Factors Affecting Travel Demand by Bus: An Empirical Analysis at U.S. Metropolitan Statistical Area Level

Abstract

Understanding the nature of transit travel demand is at the heart of transportation policy making and the success of transit systems. Unfortunately, most existing studies focus narrowly on a single or a few transit systems or metropolitan areas to analyze the determinants of transit travel demand. It is, therefore, difficult to generalize their findings over a broader geographic scale. This paper examines and explores the determinants of transit travel demand by bus generalizable to the entire U.S. by investigating internal and external factors at the U.S. metropolitan statistical area (MSA) level for 2010. The authors collected and used data for all 358 MSAs nationwide. Internal factors are those that the transit managers and operators have some control over (e.g., transit supply, transit coverage), whereas external factors are those that they do not have control over (e.g., population density, vehicles per household). Several variables came directly from the U.S. Census Bureau and National Transit Database, whereas the authors constructed others using data from these two sources. The paper employs a log-log regression model to address the issues of non-normality, non-linearity, and heteroscedasticity. Findings suggest that internal factors tend to be significant predictors of travel demand by bus mode compared with external ones. As such, transit authorities have the ability to attract ridership by adjusting the influential internal factors under their control.