Previous contextual research relies heavily on data collected by the U.S. Census Bureau due to its widespread availability, but this incorrectly assumes that populations are equally distributed within the arbitrarily-drawn Census boundaries. Areal interpolation research suggests that methods employing ancillary data assist in correcting the problem. This thesis posits that the augmentation of census variables through Intelligent Dasymetric Mapping, and the redefinition of neighborhoods as circular or road network-based buffers around interview locations, provides more accurate contextual information for modeling neighborhood effects. This hypothesis is tested through the multivariate regression analysis used to explain individual perceptions of community collective efficacy which is theorized to mediate specific of criminal behavior. Effects of neighborhood size (i.e. buffer distance) are also explored. Results indicate improved model explanatory power compared to that of models using traditional measures of context while significance of model parameters estimates varies by scale.