This paper, which draws on follow-up data collected from English subjects first tested as adolescents and then, 1I years later, as young inen of 24–25, falls into three parts.

Part I examines the structuring of responses to authoritarian statements. No general authoritarian factor could be isolated in either adolescence or in adult life. Instead, four separate authoritarian response tendencies, each with its own antecedents, significance and predictive value, were obtained. Only the adolescent measures correlated significantly with ability level. This, together with the greater stability across time of the responses of the more able 13–14 year olds, led us to hypothesize that some of the variance in adolescent scores was cognitively, not motivationally, determined.

Part II reports a series of experimental studies (using additional data from the follow-up investigation) testing the cognitive and developmental hypothesis, which received support.

Part III proposes a general model for the consideration of attitudes, in which any given attitudinal response is located in a three-dimensional space of cognitive complexity, personality needs and social structure.