This weekend, I uncharacteristically listened to This American Life, which had an interesting show called "The Facts Don't Matter". In the first segment, the essay's author -- I can't recall her name -- goes to visit Zogby International while they are doing a poll before the Wisconsin primaries. She is troubled that the Zogby methodology nearly requires respondents to provide specific answers to questions -- even if they don't know what they think, or don't know anything about the issue. There's a specific example of a man who voted Republican, but the pollster pushed him -- "What Democrat would you pick if you had to vote for a Democrat?" -- so the guy randomly says "Dean". So this man is misleadingly represented as a Dean voter. The pollster is allowed to push the "I don't know" button, but only if the respondent says "I don't know" -- you can't offer that as a choice. In other words, the poll is structurally laid out to prevent ambiguity.
The segment had a brief interview with Peter Yankelovich, a competing pollster, who has a methodology for collecting "mushiness" of opinion. It's essentially a sliding scale that gives you a numeric range for how committed the respondent is to the particular answer. Not perfect, of course, but it gives a sense of how "hard" support is on an issue. Yankelovich developed this system in the 1970s, but found that journalists wouldn't cover it. It just took too much time and effort to report the mush factors.
It's almost a metaphysical requirement in journalism -- there must be no ambiguity, and there must be a grand narrative. The David Brooks column from the other day is an example of this metaphysics in action: according to the media mythology, nuance is risible, and moral conviction can only come in the form of pure black-and-white perception, never mind the facts. The whole methodological structure of polls -- and how they are used to shape the coverage -- is designed to provide a "scientific" elimination of vagueness and complexity.
Posted by Chris at March 15, 2004 09:45 AM