ABSTRACT
Political communication experimentalists rely heavily on single-shot outcomes as opposed to longer-term effects. This is unfortunate given the instability of political attitudes. More than 50 years ago Converse (1964) argued that most people’s political attitudes are superficial and, hence, unstable. As a result, their responses vary when asked the same questions at two or more points in time. Converse (1964) coined the term, “nonattitudes” to characterize the political beliefs of most American adults. Asher maintains that “the presence of nonattitudes is one of the . . . most perplexing problems in public opinion polling” (1998, p. 26). The problem of instability in people’s responses to queries about their preferences among candidates or their attitudes about issues has vexed public opinion scholars for decades and explains the seeming volatility found in public opinion poll results (Asher, 1998; Bishop, 1990; Erickson, Luttbeg, & Tedin, 1980). The problem may stem from the fact that most people do not have stable beliefs about political issues, which is Converse’s position. It may stem from measurement error, which is an intrinsic shortcoming for many social science measurements, as Achen (1975) argues. Or, the problem may be a product of both considerations. Zaller (1994) provides an explanation based on “ambivalence.” Even if people possess “meaningful beliefs” about political issues, they respond to questionnaires based on “top of the head considerations” and, therefore, our instruments only capture part of their beliefs (1994, pp. 278, 280).
