Human exposure research has consistently shown that, for most volatile organic compounds (VOCs), personal exposures are vastly different from outdoor air concentrations. Therefore, risk estimates based on ambient measurements may over- or underestimate risk, leading to ineffective or inefficient management strategies. In the present study we examine the extent of exposure misclassification and its impact on risk for exposure estimated by the U.S. Environmental Protection Agency (U.S. EPA) Assessment System for Population Exposure Nationwide (ASPEN) model relative to monitoring results from a community-based exposure assessment conducted in Baltimore, Maryland (USA). This study is the first direct comparison of the ASPEN model (as used by the U.S. EPA for the Cumulative Exposure Project and subsequently the National-Scale Air Toxics Assessment) and human exposure data to estimate health risks. A random sampling strategy was used to recruit 33 nonsmoking adult community residents. Passive air sampling badges were used to assess 3-day time-weighted-average personal exposure as well as outdoor and indoor residential concentrations of VOCs for each study participant. In general, personal exposures were greater than indoor VOC concentrations, which were greater than outdoor VOC concentrations. Public health risks due to actual personal exposures were estimated. In comparing measured personal exposures and indoor and outdoor VOC concentrations with ASPEN model estimates for ambient concentrations, our data suggest that ASPEN was reasonably accurate as a surrogate for personal exposures (measured exposures of community residents) for VOCs emitted primarily from mobile sources or VOCs that occur as global "background" source pollutant with no indoor source contributions. Otherwise, the ASPEN model estimates were generally lower than measured personal exposures and the estimated health risks. ASPEN's lower exposures resulted in proportional underestimation of cumulative cancer risk when pollutant exposures were combined to estimate cumulative risk. Median cumulative lifetime cancer risk based on personal exposures was 3-fold greater than estimates based on ASPEN-modeled concentrations. These findings demonstrate the significance of indoor exposure sources and the importance of indoor and/or personal monitoring for accurate assessment of risk. Environmental health policies may not be sufficient in reducing exposures and risks if they are based solely on modeled ambient VOC concentrations. Results from our study underscore the need for a coordinated multimedia approach to exposure assessment for setting public health policy.