This paper reviews the literature on the relationship between the built environment and roadway safety, with a focus on studies that analyse small geographical units, such as census tracts or travel analysis zones. The authors review different types of built environment measures to analyse if there are consistent relationships between such measures and crash frequency, finding that for many built environment variables there are mixed or contradictory correlations. They turn to the treatment of exposure, because built environment measures are often used, either explicitly or implicitly, as measures of exposure. They find that because exposure is often not adequately controlled for, correlations between built environment features and crash rates could be due to either higher levels of exposure or higher rates of crash risk per unit of exposure. Then, they identify various built environment variables as either more related to exposure, more related to risk, or ambiguous, and recommend further targeted research on those variables whose relationship is currently ambiguous. (Author/publisher)
Abstract