Beware the power of incentives

If you had to choose between (1) teaching two weeks of difficult AP course content or (2) prepping your students for a broader but less rigorous state test in your subject, which would you choose?

Before you decide: Your pay is tied to student performance on the state test, but not the AP test.

One of my friends faced that dilemma this fall, and of course he chose the state test.

It’s kind of sad — his students will be at a disadvantage for the more rigorous AP test — but totally predictable, given the current incentive structure in public education.

At least my friend and his school will look good.

Like teachers who use tests and the threat of failure to “motivate” their students (which I wrote about in early November ), many states use tests and the threat of failure to “motivate” their teachers. Both ignore the unintended consequences.

I understand why states and school districts want to judge teachers by student test performance. (Caveat: My state does not do this.)

Using testing to judge teachers was the natural outgrowth of two decades of research. Stanford economist Eric Hanushek, among others, found that the best teachers could achieve 1.5 years of academic growth with their students, while the worst teachers achieved only 0.5 years of academic growth.

These effects held true at both the best and worst schools, with middle class students as well as students in poverty. Obviously, the “teacher effect” is huge, and schools want their teachers to be good ones.

Other research, which sought to determine which variables produced these super-teachers, was less successful. No one could isolate any one factor – whether it was teacher education, teacher experience, teacher pay or class size – that predictably resulted in better student performance.

So, the thinking was, if we can’t figure out how to alter inputs to make teachers perform better, let’s just judge them on their students’ output. We’ll reward the best teachers and drive out the low performers, and we can solve the problem of “bad teachers” without having to know what makes a teacher “good.”

I could write a book about the unintended consequences of these testing schemes, but for now I just want to consider the incentives and motivation they create for teachers and students.

Decades of research – starting with Edward Thorndike and B.F. Skinner in the 1950s — have demonstrated that when you create a simplistic, behaviorist reward structure, you tend to get the behavior you ask for – and nothing more.

Many critics have already lamented how this single-mindedness has squeezed out art, music, phy ed, public speaking and creative play from public school curricula. You reward schools for teaching reading and arithmetic? They will make sure they teach nothing else.

My friend’s story was a bleak reminder that with the crude, bureaucratic systems we have in place, a rigid testing scheme can also squeeze out rigorous college-level instruction in favor of boring review activities.

Imagine your child in a Calculus class, and suddenly the state decides that her teacher will be paid based on students’ arithmetic skills. We want students to be able to add, subtract, multiply and divide, right? The teacher will put derivatives on hold to make sure everyone can do long division and multiplication by hand.

Or imagine a college-level English class, where a unit on Julius Caesar has to be cut because the state test emphasizes labeling parts of speech and proper comma placement?

The problem here isn’t the principle of tying teacher pay to student performance; the problem is the implementation.

If we pay teachers more for improved state test scores, we probably will get higher state test scores, but the students might not be able to write a coherent paper or perform an experiment.

Similarly, if we paid teachers more for better student evaluations, they would find a way to make students like them better, but the students might have more time on their phones and less time doing actual work.

For that matter, if we paid teachers more for cleaner classrooms, they would probably even wash the desks after school.

Whatever we identify as the “golden ticket” – whether for teachers or their students – we had better be certain that it is the outcome we actually want. Incentives are far more powerful than we realize, so we should use them with care.

My friend’s dilemma is yet another reminder that too often, thoughtful, rational consideration of our goals is discarded in favor of easily measurable targets. I’m certain no one ever “decided” to incentivize teachers to replace AP instruction with basic test prep, and yet, that is exactly the system they created.

If we really want our teachers to replace boring, uninspiring content with rich, engaging, rigorous content, our task is clear: We must figure out to incentivize that.