NCLB Debate

Over at newtalk, they are having a little NCLB debate. Here is a response from Richard Rothstein:

NCLB is a failure, and should be scrapped, for these reasons:
a) By requiring accountability only for math and reading, it distorts the goals of education. In any institution, if principals hold agents accountable only for some of the institution’s goals, agents will distort their behavior to accomplish only those things for which they are held accountable. This is rational behavior. It is the reason the Soviet command economy was so inefficient (textile mills required to make adequate yearly progress in the production of yards of cloth produced only useless narrow widths; trucking firms required to cover more miles drove around in circles, etc.); it is also the reason that NCLB-type accountability programs in job training, welfare reform, and health care have been abandoned. Local agencies held accountable for placing workers in jobs concentrated on placing workers most likely to be placed without agency assistance, not those most in need of assistance; hospitals required to improve the survival rates of cardiac surgery patients refused to operate on the sickest patients, etc. The biggest tragedy of all this is that goal distortion has been most severe for disadvantaged children, because these are those for whom there is the biggest payoff for substituting math and reading drill for instruction in other subjects and behavioral traits.
b) Leave aside any racial or socioeconomic differences in achievement. There is sufficient variation among youth, irrespective of social background, that a single standard of proficiency cannot possibly be “challenging” (to use the NCLB term) to below-average, average, or above-average children. The problem is not whether proficiency is defined too high, or too low, or differently among states. Any standard of proficiency is irrational if applied to all students.
c) NCLB permits no adjustment for socioeconomic differences, although these are the most important determinants of student achievement, on average. (Yes, I know, some poor students excel, just as some smokers don’t get lung cancer, but on average, socioeconomic disadvantage has a big impact. As you know, I’ve written a book on this topic, but forget me as an authority - Janet Currie has estimated that 25% of the black-white school readiness gap can be explained by differences in a few measurable health characteristics of children and their mothers; our own Rick Hanushek and colleagues have estimated [for Texas] that about 14% of the racial achievement gap can be explained by differences in student mobility rates. Etc.)
d) An accountability system based on test scores alone has stimulated nationwide score inflation, as teachers and administrators have naturally learned how to game the system, both by legal means and by unlawful cheating. The nationwide improvement of state test scores, relative to NAEP scores, is evidence alone that NCLB and its required state accountability systems have resulted more in gaming than in improved instruction. We are back to the “Lake Wobegon effect” of the minimum competency movement of the late 1970s.

The above does not mean that, in principle, the federal government could not design and administer a satisfactory accountability system. But federal micromanagement of education (NCLB) has now proven to be so incompetent, for reasons stated above, that it would be foolish to rush headlong into another federal system, as poorly thought through as NCLB was. Why not let the states struggle with these very difficult challenges? Some may design better accountability systems, some may stick with NCLB-type systems. But after the NCLB fiasco, we’ve got little to lose by letting states experiment with alternative accountability systems.

Go on over and read the crazies defend NCLB, and the more reasonable explain why it should go away.