Date of Award

Fall 12-1-2015

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Criminal Justice

First Advisor

Dean Dabney

Second Advisor

Brent Teasdale

Third Advisor

Mark Reed

Fourth Advisor

John Jarvis


Approximately a third of homicide cases go unsolved each year. Research focused on understanding what affects homicide clearance rates is often methodologically underdeveloped and has produced mixed findings. These deficiencies compromise the ability of researchers to provide important guidance to police practitioners seeking to develop best practices. Under-specified modeling and limited access to accurate sources of homicide investigation data are two potential and interconnected reasons for the inconsistencies found in previous studies. The purpose of this study was to expand the literature on homicide case outcomes as follows: 1) to organize predictors into five substantive domains (involved subjects, event circumstances, case dynamics, ecological characteristics, and investigator factors) and operationalize multiple measures of each as viable predictors of clearance outcomes; 2) to explore the utility of using original and verified police data with a larger number of nuanced data points than previously documented in modeling efforts; and 3) to forward a unique multi-method account of the factors that predict homicide case outcomes that can be readily replicated in future studies. Data were collected from one Southern metropolitan police department's 2009 to 2011 homicide investigations (N = 252). Access to official homicide case files allowed for key subject, incident, and evidentiary information to be obtained. Critical investigation details and context were added to the case file data via interviews and survey administration efforts involving the lead detectives that worked the cases. The dataset was further supplemented with Census data. Subsequent analyses included examination of the data quality and multivariate logistic regressions. A comparison of the dataset after the first stage of data collection to the final product was conducted to understand the extent to which the dataset were improved. The multi-method process resulted in more precision to the data recorded from case files, significant reductions in missing data, and heightened detail on key variables. Consequently those data allowed for specification of a multivariate model that included multiple measures from all of the homicide investigation domains. Those results suggest the expanded data more accurately captured the factors that predict clearance outcomes as measures within all five domains were significant predictors of investigation closure.