Discrepancies between the pessimistic national evaluations and the more encouraging state and local evaluations of compensatory education programs are explored in this paper. A conventional explanation for the discrepancies, namely, that the averaging of test scores from a large national sample "cancels out" the results of successful programs, is challenged. It is held, rather, that the quality of the evaluations themselves, as well as vested political interests may be responsible for the findings. Well publicized "exemplary" programs are revealed to have enjoyed only short term success. It is also suggested that a decline of initial achievement gains may occur while children are enrolled in enrichment programs. (Author/GC)