This descriptive study of the indicators from the National System for the Evaluation of Higher Education-Sinaes/Inep/MEC. There is evidence of methodological problems in the construction of the instruments for the National Examination of Students' Performance (Enade). There was a variation in the reference matrices in 2004 and 2007, changes in the formulas, and multi-colinearity in calculation of the Preliminary Classification. In 2004, 65% of the questions were classified as "difficult", as compared to only 35% in 2007..Nearly half of the questions evaluated the capacity for memorization (47% and 42%). Cart analysis pointed to the following factors associated with better performance on the graduating student's: higher socioeconomic status, public institution, and high supply of complementary training by the institution. Both the graduating student's specific component grade and the difference between them and first-year student's overall classification appear as robust performance indicators. This study proposes a multiple-criteria indicator which proved capable of discriminating between highly rated institutions in the medical field. Consistent and robust reference matrices are necessary for monitoring courses that train excellent professionals who are aware of their social responsibilities.