Predicting behaviour from attitudes

At the beginning of Chapter 2, I said that the study of attitudes is a dangerous starting point for a book on social psychology. Probably the greatest danger concerns the predictive utility of the attitude concept. Sophisticated techiques of attitude measurement can be devised, impressive theories of attitude organization can be proposed, but if, at the end of it all, we are no more able to predict what a person will or will not do in a given situation, what use are our measurements and theories?

Some of the most surprising studies in social psychology, therefore, are those that have claimed to find little or no relation between people's behaviour and their verbally expressed attitudes. When one looks at those studies that have attempted directly to compare verbal expression of attitude towards a group or issue with other attitude-relevant behaviours, a rather confused picture emerges. Sometimes the verbal measures provide quite good predictors of the specific kinds of behaviour under investigation, but very often they seem to allow no such prediction at all. In his review of such research, Wicker (1969) concluded that only in a minority of cases was a close relationship found between verbally expressed attitudes and overt behaviour, the typical result being one of only a slight association, or no association at all.