UC Berkeley

This dissertation analyzes the effects of two correctional policies on prison misconduct. Chapter 1 briefly frames prison as a policy built environment and provides an overview of mass incarceration in the United States. Chapters 2 and 3 provide causal estimates of the effects of two correctional policies on prison misconduct.

Chapter 2 estimates the relationship between prison visits and self-reported inmate misconduct using the 2004 Survey of Inmates in State Correctional Facilities (SISCF). This paper contributes to the extant literature by broadening the scope of the conversation about the determinants of inmate behavior to include influences from outside of the prison, namely prison visits, as opposed to limiting the discussion to individual or prison-specific influences. By employing an instrumental variables approach to estimating the relationship between prison visits and inmate misconduct the paper is the first to address the threats to internal validity posed by direct estimation of the effect of visitation on prison misconduct. The intuition behind my identification strategy is that distance between an inmate's home and place of incarceration isolates quasi-random variation in prison visitation, in effect assigning prison visits to inmates in a given state at random. The results suggest receiving visits reduces certain types of misconduct and the findings suggest the potential to reduce prison misconduct without resorting to increased isolation.

Chapter 3 estimates the relationship between facility security level and prison misconduct using an administrative data set from the California Department of Corrections and Rehabilitation (CDCR). The different levels of prison facility are designed to recognize heterogeneity in the inmate population and to appropriately house inmates during their incarceration to minimize risk of misconduct and escape. Prison facility security levels vary in physical characteristics, average levels of violence and other misconduct and staff perceptions of safety. An increase in facility security level could result in a suppression effect on misconduct and/or a peer effect which could positively or negatively effect misconduct. In this chapter, I employ a regression discontinuity (RD) design that exploits cutoffs in the security classification score to characterize the relationship between security classification and prison misconduct. The results of the paper suggest that inmates placed in a Level III facility are 8 percentage points less likely to incur a RVR than inmates placed in Level II, and that this result is driven almost entirely by a lower likelihood of write ups for Division E or F violations, which are the lowest level of violations eligible for write up as RVRs. I hypothesize that this result may be a result in differences in the priorities of custody staff as opposed to lower numbers of these types of violations at Level III prisons. In contrast to the findings between Levels II/III, I do not find an effect of facility security classification on the incidence of serious RVRs at the Level III/IV cutoff.

Overall, the goal of the dissertation is to contribute to the extant knowledge about the effects of correctional policies on inmate outcomes by describing how certain correctional policies shape the in-prison behavior of both inmates and custody staff. Since the effects of incarceration most likely reverberate to those who interact with inmates during their incarceration and persist after an inmate is released, understanding the effects of correctional policies on in prison behavior contributes to our understanding of how incarceration affects individuals, their families and their communities post-release and, in doing so, contribute, in some part, to a better understanding of what it means to use incarceration so extensively in the United States.