In The Rational Public, Benjamin Page and Robert Shapiro argue that "collective policy preferences of the American public are predominantly rational, in the sense that they are real…coherent…and that when collective policy preferences change, they…do so in…predictable ways."[1] I will argue that the authors' use of aggregate, or collective, data is convenient and at times convincing, but that the use of such aggregate data obscures an internally inconsistent argument. In short, I will show that their ultimate conclusion—that public opinion is rational—is flawed because the defining elements of their argument are flawed. In this discussion, I will draw upon the work of John Zaller and his book, The Nature and Origins of Mass Opinion, to highlight the authors' inconsistencies. I will also look at specific examples used in The Rational Public that lead one to contradictory conclusions.
Are policy preferences real?
To support their argument that collective policy preferences are real, Page and Shapiro assert that "at any given moment an individual has real policy preferences," that "over a period of time," show a "central tendency of opinion, which might be called a 'true' or long-term preference." Therefore, an "individual's opinion…over a sustained period of time…will be stable."[2] Unfortunately, the authors do not show any empirical evidence that individuals do have such long-term or stable preferences. They assume this to be a given and refer to an endnote that states, "there seems to be a virtual consensus on this general point, though with important variations."[3] The authors fail to provide any longitudinal analysis of individual opinions, and thus, leave the reader having to make a "leap of faith" that, over time, individuals do, in fact, have true or fixed opinions.
John Zaller further argues against the authors' notion that individuals have "fixed" opinions. Zaller states that individuals "do not typically carry around in their heads fixed attitudes."[4] Instead, individuals "respond on the basis of whatever considerations are most immediately salient in their minds."[5] Zaller expresses concern about this trend, saying "instability in people's attitude reports over time is one of the most deeply worrisome, if not routinely emphasized, empirical findings of modern survey research."[6] Though Page and Shapiro dismiss individual response variation because they believe in a "central tendency of opinion," Zaller's research shows that "response variation is rooted in an important substantive phenomenon, namely the common existence of ambivalence in people's reaction to issues."[7]
Page and Shapiro use their argument that individuals have real opinions to conclude that the public has real opinions, stating "if this picture of individuals' opinion (as stable and ascertainable) is correct, then at any given moment the public as a whole has real collective policy preferences."[8] Because the authors did not show that the individual has real policy preferences, then they cannot use this premise to reach their conclusion that the collective has real policy preferences. In short, the first element of rational fails for not being real.
The incoherent collective
Page and Shapiro posit that collective policy preferences form "coherent and mutually consistent (not self-contradictory) patterns, involving meaningful distinctions; that these patterns make sense in terms of underlying values and available information."[9] John Zaller's research and two examples in The Rational Public make this argument difficult to believe. First, in chapter five of Zaller's book, "Making it up as you go along," Zaller describes "response effects" as a term that "refers to cases in which seemingly irrelevant features of the questionnaire design affect the responses given."[10] In other words, how one asks the question often determines how the individual will respond. This phenomenon directly counters Page and Shapiro's view that the collective can be coherent, consistent, and able to make meaningful distinctions. Further, this wording problem is especially troubling given that Page and Shapiro's research is based on the accumulation of data from "thousands of questions asked in many hundreds of surveys," from which the questions were "identical" but not the same.[11]
Zaller provides several examples that lead him to believe that "the differences in wording of questions can determine how people think about and hence respond to issues even when, as here, the denotative meanings of the competing wordings are exactly the same."[12] Zaller provides a clear demonstration of this occurrence in a survey question that asked respondents which program they would favor upon hearing about an unusual disease that is expected to kill 600 people. In one survey, 72 percent of the respondents selected program A, the program that would save the lives 200 people. In a second survey asking the same question but with differently worded responses, only 22 percent of respondents favored the (same) program in which 400 people would die. This 50 percent reduction in opinion hardly reflects a public that can make meaningful distinctions.
Page and Shapiro also provide examples in their book that make the reader question the public's "mutually consistent pattern." In chapter four, the authors allege to provide "concrete examples of how public opinion meets our criteria for rationality."[13] Their example of public opinion on education, however, shows that from 1970 to 1980, the percentage of respondents that said we should "do more on education" declines by fifteen percent, while those saying we should "spend more on education" increases by ten percent.[14] Next, the authors introduce an example about "cities and race" and describe the "urban crisis" of the late 1960s, "which came to be associated with the poor conditions under which blacks and other minorities lived."[15] Their data shows a consistent 20 percentage point difference between those that said that government should spend "more on problems of big cities" than the percentage that said that government should spend "more on improving conditions of blacks." This difference remains constant in nearly every year from 1972 to 1989. Despite the clear similarity of meaning in each question about the "urban crisis," the public does not respond similarly to the questions.
Zaller's demonstration of "response effects" and two of Page and Shapiro's examples illustrate the difficulty, not the ease, with which the public can make clear distinctions among policy alternatives. Further, these examples display not only a lack of consistency in opinion, but also an often self-contradictory pattern in opinion. In sum, Page and Shapiro fail to show that the public has coherent patterns of preferences. Therefore, the authors' second element of rational fails since the public is not coherent.
A predictable public?
Page and Shapiro argue that "when collective policy positions change, they almost always do so in understandable and, indeed, predictable ways."[16] In chapter eight, they summarize numerous examples of when public opinion predictably changed in response to major international and domestic events. Within this chapter, they discussed the impact of wars, because "events that have most frequently and most strongly affected American public opinion have involved war."[17] Though I would not dispute their claim that public opinion changed in response to war, I find that the authors are overly generous in saying that such change is "understandable" and "predictable." The authors mention World War II, the Korean War, and the Vietnam War to show that opinion changed in "comprehensible ways."[18] However, public opinion changed quite differently prior to and during each war, leaving the reader to wonder: what is an "understandable" or "predictable" change in opinion?
In their discussion of World War II, the authors state that their data showed strong support for the mobilization of war and strong support during the war. Next, they say that the War in Korea produced "analagous, though more limited" changes in public opinion, and that the public continued to favor the war throughout the conflict.[19] Finally, the authors cite the Vietnam War and the initial increased public willingness to fight. Yet, within the first several years, public support for the War eroded, and antiwar demonstrations became commonplace across the nation. Despite the juxtaposition of these three seemingly similar examples, one cannot see "predictable" trends in public opinion. Indeed, public opinion changed, but in each instance, the degree of support prior to and during the wars markedly varied, especially in the case of Vietnam.
Though they believe changes in opinion occur in predictable "ways", the authors do not suggest how much of a response can be expected, or in which direction it should go. Following the truck bombing of the U.S. Marines in Lebanon, the public "provoked desires for the United States to remove its forces from that country."[20] Is that response predictable? How strong were the "desires"? One might argue that a predictable public response would have been to retaliate. In short, the authors' assertion that the collective policy preferences change in "understandable, predictable ways" and that "the public reacts consistently, in similar ways to similar stimuli" is weak. Though I concede that opinion does change during such major events, I would hesitate to predict how or in what ways the opinion would change. Therefore, I assert that the authors' third element of "rational" fails since change in public opinion is not "understandable, predictable," or, in any way, "consistent."
Conclusion
Benjamin Page and Robert Shapiro did not prove that "collective policy preferences of the American public are predominantly rational." Their argument failed because each of the primary defining elements of a rational opinion—that it is real, coherent, and predictable—failed. The massive research effort by the authors should be lauded, but the inability to control the wording of the questions and the lack of longitudinal analysis significantly hampered their attempt to make such bold generalizations about collective opinion. On the surface, their conclusions based on the aggregation of data seem at once impressive and plausible. However, close examination of the data reveals that the public is not rational.
Bibliography
Page, Benjamin I. and Robert Y. Shapiro. The Rational Public. Chicago: Univ. of Chicago Press, 1992.
Zaller, John. The Nature and Origins of Mass Opinion. New York: Cambridge University Press, 1992.
[1] Benjamin I. Page and Robert Y. Shapiro, The Rational Public
(Chicago: Univ. of Chicago Press, 1992) xi.
[4]
John Zaller, The Nature and Origins of Mass Opinion (New York: Cambridge
University Press, 1992) 1.
[20] Ibid. 333.