A bad metric can indeed be worse than nothing, because people optimize for the metric, and if it's a poor proxy for the real thing you want, you can actually ge

A bad metric can indeed be worse than nothing, because people optimize for the metric, and if it's a poor proxy for the real thing you want, you can actually get less of it. Imagine what would happen if you evaluated software engineers by lines of code written, and you get some idea of the potential downside. Other potential ways to assess teacher effectiveness: - auditing classes & lesson plans - feedback from peers, parents, students and principals