Abstract

In this study it was assumed that the effect of attitude upon behavior is determined by interaction between the pressure of nonattitudinal variables which is discrepant with attitude and the attitude strength which resists against the pressure of nonattitudinal variables so that the behavior is consistant with the attitude. Attitude strength was considered as the individual's convingsness about his attitude resulted from intra-attitude variability of the attitude structure. The smaller the intra-attitude variability the more the individual's convingsness about his own attitude, and in turn the more is the behavior consistant with the attitude. And it was also assumed that attitude strength is different attribute of attitude from affective extremness which has been defined as attitude operationally, because even same attitude values can lave different intra-attitude variability. Three hypotheses were tested: First, the more the intra-attitude variability the more the error of predicted behavior from the attitude. Second, even though attitudes have same values, the more the intra-attitude variability, the more the error of predicted behavior from the attitude. Third, attitude extremness which is independant from intra-attitude variability will not correlate with the error of predicted behavior from attitude. This hypothesis implies that attitude extremness by itself is not an determinant of the error of the behavior prediction, whereas attitude strength determine the prediction error. This study performed a serial repeated measure of attitude, intra-attitude variability and behavior intention. To test above hypotheses, it was calculated correlation of intra-attitude variability with error of predicted behavior intention from attitude, partial correlation intra-attitude variability with the error of predicted behavior intention from attitude controlling for attitude extremness, and correlation of attitude extremness with error of the predicted behavior. Supporting all hypotheses, the results showed that all the correlations but last one were positive and statistically significant. The results illustrate that attitude strength which can be inferred from intra-attitude variability of attitude structure is a different attribute of attitude from attitude extremness and that behavior-predictability of attitude can be differentiated and predicted by the intra-attitude variability.