Since the first two charter schools opened in St. Paul, Minnesota in 1992, a reliable study of the effects of charter school attendance on student achievement has been called for. The question in evaluating the effects of charter school attendance is whether charter school students perform better or worse, on average, than similar students in traditional public schools (TPS).
In the literature on effectiveness of school attendance, several outcomes have been suggested as measures of school performance such as achievement test scores, graduation rates, dropout rates, and earnings after graduation. No matter what outcome is measured, there are several logistical and methodological obstacles to overcome. Difficulties arise in evaluating the effectiveness of charter schools because we cannot implement a social experiment by randomization. If we could assign students into two sectors randomly, randomization would be 'balancing out' differences with respect to all observable and unobservable covariates between the treatment group, students who enrolled in charter schools, and the control group, students who enrolled in traditional public schools, so we could attribute difference in means of outcomes solely to the effects of charter school attendance.