Date of Degree


Document Type


Degree Name



Criminal Justice


Jeffery Mellow

Subject Categories

Criminology | Criminology and Criminal Justice | Public Policy


Escape, Jail, Prison, Situational Crime Prevention, Violence


Introduction: Preventing escapes from custody is a critical function of prisons, jails, and the individuals who run these correctional facilities. Escapes are a popular topic in the news, among lawmakers, and in public discourse. Much of this interest stems from the widespread notion that escapees pose a serious threat to public safety, as well to the safety of correctional staff and law enforcement officers tasked with preventing and apprehending them. However, despite the importance of preventing escapes and minimizing violence, there has been very little empirical research on these issues in the past several decades. Extant research has also been limited in terms of its depth, breadth, and methodological rigor. Thus, the current dissertation seeks to address the following research questions:

1. What jail-level factors are related to escape-proneness?

2. What prison-level factors are related to escape-proneness?

3. What inmate-level characteristics are associated with escape behavior?

4. How often and at what point does violence occur during escapes?

5. What facility-level factors influence the likelihood of an escape being violent?

6. What incident-level variables influence the likelihood of an escape being violent?

7. What characteristics of the escapee influence the likelihood of an escape being violent?

Methods: To address these research questions, this study explores the degree to which facility-, incident-, and inmate-level factors are associated with two overall outcomes: 1) escapes from custody and 2) violent escape outcomes. To accomplish this, a series of analyses were conducted using several different sources of data. Specifically, the first two analyses used data from the 2011 Annual Survey of Jails (n=366) to examine how jail-level variables impact the number of escapes and attempted escapes from jails, and from the 2005 Census of State and Federal Adult Correctional Facilities (n=1821) to examine how prison-level variables impact the number of escapes and the number of walkaways from prisons. The third analysis used the 2008 and 2009 iterations of the National Corrections Reporting Program (n=7,300) to test whether relevant inmate-level characteristics were associated with the likelihood of an individual being an escapee. The final set of analyses examined the degree to which facility-, inmate-, and inmate-level factors were able to predict four violent escape outcomes: violence at the breakout, in the community, during recapture, and overall. These analyses used data from the Correctional Incident Database, 2009 (n=610).

Findings: Several jail-level variables--including rated capacity, ethnic heterogeneity, percent noncitizens, and privately operated--were significantly associated with the number of escapes and escape attempts from jails. There were also many prison-level variables associated with the number of escapes and the number of walkaways from a facility, including measures related to the facilities' administration and management (e.g., rated capacity, percent capacity, inmate-staff ratio, inmates from other authorities, court order, secure perimeter, security level, region), inmate populations (e.g., percent male, percent noncitizens), and treatment and programming options (e.g., percent on work assignment, percent on work release, alcohol or drug treatment, inmates permitted to leave).

At the individual-level, information about inmates' demographics (e.g., age, sex, race), criminal histories (e.g., prior time in prison and jail, prior escape), and current sentence (offense type and counts, sentence length, percent of sentence served) were associated with individual escape behavior.

Finally, findings indicate that violence is, overall, a relatively rare outcome in escape incidents, though when it does occur it is precipitated by certain situational factors. Incident-level factors were the best indicators of violence, including whether the escape occurred in secure custody, the location of the incident, and the start time of the escape. The classification of the facilities was also associated with violence (i.e., escapees from higher security prisons and jails were more likely to use violence than escapees from minimum security facilities). Inmate-level factors were the least important for understanding when an escape would result in a violent outcome, though some of the findings indicate that young, male escapees, who were in custody for a violent offense and had a history of escaping, were more likely to use violence during their escapes than other escapees.

Discussion and Implications: These findings demonstrated that opportunity- and place-based theories of criminal behavior, such as the situational crime prevention and routine activities frameworks, are most useful for understanding when escapes are likely to occur and when they are likely to result in violence. For example, higher security prison facilities had fewer escapes than lower security prisons, but prisons that permitted inmates to leave the facility (e.g., to study, participate in a rehabilitative program, or work) had a greater number of walkaways. At the individual level, inmates who were on community release were much more likely to have been escapees than those who were not on community release. Finally, inmates who escaped during transport were more likely to use violence than those who escaped under other circumstances.

Based on these findings, this dissertation provided several recommendations for policy and practice. For example, it was recommended that correctional administrators adopt strategies for preventing escapes that are rooted in the situational crime prevention framework. These might include modifying the environment and enhancing certain types of security features, but could also include providing counseling to inmates, allowing more home visits and furloughs, offering more programming in the prison, and protecting inmates when their safety is threatened. It was also recommended that administrators identify and implement best practices for situations in which violence is most likely to occur, such as during inmate transport. Finally, given that most escapes are nonviolent and relatively minor incidents, it was recommended that administrators consider expanding their practice of punishing escapees internally rather than charging them with a new crime that could potentially add years to their sentence.

Conclusion: Though there are several substantive and methodological limitations to the current dissertation, this research contributes to the literature by: analyzing the impact of a range of facility-, incident-, and inmate-level factors on escapes from custody and violence; examining a broader range of escapes from across the country; using more recent data and more rigorous analyses; clarifying some of the contradicting and confusing findings from previous studies; and providing a thorough analysis of the amount, scope, and predictors of violence escape outcomes.