Determinants of sustainable mode choice in different socio-cultural contexts: A comparison of Rome and San Francisco

ABSTRACT

This paper is a part two of a study investigating the relative importance of the built environment, socio-demographic, and attitudinal factors on mode choice. A semi-experimental approach that aims to measure causal effects of the built environment is utilized. This paper reports spatial analysis, survey and modeling results for San Francisco, CA, USA and compares the results with a previous similar study in Rome, Italy. Results reveal that the local street network's integration is important in both cities and that in both cases built environment seems to have higher impact on mode choice than attitudes and socio-demographic factors. Built environment is especially impactful when diversity, design quality, density and syntactical accessibility are combined. In San Francisco willingness to spend time walking, biking or taking transit is lower than in Rome, and residents are more sensitive to concerns about safety and security. Work travel is more affected by demographic and attitudinal factors in San Francisco than in Rome implying that in San Francisco, nonwork travel behavior may have slightly higher potential to respond positively to improvements in the built environment than work trips. In Rome, peer pressure, cost sensitivity, and probiking attitude can compensate for lack of some built environmental characteristics, but not in San Francisco, where only protransit attitude has this effect. Moreover, lack of any built environmental characteristics reduces the possibility of sustainable mode choice more dramatically in San Francisco pointing to the higher importance of investments on improving the built environment rather than marketing efforts to change attitudes.

KEYWORDS:

• International comparative research
• modal choice
• personal traits
• population density
• space syntax
• street network integration

Notes

¹ Including categories such as urban and suburban in mode choice models (common in mode choice models in the literature which include built form as neighborhood type categories) is not able to capture the effect of different built form characteristics of different neighborhoods especially when comparing two completely different contexts. For example, two urban neighborhoods in San Francisco (or in Rome) may have the same population density but not the same street network characteristics and/or the same street design quality. It is even more problematic when we compare a suburb in Rome to a suburb in San Francisco. We therefore, try to make the comparison between street types more reliable by categorizing them based on the way different urban form factors are combined in each type.

² The public right of way surface area for each street segment was calculated by multiplying the entire distance from building edge to building edge (as the width) by the length of that street segment. The planted area was also calculated for each street segment. Total measures were then calculated for each buffer zone to determine the percentage of planted surface in each.