One begins to feel that those testing merit pay for teachers are deliberately missing the point — at least by the time their findings filter down through the mainstream media. Here’s the latest:

Offering big bonuses to teachers failed to raise students’ test scores in a three-year study released Sept. 21 that calls into question the Obama administration’s push for merit pay to improve education.
The study, conducted in the metropolitan Nashville school system by Vanderbilt University’s National Center on Performance Incentives, was described by the researchers as the nation’s first scientifically rigorous look at the effects of merit pay for teachers.
It found that students whose teachers were offered bonuses of up to $15,000 a year for improved test scores registered the same gains on standardized exams as those whose teachers were given no such incentives.

In absorbing the implications of these results, we must first absorb the relevance of this factor, offered at the end of the article (and not printed at all in the Providence Journal reprint, by the way):

Only about half of the 300 teachers originally in the Nashville study were left at the end of the three years because some retired, moved to other schools, or stopped teaching math. About 40 teachers got bonuses each year. Overall, the researchers said, test scores rose modestly for both groups of students during the three-year study, suggesting that the financial incentives made no difference.

Now, let’s ignore the possibility that the results of a study that loses half of its participants is of questionable validity. And let’s put aside practical questions that arise from that datum, such as:

• Did teachers from both groups exit the study at equal rate?
• If not, could the possibility of receiving a bonus have helped to retain teachers, or the probability of not earning a bonus have inspired exit?

Let’s also take off the table the possibility that teachers in the control group, in keeping with their unions’ strong and frequently stated distrust of merit-pay systems, tried especially hard so as to disprove the thesis that they’d try harder if it meant a bonus (thereby effectively confirming the thesis in an unmeasurable way).
At least some of us who support them see merit pay as the palpable incentive at the end of an entire reworking of the system to increase accountability up the education and administration chain and as a long-term process of changing the way in which teachers think about their jobs.