Our university, in comparison with other medical institutions, is proud of a very high ratio of the students who graduate in due course of six years, and that a high ratio of the successful applicants in the national examination for medical license (hereafter N E). These results may be due to teachers' enthusiastic instruction and students' diligent response as well. Of course it is not the sole objective of medical education to make much more, say even one more, successful applicants in N E. However, it must be certainly one of the main objectives of that education. So far, the author made several reports, to pursue the main factor of raising the ratio of the successful applicants in N E, assuming the students of the same year as one group and comparing some factors among these groups. In this essay, in order to ascertain the correlation between each student's score of each subject in the final examination for graduation and the success or failure in N E, the author examined the existence or nonentity of the differences between the mean orders and the mean scores of the successful group and the failed, and then searched out and analyzed the principal components which comprise the scores of twenty subjects. As a result, the author found the primary principal component to be no less better than the score in the comprehensive graduation test and the mean score in the test of the whole subjects as the data for forecasting the success or failure in N E. The author also tried to make the same kind of analysis on the data in the time of matriculation, but it resulted in recognizing the strong influence of the effect as a result of selection (the effect which causes the phenomenon where a certain correlation is not to be observed, as a result of selecting the applicants who get the similar scores in an examination).