politics

Polling Gone Wild

July 21, 2004

Anyone who has even remotely followed the campaign this year knows that the amount of polling data flowing out of each camp has been about equal to the amount of vitriol. In short, neither has offered much enlightenment.

In an article today in The Hill, Democratic strategist Mark Mellman reports that voters have been subjected to the results of five polls per week, on average. “So,” he writes, “do we know a lot more about the electorate than we used to? I hesitate to say it, but all this early polling certainly leads to confusion and perhaps a net loss of knowledge.”

Call it the white noise of the 2004 election.

Regular readers of Campaign Desk know that we share a healthy skepticism about the relevance of polls and their use by those who sponsor — and write — about them.

The key to understanding polls, according to Mellman, is to average them over the course of time. “[L]ooking at any one poll or at any pair of polls and trying to draw conclusions can be a fool’s errand.”

Which leaves us wondering: Why do a dozen or so prominent press outlets keep pollsters in business by shoveling so much money at them? Here’s an idea, gratis, from Campaign Desk: Use that dough to hire a few birddog reporters, instead of buying yet another poll that doesn’t square with yesterday’s and won’t square with tomorrow’s.

Sign up for CJR's daily email

–Susan Q. Stranahan

Susan Q. Stranahan wrote for CJR.