Charter School Demand and Effectiveness, A Boston Update

Discussion Paper

Joshua Angrist, Sarah Cohodes, Parag Pathak, Elizabeth Setren, Christopher Walters

October 2013

As in the 2009 report, Informing the Debate, we find that attending a charter school in Boston significantly boosts MCAS scores and proficiency levels. Positive test score effects from the most recent years where our lottery sample coverage is nearly complete are  of similar magnitudes. Non-lottery results confirm the lottery results for charters from which we were able to collect lottery records, and point to lower performance for closed  charters and those without complete records. However, test scores are only one part of the story. This report also provides evidence on the demand for charter schools.

Many students in Boston apply to a charter, with application rates rising in the past few years,  especially for middle schools. A majority of students who apply get an offer to at least one  school, but not all students accept these offers. A third of middle school students and 60 percent  of high school students choose other options. Many of these offers arrive after the lottery, a  contributing factor to low take up rates, along with the many school options available in Boston, especially for high school. Offer rates at Boston charters are broadly similar to the offer rates for first choice schools in the BPS assignment mechanism.

Charter school students tend to have somewhat higher early test scores than the general BPS population. This most reflects that higher  scoring students are more likely to apply in the first  place. The proportion of students with special needs and English language learners is also lower in  the applicant group than in the general population. Importantly, however, gaps between charter applicants and non-applicants are shrinking. In the most recent year, we see almost as many special education students applying as exist in the BPS population. At the same time, some gaps remain. This is important because our analysis of charter effectiveness (here, as in earlier work) uncovers substantial differences in impact. Students from groups least likely to apply, including English language learners and students with low achievement scores, are those for which achievement gains are likely to be the largest.