A two-year evaluation of William Glasser's Schools Without Failure (SWF) program was carried out in the New Castle School District in Pennsylvania. Ten elementary schools were paired on the basis of size, socioeconomic status, and past achievement of pupils. In the first year, one school of each pair was randomly assigned to begin teacher training and implementation of the SWF program; in the second year of the study, both groups of schools received training in SWF methods and implemented the program. Data were collected and analyzed to determine whether the second year or the first year produced stronger changes and how schools that received two years of training differed from traditional schools. Tests and observations were conducted at the beginning and end of the first year and at the end of the second year. Measurements were taken of pupil achievement and attitudes, teacher and parent attitudes, disciplinary referrals to principals, and classroom interactions. (Author/JG)