Daniel Henninger wrote favorably in yesterday's Wall Street Journal about linking teacher pay to standardized test scores. He notes the success of grade schools in Little Rock, Arkansas that have moved to such a system.

I'd like to believe that student learning is actually increasing in those schools, but whenever I hear about such pay-for-performance schemes, I am struck by two ugly realities. The first is something my wife observed during her years of teaching in Detroit — even in the absence of bonuses, teachers in her school engaged in systematic cheating on standardized tests. They did this primarily by copying sections of the test in advance and helping students memorize the answers. They also erased incorrect answers and penciled in the correct ones.

The second ugly reality is that teaching performance is largely not a function of effort. Poorly performing teachers don't fail students because they aren't hustling. They fail students because they are unintelligent, and not competent to teach. They have not been adequately trained to teach. They do not have a coherent methodology or pedagogy for teaching. As a consequence, they busy themselves with a combination of babysitting and slogging through poorly designed materials and curricula.

Exceptional and even competent teachers, in other words, can't help but educate their students. True, at the margin, a bonus may motivate additional behaviors that may enhance education. A teacher may think longer, say, about whether her lesson plan is adequate, or work just a bit more with a struggling student. It is simply not the case, however, that poorly performing public schools are filled with competent teachers who currently refrain from making the effort because the money isn't there. They don't teach well because they can't.

Throw some money at them, however, in a system where they control the measurement instrument (to wit, penciled-in standardized test bubble sheets), and you will see improvements in test scores. Just don't look too closely.

A modest proposal, because I would love to be proven wrong: some well-meaning foundation out there should fund a study, in which objective monitors — not school employees — administer the tests, collect the answer sheets, and remove them to a neutral location for machine grading. I'd be interested to see how the "turnaround" schools profiled by Henninger respond to such a proposal.