Jun 12, 2009

Experiments with Merit Salaries Under Attack

by Laura Starita

In these times of near double-digit unemployment levels, it is hard to believe that job recruiters actually compete with one another. And yet employers insist that it is difficult to find the best people, and to do so—and keep them—companies need to have any number of ways to attract them, starting with competitive salaries. So it does seem strange that the public education sector has so staunchly resisted making salaries in any way dependent upon a person’s performance of the job. Indeed, more than 95 percent, by most counts, of public sector teachers are paid a salary based on a combination of factors, including tenure and credentials, that have nothing to do with their performance as it relates to student learning.

There are a number of places that are trying to change that. Geoffrey Canada of the Harlem Children’s Zone promotes merit pay for teachers as a way of attracting high potential talent (and culling out ineffective, and even damaging, teachers); the Milken Foundation has also donated to an initiative in Dallas aimed at experimenting with merit pay for teacher’s programs. A number of schools opening in New York City—including this one profiled in a recent New York Times piece—are designed with the premise that excellent teachers paid well will successfully help kids learn.

It should be noted that there is no conclusive evidence that paying teachers for their performance in the classroom in fact results in better learning for students. These experiments are driven by intuitive sense, and the fact that the non-experimental research available is positive. A useful 2008 paper by economist Michael Podgursky of the University of Missouri, and education specialist Matthew Springer of Vanderbilt, provides an overview of the few efforts that have taken place and the evidence is suggestively positive, though not conclusive. The authors say that empirical results are thin on the ground because too few comprehensive initiatives have taken place, and those efforts that are underway were not designed to be rigorously evaluated (for example, programs such as KIPP do a lot of different things, including pay their teachers well, and they are ‘schools of choice’ meaning parents and students have to both choose and be chosen, which creates a non-random study population).

A few interesting points are worth nothing from this work however: First, the available research shows no relationship between a teacher’s credentials and how well students learn—certified or not, Ivy-league educated or not, it seems teacher training does not directly correlate with teacher skill. This alone suggests that paying teachers based on their educational path doesn’t make any sense. The second useful point is that the available research suggests that performance-based pay structures for teachers would potentially work in two ways to improve learning. First, by paying teachers salaries comparable to what they may earn in the private sector, more talented people may opt for careers in education; second, under-performing teachers will have an incentive to improve their performance (assuming any program provides a means for that improvement), or they will be culled from the ranks, resulting in a higher percentage of the teacher corps being high performing. The suggested improvements based on these concepts alone seem worth exploring.

Hopefully, those explorations will be permitted to happen in significant enough numbers to be meaningful. Unfortunately, school experiments have come under particular attack from teachers unions. The Wall Street Journal reported last week that two school choice initiatives are under siege by public school teachers unions in Milwaukee and New York City—in the case of New York, the union is objecting to the growth in the number of charter schools since Mayor Michael Bloomberg took over the department of education. Charter schools are often at the center of merit pay experiments. Allowing such experiments to take place can only benefit everyone: teachers, unionized or not, will have more information on what really works. If it turns out that merit pay is ineffective, the unions can say “I told you so.“ If effective, though, unions will be able to be more actively involved in forging policy that makes sense regarding wages, benefits and even—shocking—when a teacher should be fired. Permitting in the teacher corps the kind of career dessication enjoyed by French civil servants is a guarantee of stagnation. And that benefits nobody.