Abstract

On the assumption that attitude statements are differentiated into descriptive and evaluative dimensions, this study reexamines previous results that the more pro a judge's attitude, the more polarized his judgments of attitude statements. In experiment I, it is revealed that the results of previous researches are due to the fact that only "pro" statements are selected. When both sides on the descriptive dimension ("pro" and "anti" statements) are included in attitude statements, differential polarization between pro and anti judges disappears. In experiment II, it is revealed that some subjects cannot differentiate descriptive dimension independently from evaluative dimension of attitude statements. These subjects misinterprete the task as the judgment on the evaluative dimension rather than on the descriptive dimension. As its result, the lower a subject's level of dimensional differentiation, the more polarized his judgments of attitude statements.