In 2009, CREDO released an expansive study of charters that, let us just say, made some waves. It showed that among the 16 states studied, there was wide variation in charter quality, and that while lots of charters were doing well, lots were doing worse than local district schools. Ever since, charter antagonists have gleefully used this report to make all types of unflattering claims about chartering.

They, I suspect, are less buoyant today.

Four years after the first report’s release, CREDO is out with an update. It includes all of the previous participating states and a slate of new ones. In total, the states covered by the new report educate over 95 percent of charter students nationwide.

There are lots of articles out  about the study. But many miss some of the most important findings—both in terms of the sector’s basic descriptive statistics and the quality of its schools.

I encourage you to read the full report, not just the executive summary, because some of the most interesting and important findings are lost in the shortening process.

Here are my big takeaways in no particular order:

• The study expanded from 16 states (in 2009) to 27 in 2013; now 1.5 million charter kids are included.
• Contrary to public perception, only 56 percent of charter kids live in urban areas; the rest are in suburbia, rural areas, and small towns.
• More than half of the charter kids studied live in poverty—higher than the traditional public school rate. A growing number of Hispanic and low-income families are choosing charters.
• Among the 16 initially participating states, charters have improved significantly in both reading and math, with charters now outperforming TPS peers in reading (about 8 days more of learning per year).
• Figure 22 nicely shows how the charter sector has improved in both subjects over time.
• The charter sectors in most states improved during these four years.
• The performance of the newest charters lags behind that of longer-standing charters.
• Looking at all 27 states, the average charter student starts further behind academically than TPS peers.
• Importantly, charters are producing large academic gains for the most historically disadvantaged students: e.g., Hispanic and African American students living in poverty and ELL kids.
• Enrollment figures among these groups is growing; in other words, the groups of kids most benefitting from charters are increasingly finding their ways into charters.
• However, there continues to be wide variation in performance within the national charter sector. Some charters outperform peer district schools, some do about the same, and some underperform. But these proportions have been trending, since 2009, in the right direction for charters.
• One continuing disappointment: CMO-run schools aren’t doing as well as they should. We’re all aware of the great CMOs, but too many states have allowed middling or poor charters to replicate, causing overall CMO performance to lag.
• Maybe most notable is the enormous variation in performance of the charter sectors of various states. Some state charter sectors are helping their kids learn dozens of extra days worth of material annually. The best results are in D.C., LA, MA, MI, NJ, NY, RI, and TN.
• The performance of Nevada’s charters is galling.
• I suspect it’s no coincidence that the states with great charter performance have independent authorizers (no more district authorizing!).
• The report compares the additional learning kids get from their states’ charters to their respective states’ NAEP scores. Two major findings jump out: In MA and NJ—states with high statewide NAEP scores—the charter sectors are still remarkably outpacing their TPS counterparts.
• In D.C., where NAEP scores are staggering low, the charter sector is offering kids remarkably higher-performing options.
• Per Figure 42, the longer kids stay in charters, the more they learn compared to TPS peers.
• Lastly, the report offers five scenarios for closing different sets of low-performing charters. The upshot is clear: Get serious about shuttering failing charters, and the sector’s overall performance skyrockets. (It would’ve been nice had they shown the sector-wide positive influence of various scenarios for growing the best charters.)

All in all, the results are encouraging for charter supporters. It appears that the systemic elements of chartering are working as many predicted. In other words, chartering is a continuous improvement process for a system of schools: When you build a strategy around closing bad schools, enabling great ones to grow and enabling promising new schools to start, you shift the quality distribution to the right year after year.

This of course requires both the right policies and smart practices, especially among authorizers. But if we keep about this with fidelity, we should see the charter sector continually improving. That means more great seats for kids in need.

—Andy Smarick

This post originally appeared on the Fordham Institute”s Flypaper blog.