Charter school performance is a mixed bag: some charters outdo their neighborhood district schools, others show no difference, and some do worse. A new Mathematica meta-analysis attempts to identify the characteristics common to each of these groups. What, in other words, makes a high-performing charter schools so effective?
As author Phillip Gleason notes, it is difficult to carry out studies of this nature. Much of the data are based on observation, so determining causation is essentially impossible. Observation also takes time and costs money, which usually necessitates small sample sizes. And many of the “practices” being studied are abstract concepts, such as principal quality, that are difficult to measure quantitatively and objectively.
To mitigate these impediments, Gleason compiled seven studies that used different methods—including observational study, survey administration, and lottery-based designs (comparing students who won a spot via charter lotteries to those who did not)—to study charters schools around the country. The sample sizes in each of these studies range from twenty-nine to seventy-six schools.
Three charter characteristics were found to be linked to high student achievement in many studies (therefore showing a ‘strong association,” according to Gleason—a term he never defines quantitatively): longer school days and/or school years; a...