Problems with "Manifesto" thinking

In yesterday's Weekend Quotables, Rich Rothstein and Pres. Obama get it mainly right re. the importance of having a good teacher in every classroom. Rothstein is quick to point out however that, "differences in the quality of schools can explain about one-third of the variation in student achievement. But the other two-thirds is attributable to non-school factors."

But even here, there are limits to how much we can claim when we draw on statistical social-science or education research. While Rothstein's scores a point against "Manifesto" thinkers like Rhee and Klein, who put the entire burden of closing the so-called "achievement gap" on the schoolhouse, those at the school level are often trapped by the data. Statistical analysis is useful in looking at the big picture and for policy analysts. But when taken too far, when used the wrong way, even the best analysis can be misleading.


Yes, in general, good teaching trumps other in-school factors when it comes to student learning. But not always. For example, in many overcrowded urban schools, where class size mushrooms to 50 students, replacing one teacher with another, better one, may make a negligible difference whereas radical changes in the learning environment can assume much greater importance.  The same can be said of school leadership. In a school with a dysfunctional or badly inexperienced leadership team, even the best classroom teacher can become grossly ineffective, especially when effectiveness is measured only, or mainly by student test scores. The point here is that at the school level the most necessary changes can be the ones least considered in the research, the ones figured out by the knowing eye of the educator or school leader.

Policies and legislation that are based on overblown research claims have become disastrous for school-based educators. Race To The Top is the best example, although it's doubtful that many RTTT policies, including school closings, private management, or merit pay, have any basis at all in the research. The worst cases can be found with the L.A. Times publication of the names of individual teachers connected to their so-called "value-added" scores or the mass firings of teachers at Central Falls High School in Rhode Island.  And now Democrats are pushing legislation based on such research, that would mandate this faulty use of data in teacher evaluation, nationally for all schools.

The most important school decisions should be made closest to the learning point and start with a concrete analysis of local conditions. The farther away you get from that point, the worse the decision-making gets (I don't know whose law that is, but it's a good one). Big-picture statistical research can be useful. But it's power diminishes when it claims too much.