A critical question in the use of value-added models is whether we’re really measuring how good a teacher is or whether we’re in part picking up that some teachers are in schools where all teachers do better. If the latter is really important, then the quality of individual teachers matters less than otherwise. Such a finding would argue in particular that efforts to attract top-rated teachers to low-performing schools might be misguided. If apparent teacher quality is really a school rather than a teacher characteristic, then bringing a highly-rated teacher into a low-rated school will merely lower the teacher’s without rating doing much for the kids in the receiving school.
A nice paper, Portability of Teacher Effectiveness, argues pretty convincingly that quality does stick with the teacher. The authors separated teachers in North Carolina and Florida into “high-performing” (top 30%) and “low-performing” (bottom 30%). Then they looked at teachers who switched from high-ranked schools to low-ranked schools or vice versa.
Basically high-performing teachers who switched from strong to weak schools stayed good teachers and low-performing teachers who went in the other direction stayed low-performing. That’s pretty strong evidence that teacher quality really is attached to the teacher rather than being a property of where the teacher teaches.
I’ve oversimplified a little what the research team actually checked. Although I’ve given you the bottom line, the authors did something more sophisticated. Any test score, including teacher VAM rankings, includes some measurement error. That means that a high score probably overstates the teacher’s quality a little and a low score probably understates the teachers quality. When we see a second score for a teacher, whether the teacher switched schools or not, we expect to see “regression toward the mean.” A second score for a high-ranked teacher will probably fall and a second score for a low-ranked teacher will likely rise some. So the authors didn’t actually check whether VAM scores of teachers remained unchanged after a school switch. Instead, they checked whether scores rose or fell by the same amount whether a teacher switched from a high-ranked school to another high-ranked school or to a low-ranked school instead. That’s exactly what happened.
Next time more on teachers switching schools.