![Sample our Law journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5942/TOC-Subject-Sample-2021-Law.jpg"})
Abstract
Recent scholarship has highlighted the potential implications of in-prison experiences for prisoner reentry and, in particular, recidivism. Few penological or reentry studies, however, have examined the relationship between one experience that may be especially consequential, inmate misconduct, and recidivism. The goal of this study is to address this gap in the literature by employing a matching design that estimates the effect of inmate misconduct on reoffending, using data on a release cohort of Florida prisoners. The results indicate that inmates who engage in misconduct, violent misconduct in particular, are more likely to recidivate. Consistent with prior scholarship, we find that this relationship holds only for adult inmates. These findings underscore the importance of prison experiences for understanding recidivism, examining youthful and adult inmate populations separately, and devising policies that reduce misconduct.
Acknowledgements
We thank the FDOC for permission to use their data. The views expressed here are those of the authors and do not reflect those of the Department of Corrections. We also thank the Editor and the anonymous reviewers for their constructive suggestions and ideas.
Notes
1. We excluded inmates who served one month or less in prison (e.g. one or two days), because the first month typically constitutes a transition period for screening and assessment prior to a final prison placement. Ancillary analyses that included all of the inmates in the sample (i.e. did not include this restriction) produced results that were substantively identical to those shown here.
2. We include, as an added control, fixed effects for the 67 counties in Florida but, to conserve space and because they are not central to the discussion, do not report results for them here.
3. We used a .005 caliper and 1:1 nearest neighbor matching without replacement. Using a narrow caliper setting ensured that treated individuals were matched to subjects from the comparison pool who had nearly identical probabilities of having engaged in misconduct (DiPrete & Gangl, Citation2004). In addition, we used the non-replacement option to ensure that only unique pairings were allowed.
4. As recommended in the propensity score literature (see, e.g. Becker & Ichino, Citation2002; Rosenbaum & Rubin, Citation1983, Citation1984), we incorporated polynomial specifications in the propensity models to attain balance on the covariates.
5. Prevalence estimates for prisoner misconduct in this study are 40% for any misconduct (7,810/19,594) and 9% for violent misconduct (1,782/19,594). These estimates accord with those in other studies (see, e.g. Bales & Miller, Citation2012; Sorensen & Cunningham, Citation2010; Steiner & Wooldredge, Citation2008; see also Sorensen, Cunningham, Vigen, & Woods, Citation2011; Wolff & Draine, Citation2009).
6. These analyses are available upon request.
7. In ancillary analyses, we created a matched group from individuals who committed a non-violent infraction to determine if using a different counterfactual would alter the results. The findings were similar to those here, save that the effect was not quite as large. For example, prior to matching, the general recidivism rates of the violent misconduct group and the matched non-violent misconduct group were 54 and 48%, respectively. After matching, they were 54 and 49%, respectively. The analyses are available upon request.
8. The reviewers suggested conducting analyses that examined other types of misconduct. To this end, we examined several other types, including property, disorderly, drug, other, and a variety score that included these types and also violent misconduct. The matching analyses produced findings that were substantively similar to those shown in the paper—that is, they consistently found a positive and significant effect of misconduct on recidivism. Because of the literature’s focus on the possibility that violent misconduct may exert a stronger effect on recidivism and the consistent findings regardless of type of misconduct, we present only the results shown in the paper. The additional results are available upon request.
9. One reviewer suggested examining the approach used by Spivak and Damphousse (Citation2006). This approach consisted of creating a percentage of time served vs. assigned sentence length. Ex-prisoners with smaller percentages could be inferred to have behaved better while in prison because of reduced time due to “good conduct.” This approach may be useful in contexts where information about inmate misconduct is not directly available. However, in this study, the association between misconduct and this percentage-of-time-served measure was weak (r = .25), suggesting that the direct measure of misconduct may be preferable to this indirect measure.
10. DiPrete and Gangl (Citation2004, p. 291) have emphasized that a gamma of any given value “does not mean that there is no true effect.” Instead, it “means that the confidence interval would include zero if an unobserved variable caused the odds ratio of treatment assignment to differ between treatment and control groups by(gamma) and if this variable’s effect on [the outcome] was so strong as to almost perfectly determine whether the [outcome] would be bigger for the treatment or control case in each pair of matched cases in the data” (p. 291; emphasis in the original).
11. We thank an anonymous reviewer for the recommendation to pursue this line of analysis. Results for the full set of models for the three analyses are available upon request.
Additional information
Notes on contributors
Joshua C. Cochran
Joshua C. Cochran, MS, is a doctoral candidate at Florida State University’s College of Criminology and Criminal Justice. His research interests include theory, international comparative analyses of criminal justice systems, and prisoner reentry. He has published in Justice Quarterly, Crime and Delinquency, and other crime and justice journals.
Daniel P. Mears
Daniel P. Mears, PhD, is the Mark C. Stafford professor of criminology at Florida State University’s College of Criminology and Criminal Justice. He conducts basic and applied research, and his work has appeared in leading crime and policy journals and a book, American Criminal Justice Policy (Cambridge University Press 2010).
William D. Bales
William D. Bales, PhD, is a professor at Florida State University’s College of Criminology and Criminal Justice. He focuses on a range of crime and policy topics, including the effectiveness of electronic monitoring and tests of labeling theory. He has published in Criminology, Criminology and Public Policy, and other crime and policy journals.
Eric A. Stewart
Eric A. Stewart, PhD, is a professor at Florida State University’s College of Criminology and Criminal Justice. He is a member of the Racial Democracy, Crime and Justice Network. His research interests include racial inequality, crime throughout the life course, and the effects of multilevel processes on adolescent development.