Evidence-Based Policing Matrix

Home › Evidence-Based Policing › The Matrix › Research on Individuals › Individuals – Goldstein et al. (2021)

Individuals – Goldstein et al. (2021)

Study Reference:

Goldstein, N. E., Kreimer, R., Guo, S., Le, T., Cole, L. M., Nemoyer, A., Burke, S., Kikuchi, G., Thomas, K., & Zhang, F. (2021). Preventing school-based arrest and recidivism through prearrest diversion: Outcomes of the Philadelphia police school diversion program. Law and Human Behavior, 4(2), 165-178.

Location in the Matrix; Methodological Rigor; Outcome:

Individuals; Focused; Reactive; Moderately Rigorous; Mixed effects

What police practice or strategy was examined?

This study evaluated the Philadelphia Police School Diversion Program that was administered by the Philadelphia Police Department starting in 2014. This program aimed to dismantle the school-to-prison pipeline that overwhelmed the School District of Philadelphia. Children at least 10 years of age who lacked previous adjudication or pending cases and committed a minor offense on school property qualified for the diversion program. Officers were prohibited from arresting the individual when they received a call to respond to a qualified incident at a school. Instead, the diverted juvenile would receive a home visit from a Philadelphia Department of Human Services (DHS) social worker. The social worker would recommend free services to the family and/or youth that were funded by DHS; however, there was no obligation for the youth to participate in the programs (to prevent a net-widening effect). Once a juvenile was enrolled in the diversion program, they were no longer eligible to be diverted in future incidents.

How was the intervention evaluated?

A quasi-experimental design was employed to evaluate the diversion program’s impact on school-based arrests, serious school behavioral incidents, and recidivism rates. Youth diverted (N=1,281) under the program from September 2014 through June 2017 were compared to all youth arrested during the 2013-2014 academic year (the year prior to the diversion program; N=1,021) and to a quasi-control sample (N=531). This quasi-control sample included youth in the 2013-2014 arrested sample that would have qualified for the diversion program had the program existed during the 2013-2014 academic year. A propensity score matching technique was also used to match youth in the quasi-control arrested sample to youth in the diversion sample based on the individual’s race/ethnicity, gender, age at initial incident, and the initial incident type. The researchers used school-based data to assess differences in school-based arrest rates and serious behavior incident rates. All arrests incurred by the juvenile in the two years following the initial date of arrest/diversion were used to assess recidivism rates.

What were the key findings?

This study found a notable reduction in the annual number of arrests during the diversion program, with the Philadelphia schools experiencing 84% less school-based arrests by year five of the program. Behavioral incidents also decreased, with schools experiencing a 34% total reduction in behavior incidents during the fifth year of the diversion program. When comparing the diverted youth sample to the arrested youth sample and the quasi-control youth sample, the study finds that in most months diverted youth were significantly less likely to experience a rearrest. Among the youth that did recidivate, diverted youth experienced a longer time to rearrest than youth in the quasi-control group and the full arrest group. However, once the quasi-control and diverted youth were matched using propensity scores, there were no significant differences between these two groups for the likelihood of rearrests and the time to rearrest.

What were the implications for law enforcement?

The authors conclude that the diversion programs reduced school-based arrests without increasing behavioral incidents. This suggests that it is possible to respond to juvenile misconduct using alternatives to arrest without compromising school safety. However, the study’s most rigorous analyses (those comparing diverted youth to a matched sample of comparison youth) raise some caveats about the program’s effectiveness. The authors suggest the relationship between diversion and recidivism is likely complex and future research should explore potential confounding variables in their evaluations.

Where can I find more information about this intervention, similar types of intervention, or related studies?