Cindi Ross Scoppe

August 11, 2013

Scoppe: The science, and art, of grading schools

How could one set of student test scores produce two very different sets of school grades?

FOR YEARS, school test-results season has been a confusing time, marked by a combination of useful data about how well our schools are performing, useful data that is grossly misused and useless data.

Last year, Education Superintendent Mick Zais got permission from the federal government to eliminate the useless data — the all-or-nothing scores required by the federal No Child Left Behind law, which label a school a failure if just one of more than a dozen demographic subgroups falls short of that year’s goal, by even a single point. But while he has replaced that with a much better system, the new system isn’t perfect.

And having this better system has injected a new level of confusion: Whereas the old federal grades were so flawed that any reasonable person would dismiss them, the new ones can’t be so easily dismissed. Yet the school grades Dr. Zais released earlier this month still paint a very different picture, a much more negative picture, than the ones the Education Oversight Committee will release this fall — even though they are based on the same student test scores.

There are good and bad aspects of both systems — one designed by the Legislature, approved by the governor and overseen by the Education Oversight Committee and the other designed by Dr. Zais and approved by the federal government.

But you wouldn’t know that by listening to the superintendent and the Oversight Committee, each of which is defending its own system and attacking the other’s, rather than working together as they did from the time South Carolina passed its Education Accountability Act in 1998 until the time Dr. Zais took office in 2011.

What we need is a single grading system, which wasn’t possible before the Obama administration decided to let states apply for waivers from No Child Left Behind. Of course there’s no guarantee that the federal government would let us use the compromise system that might emerge from a cooperative process between Dr. Zais and the committee, but we won’t know until we try.

There are three main differences between the two systems:

•  The Zais/federal system raises the bar each year, so that, for instance, a school that got a “C” this year likely would fail next year if it produced the same results. The state system initially raised the bar each year, but that requirement was stripped out when the Legislature overhauled the law in 2008.

We’re never going to make the progress we need unless we demand an increasingly higher level of performance, but we need to make sure everyone understands that, rather than mistakenly believing that lower school scores mean schools are doing worse. Even more, we need to figure out how fast to increase the requirement. The EOC says Dr. Zais’ plan is unrealistically ambitious; Dr. Zais criticizes the EOC for not raising the bar at all.

•  The Zais system judges schools based not just on overall scores but on scores of demographic subgroups. The state system merely recognizes schools that do a particularly good job of closing the performance gaps between poor and middle-class students, black and white students and other subgroups.

Actually requiring that each subgroup meet expectations is important, because without that, schools can ignore the difficult-to-teach students, knowing that their low scores will be masked by the high scores of easier-to-teach students. Of course this can be taken to the extreme, as No Child Left Behind does, labeling a school a failure if even one demographic subgroup falls short. The key is determining how far is too far, and as with the increasing standards, the EOC says Dr. Zais has gone too far, and he notes the absence of such a requirement in the state system.

•  The Zais system calculates school grades based on the average of student test scores. The state system grades schools based on how many students pass. So if the passing score was 70 percent, in a hypothetical school with four students who scored 70, 70, 70 and 50, the grade under the Zais approach would be 65 (70 plus 70 plus 70 plus 50, all divided by four) — and the school would fail. But three out of four students passed, so using the individual student approach, the grade would be 75 percent, a passing score.

There are legitimate reasons for each approach, and legitimate concerns about each. The state system can encourage schools to focus on helping students who almost passed, while ignoring the kids who fail miserably. The Zais system can encourage schools to focus on improving the grades of kids who already do well, to pad the numbers of the kids at the bottom. I don’t know which approach I would pick if I had to do that right now.

What is not a significant difference between the two systems — though you wouldn’t know that if you just listened to the superintendent — is that Dr. Zais’ system gives the schools a grade of A, B, C, D or F and the state system uses labels them excellent, good, average, below average or at risk.

I prefer the letter grades, and I find substituting “at risk” for failing a little insulting, but really, it doesn’t matter. If this were the biggest thing we had to disagree over, we’d be in great shape.

Unfortunately, it’s not.

Ms. Scoppe can be reached at

Related content



Editor's Choice Videos