University of California Transportation Center

Intensified critiques of urban sprawl and its associated ecological and social impacts, have led to a broad movement - New Urbanism - that calls for incorporating the elements of compact urban form, mixed land uses, public spaces, and greater pedestrianization into the design of urban spaces. Proponents of this program argue that, among other benefits, increased pedestrian activity will reduce the adverse impacts of automobile travel, while stimulating greater public interaction and a higher quality of life overall. This dissertation evaluates the accuracy of these claims, using four case studies in the San Francisco Bay Area to examine the interactions of urban form with residents' travel behavior and use of their neighborhoods, and to develop models of walking frequency based upon these relationships.

The research methodology integrates both qualitative and quantitative analytic approaches. It conducts analyses across a range of scales, from regional land use, demographic, and access data that are more typically associated with transportation policy studies, to fine-grained urban form and travel behavior characteristics at the block and parcel levels, that are generally addressed only in studies with a strong urban design of architectural emphasis.

A detailed urban form survey evaluates the specific urban form characteristics of each case study neighborhood, in order to develop a hypothesized "walkability" spectrum for the four case studies based on urban form features. An extensive mail-back survey targets residents' travel behavior and use of local neighborhood, providing data that is linked to the analysis of urban form in each of the case studies.

The results of these two surveys are integrated to determine the effect that variations in urban form have on the perceived walkability of neighborhoods, residents' use of their neighborhood centers, and ultimately, whether this affects their perceptions of the neighborhood's overall livability. A combination of univariate, bivariate, and multivariate statistical analyses are applied to develop predictive models of walking frequency as a function of neighborhood urban form.

Research findings indicate that the walkability spectrum hypothesized for the four case studies based on urban form surveys, is confirmed by actual travel behavior and residents' perceptions of their neighborhood. Walking frequency models are influenced most highly by the variable of distance, but retain a residual neighborhood effect as well, which appears to correlate with the urban form variations related to walkability in the four case studies. Findings also suggest that some individuals self-select into neighborhoods due in art to the perception that they offer good pedestrian, bicycle, and transit amenities.