M.B. Rodriguez-Gonzalez, J. Aguero-Valverde

Pages: 63-74

Abstract
The distance pedestrians are willing to walk from their homes to the bus stop is of great importance for planning, modeling, design and evaluation of public transport systems. The distributional assumptions used for modeling walking distances are relevant because the empirical distributions have characteristics that are not well represented by standard distributions. There is also the issue of perceived versus the actual distance pedestrians walk to the bus stops since user perception of transit characteristics such as distance to the bus stops is important for patronage. The purpose of this paper is to propose an easy and reliable way to estimate the walking distances to the bus stops. It also explores which statistical distribution better fits the empirical distribution of walking distance and compares the real, perceived and stated preference distances. The results show that a small sample size (about 300) is enough to estimate the average distance walked to bus stops within 25 m with a 95% confidence. The real distances users walk to bus stops in San José Costa Rica are approximately 255 m, 525 m and 740 m for the 50th, 85th , and 95th percentiles respectively, based on the estimates of the gamma distribution. The users of the transit system in San José perceived a distance that was in general higher than the real distance they walked, also they declared to be willing to walk further distances than those they are walking presently.
Keywords: gamma distribution; bus stop location; service coverage; walking distance; user survey