ABSTRACT
Early pessimism about the idea that attitudes guide behaviour was nourished in an article by LaPiere. Although measures of general attitudes toward objects are poor predictors of single, specific behaviours, they fare better at predicting behaviour over a wider range of situations and contexts. Attitudes and behaviour are more closely related if the aspects of the attitude that are highly accessible at the time of attitude measurement are also accessible at the time the behaviour is performed. LaPiere concluded from the enormous discrepancy between stated attitude and overt behaviour that questionnaire responses are not valid indicators of a person's true attitude. Eckes and Six reported the most comprehensive meta-analysis of attitude-behaviour studies, covering research conducted between 1927 and 1990. Studies were categorised according to the behavioural domain that they examined, and the mean correlation between attitude and behaviour was computed separately for studies in each category.
