With almost every poll I’ve seen in the past two weeks showing Bush either gaining (in some cases pretty significantly) or holding steady with Kerry, MSNBC headlines their site with


The story shows Bush with a 45%-44% LEAD, for crying out loud. Never mind that odds are good that Bush will lose the popular vote even if he wins in the electoral college.

Luckily for us, they don’t mention anything about recent polling numbers. We’d hate for any context for these numbers to be available. We can’t really compare these number to other polls. If they’re meaningful at all, it’s only when compared to the same (or very, very similar) poll conducted by the same organization.

It’s sort of like the cheap scale I have in the bathroom closet. I don’t even pretend to believe that it’s accurate. But if the same scale shows me two pounds heavier this week than last week, then I need to lay off the junk food and run a few extra miles. If the neighbor’s scale shows me down five pounds, it doesn’t mean a thing.

As for the “down but not out” headline, it’s a quote from a Democratic pollster. How’s that for objective reporting? Here’s the graph:

“Kerry has a ways to go to be able to take advantage of Bush’s weaknesses,” said Peter D. Hart, a Democratic pollster. Still, he noted, the survey numbers suggest Bush “has to come back and win this election. He goes into the summer period down but not out.”

Wow. That’s a real endorsement. Kerry has a “ways to go” and the leader of the current poll “has to come back” to win.

I’m not suggesting that Bush is in good shape. I’m not suggesting that Republicans should feel assured. What I’m suggesting is that this analysis is about worthless, especially considering it’s headline material on MSNBC.

The quote that’s been going around a lot lately (I think from Mort Kondrake, but I’m not sure) is “The media can’t tell you what to think, but they can tell you what to think about.”

By and large, I agree with that statement. But the headline games and el-cheapo “analysis” like this can tell people WHAT to think.