The use of statewide tests of student achievement as one component of accountability are certainly not new. An increasing number of states have mandated statewide testing through legislation aimed at tying financial incentives to a variety of accountability indicators including student achievement. These initiatives have generated several side effects, both positive and negative: (1) there has been a renewed interest in research on factors that influence student achievement; (2) the general public, ever-wary of tax increases, has been given a concrete measurement (however controversial) by which to gauge student success; and (3) teachers, administrators, and other professional educators have become increasingly aware of the public policy implications that quantitative data can have on schools, personnel, and school programs. As the result of a school funding equity lawsuit in the state of Tennessee, new legislation mandating revised school funding formulae and accountability procedures was implemented in 1991 for all K-12 public schools. Part of the accountability procedure includes mandated annual testing of all students in grades 2 through 8 in the areas of science, math, language arts, reading, and social studies. The goal of the Tennessee Education Improvement Act (EIA) of 1991 is to reduce variability among scores in school systems across the state regardless of socioeconomic status (equity), and to ensure that all students are progressing (or value-added) from one year to the next in each of the key subject areas. These goals are reflective of the national trend toward increased accountability in education. This research, focusing solely on the area of science, addresses the following questions: (1) is there evidence of more equity and value-added in student scores?; (2) was variability in scores decreasing?; (3) how do scores compare across years and grade levels?; and (4) what are the implications for curriculum and assessment reforms? The data set for this study consisted of scale science scores in 133 Tennessee public schools, grades 2-8, for the years 1990-1994. The null hypothesis of the investigation was that there is no difference in science scale scores across years or grade levels. (Contains 11 references.) (Author/DKM)