Over 400,000 people visited Farnam Street last month to learn how to make better decisions, create new ideas, and avoid stupid errors. With more than 100,000 subscribers to our popular weekly digest, we've become an online intellectual hub. To learn more about what we do, start here.

Incentives and pay for teachers

A study by Richard Murnane at Harvard University found that, from 1917 to 1980, merit pay had been tried in more than 7,000 U.S. school districts, but the only plans that survived were those in which pay was token.

Rewarding teachers with higher pay when their students score higher on tests represents a seductive thought in our pay-for-performance culture. However, paying a teacher for performance assumes, among other things, that at some level, the teacher is directly responsible for student test scores and that by only trying harder their students will achieve higher test scores. I have several friends who teach and they tell me that the biggest difference between students who excel and those who don't boils down to parent interest and involvement.

Framed differently, merit pay for teachers implies that a suburban teacher in Orange County will be directly responsible for students who perform better than students of a similar inner city school who can barely speak English. Give me that job in a merit pay scenario. Teachers, in this situation, are given the incentives such as (1) only work at the best schools where they can easily perform above average with no extra effort; (2) avoid low income neighborhoods where parents are likely to be less involved; and (3) help students pass a test rather than learn the material — Freakonomics provides the example of teachers recording books on tape for students who can't read which never really solves the underlying problem. These side effects of merit pay are all undesirable from a societal point of view.

Merit pay is hard to get right in business and perhaps even harder in teaching.

An article in the Washington Post over the weekend argues that merit pay doesn't work. A response to this article argues that it does. I argue that you need to think through the seductive argument and look at behavior you are encouraging.

Gladwell talks about teacher quality in this video. As Gladwell points out that teacher credentials mean zip when it comes to student performance.