THE BLOG

Obama's Trip Bounce: The Media's Obsession with Polls Leads to a Bad Case of Premature Pontification

05/25/2011 12:40 pm ET

Isn't it strange that Barack Obama didn't get a bounce from his wildly well-received overseas trip? Oh wait, maybe he did. But, hey, it was just a small bounce. Or was it more of a bump? Perhaps a bouncelet? A hop? A ricochet? A swelling? Or was it a rash? In which case, if it persists for more than two weeks, should he see an electoral professional, or just declare victory?

Of course, almost all of this analysis is based on polls taken before the end of Obama's trip -- a serious case of premature pontification.

But that didn't stop many in the media from weighing in, building edifices of soaring opinion on the shakiest of data.

By July 23, four days before the end of the trip, U.S. News and World Report's Michael Barone had seen enough to render his bounce-analysis: "The assumption among most observers seems to be that Barack Obama will get a bounce in the polls from his trip to Afghanistan and Iraq, the Middle East, and western Europe. But it's not apparent in the polls that have come in to date."

He doesn't say who the "most observers" are that made this "assumption." And while he cites several polls and lets us know that he puts "more weight on the Rasmussen poll," he doesn't say why. Perhaps he likes that Rasmussen polls "likely voters" while others, such as Gallup, poll "registered voters." He does however advise: "Stay tuned. I don't think the voters' decision-making process is complete yet." I'll make a note of that.

Fox News' poll, meanwhile, taken over the two days Obama was in Jordan and Israel, revealed that "the significant news coverage Barack Obama is receiving on his foreign trip has not translated into a bounce in his numbers." You can hear their disappointment in every word, can't you? This lack of bounce translation (perhaps it was done by Maliki's interpreter) was based on 900 registered voters willing to answer the phone and talk to Fox's pollsters "in the evening."

Over at Time, also four days before the trip was over, Joe Klein bypassed the question whether Obama did or didn't get a bounce -- why bother, it had already been decided by the conventional wisdom board of directors that he didn't -- and got right down to laying out the theories of why he didn't:

"Lots of speculation on the web, and in whispering circles, about why Obama's foreign trip -- a slam-dunk success substantively and in photo-op terms (Obama laughing with Petraeus in the helicopter was the best)--hasn't resulted in a polling bump. The emerging conventional wisdom seems to be that the trip is a bit too grand, too...presumptuous and voters are wary of that."

Klein doesn't buy the conventional wisdom explanation (trip was presumptuous) of why the conventional wisdom about the trip (no bounce) was right. He's got an unconventional explanation of why the conventional wisdom is right:

"People may be thinking, what on earth is Obama doing over there when we have so many problems back home? Why isn't he talking about the economy?"

Perfectly reasonable -- if Obama didn't get a bounce. But not if he did. Comparing two sets of daily tracking polls -- Rasmussen and Gallup -- taken before Obama's trip started and (hey, here's a novel idea) after the trip actually ended, we see that Rasmussen shows a gain of 5 points for Obama, while Gallup shows a gain of 7. Doesn't that sound like a bounce to you?

Media insiders now talk about polling brands the way average Joes talk about their favorite beer. "My producer is a Quinnipiac person, but I'm more into Gallup." I suppose if they married, they'd have to raise the kids CBS/New York Times.

Not surprisingly, most journalists' favorite polls are the ones that legitimize their favored campaign narrative -- which they've often simply pulled out of their Rasmussen.

After all, nothing is more fun for journalists -- and better for ratings -- during campaign season than constantly changing the narrative: "Obama is up!" "McCain is surging!" "Obama is coming back!"

Fortunately for the new-narrative seekers, polls in America have the shelf life of bread in France. They get their freshly baked daily baguettes; we get our freshly baked daily polls.

And now we even have superpolls -- polls of polls. The CNN Poll of Polls has Obama with a lead of 44 to 41. The Real Clear Politics poll average has Obama leading 46.7 to 41.7. And pollster.com has him up 44.7 to 41.5.

Maybe what we need is a Huffington Post Poll of Polls of Polls, which will have a margin of error of +/- pi divided by Schrödinger's cat.

In about 13 weeks, the only poll that really matters will be taken -- among 142 million registered voters, and not 900 people so bored with life they are willing to interrupt their dinner to talk to a pollster.

But until then, since the media will no doubt continue to subject us to day after day of polling results and highly suspect analysis, can they at least make it clear precisely what kind of polling data they are basing their analysis on? What is the response rate on the poll? Is it a poll of "likely voters" or "registered voters"? What is the breakdown of Democrats, Republicans, and independents?

I'm fine with people using polls as an amusement or as a fun way to start a dinnertime conversation, but to run around making grand statements based on dubious data is the height of foolishness. It's beyond time to break the cycle.

For more polling absurdity, check out this terrific Seth Colter Walls piece.