ABSTRACT
Nearly all contemporary public opinion research rests on the assumption that sociopolitical attitudes are unidimensional and bipolar-i.e., positive, negative, or neutral evaluative responses (e.g., Eagly & Chaiken, 1993; Green & Citrin, 1994; Lodge, McGraw, & Stroh, 1989). In studies of mass belief systems and electoral behavior, for example, attitudes toward policies, candidates, and groups are typically operationalized as summary statements that range from “unfavorable,” “oppose,” “cold,” or “negative” at one end of the continuum to “favorable,” “support,” “warm,” or “positive,” at the other. This view implies that positive attitudes are the diametric opposite of negative attitudes, such that the more one likes a political object the less one dislikes it. Unfortunately, this structural assumption masks a fundamental and readily acknowledged aspect of belief systems, namely, that individual opinions are not simply positive or negative evaluative tendencies, but instead are often simultaneously positive and negative (Alvarez & Brehm, 1995; Cacioppo, Gardner, & Berntson, 1997; Feldman & Zaller, 1992; Hochschild, 1981; Huckfeldt & Sprague, 1998; Lavine, 2001a; Lavine, Borgida, & Sullivan, 2000; Lavine, Thomsen, Zanna, & Borgida, 1998; Nelson 1999; Thompson, Zanna, & Griffin, 1995; Zaller & Feldman, 1992). That is, rather than endorsing one side of a political debate and refuting the other, individuals often embrace central elements of both sides.
