An Examination of Student Achievement in Michigan Charter Schools

Abstract

This paper examines the effect of charter schools on student achievement in Michigan using a matched student dataset. Proponents of charter schools argue that by applying market pressure to traditional public schools, having the freedom and incentives to apply innovative curricular and instructional ideas, and offering students a choice in the schools they attend, charter schools can raise student achievement. Studies of the effect of charter schools on student achievement have been mixed, however. Methodologies vary widely depending upon the availability of data. Some studies track the same students as they transfer between charter schools and traditional schools; others rely on cross-sectional student or building-level data. We construct a dataset that matches the scores of the same student taking tests in two consecutive years. Estimating a value-added education production function, we find that charter schools are at a disadvantage to traditional public schools by an average of 0.2 standard deviations. These findings depend upon proper matching of students across school types, which in this case is accomplished by using prior test scores as a control variable and as a way to segment the sample. We also find that charter schools run by for-profit companies have an advantage over those run by not-for-profits and that charter schools improve the longer they are in operation.