Corrections_Today_September_October_2020_Vol.82_No.5

Recidivism

Table 1: First-Time Inmates Exiting from Prison for the First Time During FY2010-FY2015 by Three-Year Return to Prison Rates

return rate at 11%. Those paroled to the street had the next highest rate of return at 28%, while those released to an ACC had the highest return to prison rate at 40%. Addi- tional statistical tests performed on these data confirm these differences while controlling for race, gender, age and Equivant COMPAS general recidivism risk score. 3 Interestingly, these results run counter to recent meta-analysis research indicating that halfway house participants gen- erally demonstrate better recidivism rates than those released to parole or released without supervision. 4 Given the stark differences shown in Table 1 among the two post-prison supervision options, the simplistic decision to make at this point is to conclude that the ACC pathway is inferior to street parole. There is no denying that the recidi- vism rate is substantially higher for the halfway house option. However, is the only goal to ensure a lower comparative recidivism rate or eval- uate the actual resources consumed among groups? The fallacy that has been perpetuated over the years is that these two questions are one in the same. Taken a step further, the data will show that, at worst, there is really no difference between the ACC or street parole programs in

Three-Year Return Rate

Three-Year Return to Prison Any Reason

Count

Adult Community Corrections Facility (ACC)

No

528

Adult Community Corrections Facility (ACC)

Yes

347

40%

Total

875

Street Parole

No

592

Street Parole

Yes

227

28%

Total

819

No Supervision

No

432

No Supervision

Yes

51

11%

Total

483

All

No

1,552

All

Yes

625

29%

Total

2,177

group told a vastly different story. For the entire observed cohort, Table 2 shows that ACC participants spent fewer total days in prison during the study period than those who parole to the street by an average of 80 days (942 days compared to 862 days). This difference was confirmed as statistically significant. 3 Addi- tionally, ACC participants spent, on

terms of resource consumption, and surprisingly that the ACC actually provides significant utility. Prison days results With the “damning” recidivism evidence in one hand, an analysis from the same dataset of the number of prison days consumed by each

Table 2: Mean Total Prison Days by First-Time Inmates During FY2010-FY2015 by Supervision Type

ACC Modeled (10 percentage point reduction in recidivism)

No Supervision

Street Parole

ACC

Mean Days Tracked

Mean Prison Days

Mean Days Tracked

Mean Prison Days

Mean Days Tracked

Mean Prison Days

Mean Days Tracked

Mean Prison Days

First Release

967

967 967

854

854 854 942

738

738 738 862

738

738 738 830

Return to Prison

2010 2053

1755 1843

1546 1670

1546 1670

End of Study

1010

58 — September/October 2020 Corrections Today