Abstract:A statistical analysis of commuters' route choice is presented. A binary logit model with normal mixing distribution using stated preference repeated-measurement data is estimated. General descriptive statistics are initially introduced in the paper to explore various route choice criteria and provide the basis for model estimation. The analysis is based on mail-out/mail-back surveys that were customized using routes generated by a geographic information system and information gathered from two previous route choice surveys. The results indicate the significance of travel time, travel time reliability, traffic safety, and roadway characteristics on route choice. The estimation results also underscore the influence of traffic information on route choice.