This study explores the relationship between vehicle occupancy and several other variables in the National Highway Traffic Safety Administration's (NHTSA) Fatality Analysis Reporting System (FARS) database and a 15-passenger van's risk of rollover. A univariate analysis is used to demonstrate the effect of selected variables on single-vehicle rollover crashes. Variables used include speed, number of occupants, driver experience and avoidance maneuvers. Also, a logistic regression model is constructed using data from NHTSA's State Data System - a collection of all police reported crashes for that state. The resulting model permits jointly estimating the effect of these variables on the odds and rate of rollover occurrence, conditional on being in a single-vehicle police-reported crash.