There’s a Difference Between Relative Performance and Absolute Performance
September 18, 2012 Leave a comment
The effects of poverty are often used as an argument against using test scores to evaluate teachers. The reasoning goes that because low test scores are a result of bad environments and not bad teachers, it’s unfair to use them to judge teachers. I’m amazed how many people make this argument because it commits the cardinal sin of conflating absolute performance with relative performance. For example, here’s Eugene Robinson in the Washington Post:
According to figures compiled by the College Board, students from families making more than $200,000 score more than 300 points higher on the SAT, on average, than students from families making less than $20,000 a year. There is, in fact, a clear relationship all the way along the scale: Each increment in higher family income translates into points on the test…
The brie-and-chablis “reform” movement would have us believe that most of the teachers in low-income, low-performing schools are incompetent — and, by extension, that most of the teachers in upper-crust schools, where students perform well, are paragons of pedagogical virtue…
It is reasonable to hold teachers accountable for their performance. But it is not reasonable — or, in the end, productive — to hold them accountable for factors that lie far beyond their control. It is fair to insist that teachers approach their jobs with the assumption that every single child, rich or poor, can succeed. It is not fair to expect teachers to correct all the imbalances and remedy all the pathologies that result from growing inequality in our society.
Robinson’s point is that it’s unfair for a teacher in the South Bronx to be rated poorly because their students fare worse on tests than students in Park Slope. And he’s right…if that were actually happening. But that’s not happening. Proposed evaluation systems all deal with relative performance — that is, how a teacher’s students do compared to other students of similar demographics, socio-economic status, and prior achievement levels.
Almost nobody is clear about this and thus people assume that what’s being proposed is a system where a teacher in a poor neighborhood is being evaluated purely based on how their students perform compared to kids in wealthier areas. In reality, the proposal is for 20%-40% of a teacher’s rating to be based on how their students perform relative to other similar students. It is possible that there is slightly more test score variance among students in poverty, but this nuance is not what’s being argued about.
The relevance of poverty to the debate on education reform cannot be understated, but it has become a red herring in the debate over test-based evaluation. Either test-scores are good way to evaluate a teacher compared to teachers who serve similar students, or they’re not. But the nominal level of achievement of those students doesn’t matter in making that decision.
Education-only Feed. Subscribe.