Urban form and drivers' behavior: Transportation safety implications of residence in urban, suburban and rural environments.

Abstract

This dissertation research addresses the relation of transportation and land use planning, but differs from most work in this area in that it focuses on the outcome of transportation safety. Two major elements distinguish the research design of this dissertation from that of most other traffic safety studies. First, the dissertation defines a neighborhood's traffic safety performance as the number of crash-involved drivers per 1,000 licensed drivers residing in each neighborhood, regardless of the actual locations of crashes. Therefore, a neighborhood's traffic safety is considered improved only when local drivers face a lower probability of traffic crash involvement. This view transforms the focus of traffic safety from places---e.g., troublesome intersections---to people. Second, the study area in the dissertation covers neighborhoods throughout the entire Detroit metropolitan region, instead of pre-selecting case study neighborhoods. In order to identify the alternative neighborhood types, neighborhoods throughout the region are characterized using an innovative technique for defining neighborhoods, relying on geographic information system-based spatial analysis and statistical clustering. Neighborhoods are grouped into six neighborhood clusters: CBD, traditional urban, inner suburban, outer suburban, rural, and employment center, based on their (dis)similarities of physical environment settings and land use patterns. The findings lead to the conclusion that although higher density neighborhood settings raise local residents' traffic crash risk overall, the risk attributable to certain hazardous actions that are more likely associated with aggressive driving behaviors (e.g., speeding, following too close, disregarding traffic control, or failing to yield the right-of-way) by residents of these areas is lower. This suggests that, on the one hand, promoting more compact development alternatives and a wide range of alternative transportation mode choices would reduce the occurrence of aggressive driving behaviors, and thus result in desirable transportation safety outcomes. On the other hand, to the extent that planning is currently interested in enabling denser land use alternatives such as new urbanism, transit villages, or smart growth generally, greater attention needs to be paid to design for traffic safety. Such attention to design would help mitigate the potentially higher overall traffic risk associated with denser neighborhoods, especially those neighborhoods which are designed solely for high residential density without being accompanied by greater accessibility and diversified land use patterns.