The goal of this paper is to determine the recent annual occupant populations and trends related to rollover injuries and fatalities, and to assess the risk factors that may have significantly contributed to occupant injuries and fatalities when rollovers and ejections occur. Fatality Analysis Reporting System (FARS) 2004-2017 data were used to obtain the recent occupant fatalities related to rollovers and ejections. National Automotive Sampling System (NASS) – Crashworthiness Data System (CDS) 2013-2015 weighted survey data were used extensively to estimate the occupant injury severities and occupant ejection details associated with rollovers. For rollover cases, the injured body regions (e.g., head, neck, shoulder and back, and chest) and injury contact sources (e.g., vehicle roof, side door, or seat back) were investigated in detail. This study paid close attention to the interaction between the vehicle roof and occupant injuries for the consideration of the requirements from Federal Motor Vehicle Safety Standards (FMVSS) 216a and 226. Finally, occupant injury risk and key risk factors were evaluated using methods of relative risk, including multiple logit model and case-control study. The data analysis using FARS showed a decrease in annual fatalities from approximately 10,500 during 2004-2006 to approximately 7,000 during 2014-2017. Approximately one thirds of all occupant fatalities of light passenger vehicles (Gross Vehicle Weight Rating, GVWR <=10,000 lbs.) are related to rollovers. FARS data also provided the occupant ejection status (complete or partial ejection) and ejection path associated with rollovers. The CDS data indicated that rollovers are strongly associated with the injury sources of vehicle roof, side doors, and seat back/support; rollover crashes also resulted in primarily the injured body regions of head, neck, shoulder/back, and chest. The occupant ejection paths are usually side windows and roof opening. The analytical results also revealed that light trucks /vans, with relatively higher centers of gravity, tend to have relatively higher likelihood of rollover crashes than passenger cars, but passenger cars tend to result in a higher rate of occupant serious injuries (Maximum Abbreviated Injury Scale 3+, or MAIS 3+) than light trucks/vans, if rollovers did occur. Overall this study explored the trends and annual occupant populations related to rollovers and ejections using recent traffic data. Logistic regression model was used, with considerations of multiple risks and confounding factors, to predict the occupant injury relative risks of several key risk factors simultaneously. The analytical results using both FARS and CDS indicated that higher occupant injury risks were especially associated with higher delta-V, unbelted occupant, rollover, ejection, side impact, and older occupant age. This study, utilizing the recent crashes from three main databases of FARS, NASS CDS and NASS General Estimates System (GES), may enrich the understanding of rollover and ejection related occupant injuries.