Attitudes and Behavior

The relationship between attitudes and behavior has long been a topic of central interest within social psychology. Early theoretical and methodological approaches to the study of attitudes generally assumed a strong association between attitudes and behavior. Such an assumption can be found in many early definitions of attitudes, which frequently defined the construct in terms of its presumed influence on behavior. For example, Gordon Allport's classic definition of attitudes as “mental and neural states of readiness” that influence an individual's responses to objects and situations related to that attitude clearly implies that attitudes should exert a directive influence on behavior. Likewise, early theories of attitude structure also assumed a strong attitude-behavior association. This point is perhaps best exemplified by the highly influential tripartite theory of attitude structure, which postulated that behavior was one of the three components comprising attitudes, along with affect and cognition. Even within the attitude measurement literature, early researchers often assumed strong attitude-behavior associations. For instance, many of the early indirect measures of attitudes relied on observable behaviors as indicators of the underlying attitude.

However, by the 1960s, some researchers had begun to raise serious doubts regarding the assumption of a strong relationship between attitudes and behavior. Perhaps most notably, in a highly influential review of 46 studies examining attitude-behavior associations, Allen Wicker concluded that attitudes and behavior correlated, on average, at only .15 and rarely correlated above .30. Thus, he challenged the notion that attitudes are useful predictors of behavior and indeed raised concerns regarding the fundamental utility of the attitude construct. His criticisms and those of others, which came to be known as the “attitude-behavior problem,” ultimately did not discourage the study of attitudes but rather served as an impetus to better understand why attitude-behavior correlations were often so low. The responses to this question took a number of forms, and each of these responses has, to some degree, continued to be an ongoing theme in attitude-behavior research.