This article evaluates a residential alcohol/drug treatment program for the homeless. The process evaluation documents numerous deviations from the program as designed and other implementation problems. Foremost among these: The project was designed as a randomized experiment, but randomization...