Highway work zones present an environment that leads to opportunities for traffic crashes that may not otherwise occur. To investigate the severity of work zone-related crashes and relationships between severity and other crash variables, a database of 5,410 work zone-related crashes that occurred in Alabama from 2007-2014 was developed. The database includes information from traffic crash reports, project traffic control inspector reports, and supporting documentation from contractors. The full range of variables included in the crash reports was reduced to a manageable set of 16 independent variables whose relationships to crash severity were then explored. This analysis involved the development of an ordered probit regression model and examination of frequency distributions. The five most statistically significant variables that affect crash severity were found to be Primary Contributing Factor, Manner of Crash, First Harmful Event, Highway Classification, and Work Zone Type. Specific factors that had a highly statistically significant effect on severity include evening and overnight time periods, open country locale, rain, no-passing zones, Federal and State highways, two-lane highways, head-on, rollover, and angle crashes, pedestrian and bicyclist involvement, single-vehicle crashes, excessive speed, improper lane use, and the presence of work on the shoulder or median.