Many studies have found that residents living in suburban neighborhoods drive more and walk less than their counterparts in traditional neighborhoods. This evidence provides support to the idea of using smart growth strategies to alter individuals' travel behavior. However, the observed differences in travel behavior may be more of a residential choice than a travel choice. Applying seemingly unrelated regression to a sample from Northern California, we explored the relationship between the built environment and nonwork travel behavior, controlling for measures of residential self-selection. This study shows that, at the neighborhood level, individuals' non-motorized travel is greatly influenced by residential self-selection, and residential preference and travel attitudes provide an incremental contribution in explaining the variation in auto and transit travel. After accounting for the influence of self-selection, we also found that neighborhood characteristics themselves affect individuals' travel choices. Therefore, if cities use land use policies to offer residents opportunities to drive less and use alternative modes more, the evidence suggests that they will tend to do so.
Abstract