Recent environmental justice studies have emphasized the growing need to analyze the health impacts of disproportionate exposure to multiple pollution sources and incorporate geostatistical techniques that are suitable for analyzing spatial data. These objectives are addressed in a case study that evaluates spatial and social inequities in potential cancer risk from inhalation exposure to hazardous air pollutants (HAPs) from four types of emission sources in the Tampa Bay Metropolitan Statistical Area, Florida. This study utilizes modeled estimates of lifetime cancer risk from the 1999 National-Scale Air Toxics Assessment and socio-demographic data from the 2000 US Census. Statistical analyses are based on conventional multiple regression and locally derived spatial regression models that account for residual autocorrelation. Race, ethnicity, and home ownership are found to be significant predictors of cancer risk from ambient exposure to all four HAP source categories, after controlling for other relevant explanatory factors and spatial dependence in the data.