Abstract: Understanding Predictive Factors in Prison Misconducts among Juveniles Serving Time in Adult Prisons (Society for Social Work and Research 14th Annual Conference: Social Work Research: A WORLD OF POSSIBILITIES)

13358 Understanding Predictive Factors in Prison Misconducts among Juveniles Serving Time in Adult Prisons

Thursday, January 14, 2010: 2:30 PM
Garden Room A (Hyatt Regency)
* noted as presenting author
Karen Kolivoski, MSW , University of Pittsburgh, Doctoral Student, Pittsburgh, PA
Jeffrey Shook, PhD, JD , University of Pittsburgh, Assistant Professor, Pittsburgh, PA
Background and Purpose:

Juveniles serving time in adult prisons are an especially vulnerable population, yet few studies have addressed what happens to them as they serve time. Looking at one specific outcome measure—prison misconducts—and the variables that are predictive of this behavior can lend further understanding regarding this population.

Only two known studies have addressed misconducts of juveniles in prison (Mcshane & Williams, 1989; Kuanliang, Sorensen, & Cunningham, 2008), and have left important gaps to fill regarding knowledge in this area. One piece that has yet to be fully addressed is the role of individual characteristics of juveniles in predicting prison misconducts, particularly in multivariate models. Thus, this study seeks to answer the question of the role of individual factors in predicting misconducts among juveniles in prison.


Administrative data (n = 763) from the Michigan Department of Corrections on juvenile inmates in adult prisons in Michigan from 1984-2003 were used for this study. Inclusion criteria were those who are male and had their first commitment to prison for an offense before age 17. Descriptive analyses and a Poisson regression model with a scaling adjustment were run. Individual-level variables that were entered into the model included race, age at commitment, known mental health and/or drug use, prior history in the system, and offense categories to look at the relationship to total number of official prison misconducts.


Results of the Poisson regression show significant differences on a number of predictors: age, race, mental health, prior system involvement, and type of offense. Specifically, consistent with previous research, age had a negative relationship to prison misconducts in that younger inmates had more misconducts. African-American juveniles in the sample had 32% more total misconducts than whites β = -.278, exp(β) = 1.320, Wald chi-square(1) = 6.391, p = .011. Youth with known mental health issues had 40% more total misconducts than those without β = .337, exp(β) = 1.401, Wald chi-square(1) = 10.871, p = .001. Using juveniles with the most serious crimes as the reference category, differences between offenses emerged as well. Notably, property offenders had 88% more total misconducts than serious person offenders β = .636, exp(β) = 1.888, Wald chi-square(1) = 24.915, p < .001.

Implications and conclusions:

Unlike prior research (McShane & Williams, 1989), this study identifies significant, individual-level predictors of official misconducts in a sample of juveniles in prison. The results have important implications regarding identifying those young people who are most likely to commit misconducts. Moreover, however, study findings bring attention to broader issues regarding the need for comprehensive and appropriate mental health services and raises questions regarding the use of current prison sanctions for some offenders. Ultimately, implications of this research lie in the need for dialogue around confinement of juveniles in adult facilities.