The Effectiveness of Field Contacts in Community Supervision: A Multi-site, Multi-year Evaluation Using Matched Samples

Abstract

Field contacts are a core practice within probation and parole. While criminological theory suggests field contacts may reduce recidivism, there is very little empirical evidence of this impact, and the costs of field contacts to officers and agencies are high. To date, only two studies have investigated the relationship between field contacts and recidivism, and both studies rely on a single site and regression modelling to address issues of endogeneity. The goal of the current study is to improve the strength of evidence of the effectiveness of field contacts in reducing recidivism through a multi-site study that addresses previously unaddressed methodological issues. This study employs coarsened exact matching, numerous variations in model specification, and alterations of sample specification to address issues of endogeneity. When accounting for other differences between supervisees who receive field contacts and those who do not, the use of field contacts is accompanied by reductions in recidivism. These findings provide vital information for understanding what works within community supervision. This study was limited to examining the use of one or more field contacts; future research should explore the impact of varying dosage as well as other forms of contact.

Keywords:

• Coarsened exact matching
• community supervision
• field contacts
• parole
• probation
• recidivism

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 Detail on data exclusions available upon request.

2 Because individuals can serve supervision terms for multiple sentences at once, the percentages for offense type and supervision type sum to greater than 100% as some individuals may be, for example, serving both a probation and supervised release sentence or sentences for both a felony and misdemeanor offense at the same time.

3 Detail on data exclusions available upon request.

4 Greater detail on the variables and values used in the matching process before and after matching is available upon request.

5 Greater detail on the results of the Rosenbaum bounds tests available upon request.

6 As a sensitivity analysis, we also re-run all analyses with the full sample (including those who fail before seven days).

7 Full regression output available upon request.

8 Greater detail on sensitivity analyses available upon request.

9 Sensitivity analyses were also run on the full sample (leaving those in who recidivate before seven days). The results are nearly identical to those observed for the seven-day sample.

Additional information

Funding

The work was supported the National Institute of Justice under Grant #2013-IJ-CX-0103.