Expected utility theory, proposed as a model to explain individuals' preference behavior, has received much criticism since Allais. Previous research in this field has focused on developing new theories by assuming that the cause of violations lies in the nonlinearity between probabilities and individuals' perceived likelihood of occurrence. This paper adopts a micro-level approach to investigate, if violations are universal, which types of individuals exhibit higher rates of violations, thereby seeking to identify the causes of violations. In particular, the relationship between individuals' information processing ability accumulated through learning and the violation phenomenon was investigated. To examine these issues, problems similar to those employed in previous studies were used. The proportion of subjects exhibiting preferences violating expected utility theory, while not sufficiently high to reject the null hypothesis except for the common ratio effect problems, was still at a high level. In the independence problems, 40% showed theory-violating preferences, and in the common ratio problems, 47% did, yielding results similar to those in many prior studies. To examine how subjects' individual characteristics relate to these violation phenomena, this paper investigated differences in violation rates between two groups classified by their information processing ability accumulated through education and learning. In all problems used in the experiment, the lower group exhibited higher violation rates than the upper group. Examining the D-statistic for testing differences in violation rates between the two groups, the null hypothesis of no difference between the two groups could be rejected at the 5% significance level in nearly all problems. Therefore, we can state that information processing ability has a significant relationship with axiom violations, and that higher information processing ability tends to be associated with preferences more consistent with expected utility theory. That is, one cause of axiom violations can be found in individuals' information processing ability. We had hypothesized that individuals' risk attitudes would be related to axiom violations. The investigation revealed that the group with lower information processing ability was more risk-averse, but the differences in violation rates between groups classified by risk attitudes were not significant except for the common ratio effect. Therefore, we can conclude that information processing ability has greater explanatory power for violation phenomena than individuals' risk attitudes.