Examining when children form logical capacities to infer causal attributions has been neglected despite the fact that elementary school children were used as subjects in quite a number of studies of causal attributions. Thus, the purpose of the present study was to make clear the development of children's perceptions of the causes of academic achievements. In the first study, age-related change on information processing relevant to inferring causal attributions was examined. To measure the degree of logical capacities of such information processing, a new test designated as ICA, consisting of 10 items describing hypothetical persons to get successful or failing points on a mathematical examination was made. Subjects were provided with information such as whether the person's point would be higher or lower than the average point of other classmates, whether the examined point would be consistent with that of past examinations and whether the teacher conducting the examination would be different from the one conducting past examinations, etc. Subjects were 34 third graders, 82 fourth graders, 116 fifth graders in an elementary school, 173 junior high school students and 69 university students considered as adults. Upon reading each information, subjects were requested to select the most suitable cause among five given choices: ability, effort, difficulty of test questions, teacher's teaching method and luck. The Main results were as follows. 1.Even lower graders could infer effort more easily than other causes. 2.Inferring ability as the most suitable cause, especially from successful information, was found difficult at every age; particularly, elementary school children frequently selected effort instead of ability, presumably because they had no formed concept of ability. 3.Most of the elementary school children inferred easily teacher's teaching method as a cause of success, but they did not infer it as a cause of failure. The second study aimed at examining the development of attribution-affect linkages. In adults, affective reactions were believed to be maximized in the case of internal causal attributions for success and failure, and minimized in the case of external causal attributions. In order to measure the degree of the linkages, we made a pair comparison test of 12 items designated as CAAF. In each item, subjects were assumed to get a successful or failing result of an examination due to both internal and external causes. Then, subjects were asked to choose in which case they would elicit stronger affects. As in the first study, ability and effort were used as internal causes, and teacher's teaching method, difficulty of test questions and luck as external causes. Furthermore, proud, ashamed, sure of and hopeless were used as affects. The Main reults were as follows. 1.Proud or ashamed-effort linkages was stronger than sure of or hopeless-ability linkages at every age, and the difference between these two linkages proved outstanding especially in elementary school children. 2.Generally, the degree of the linkages increased with age, and the most remarkable developmental difference was found between fifth graders and junior high school students. However, no significant difference was shown between junior high school students and university students.