Paul E. Peterson
Kennedy School of Government, Harvard University
April 2007
Soon after the RAND Corporation released its mixed review of Philadelphia's school management experiment, the Wall Street Journal ran an op-ed by Paul Peterson that pointed out serious flaws in the study. Not satisfied with an abridged column's worth of criticism, Peterson has now released an in-depth review of RAND's study, along with his own opinion of Philly's private education providers (such as Edison Schools). Peterson has two main gripes with RAND's methodology. First, for many of the students it tracked, RAND compared only two sets of test scores, not the three required for true "quasi-experimental" studies (a quality standard in the field of statistics). Second, RAND compared scores from three different tests--the Pennsylvania System of School Assessment (PSSA), the Stanford-9, and the Terra Nova--side by side, assuming that "the standard score obtained... on any one of the three tests was the same as the students would have received had they taken the other one." Such an assumption is flimsy at best, says Peterson. In his own evaluation (for which he didn't have access to the student-level data that RAND did), Peterson measured the growth of PSSA scores for two cohorts of students over three grades. His results contradict the RAND study: In math, privately-managed schools made the greatest gains in getting students to (or above) proficiency in both cohorts. In reading, they did so in Cohort 2 and slightly trailed the district's restructured public schools in Cohort 1. Peterson recognizes that his own methodology isn't perfect either. But you'll want to check out his study here.
