The Illusion of Public Opinion: Fact and Artifact in American Public Opinion Polls. George F. Bishop. Lanham, MD: Rowman & Littlefield, 2004. 248 pp. $80 hbk. $27.95 pbk.
The Voter's Guide to Election Polls, 3d ed. Michael W. Traugott and Paul J. Lavrakas. Lanham, MD: Rowman & Littlefield, 2004. 216 pp. $65 hbk. $19.95 pbk.
Each of these volumes provides timely and accessible reminders of the ways that polls can perplex pollsters, reporters, and voters. Neither book sifts the last election's crop of polling problems. But the social, political, and methodological issues raised in each volume will be fresh and familiar. Take "moral values," for instance. Remember the full-page, pre-election ad in the New York Times that questioned the Gallup Poll's alleged religious bias based on George Gallup Jr.'s outspoken evangelical Christianity and his polling organization's alleged pro-Bush bias? Or the closed-end exit-poll questions about "moral values," which mainstream U.S. news media cited to explain the incumbent's victory before foreign media noted the factor's decline from previous presidential election polls? The underlying issues, if not those particular events, are all here.
Such questions attach to mass opinion polls dating back to the 1930s, as Traugott, a communications professor and senior research scientist at the University of Michigan, and Lavrakas, a former journalism professor and survey director now at Nielsen Media Research, recount in an early chapter in their popularized approach. George Gallup Sr. got his start by preaching the gospel of polls as "the voice of the people" and by showing that the dominant Literary Digest poll was biased in favor of well-to-do telephone users and automobile owners in the middle of the Great Depression.
Problems still persist, as the authors summarize in ten chapters of a readable reference organized in question-andanswer format. The twelve-page chapter 8, for example, addresses the question, "How do media organizations analyze polls?" (Answer: Much data collection, little data analysis. Television? Even less.) Concise descriptions of marginals and frequencies-and their misuse by news media-are here. So are brief primers on the overuse of polls in political horse races and underuse or misuse on political issues.
This guide to a pervasive phenomenon (the authors estimate that between 5,000 and 10,000 polls were conducted during the 2000 election cycle) serves a practical purpose for private citizens and news practitioners alike. It is a convenient digest for evaluating polls, whether scientific or pseudo-scientific. And this digest is a timely tool for coping with increasing waves of polls proliferating in media old, such as newspapers embracing pseudo polls for marketing purposes, and new, with Internet pop-up polls like spam.
This second edition includes polling developments since the 2000 election, problems with telephone surveys in the era of the cell phone, and descriptions of problem-plagued media partnerships in election polling.
For a more substantive critique of polling, The Illusion of Public Opinion dishes the dirt in a style that is as readable as it is substantive. Bishop, a professor of political science at the University of Cincinnati, includes nine footnoted chapters, eleven pages of references, lots of charts and tables, and two bookmark-worthy Web sites (http://www.pollingreport.com and http://brain.gallup.com). Much of the cited research is Bishop's own, supplemented by a rich literature review of others' work. His voice and writing style at times approach the tone of journalistic scoop, but it is more readable than sensational. His conclusions form an indictment of mass opinion polling and of news media reporting and analysis of the same. …