Charter schools, first launched in the 1990s, are an important and growing component of the public school system in the United States. As of November 2009, more than 5,000 charter schools served over 1.5 million students—approximately three percent of all public school students—in 40 states and the District of Columbia (Center for Education Reform 2009). Charter schools are intended to play a key role in school improvement under the existing Elementary and Secondary Education Act (No Child Left Behind) as well as the programs established under the American Recovery and Reinvestment Act of 2009. However, there remains considerable debate as to whether, how, and under what circumstances charter schools improve the outcomes of students who attend them. This report summarizes the results of a new study: the Evaluation of Charter School Impacts, a large-scale randomized trial of the effectiveness of charter schools funded by the Institute of Education Sciences and conducted by Mathematica Policy Research and its partners.1
The evaluation, which we conducted in 36 charter middle schools across 15 states, compares outcomes of students who applied and were admitted to these schools through randomized admissions lotteries (lottery winners) with the outcomes of students who also applied to these schools and participated in the lotteries but were not admitted (lottery losers). This analytic approach produces the most reliable impact estimates. But because the study could only include charter middle schools that held lotteries, the results do not necessarily apply to the full set of charter middle schools in the U.S.
Key findings from the evaluation include:
• On average, charter middle schools that hold lotteries are neither more nor less successful than traditional public schools in improving student achievement, behavior, and school progress. Participating schools had no significant impacts on math or reading test scores either a year or two years after students applied, other measures of academic progress (such as attendance or grade promotion), or student conduct within or outside of school. Being admitted to a study charter school did significantly and consistently improve both students’ and parents’ satisfaction with school.
• The impact of charter middle schools on student achievement varies significantly across schools. Across 28 sites (covering 32 schools), the effects on reading scores after two years were estimated to be greater than zero in 11 sites and less than zero in 17 sites (with magnitudes ranging from -0.43 to +0.33 standard deviation units), with 4 of the individual site estimates statistically significant. The estimated effects on math scores were greater than zero in 10 sites and less than zero in 18 of the 28 sites (-0.78 to +0.65 standard deviation units), with 10 of the site estimates statistically significant.
• In our exploratory analysis, for example, we found that study charter schools serving more low income or low achieving students had statistically significant positive effects on math test scores, while charter schools serving more advantaged students—those with higher income and prior achievement—had
1 The evaluation team also included Optimal Solutions Group and Paul Hill of the University of Washington’s Center on Reinventing Public Education.
significant negative effects on math test scores. Charter middle schools in large urban areas also had significant positive impacts on math achievement compared to negative impacts in other locales, although urbanicity was no longer an influential factor once such characteristics as students’ demographics and income levels were controlled for. There were also differential effects on reading achievement, with negative and significant impacts for study charter schools serving more advantaged students and no impacts for study charter schools serving fewer advantaged students.
Some operational features of charter middle schools are associated with more positive (or less negative) impacts on achievement. These features include smaller enrollments and the use of ability grouping in math or English classes. Although impacts differed for study charter schools with longer- versus shorter- hours of operations or higher versus lower revenue per student, these features were no longer significant once other school and student characteristics were controlled for. We found no statistically significant relationships between achievement impacts and the charter schools’ policy environment, including the extent of its decision-making autonomy, the type of authorizer and how the authorizer held the school accountable, and whether it was operated by a private organization.