On the predictive validity of attitudes: The roles of direct experience and confidence1

Previous research has demonstrated that behavior is more accurately predicted from attitudes formed via direct, behavioral interaction with the attitude object than from attitudes developed via indirect, nonbehavioral experience. The present research examined the hypothesis that the confidence with which an attitude is held may be a mediating variable in the observed relationship between the manner of attitude formation and attitude-behavior consistency. In the first experiment, it was demonstrated that subjects who formed their attitudes through direct experience held those attitudes more confidently and behaved more consistently with those attitudes than did subjects who formed their attitudes through indirect experience. In the second experiment, it was found that, regardless of the manner of attitude formation, subjects who were led to believe that they held their attitudes confidently displayed greater attitude-behavior consistency than did subjects led to believe that they held their attitudes with little confidence. Taken together, the results suggest that it may be fruitful to view confidence both as a variable which mediates the effect of the manner of attitude formation on attitude-behavior consistency and as one which, independent of how an attitude is formed, acts as a determinant of attitude-behavior consistency.