Today is NCLB’s 6th birthday. NCLB is, at its core, a policy predicated on the idea that schools vary widely in their ability to improve students’ test scores. By holding schools accountable, the hope is that “bad” schools will become more like “good” ones. (Note - this is a post about NCLB on NCLB’s terms, so I’m going to focus on test scores. For more posts on NCLB, take a look here.
However, as I wrote yesterday, once we take into account students’ background characteristics, school effects on standardized test scores are pretty small. The good news is that teacher effects on test scores are quite large (you can find more posts on teacher effectiveness here). In short, the differences between teachers in improving test scores are much larger than the differences between schools. This finding has significant implications for the potential success of school-based efforts to improve test scores, as Barbara Nye, Spyros Konstantopoulos, and Larry Hedges wrote in their paper, “How Large Are Teacher Effects?”:
Many policies attempt to improve achievement by substituting one school for another (e.g. school choice) or changing the schools themselves (e.g. whole school reform). The rationale for these policies is based on the fact that there is variation in school effects. If teacher effects are larger than school effects, then policies focusing on teacher effects as a larger source of variation in achievement may be more promising than policies focusing on school effects.
(You can click to enlarge the picture above - courtesy of the Halloween Edu-Parade, Rod Paige is Armstrong Williams.)