What Can We Learn from Charter School Lotteries?

Abstract

We take a closer look at what we can learn about charter schools by pooling data from lottery-based impact estimates of the effect of charter school attendance at 113 schools. On average, each year enrolled at one of these schools increases math scores by 0.08 standard deviations and English/language arts scores by 0.04 standard deviations. There is wide variation in impact estimates. To glean what drives this variation, we link these effects to school practices, inputs, and characteristics of fallback schools. In line with the earlier literature, we find that schools that adopt an intensive “No Excuses” attitude towards students are correlated with large gains in academic performance, with traditional inputs like class size playing no role in explaining charter school effects. However, we highlight that “No Excuses” schools are also located among the most disadvantaged neighborhoods in the country. After accounting for performance levels at fallback schools, the relationship between the remaining variation in school performance and the entire “No Excuses” package of practices weakens. “No Excuses” schools are effective at raising performance in neighborhoods with very poor performing schools, but the available data have less to say on whether the “No Excuses” approach could help in nonurban settings or whether other practices would similarly raise achievement in areas with low-performing schools. We find that intensive tutoring is the only “No Excuses” characteristic that remains significant (even for nonurban schools) once the performance levels of fallback schools are taken into account.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.