E-Mail This Article to a Friend Waves of anxiety (continued)

Nor is Soto’s critique restricted to bloggers and liberal activists. Pollster John Zogby, who never fails to tout the fact that his surveys in 1996 and 2000 came the closest to predicting the final outcome, wrote a piece for the Financial Times on September 7 criticizing a Newsweek poll, taken right after the Republican National Convention, that had given Bush an 11-point lead among likely voters. Zogby, by contrast, had placed the lead at just two points. The difference, Zogby explained, was that — similar to Gallup — respondents to the Newsweek poll were 38 percent Republicans, 31 percent Democrats, and 31 percent independents.

"While party identification can indeed change within the electorate," Zogby wrote, "there is no evidence anywhere to suggest that Democrats will only represent 31% of the total vote this year." He asserted that his own polls use "a party weight" of 39 percent Democrat, 35 percent Republican, and 26 percent independent — far more likely to reflect actual turnout on Election Day, he argued.

At first glance, MoveOn, Soto, and Zogby would appear to have an open-and-shut case. But there is a flaw in their argument. The truth is that Gallup — and Newsweek — did not actually do any party-weighting. Rather, they contacted registered voters and determined who among them would be likely voters by asking such questions as whether they had voted in the last presidential election, how closely they were following the current campaign, and the like. These likely voters were then asked for whom they planned to vote and with what political party they identified.

For most people, party identification is a fixed part of their identity, as much as gender or ethnicity. But for a small but statistically significant subset, party identification shifts depending on their political mood of the moment. For instance, some people — upon telling a pollster they’re planning to vote for Bush — will follow that up by saying that they’re Republicans. Two weeks later, these same people might switch to Kerry — and start calling themselves Democrats.

"Many people think of party identification as allegiance to the candidate," David Moore, senior editor of the Gallup Poll, told me. By this line of argument, Zogby’s party-weighting system understated Bush’s lead by rigidly sticking with a fixed percentage of self-identified Democrats, even though the actual proportion of people calling themselves Democrats was dropping as Bush’s lead was rising. Similarly, this week, with Kerry’s numbers rising after his strong debate performance last Thursday, the percentage of self-described Republicans will almost certainly drop. (The latest Gallup poll, released this past Monday, finds that Kerry and Bush are now tied, with 49 percent each.)

Responds Soto: "I’m not going to disagree with David Moore, because he’s doing it for a living and I’m not." But, he adds, "On its surface it sounds ludicrous to me. There’s nothing wrong with being a lifelong Democrat who’s voting for Bush and telling a pollster that."

Andrew Smith, director of the University of New Hampshire Survey Center, says party-weighting is a hotly debated topic among professional pollsters. Zogby’s methods, Smith contends, have failed as often as they have succeeded. As for Gallup, Smith says, "they have historically been, if not the most accurate, then one of the most accurate."

If nothing else, MoveOn.org could have spent its members’ money more wisely. Even Soto wishes the organization had targeted local media in a few swing states rather than lavishing its cash on the Times. Soto also isn’t thrilled about the sneering tone regarding George Gallup Jr.’s religious beliefs. "I think it went overboard, simply because I try not to be a conspiracy nut," Soto told me. "I’m not sure bringing Mr. Gallup’s family or religious background into the matter was helpful." (Gallup’s remarks about "responding to God" also reportedly were taken out of context: now retired, he is involved in surveys about Americans’ religious beliefs.)

MoveOn.org spokesman Trevor Fitzgibbons did not respond to inquiries by phone and e-mail this week.

AIR AMERICA Radio, the five-month-old liberal network that’s been only a rumor in Greater Boston, has taken a significant step closer to competing for local listeners. This past Monday, WKOX (AM 1200) and WXKS (AM 1430) began simulcasting a line-up of progressive political talk shows, mixing Air America programs such as Morning Sedition (6 to 9 a.m.) and The Al Franken Show (noon to 3 p.m.) with syndicated liberal talkers Stephanie Miller (9 a.m. to noon) and Ed Schultz (3 to 6 p.m.).

Unfortunately, neither station’s signal is all that great, and both are required to power down to such low levels after sunset that they can be difficult to tune in, even inside Route 128. That’s certainly going to hinder Air America’s The Randi Rhodes Show, which runs from 6 to 10 p.m.

"I’m really happy we’re on in Boston. Now the calls will change from ‘How come we’re not on Boston?’ to ‘How come the signal’s not better in Boston?’ And that will be a good thing for me. It’s time," says humorist Barry Crimmins, a Phoenix contributor and former Boston comedian who writes for and occasionally appears on Rhodes’s program. (Crimmins’s new book, Never Shake Hands with a War Criminal, will be published next month by Seven Stories Press.)

And if you liked that joke, you’ll love this one: WKOX and WXKS are owned by Clear Channel, the Texas-based radio conglomerate best known for yanking the Dixie Chicks off its country stations after lead singer Natalie Maines had the temerity to criticize George W. Bush.

"We live in America, where it’s all about the best business model, I guess. But it does seem a little ironic," agrees Joe Mazzei, marketing director for the two stations and for KISS 108.

Dan Kennedy can be reached at dkennedy[a]phx.com. Read his Media Log at BostonPhoenix.com.