Pink Post-It Notes and Education Reform

In an editorial on November 7, the Salt Lake Tribune suggests using students'
evaluations of teachers to determine, in part, teacher pay. The
reasoning goes like this: We're all former students. And most of us
can recall a handful of teachers who really impacted our lives. So
students can identify excellent teachers, and this means student
evaluations will be a good measure of teacher performance.

This reasoning is not quite right, and I'll try to explain why in a
minute. But first I'll explain why a business professor, of all
people, might know something about this. It's because this problem
--- how to reward good job performance by employees --- is faced by
pretty much all organizations. Economists study financial incentives,
and so business economists have put a lot of thought into whether ---
and if so, when --- financial incentives work inside organizations.
I've been trying to help MBA students here and elsewhere think through
these issues for nearly 15 years.

The Tribune assumes that if a measure is correlated with good job
performance, then it's a good performance measure. This is not
necessarily true, for the following reason: Once a performance measure
is used to set pay, employees start to think about ALL of the possible
actions that might make the measure go up. And in some settings,
employees can do lots of things to make measured performance improve,
but that aren't actually associated with good work.

Here's an example from the private sector: 3M Corporation, inventor of
the ubiquitous Post-It Notes, used to require that each of its
operating divisions earn 25% of its sales from "new" products, i.e,
products that had been introduced in the past four years. The firm
felt that innovation was a key driver of success, so this measure was
in place to drive employees to innovate. The result? Employees
innovated. But employees also began to think about all the possible
ways to achieve the 25% target. One division found that changing the
color of an existing product was enough to qualify as a new product:
thus, Pink Post-It Notes were born. When a new CEO took the reins of
3M in 2001, he dumped the 25% rule, citing high costs of employees'
attempts to boost the measure without actually innovating.

Of course, 3M's goal is to generate profits, while our education
system's goal is to generate, well, "education." Does the same
reasoning apply?

Let's ask: Are there actions a teacher can take that might boost
student evaluations, but that we wouldn't call "good teaching"? How
about offering an easy class? High grades? Light homework? Lax
classroom discipline? Movies in place of boring readings? Pizza
every Friday?

The fact is that learning new things can be hard. And while there are
some teachers who can inspire us at the same time they drive us to new
achievements, these are likely the exceptions rather than the rule.
Unless we're very certain that students know what's best for
themselves, we may not want to pay teachers based on how well they
cater to student desires.

More broadly, performance measurement issues are central to education
reform. Measuring the job performance of educators is hard, and as a
result the financial incentives for outstanding job performance have
historically been weak. But tying teacher pay to poor measures of job
performance --- ones that reward "good teaching" but also actions that
we wouldn't call "good teaching" --- might be even worse.

I agree with the Tribune that rewarding good teaching is a worthy
social aim. But any plan to do so must carefully weigh the benefits
and costs of the selected measures of teacher performance.