Sensitivity analyses using only temporal estimates showed that the spatiotemporal models showed little difference in the effect estimates but showed slightly higher statistical significance for the spatiotemporal models. This confirmed the findings by Brauer et al. (2008) who compared ORs from traffic-based land using regression models and analyses based on exposures estimated from the ambient monitoring network. Although the effect esti- mates based on land use regression models did have somewhat smaller CIs than those based on monitoring network data, they did not indicate increased estimates of the effect, as one might expect given improved spatial resolution and the potential for reduced exposure misclassification.