This study examines the New York Times' conformity to the guidelines adopted by the American Association for Public Opinion Research (AAPOR) regarding minimal information that should be disclosed about a poll in any report of the poll results. The study reveals that stories based on polls conducted by the Times, regardless of length, are more likely to comply with the disclosure standards than stories on polls from other sources. A composite compliance score created to represent the sum of what was considered the positive score for each standard shows that about half of the articles in the study met five or fewer of the twelve standards studied, and about half met more than five.
From the early polling professionals who believed that polls can strengthen democracies' to critics who caution about the media's agenda setting influence,' public opinion polling and its uses by the media have been the focus of an intensifying debate.
According to former executive vice president of the Gallup Organization, Irving Crespi, the first professional polling organizations wanted to serve both the government and the people by providing information on how the public views specific issues.3 Journalists who have been conducting their own unscientific polls even before the appearance of professional polling organizations embraced these new sources of information early on and argued that increasing reliance on social science methods will enable them to provide information that is more reliable to the public and can contribute to improving the credibility of the profession.
Critics, however, have warned about the possible shortcomings in attempts to reduce the complexity of public opinion to quantifiable data.4 Others argued that a news organization conducting its own poll takes journalism from the realm of reporting news to creating news, which can be unethical.5 The quality of the research and reporting on the findings raise further questions.
The relevance of research evaluating the press' performance in reporting public opinion polls is implied in two recent contributions to the literature on public life and public journalism. In The Good Citizen: A History of American Civic Life, Michael Schudson reasons that the legacy of George Gallup [and by extension, the legacy of scientific polling] "both as a critique of other imperfect measures of 'what the common man thinks' and as a positive resource for democracy, [the legacy of George Gallup] is great. It may be one of the enduring fruits of the nineteenthcentury idea of mass democracy; however transformed and tamed. "6
It is important to note that Schudson bases his rather positive report card on the health of American civic life, in part, on how publicopinion pollsters have measured concepts such as "trust in major institutions." For example, he reasons that to rush to judgment about citizens' lack of trust in major institutions is problematic because pollsters have not frequently linked these attitudinal measures to behavioral indices:
Polling that indicates a lessening of trust in major institutions is a much more equivocal measure of civic health. We do not know what it means. How does the answer people give to an abstract question about their level of trust relate to actual behavioral indices of trust--compliance with the Internal Revenue Service or willingness to defer to the authority of a medical doctor, a government bureaucrat, a school administrator, or a court order?7
He says that "polling questions are still designed not to afford general knowledge of public opinion, so much as to generate news," reminding readers that 125 daily newspapers financed Gallup's American Institute of Public Opinion.'
In this book, Schudson reasons that Gallup, and implies that most pollsters,
presuppose a view of opinion as virginal, presconstituted, prepolitical, presocial... this view has dominated scholarly discourse in political science and economics. …