This project is part of a research agenda to discover ways to plan and implement urban forms that reduce dependence on the single occupancy vehicle (SOV). The purpose of this project was to empirically test the relationship between land use density, mix, jobs-housing balance, and travel behavior at the census tract level for two trip purposes: work and shopping. This project provides input into policies at the national, state, and local level targeted at the reduction of SOV travel and for urban form policies.
This research employed a correlational research design in which urban form (e.g., density) and travel behavior (e.g., mode choice) relationships were analyzed while controlling for non-urban form factors (e.g., demographics). Data for travel behavior variables (model choice, trip generation, trip distance, and travel time) were obtained from the Puget Sound Transportation Panel. Data for the urban form variables (employment density, population density, mix, and jobs-housing balance) were obtained from the U.S. Census Bureau, the Washington State Employment Security Department, and the King County Assessor's Office. The databases developed for this study were composed of these data sources, matched together by one common variable: the census tract. The databases were structured around two separate units of analysis: the trip and the tract. Relationships between urban form and modal choice were analyzed at the tract level, while urban form relationships with trip generation, distance, and travel time were analyzed at the trip level.
Simple statistical analytical methods were used to identify relationships between urban form and travel behavior variables, including T-tests, linear correlation, partial correlation, multiple regression, and cross-tabulation. Findings from the application of these methods indicated that employment density, population density, and land-use mix were negatively correlated with SOV usage and positively correlated with transit usage and walking for both work and shopping trips. Employment density, population density, and land-use mix were negatively correlated with trip distance and positively correlated with trip generation for work trips. Travel time was positively correlated with employment density and negatively correlated with mixing of uses for work trips. The jobs-housing balance was negatively correlated with trip distance and travel time for work trips. Transit, walking, and SOV usage were found to have non-linear relationships with population and employment density for both work and shopping trips. An analysis of density thresholds was conducted to identify levels of population and employment density, where significant decreases in SOV travel and increases in transit and walking occurred.