This report results from an investigation of the extent to which differential attrition form the research sample in the Channeling Demonstration might have led to biased estimates of program impact.