DRAFT Technical Background Document on the Impacts of Residential Density Based on a Review of the Empirical Literature

There have been scores of studies of land use and travel behavior over the past two decades. Extensive reviews are in Badoe and Miller (2000), Boarnet and Crane (2001, chapter 3), Brownstone (2008), Crane (2000), Ewing and Cervero (2001), and Handy (2005), and NRC (2009, chapter 3). There is active debate about appropriate methodology, and only a handful of studies use methodologies that can withstand various criticisms that have been posed in the literature. Beyond that, there is also active debate about how to choose studies. The results, in terms of the elasticity of vehicle miles traveled (VMT) with respect to density, are quite stable over a range of study types and methods. That does not imply that any methodology is appropriate, as key issues such as causality and residential selection (described below) are better handled with individual data. Some studies, mostly from the 1990s, use data aggregated to geographic observations, such as census tracts or transportation analysis zones. In those studies, the unit of observation is the geographic area, not an individual traveler. This makes it difficult to link those results to behavioral theories of travel. A more recent generation of studies has focused on individual travel diary data, allowing more clear links to behavior and, depending on the study, better inferences about causality.