Pete Langr: Teacher incentives might not work as originally planned

Merit pay for teachers sounds like a great idea; like Pavlov’s dog, people respond to incentives. We naturally expect that teachers who are promised a few thousand dollars in bonus money will find ways to improve the not-so-good test scores of many school students.

The concept is so attractive that, for much of the public, the efficacy of merit pay is accepted as a matter of faith. Witness the writing of one merit pay proponent, Kerri Toloczko, who rhetorically asks us “Does merit pay give us brighter children?” and then replies to herself: “What a silly question — of course it does!”

In Toloczko’s happy world, there’s no need for evidence, no need for a study … no possibility that she might be wrong. She tells us what she likes to believe, and that’s that.

For those of us who like a little reality to go with our beliefs, there’s the recent study from researchers at Vanderbilt University. It is being described as the first rigorous look at merit pay for teachers. (Imagine, years of argument about merit pay and not one good study to say whether or not it works.)

The results are sobering: In a study of math teachers, there was no increase in test scores for those students whose teachers received bonuses, as compared to a traditionally paid control group.

This shouldn’t really come as a surprise. Advocates of merit pay tell us that a teacher’s compensation should be more like the private sector, but the reality is that most people in the private sector don’t get the kind of merit pay that’s suggested for teachers. Nurses, snowplow drivers, Walmart clerks, airplane pilots and laboratory workers don’t get bonuses for such things as ensuring that patients get the right medication or that piles of snow don’t get left in the street — or that planes don’t crash.

Nor will I get a bonus if you all love this article.

Americans have taken the merit pay argument on faith for so long that few seem to have examined those who actually work for something resembling merit pay. Salesmen who earn bonuses are ubiquitous, but the cost of that system is often a lack of trust and sometimes unmitigated fraud (as we recently saw with some home mortgage sellers). Business executives may be paid huge bonuses for high profits, even though the result of risky decisions may be bankruptcy or a taxpayer bailout. Doctors may be paid by how many procedures they perform, and the result may be higher medical costs or poorer patient outcomes.

At the very least, playing with high-stakes testing and bonuses is playing with fire, and should be done cautiously.

Even President Obama, an advocate for merit pay, should know this. His secretary of education, Arne Duncan, once fired teachers for cheating on high-stakes tests in Chicago. Estimates were that a minimum of 5 percent of teachers were involved.

Most workers — those nurses, Walmart clerks, snowplow drivers, laboratory workers and pilots — get something far different from bonuses when they do quality work. If they do their jobs reasonably well, their reward may be to not get fired. If they do their jobs exceptionally well, then they are likely to be promoted. If they work extra long hours, they get overtime pay.

Yet these incentives rarely exist in our schools. Some underperforming teachers have little fear of being removed, even though districts such as Duluth lay off high performers every year. Other teachers might be excellent candidates for promotion to mentoring and evaluating colleagues — a job which principals often don’t have time to do well — but a lack of available money makes this rare. And overtime pay doesn’t make sense in a profession in which most teachers can never really do enough (and in which long hours outside of the school day is a fact of life for many).

Should teacher contracts be changed to allow for student growth to be part of teacher evaluations? Certainly. But cautiously and fairly.

Should school districts be allowed to consider performance, rather than just seniority, when forced to lay off teachers? Absolutely.

Should we conduct experiments with different models of teacher compensation, including merit pay? You bet.

Just don’t go thinking we know exactly what works. We don’t.