������Ƶ

 from Stanford’s Center for Research on Educational Outcomes is—like CREDO’s last major report, in 2009—inspiring a host of talking points. With 95 pages of findings to sift through, there is something for charter school friends and foes alike—which is why it is a mistake to use this, or any CREDO study, as an empirical justification for or argument against the charter school sector in general.

The report indicates that charter schools serve roughly double the percentage of African-American students, a higher percentage of Hispanic students, and a higher percentage of students in poverty than traditional public schools. But CREDO also found that charters serve fewer English language learners and special education students than both traditional public schools as a whole and charter school “feeders”—demographically similar schools that students may attend before enrolling in charters.

The study found that charter school students are, on average, making bigger reading gains than their traditional public school peers, amounting to about eight days of additional learning time. In math, charter school students are making learning gains equivalent to their traditional school peers. In both areas, the charter school sector has improved its performance since 2009.

CREDO’s school-level achievement numbers are particularly amenable to ideological massaging. On one hand, they clearly show that a majority of charter schools perform the same or worsethan traditional public schools in math (40 percent the same + 31 percent worse) and reading (56 percent the same + 19 percent worse). Hence the  about the study: “Charters not outperforming traditional public schools, report says.” It’s settled: charter schools are a disaster!

But wait! The data ����������clearly show that a majority of charter schools perform the same or better than traditional public schools in math (40 percent same + 29 percent better) and reading (56 percent same + 25 percent better). Hence , which references “general gains” at charters. It’s settled: charter schools are a triumph!

What’s that old line about lies, damned lies, and statistics?

The real lesson from the new CREDO study is less dramatic, and thus gets less attention. The national aggregate data on charters mask wide variance in school quality; charters in Washington, D.C. are good and getting better, for example, while Nevada charters are weak. In other words, “charter schools” alone aren’t the solution to our educational ills, though high-quality charters can make a big difference in students’ lives.

The CREDO study shows that the rules states set for their charter school sectors affect the quality of their schools. Indeed, the paper states, “The rise in average student growth across the continuing schools is due in no small part to the closure of low-performing schools, which amounted to about 8 percent of the 2009 sample of schools.” This ought to be intuitive. Charter schools work better in states (and the District of Columbia) with laws that maintain strict school accountability standards. The study found that “the use of the option to close schools represents the strongest available tool to improve overall sector quality for the time being.” (For a more comprehensive look at the promise and challenges of closing charter schools, see .)

To repeat the analogy I offered on Twitter yesterday: “Charter schools don’t solve U.S. education’s quality problem any more than oatmeal solves the problem of eating a quality breakfast.” Like oatmeal, charter schools are just one, potentially beneficial option—but by no means a guarantee of educational quality.