The Unappreciated Success Of Charter Schools

By Adam Ozimek

I think the conventional wisdom on charter school evidence could be summed up thusly:  ”some charter schools appear to do very well, but on average charters do no better and no worse than public schools”. But I would like to propose a better conventional wisdom: “some charter schools appear to do very well, and on average charters do better at educating poor students and black students”. If the same evidence existed for some policy other than charter schools, I believe this would be the conventional wisdom.

Two of the most widely cited charter studies are a 2009 and 2013 analysis of charters in 16 and 27 states respectively by the Center for Research on Education Outcomes (CREDO). The results that many cite are the charter schools do no better or worse than nearby public schools on average, which was the conclusion of the 2009 study. However, I think this claim really missed the bigger picture. While overall charters and public schools compare relatively closely, both the 2009 and 2013 study found that charters did better for students in poverty. In addition, performance gap is growing over time:

Charter school impacts with students in poverty and English language learners were positive in 2009 in both reading and math. These positive results have sustained and in fact increased in 2013.

And the results are especially strong for black students in poverty. As the CREDO study reports:

“Black students in poverty who attend charter schools gain an additional 29 days of learning in reading and 36 days in math per year over their [traditional public school] counterparts (see Figure 30). This shows the impact of charter schooling is especially beneficial for black students who in poverty.”

You see this result repeated on other studies as well. Using randomized study results from charter school lotteries in Massachusetts, Angrist, Pathak, and Walters find that non-urban charters don’t outperform public schools and may even do worse, but urban charter schools benefit black students and poor students:

Black and Hispanic students benefit considerably from urban charter attendance in middle school, but the estimated math gains for whites are smaller, with no increase in whites’ ELA scores. Urban charter middle schools appear to produce especially large achievement gains for students eligible for a subsidized lunch and for those with low baseline scores.Attendance at urban charter high schools increases math scores in every group and raises reading scores for everyone except whites, though estimates for small groups are imprecise.

It’s hard to imagine it another policy being called a failure because it only benefitted poor students and black students but the overall scores were held down by non-urban schools and white students.

The improvement of the charter sector over time is not surprising and has been documented elsewhere. A recent NBER paper found this was true for charter schools in Texas. The abstract reports:

We study quality changes among Texas charter schools between 2001 and 2011. Our results suggest that the charter sector was initially characterized by schools whose quality was highly variable and, on average, less effective than traditional public schools. However, exits from the sector, improvement of existing charter schools, and positive selection of charter management organizations that open additional schools raised average charter school effectiveness over time relative to traditional public schools.

As a result, I think charter critics who draw on empirical research that compares outcomes are fighting a losing battle. The charter sector is outperforming public schools by some measures already, but more importantly they are getting better over time. I have little doubt that the next CREDO study will show charters making even more gains. Critics determined to oppose charters should start to pivot now, because they are standing on a leg that will give out eventually.

The charter sectors’ ability to do better for poor students and black students is important given that they disproportionately serve them. I remember when I was an undergrad in the early 2000s, the debates on charter schools were far more theoretical than they are now. Back then I frequently heard the concern that charter schools were just going to engage in “cream skimming”, be a way for middle class white families to escape urban school systems, and thus serve as one more form of segregation in this country. This concern has not come true, and currently 53% of charter students are in poverty compared 48% for public schools. Charters also serve more minority students than public schools: charters are 29% black, while public schools are 16%. So not only do they serve more poor students and black students, but for this group they relatively consistently outperform public schools.

What’s odd is how often these facts go ignored. If the opposite were true, and charters served less minority or low-income students than public schools then it this would be trumpeted constantly and presented as perhaps the most important evidence in this debate. Or if charters showed strong positive results overall but didn’t benefit poor students or black students they would be condemned as institutions that further inequality. I’m not accusing anyone of conscious bias here, but I think if the empirical research on any other policy showed similar results that charters do for poor students and black students it would be far more widely embraced, and the average effects would be downplayed as less important.

ADDENDUM: Via Twitter TWTR -0.25%, Matt Barnum reminds me of a 15-state randomized study from Mathematica:

This paper presents findings from the first national randomized study of the impacts of charter schools on student achievement, which included 36 charter middle schools across 15 states. The paper compares students who applied and were admitted to these schools through randomized admissions lotteries with students who applied and were not admitted. It finds that, on average, charter middle schools in the study were neither more nor less successful than traditional public schools in improving student achievement. However, impacts varied significantly across schools and students, with positive impacts for more disadvantaged schools and students and negative impacts for the more advantaged.