Isn't that likely driven by a selection bias? "Everyone in school A takes the test in-school" vs "School B does not offer the test, so a subset of school B's students take the test on a weekend."

From that alone, I'd expect school B's test results to substantially outperform school A's results.