Alex Cohen is the Director of Learning and Evaluation for the Richard M. Fairbanks Foundation.

Stanford’s Center for Research on Education Outcomes (CREDO) recently released its latest study on academic performance in Indianapolis schools. The study aims to answer an important question: How do charter schools and Innovation Network Schools (which are part of the school district but have full autonomy when it comes to curriculum, hiring, etc.) compare to traditional public schools in their effectiveness at raising student achievement?

の results provide strong evidence that Indianapolis charters substantially raise performance—leading to more than 60 additional days of learning in English/language arts and more than 90 additional days of learning in math, relative to traditional public schools, in the 2016-2017 school year. These findings are consistent with positive results for charters from the previous CREDO study in Indianapolis from 2015.

The recent CREDO study also finds that Innovation Network Schools led to higher growth in the 2016-2017 school year, relative to traditional public schools, though this effect is not statistically significant.

To come up with these estimates, CREDO compared test score growth between students in different schools, after controlling for differences in student characteristics. What this means is that CREDO was able to look at, for example, two fifth graders with the same race, gender and income and similar ISTEP+ scores in fourth grade. But one attended a charter school and the other attended a traditional public school. How did these two students fare on the ISTEP+ in fifth grade? By completing this type of comparison with all tested students in grades 3 through 8, CREDO was able to generate estimates for growth across schools and school types.

The CREDO findings provide good news for the children enrolled in charter schools and Innovation Network Schools in Indianapolis and their families, as well as the teachers and other staff who work there. They also raise important questions about the best way to measure school quality.

The CREDO approach isn’t perfect. But more rigorous approaches, like lottery-based randomized controlled trials, may not be broadly applied. Lotteries can only be used with oversubscribed charter schools (i.e., those with more students who want to enroll than there are seats available). Not all charters are oversubscribed, and those that are may not be representative of all charter schools.

Another approach is to compare students who began in traditional public schools but then switched to charter schools, though this has its own issues. For example, because this approach relies only on students who switch from traditional public schools to charters, it misses students who begin at charter schools in elementary school and might benefit differently from charter schools.

Finding the right approach to evaluating schools is critical because parents, policymakers and funders rely on measures of school quality to make big decisions. This is especially true for parents as Indianapolis continues to expand its school choice options. How will we know which schools are good for our kids?

Future work might see how the CREDO approach performs when compared to alternative approaches to measuring school performance in Indianapolis—or how this approach compares to the current A-F system. Until then, the CREDO study provides support for efforts to expand autonomy among Indianapolis schools as a way to improve outcomes for students.