The self-selection problem in program evaluation is explored, and two ways of addressing it, covariate control approaches and two-stage sample selection models, are described. A longitudinal study of the effects of participation in Alcoholics Anonymous illustrates both approaches and shows the advantages of the two-step procedure. (SLD)

An original study with a 96 percent questionnaire return rate and four replications with high return rates were compared in terms of populations sampled, implementation, and results. Future use of this procedure, which includes advance mailing, telephone contact, telephone explanations, and followups, is discussed. (SLD)

Dissemination approaches of the U.S. Department of Education are explored in the context of federal efforts to recognize the best, considering relatively systematic approaches to recognition and the five vulnerable evaluation areas of (1) undefined universes; (2) noncomparable goals; (3) identifying reproducible conditions; (4) saturation; and (5)…