The purpose of this study was to increase yielding on a city wide basis in two cities located in large metropolitan areas, Ann Arbor, MI and Saint Paul, MN. A multifaceted treatment program was used to increase yielding in both cities that consisted of highly visible enforcement, public posting of the percentage of drivers yielding to pedestrians each week along with the record, and low cost engineering treatments. Enforcement and the low cost engineering components were only introduce at treatment sites, while measures were obtained at generalization site to assess whether changes were taking place at untreated crosswalks. The percentages posted on the feedback sign was only based on data collected at the treated sites. The program produced a large change in yielding at the treatment sites and a more modest change at the generalization sites. All data were collected by research assistants who followed a staged crossing protocol. Probe data were collected on naturally crossing pedestrians at all sites. Increases in yielding to naturally occurring pedestrians were larger than those for staged crossings at both the enforcement and generalization sites. This replicates data observed in other studies and appears to be related to naturally crossing pedestrians crossing more aggressively than staged crossings.