for the support and advice received from colleagues at the Research and Evaluation Center who contributed to this report and the larger project of which it is a part. Finally, thank you to all who participated in Osborne’s Prepare program.
Recommended Citation
Tomberg, Kathleen A. and Gina Moreno (2024). The Osborne Association Prepare Program: Recidivism Analysis. New York, NY: Research and Evaluation Center, John Jay College of Criminal Justice, City University of New York.
Data analyzed in this document were provided by the New York State Division of Criminal Justice Services (DCJS). All opinions, findings, and conclusions expressed in this publication are those of the authors and not those of DCJS. Neither New York State nor DCJS assumes liability for the contents or use thereof.
Copyright
Research and Evaluation Center at John Jay College of Criminal Justice
City University of New York (CUNY)
Research and Evaluation Center
524 59th Street, Suite 605BMW
New York, NY 10019
http://www.JohnJayREC.nyc
March 2024