Thanks to an excellent speech by Dan Fallon (a former colleague of mine, now at Carnegie Corporation), I understand education policy much better. Dan shows that in the 1960s, experts and policymakers were much influenced by James Coleman’s massive studies, which were later confimed by Chrisopher Jencks. Using the data they had, these scholars found that you could predict academic success very accurately by looking only at the home and the neighborhood from which a kid came. In other words, schools didn’t matter; society did. Coleman in particular expressed caveats about this finding, but it was the simple summary of his work.
Then (as Fallon explains it) Southern governors in at least five states mandated tests as part of their efforts to improve eduation. These tests for the first time collected data on students, teachers, and schools. A Tennessee agricultural statistician named William Sanders realized that the data would allow him to find out whether it makes a difference which teacher you have. In short, the answer is yes. Numerous subsequent studies have confirmed that differences in teachers make a huge difference for kids. In particular, teachers’ own education and their teaching styles are extremely important.
In the little part of the education world that I know something about–civics–exactly the same pattern occurred. In the 1960s, political scientists looked at available data and concluded that you could fully predict an adolescent’s civic and political behavior and attitudes based on his or her family and community; schools made no difference. As a result, scholarship on civic education virtually ceased. Then the National Assessment of Educational Progress (NAEP) civics exam collected data on civic knowledge. Lo and behold, it turned out to matter whether and how young people are taught civics.