Many studies have employed machine learning approaches to examine the land use-travel relationship. However, they often emphasize nonlinear and threshold associations, lackluster because of the amount of studies. This paper offers three possible ways to frame research questions.