Bill Gates spent $45 million trying to find out what makes a school teacher effective. I’ve studied his Measures of Effective Teaching(MET) project, and think it ignores a matter of fundamental importance.

Consider: What makes an effective lawyer, carpenter, baseball player, surgeon?

The answer is that it depends—depends on what they’re being asked to do. An effective divorce lawyer isn’t necessarily an effective criminal defense lawyer. A good framing carpenter isn’t necessarily a good finish carpenter. A good baseball catcher isn’t necessarily a good third baseman. A good heart surgeon isn’t necessarily a good hip-replacement surgeon.

Put lawyers, carpenters, baseball players, and surgeons in wrong roles, test them, and a likely conclusion will be that they’re not particularly effective. So it is with teachers. Put them in wrong roles, and they probably won’t be particularly effective.

Gates’ faith in test scores as indicators of effectiveness makes it clear that he buys the conventional wisdom that the teacher’s role is to “deliver information.” But what if the conventional wisdom is wrong?