Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

William Fremouw

Committee Co-Chair

Nicholas Bowman

Committee Member

Thomas Horacek

Committee Member

Kevin Larkin

Committee Member

Aaron Metzger

Committee Member

Natalie Shook


With 2.2 million adults incarcerated throughout the United States (Glaze & Herberman, 2013), prisons are crowded, volatile environments susceptible to violence. Prior research has identified demographic and criminal variables that consistently predict prison violence. The purpose of this study was to investigate the utility of psychological variables to predict prison violence among 180 inmates from a state prison in West Virginia. The psychological variables studied included history of mental illness and results from psychological assessments (Beta-III, MMPI-2-RF, TCU Drug Screen II). Using a logistic regression analysis, history of mental illness and the MMPI-2-RF scale of Psychoticism accurately predicted violent inmates from non-violent inmates while controlling for demographic and criminal variables. A weighted measure of violence severity was calculated and a multiple regression analysis was conducted. In the final model, mental health history significantly predicted Violence Severity Index scores. Hierarchical cluster analysis identified three meaningful groups of violent inmates based on the five MMPI-2-RF personality psychopathology scales. Eleven inmates were characterized by high scores on Psychoticism and Negative Emotionality/Neuroticism. Another cluster of 15 inmates were defined by high scores on the Aggressiveness and Disconstraint (e.g., risk-taking, impulsivity) scales. The third group had low scores on these four scales and had moderate scores on Low Positive Emotionality/Introversion. The results of this study support the conclusion that psychological variables have utility in predicting prison violence, but more research is needed to continue understanding this relationship.