Walt Gardner takes a peculiar take on the good data from schools on military bases:
Yet there are several factors that militate against extrapolation. First, the standardized tests that are mandated by No Child Left Behind are not used at military base schools for the same purpose. Instead of judging the effectiveness of teachers, the tests are used strictly for diagnostic purposes. Second, the average class for kindergarten through third grade there has 18 students. This compares with 24 students in New York City. Third, military parents are provided health care for their children and housing. The availability means that students do not attend classes with ailments afflicting many of their counterparts in public schools. Finally, the military puts a premium on education, providing parents with time off from work to participate in their children's schools.
None of these factors applies to public schools. As a result, they'll never be able to match the success of schools on military bases.
This is an overly realistic "reality check." There is no reason all of the above can't be done nationwide, as they are in many advanced countries. Even providing parents with more time off is actually pretty easy. Just write a law to that effect. Lots of countries have laws that mandate people get more time off work.
Undoing the use of testing for teacher evaluation could be done tomorrow. Heck, it is barely implemented.
If it isn't realistic to consider those four measures, why do we bother with the pretense of democracy at all?