There are many ways to evaluate the built environment, including measures of observable individual characteristics (such as activity density), continuous composite measures (such as the sprawl index), and categorically measured variables (such as neighborhood types). However, a systematic comparison of how well each of these three measurement types captures the influence of the built environment on travel behavior has not yet been undertaken. This lack presents a quandary for both researchers and practitioners who seek to quantify and describe the effects of the built environment on travel behavior. This paper assesses whether continuous, composite, or categorical measures provide more information and better-fitting models compared with measures of observable individual characteristics across four travel behaviors: vehicle miles traveled, walk trips, transit trips, and trip length. For each travel variable, four multivariate regression models were estimated with various measures of the built environment: Activity density, sprawl index, neighborhood type, and combined sprawl index and neighborhood type. Both the sprawl index and the neighborhood-Type models outperformed the activity density model. Moreover, a combined model with both the sprawl index and neighborhood types provided the best fit for all four travel behavior variables. These results suggest that both continuous and categorical composite variables provide unique and complementary information about how the built environment influences travel behavior. These findings underscore the importance of researchers' decisions on how to represent the built environment quantitatively in models, because measurement decisions influence the understanding of how the built environment affects travel behavior.
All Science Journal Classification (ASJC) codes
- Civil and Structural Engineering
- Mechanical Engineering