By Tom

By now you’ve probably heard that Washington State’s Office of the Superintendent of Public Instruction has agreed to provide the Seattle Times with a large amount of data concerning student test scores, attendance records, discipline records and demographics. They will also provide the newspaper with staff data. To be clear, this is data that is not already available to the public.

It seems clear why the Seattle Times wants this data. They are in the business of selling information. If they can get a better understanding of what’s happening in our state’s schools, they can package that information into news articles and sell it to the public. That seems obvious. What isn’t clear – at least to me – is why OSPI has entered this agreement. I honestly don’t see what they stand to gain by providing The Times with data that isn’t already released.

This bothers me at two levels. As a parent, I’m not crazy about my sons’ scholastic information being sifted through in some newspaper office by a bunch of reporters who are essentially looking for a story. It’s not that our family has anything to hide, but it’s still our information, and it should be our decision on who gets to look at it. The Times has tried to assuage those concerns by maintaining that the information will be “de-identified,” but as KUOW pointed out, it wouldn’t take much effort to use all the data to triangulate which student earned which test score or which student was suspended for which offence. That bothers me.

I’m also bothered as a teacher. It seems clear to me that The Times will have the capacity to report test scores aggregated at the individual classroom level. If you’ll remember, The Los Angeles Times did this a few years ago, and the fallout was disasterous. Whether the Seattle Times is planning this or not, we’re not sure; but it sure looks like they’ll be able to.

The problem is that when a specific teacher’s student test scores are published, they’re devoid of context, because that context would breach confidentiality. Here’s an example: two boys in our school recently lost their stepfather. He was killed violently while in the process of committing a felony. This had an adverse effect, not only on those two boys, but their teachers and their classmates. And when I say “adverse effect” I mean an effect that will probably show up in student test score data. Statisticians call this “noise,” which refers to random happenstances that push or pull data either up or down. They call it noise, because when you aggregate data, positive and negative noise tends to balance out, and the aggregated data isn’t affected. The negative effect of a homicide, for example, could be balanced by the positive effect of another family in the same school whose father got a huge promotion and raise.

But that doesn’t work so well when you drill down to the classroom level, where what remains of our privacy protections prohibit us from providing the context of our students’ test scores. Consider my classroom. I voluntarily took all eight IEP kids in our school’s fourth grade. This was a decision that worked well at the school level; by placing all eight of those kiddos in one class, I could more easily collaborate with the reading resource teacher. Instead of pulling out two or three kids from three different classrooms, each of which is at a different place in the curriculum, she can pull her whole group out from one classroom and focus on the specific skill that they’ve been working on. It works great at the school level, but it’s not going to look so great (for me) if and when my students’ test scores are published in The Times.

It’s great that the Seattle Times is taking such a keen interest in education. They don’t always get it right, but sometimes they do. And obviously, the more information they have to work with, the better. But it seems to me like this agreement gives them access to more information than they can be trusted with.

And that bothers me.