In order to clarify the impersonated factors which may affect high school achievement, the following four studies were carried out by using three hundred high school students. The purpose of the study I was to compare the psychological traits of over, normal and underachievers (OA, NA, UA) on the basis of the following five tests : (1) Yatabe-Guilford personality test (YGPT), (2) Uchida-Kraepelin performance test (UKPT), (3) Qreativity test, (4) Achievement motivation test (McClelland'S), and (5) EIPC (The estimative inventory for the problem child). The difference between OA and UA was found to be ,statistically sign:ificant with, respect to UKPT, and it waS pointed out that "will" is important as a fa<)tor. The study II was that canonical and discriminant analysis were zittempted with the same data of UKPT as used in the study I. As a result of it an unexpected clear separation among the three groups was obtained. This indicates that the subjective judgment of the profile in study I is correct as a whole, and that the application of these two analyses, to UKPT is very useful. The purpose of the study III was. to examine to what degree the multiple regression analysis using UKPT and intelligence test results could account fc;r the academic achievement variance. The result indicated that the largest correlation coefLicient was 0.368 when the measures of UKPT were four and those of intelligence test three. Therefore, the explanation was not always satisfactory. If the pre-dictor variables were used in this way, however, predictive power was generally higher for UKPT than for the intelligence test.