In the most narrow sense, compensatory education has failed; most Title I programs which seek to improve achievement have no discernible effect. This is apparently a result of the fact that the programs seek to provide for disadvantaged students more of the school resources which have never been found to affect the achievement of advantaged students. It is, on the other hand, clear that compensatory programs need not fail; highly structured preschool programs aimed at language development produce impressive short run IQ gains. The most important difference between these experiments and most other preschool programs is that the former were aimed narrowly at skills clearly related to what tests of IQ and verbal ability happen to measure. In addition, there are other models of what might be called "compensatory" education--at least their object is to redistribute educational outcomes, and to improve life chances of children from disadvantaged circumstances. These programs however, focus on more distant outcomes of schooling--college entrance and high school graduation. The difficulty with both approaches is that they rest on an extremely limited evidence about the process of schooling. For example, we have no evidence that IQ affects life chances though if it does cause longer school retention, it may. While most standard compensatory programs have failed to affect achievement, there was never any evidence that much else could have been expected. (Author/JM)