A 2-page questionnaire was mailed to State juvenile institutions in all 50 States and the District of Columbia. The sample consisted of 151 agencies. The 66 responding programs were grouped into 11 types of programs. This survey provides a way to deal with missing information when evaluating alternative juvenile justice policies and programs. This paper describes the use of best-worst-midpoint analysis, threshhold analysis, and convergence analysis. It is the first application of these evaluation methods to a national study of juvenile offender treatment programs. Tabular data and 9 references. (Author abstract modified)