Panel Paper: Estimating the Effects of a Large Network of For-Profit Charter Schools

Friday, November 3, 2017
Gold Coast (Hyatt Regency Chicago)

*Names in bold indicate Presenter

Silvia Ceballos Robles, Susan Dynarski, Daniel Hubbard and Brian Jacob, University of Michigan


School choice is one of the signature education policies of the Trump administration, which adds increasing urgency to the question of whether charter schools are effective, as well as for whom, and why. Moreover, the president has recently appointed a prominent leader of the charter school movement in Michigan as the Secretary of Education, lending special interest to the Michigan context. Additional factors make Michigan an important case study, including the prominence of the for-profit charter sector, geographic variation in charter school location, and a lax regulatory environment relative to states where quasi-experimental evidence exists on charter school efficacy.

This study examines the impacts of attending schools in a large network of K-8 charter schools in Michigan called America’s Promise Academies (APA) using lotteries between 2003 and 2013. The APA is one of the largest for-profit education management organizations in the country, operating over 80 schools nationally, and 49 schools in Michigan alone. While non-experimental evidence on for-profit networks exists, including estimates of the impact of the APA network, this study presents the first lottery-based estimates on the impact of attending a for-profit charter school. Furthermore, it increases the number of charter schools for which lottery-based impact estimates exist by 39%.

Using APA lottery data matched to state administrative records, intent to treat (ITT) estimates show that lottery winners score about 0.1 standard deviations higher on state math exams by eighth grade. On reading exams, the ITT estimates in eighth grade are similar at about 0.08 standard deviations. On both exams, lottery winners are two to three percentage points more likely to exceed proficiency thresholds. In contrast to most of the prior lottery-based charter impact literature, which has consistently found that low-income, underrepresented minorities in urban areas benefit most from attending the lottery schools, the largest impacts of attending the APA charter network are concentrated among less economically disadvantaged, female students, often in non-urban schools. The findings further highlight the importance of exploring the full range of possibilities in the large, diverse charter school sector.