Sunday, October 26, 2008

Bold Predictions

"It’s over." That’s how New York Times columnist Charles M. Blow began his column on October 17th, two and a half weeks before Election Day. His column is titled "Nov. 5, 2008" (my birthday!) and predicts that Obama will crush McCain on Election Day. He even describes a scene of Obama waking up on Nov. 5th:
"President-elect Obama (yes, get used to it) could wake up that morning as one of the most powerful presidents in recent American history. Not only is his party likely to maintain control of both houses of the Congress, it could dramatically strengthen its hand."
Blow goes on to talk about the "probability" of Obama receiving a warm international welcome and "possibility" of Obama appointing several justices to the Supreme Court. Man, this guy must be feeling pretty good about his chances to shamelessly play Oracle in the pages of America’s newspaper of record.

Even David Brooks, a conservative columnist for The New York Times predicted that Obama would win by 9 percentage points in a Q&A with students of Massachusetts College of Liberal Arts.

Let’s be real. No reputable journalist would risk being wrong so publicly unless they felt good about their information. These predictions openly rely on recent polls and other studies that consistently predict Obama’s victory over McCain by more than a few percentage points. NPR’s final survey before Election Day showed Obama had a "commanding lead in battleground states and on all key elements of campaign."

First of all, it's unsettling that that journalists buy into the hype of these polls and use them as justification for their Delphic predications. This shows they're disregarding the king caveat: polls are a snap-shot in time based on a tiny sample and have a history of being incredibly wrong more than every once and a while.

But that's not the worst of it. Journalists seem to let polls affect the general tone of their coverage. In an interview with PBS’s Jim Lehrer, Associate Director of Project for Excellence in Journalism Mark Jurkowitz talked about a recent study that concluded, "if a candidate is perceived to be and is seen as doing well in polls, if the strategic dynamic of the campaign is favoring him, then he tends to get better coverage." Polls are pseudo-events, given legitimacy by the media's coverage of them, but not terribly legitimate in their own right. Such questionable numbers shouldn't hold that much sway on a journalist's work.

Regardless, journalists shouldn’t be in the business of crystal balls and soothsaying anyway. Ideally, journalists serve their readership and if in an election season the press’s job is to educate the electorate so they can cast an informed vote, predictions serve absolutely no purpose. There is no sense in journalists telling the public who will win, when the public is the one who will end up making that decision.

If the news media insists on discussing the polls, why not use the opportunity to educate their audience about how best to consume and understand polls?

1 comment:

M. Dery said...

As usual, a tag cloud of interesting ideas, here. Scattered thoughts, riffs, nitpicks:
Yes, polls are flawed instruments, but it's a reductio ad absurdum argument to suggest they're virtually worthless, I think. It's SOP in chaos science to use snapshots of turbulent systems to predict their behavior in the immediate future. Reading the tea leaves of multiple polls may not guarantee odds a Vegas bookmaker would love, but they're valid enough to build an opinion piece on. Which brings me to my next point: you elide the difference between op-ed writers like Blow and Brooks and "journalists." These guys aren't reporters, they're thumbsuckers. Reading the mass mind is what they do for a living. Would have loved to have seen you bury them under a long list of failed predictions. An examination of their respective track records would have made your point more convincingly, I think, than questioning the inherent flaws in polling, although they're certainly there to be questioned. But I like your larger point that "horse race" coverage that spotlights the precognitive powers of pundits is less useful, to a democracy, than coverage that delves deep into the candidates' positions, records, philosophies, et. al.