Pay for performance

5/2/2011

AMONG the many groups that oppose Ohio's new collective-bargaining law for public employees, none are more adamant than teachers unions. The law would replace salary schedules that reward longevity and advanced degrees with a merit system that supposedly is based on teacher performance.

When Gov. John Kasich argued that teachers should be paid according to how much their students learn, a lot of Ohioans agreed. But it's a lot harder to determine how much children learn, and who is responsible for their progress or its lack.

Taxpayers tend to like the sound of merit pay for teachers. Many believe that teachers have it easy. According to the stereotype, they work nine months a year, usually no more than seven hours a day.

They get long breaks during the year, can retire with a full pension before they turn 60, accumulate sick leave that can be turned into cash, have better benefits than most public-sector workers, are protected by tenure, and get a raise every year just for showing up.

If you make widgets, you'll be blamed if they fail. You'll have to improve your performance or lose your job. But what if you and a dozen other people work on the same widget? What if some widget owners keep them in top shape, while others drop them, leave them out in the rain, or otherwise abuse them? You'd need a way to know who is responsible when the widgets work well or don't.

The collective-bargaining law faces an expected voter referendum in November. If it survives, state officials over the next two years would develop a mechanism for evaluating teachers, while local districts would have to figure out how to test student achievement more broadly than is done now. Some sort of value-added assessment is likely to be part of the mix.

Two years seems like a long time, even when you're talking about something as important as educating children, but this is uncharted territory. Ohio would be the first state to mandate pay-for-performance for teachers. It is more important to get it right than to do it fast.

Whether pay for performance works is in dispute. The Cleveland Plain Dealer reports that studies of pilot programs in New York City in 2007 and Nashville in 2009 concluded that merit pay did not improve student achievement.

Also problematic is a section of the new law that bases teacher evaluations in part on student and parent satisfaction, measured by surveys, questionnaires, or other feedback. Parents and students are not unbiased observers when they grade teachers. Of course, the peer evaluations favored by teachers, as well as commonly employed evaluations by school administrators, also can be used as political or personal tools to reward or punish.

There is also the danger that a truly comprehensive evaluation process would require students to take more annual standardized tests. It might require outside evaluators, adding another layer to an already cumbersome process.

And finally, none of this will measure the effect of changes in external factors on student performance. Demographic analysis won't account for students whose parents divorce, are laid off, or become homeless. Nor will it gauge the positive effect on students whose home lives become more settled.

A system that rewards good teachers, helps other teachers improve, and pushes failing teachers to find different careers is a worthy goal. Done right, it would help more students achieve greater success. Done wrong, though, it could drive good teachers out of Ohio and reduce the quality of education statewide.