I’ve mentioned The American Enterprise Institute’s Rick Hess before and how he has a lot of interesting and, to my mind, good things to say about education reform. In short, if you want to stay up on the current EdReform movement, Hess is a good resource. For instance, he has recently explained how school reformers have been led to “oversell ideas as miracle cures” in the face of critiques from those such as Dianne Ravitch, how “for schools, one size does not fit all” in modern (some would say post-industrial) America and how it’s time to re-think teacher pay.
Regarding this last, he makes several crucial points and distinctions.

Do you think that employees who are good at their work ought to be rewarded, recognized, and have the chance to step up into new opportunities and responsibilities? I do. If you’re with me on this, you embrace the principle of merit pay—whether you know it or not….First, endorsing this principle doesn’t mean signing on to the raft of slack-jawed merit-pay proposals that would-be reformers have championed in recent years. Merit pay is only useful if it’s done smart, which entails using it to help attract, retain, and make full use of talented educators.
Second, understand that there’s no proof that rewarding talented, hardworking folks “works.” You can comb through decades of economics journals and issues of the Harvard Business Review without finding any proof that paying and promoting good employees yields good results. The premise just seems like a reasonable assumption; you either buy it, or you don’t.

To be sure, Hess thinks there is value in using student test scores for evaluating teachers, but it shouldn’t be a component comprising over 50% of the total “score.” He elaborates further:

Merit pay should reward performance, value, and productivity. We can measure these in many ways—by scarcity of individuals in the labor market, annual evaluation by peers, professional observations, supervisor judgment, and so forth. The contemporary obsession with student test scores as the only metric of interest has been an unfortunate distraction.
Student achievement must be an important factor, but we should employ it deliberately, with an eye to a teacher’s actual instructional duties and responsibilities. Too often, we rely on test scores simply because we don’t have anything else. That’s not a problem specific to merit pay; that’s our peculiar failure to import widely employed practices and tools from other professions.

Second, it’s a mistake to imagine there’s one universal way to design pay systems. Why debate about whether Google, the Red Cross, or Microsoft has the “right” compensation model? There are a slew of reasonable approaches, depending on organizational context and needs. Rather than searching for proven pay models, education leaders would be better off identifying the problems they’re trying to address and asking how reconfiguring pay might help them solve those problems.
Third, the aim must be to craft systems that can evolve. The whole point of pay is to help attract and leverage talent. We need an approach that succeeds in tapping specialists, online instructors, part-time educators, and others who can best serve students. Rather than cement in place new merit-pay systems predicated on improving test scores for a teacher who spends 45 minutes a day for 180 days with the same 24 students, let’s design systems that can reward unconventional forms of excellence.
For instance, an online tutor who lives thousands of miles away but who can help struggling students make remarkable leaps in mastery of algebra is an invaluable asset. The same is true of a retired army sergeant who may be ill-equipped to teach a middle school class but who may be able to inspire and mentor 15 middle school students or of a teacher who builds a dynamic arts or science program. Today, there is little room in teacher pay scales to recognize or reward—or, sometimes, even make possible—these kinds of contributions. The attempt to superimpose rigid hierarchies atop an otherwise unchanged profession was one of the big stumbling blocks for career ladders and merit-pay proposals in the 1980s. Let’s take care not to repeat those mistakes.
Finally, today’s test-based merit-pay systems have nothing to say when it comes to productivity. They funnel more dollars to teachers who yield higher test scores. The reward is a bonus for past performance; it does nothing to amplify a teacher’s effect on students and schools. Well-designed merit-pay systems should reward teachers who choose to take up opportunities to do more good—such as instructing additional students, leveraging particular skills, or assisting colleagues—making their increased pay a pound-wise investment for their districts or schools.