With just a few weeks to go before the Iowa caucuses, polls are providing pundits and political junkies with fresh data to spin out a new round of the usual "who's up, who's down" campaign coverage. But while the press seems settled on a new narrative for the campaign, journalists should recall what the polls told them last time around about who would likely win the Iowa caucuses.
The tone of coverage of the Democratic race seemed to shift when a Nov. 19 ABC/Washington Post poll of likely caucus-goers showed a tight race among three candidates: Barack Obama, Hillary Clinton and John Edwards.
The difference from the previous survey was within the poll's margin of error, so the actual data said very little. Much of the media seemed to think otherwise. "The ground may be shifting," announced NBC anchor Brian Williams. The Los Angeles Times called it "a shift in momentum in this crucial state" - in an article that boiled the race down to just two candidates, Clinton and Obama.
The Washington Post's write-up was downright confusing - the Post mentioned the results were "only marginally different" from their poll several months prior, yet nonetheless pointed to "significant signs of progress for Obama - and harbingers of concern for Clinton."
On ABC, reporter Kate Snow mentioned something most of her colleagues seemed unconcerned with: the fact that these polls actually tell you very little about the outcome of the race. Snow recalled that "four years ago, John Kerry - who eventually was the Democratic nominee - he was polling in Iowa at 4 percent."
Indeed, campaign reporters should all remember the lesson of the 2004 Iowa caucus. A little more than a month before Iowa Democrats actually caucused in January, the poll-obsessed media had narrowed down the field to two "front-runners": Rep. Richard Gephardt of Missouri and Vermont Gov. Howard Dean.
"Two See Iowa as Crucial Battleground," announced The Washington Post on Nov. 29, 2003, billing the race as a "fight rich in substance and symbolism." A Nov. 9 Post report said that Dean was "for the first time, threatening to pull away from the pack," and even discussed his "opening for a quick-kill strategy" by winning Iowa and New Hampshire.
The polling was presumably a key factor leading reporters to reach such conclusions. A December 2003 Pew poll of likely Iowa caucus-goers showed Dean leading the pack with 29 percent, followed by Gephardt at 21 percent.
Kerry was in third with 18 percent, followed by John Edwards at 5 percent. A Zogby poll from around the same time had a closer race between Dean and Gephardt (26 to 22 percent), with Kerry and Edwards picking up 9 and 5 percent, respectively.
And what happened when Iowa Democrats actually caucused? Kerry won with 37 percent, followed by Edwards at 32 percent. "Front-runners" Dean and Gephardt finished with 18 and 11 percent, respectively.
The point is not just to note that polls at this stage are hardly predictive - though the media acknowledging as much would be a start. Nor is it to wish that the national press would simply work at finding a better method of declaring which candidates are "front-runners," and whose campaigns aren't worth your attention.
The more fundamental problem for the press - and for American democracy - is that the media's over-reliance on polls encourages a kind of political conversation that prioritizes strategic consideration and tactics over substance.
A recent study from the Project for Excellence in Journalism confirmed that much of what passes for campaign journalism focuses primarily on the tactical dimensions of the race (like poll results and fundraising) and not on the actual policy differences between the candidates.
In a recent New York Times op-ed, former ABC News political director and current Time magazine editor-at-large Mark Halperin admitted that most political coverage is built around the notion that you can judge candidates' potential to be a good president based on how well they run their campaigns.
Halperin admits he was "wrong," and suggests a change of course: Journalists "should examine a candidate's public record and full life as opposed to his or her campaign performance." What a concept. But then Halperin added a strange qualifier: "But what might appear simple to a voter can, I know, seem hard for a journalist."
Halperin seems to be saying that if you think it's hard to cover the substance of electoral politics, it's a good bet you're a campaign reporter.
That's bad news, to say the least - and makes it hard to imagine journalists are going to change any time soon.