Bad Measure

There's a lot of resistance to measurement. Legitimate resistance, because it's so often done badly, or in bad faith. The census is one example. It's stupid to go out and try to count all the noses, one by one. It's not the best way to get an accurate count: any scientist who counts populations of anything knows that.

Or take the testing of kids in public schools. I'm all for measuring how well schools are doing. It's something that's critically important to know. My objection is not to the measurement, but to the clumsy and intrusive measurement being done. There's simply no reason to test everyone repeatedly. It's like the census, only worse: it's not only inefficient, it's also interferes significantly with process we're trying to measure, and it demoralizes both teachers and students. Not only that, but it violates a cardinal rule of measurement: the people doing the measurement should not have a vested interest in how the measurements turn out. Having teachers assess their own efficacy is bad metrics.

This testing collects a huge amount of data, at a heavy cost, that we actually already have. We already know that our kids aren't learning the factual content we'd like them to learn, that they leave high school well behind European and Japanese kids. That's not in doubt. It's worth monitoring, because we hope it will change, but for policy decisions we already know what we need to know: our education system isn't good enough.

And there are other things we want to know. How good are students at problem-solving? How much initiative and entrepreneurial spirit do they have? How good are they at working in teams? How kind and responsible are they? These are things that to my mind are more important than the date of Gettysburg, or how not to dangle a participle, or the chief exports of Korea.

These things are all quite measurable, and we need to know them urgently. Our future depends on our schools getting these things right, or at least not getting them terribly wrong. If we improve the factual content scores by five percent, but do it by turning out kids that are half as kind and responsible, we have not made progress. We've gone backwards. I'm worried by the fact that a high school graduate in Singapore is more likely to know the year of Gettysburg than a high school graduate in America. But I'm far more worried about the fact that so many American kids feel that cheating on tests is okay. If you need to know the date of Gettysburg, you can look it up. But if you need a moral compass, where will you find that? And how will you even know you need it?

My point, again, is that measurement is a good thing. It ought to be in the fundamental tool-kit of every adult. But not dim-witted, test-every-kid, count-every-nose measurement. Thoughtful measurement, measurement that's designed to find out the things you really need to know.