This study investigates the influence of urban form on automobile travel using travel-diary data from Germany. Two dimensions of car use are considered: the discrete decision to own a car and the continuous decision of distance traveled. Because these decisions are likely to be influenced by factors unobservable to the researcher, we apply censored regression models to evaluate the role of biases emerging from sample selectivity. Unlike much of the literature, we find that urban form variables are a significant determinant of both automobile ownership and use, a finding that holds even after using instrumental variables to control for endogeneity.