After a brief review of recent research concerning school effectiveness, this paper proposes an approach to operationalizing school effectiveness that is based on the achievement of students in a school over a period of time. This three-level model does not follow a single cohort of students as they progress through school. Rather, it uses the achievement results of five consecutive second- and third-grade classes in a given school. The approach uses first stage modeling to obtain school posttest performance indices adjusted for pretest for each of the 5 years. In a second stage, the trend across the 5-year time span is modeled; and in a third stage, the trend parameter estimates are related to school level variables. Each grade level and subject area (reading and mathematics) are considered separately, so the school effectiveness indices (SEIs) do not produce a comprehensive index of school effectiveness. Matched student-level data files for grades two and three in South Carolina for each of the 5 school years in the span from 1984-85 to 1988-89 were used. The complete second-grade data set contained over 194,000 students whose means on the Basic Skills Assessment Program Reading Test (BSAPR) and on the Basic Skills Assessment Program Mathematics Test (BSAPM) were 832 and 810, respectively. The complete third-grade data set contained over 190,000 students whose means on the BSAPR and the BSAPM were 800 and 806, respectively. Data for 148 second-grade and 146 third-grade samples of schools are provided. Results are discussed in terms of the effects of centering, low reliabilities of slopes, and correlates of status and slope. Five data tables are included. (RLC)