The goal of this study was to explore the causality between the built environment and three transportation modes (private vehicles, mass transit, and nonmotorized modes such as bicycling and walking) while accounting for sociodemographic characteristics for an understanding of what factors may foster transit and nonmotor use. Data from the 2009 National Household Transportation Survey combined with Los Angeles County, California, land use data were used to estimate a recursive structural equations model for an examination of causality direction and for a quantification of the built environment effect and residential self-selection effect. Results suggest that urban planning strategies that promote high population and employment density, land use mix, high four-way intersection density, and road density as well as good transit service can either significantly encourage transit use, bicycling, and walking or reduce vehicle miles traveled. However, the relatively small effect of the size of the built environment on travel behavior implies that achieving the above planning goals may cause considerable residential relocation, which contributes to the major part of observed travel behavior changes.