To Walk or Not to Walk: Testing the Effect of Path Walkability on Transit Users' Access Mode Choices to the Station

Abstract

This article focuses on micro-level walkability, which is measured at the route level and thus named “path walkability,” and tests whether path walkability influences transit users' mode choices to the station. To test its impact on access mode choices, a case study is conducted in downtown Mountain View, California. A station user survey is administered to collect access mode choice, socioeconomic status, and trip origins and walking routes of 249 transit users with previous experience of walking to the station. Using a path walkability measurement instrument developed for this research, 38 path walkability indicators are extracted from each of the 249 walking routes. The 38 walkability indicators are grouped by using factor analysis yielding four path walkability factors: “sidewalk amenities,” “traffic impacts,” “street scale and enclosure,” and “landscaping elements.” The four factors are utilized as new walkability variables for modeling access mode choices. With 150 walkers and 99 habitual auto users/occasional walkers, two access mode choice models are estimated. The basic model is first estimated without walkability variables and then the four path walkability variables are introduced for the expanded model. All four path walkability variables enter the expanded model and significantly improve the predictability of the mode choice model. The model result suggests that micro-level walkability influences access mode choices in a statistically significant way and having more walking-conducive walkability available for access trips increases the chance of choosing walking over driving. This research shows that improving micro-level walkability could be a cost-beneficial incentive for more walking to the station.

Key Words:

• access mode choices
• pedestrians
• station user survey
• TOD
• walkability
• walking

Acknowledgments

The authors would like to thank Professor Michael Southworth, Elizabeth Deakin, and Robert Cervero of the City and Regional Planning Department at the University of California, Berkeley, for their guidance and support. The authors also thank Janet McGovern of San Mateo County Transit District and David Lopez of Santa Clara Valley Transportation Authority for permission to conduct the user survey at the station.

Notes

¹For further details on the user survey, please refer to Park (2008).

²The sole exception is Average Pedestrian-Level Façade Transparency measured with a graphic typology on a five-point scale.

³Including some segments within a half-mile radius from the station that are measured but not used by the 249 travelers, the final number of measured segments is 270.

⁴Asian remains in the table for a comparison with Model I. To assess the possibility of multicollinearity between Asian (categorical variable) and the four path walkability variables, a collinearity diagnostic test is performed by using Statistical Package for the Social Sciences (SPSS). With the value of variance inflation factor (VIF) at 1.0, none of the path walkability variables has multicollinearity with Asian.