Abstract

The individual difference approach to the issue of attitude-behavior consistency was examined in light of the intra-individual variability of the attitude. It was predicted that individuals whose attitudes have been relatively invariant (low intra-attitude variability) would be predicted well from their attitudes. In the first session, subjects completed the controled association of the attitude objects. Here, subjects' attitude(summation of the beliefs aroused from controled association) and intra-attitudinal variability (standard deviation of the beliefs aroused from controled association) were computed by Fishbein's attitude formulation. In the second session approximately 2 weeks later, several measures of subjects' reports of behaviors were obtained. Attitude-behavior correlations were then computed within each of both groups of subjects produced by a low-high intra-attitude variability. As predicted, only subject with low intra-attitude variability manifested high attitude-behavior correlation across all behavioral measures. The implications of the results for the individual difference variables and for the attitude-behavior consistency controversy were discussed.