ABSTRACT
Lay people and psychological theorists alike tend to think in dichotomies, or at least in bipolar terms. For example, we tend to think about good versus evil, liberal versus conservative, masculine versus feminine, and so on. So what would we say about somebody who evaluated an attitude object as both good and evil? At first glance we might be inclined to view these responses as irrational or improbable. After all, how could a person evaluate something as being both positive and negative? In the attitude domain, these sorts of inconsistencies between the experiences which underlie an attitude have been traditionally considered to contribute to error variance, and treated as barriers to exact measurement (Ajzen & Fishbein, 1980; Wicker, 1969).
