Though scholars have long been concerned about the quality of citizens' political decision making, we still know little about why citizens disagree with the best-informed opinion in society, that of public policy experts. In this article, I examine the factors that lead citizens to disagree with expert opinion on questions of public policy. I find that both elite cues and individual-level attributes of citizens lead individuals to disagree with experts. In contrast to the expectations of many recent studies of cue taking, I find that citizens are more likely to disagree with expert opinion when political elites they favor challenge this opinion. Citizens also disagree with experts as a consequence of low levels of knowledge, existing policy preferences, and life experiences. The study's results challenge the optimistic conclusions of many recent studies of cue taking and argue that there is significant value in the conventional wisdom that preceded these studies. Elite cues are not a consistent means to effective policy judgments. Instead, when it comes to reaching effective policy decisions, there is no substitute for knowledge.
Following several decades of research (Berelson 1952; Converse 1964; Delli Carpini and Keeter 1996) that had found most Americans' policy judgments "minimally consistent" (across issues), "minimally stable" (over time), and "minimally comprehended" (Sniderman, Brody, and Tetlock 1991: 2), the image of democratic citizens has been partially rehabilitated in the past decade. Rather than rejecting the finding of minimalism, recent work largely accepts it, but argues that this minimalism need not lead individuals to erroneous policy judgments. Rather, citizens can reach effective policy decisions despite these limitations because political elites provide simple and reliable cues that point toward effective policy judgments (Popkin 1991; Wittman 1995; Lupia 1994; Lupia and McCubbins 1998). Studies support this claim by demonstrating that less informed citizens who are aware of elite cues reach the same policy judgments as citizens with somewhat more factual policy information.
The validity of this optimistic conclusion rests, however, on the validity of its basic premise-that the somewhat better informed are making effective policy judgments. If not, agreement between cue takers and factual decision makers represents concurrence in questionable judgments, not effective democratic citizenship (see Kuklinski and Quirk 2000). The hazards of employing a standard of somewhat better informed opinion are twofold. We risk drawing overly optimistic conclusions about citizen performance based on mere opinion correspondence and risk drawing erroneous inferences about the factors that lead to effective policy decisions.
Our substantive interest in studies of heuristics is not whether citizens can use cues to match the opinions of the somewhat better informed, but rather, whether they can use cues to reach effective policy judgments. We should, therefore, choose as our criterion group not the somewhat better informed, but rather, the (well-informed) societal group that, on average, is most likely to make valid policy judgments. Undoubtedly, this group is public policy experts. Experts possess high levels of specialized knowledge in specific policy domains; this knowledge, no doubt, aids experts considerably in reaching policy decisions. Experts, of course, are not infallible. However, they are very likely to be less fallible than the somewhat better informed, or any other criterion group that one could examine. Yet, despite experts' high levels of policy knowledge, and the guidance this information provides for decision making, very few studies have examined why citizens disagree with expert opinion (but see Margolis 1996).
This article addresses this gap in our understanding of public opinion by examining citizen disagreement with expert opinion on four diverse issues: the Strategic Defense Initiative (SDI), nuclear energy safety, Japanese investment in the United States, and AIDS quarantines. …