Many proposals would link teachers' incentives to students' performances in standardized tests, but that is already the subject of criticism among educators, both for its potentially distorting effects on curriculum and classroom time and for its exclusive interest in certain subjects - math and science, for example, rather than drama or music.
How, too, would bonuses reflect the fact that subjects are interrelated - that a strong English teacher, for instance, makes a contribution to how well students learn math? Research also shows that good teachers tend to bring up the level of all teachers in a school. As Mr. Phillips says, merit pay may make them less likely to share best practices.
"There is so much more that goes on in classrooms, day in and day out, than simply standardized-test scores," says Sam Hammond, president of the Elementary Teachers' Federation of Ontario. "There are so many different inequities for teachers that could come out of it that it's unthinkable, quite frankly."
"It's a very challenging program to implement," says Susan Moore Johnson, a professor at the Harvard Graduate School of Education who has been studying merit pay for two decades.
She is not sure that more cash makes better teachers. "Fundamentally there's a questionable assumption at the heart of it, and that is that teachers will respond to sums of money, which often are relatively small, and that they will work both harder and more effectively."
Merit pay for teachers is not a new idea - it comes up every few decades when politicians become freshly worried about school performance. It has been a bigger issue in the U.S., where teachers receive far lower salaries than their Canadian counterparts. The last time it was seriously advanced in Canada was in Ontario under Conservative premier Mike Harris, but it was opposed by a teachers union already embittered by other education cuts, and never went anywhere.
Surveys show, though, that the next generation of teachers finds the idea more palatable. Dr. Johnson credits this in part to changes in the labour market: In the 1960s, pay was standardized in education to prevent discrimination against a largely female (and, in the U.S., minority-heavy) workforce.
"Now," she says, "people who are becoming teachers can work in other fields and know that other fields operate differently with pay - their friends get huge pay raises, and pay is more important in society.
"And I think there's just more sense that it wouldn't be a bad thing, if it were well done."
But the experiments don't always produce clear results: In Denver, teachers who opted in to a merit-pay program did raise students' performance, but they were also likely better teachers who knew they could get bonuses. And in one program in Texas, student performance didn't improve with merit pay - though research suggests that program didn't work because it distributed small bonuses among about 70 per cent of teachers in participating schools.
OVER THE NEXT HORIZON
While the idea, as Dr. Johnson puts it, "may not be ready for prime time," she points out that statistical methods of determining effective teachers have improved vastly in the past 10 years. One popular concept is to track student progress in a teacher's class over three years, using forms of standardized testing that can adjust for variables such as socio-economic status.
While such a system may identify the best teachers, it doesn't show less-successful educators ways to improve. Other merit-pay systems use a combination of testing and classroom evaluation, sometimes by independent observers rather than the principal.
Dr. Jackson insists that with the right data and expert analysis, it is possible to discern a good teacher clearly from a bad one (although many school districts don't have the funds, expertise or even data to do it properly). In particular, he proposes students' gains as a good measure - that is, tracking how much students improve over the course of the year. Research has found that teachers with poor skills may bring their students up only half a grade; the best teachers can advance their classes by more than a year and a half.