May 17, 2007

ONA Seminar II: Parsing Polls Properly

If any one seminar looks the most educational for bloggers, this would be it. It features Jim Pugh of the Wisconsin Manufacturers and Commerce group, Paul Maslin, who has worked with various political campaigns on polls, and Ann Selzer of Selzer & Co.

Polling has gotten tougher and tougher. Seventeen percent of the people in the US no longer has landlines. More people refuse to answer polls. Paul Maslin says we should be wary, and says that 75% of them will be seriously flawed to useless.

Ann asks what political coverage would look like without polling. It would reflect campaign spin much more than real data. It gives media a clue as to where to probe.

Jim says that we should also be wary of media polls. They are performed by amateurs for the most part. Campaign polls are usually more reliable, if you can get to the data. (Paul conducts the Iowa Poll, which is usually pretty reliable for the Iowa caucuses.)

This so far features a lot of generalizations and common sense. Check the sample, for instance, and the methodology. Hardly groundbreaking, but then again, maybe people don't think to do that. Everyone emphasizes the fact that professional pollsters do better jobs than amateurs, and that the pros get snapped up by campaigns -- which leaves unsaid the fact that the polls we get to see are not the highest-quality surveys on the market.

What about cell-phone respondents? Ann says they worry about the results of such polls. The most reliable polls come from door-to-door surveys. It's also important to know the sample. A large number of households are unrelated, unmarried residents, such as college roommates in apartments, and that can skew results too.

One good point: the cell phone question keeps getting more and more pertinent in every election cycle, and it has to be resolved soon if polls are going to remain reliable.

Now the discussion is drifting off a little bit; Paul is criticizing media coverage of the presidential campaigns, saying that the media is covering both parties' primary races as three-person contests. He quipped that Romney may count as three and four due to his Mormonism. A bit snarky for this crowd, which didn't laugh as much as they collectively groaned. He's predicting another bloody and tight battle for the presidency in 2008, and wants to see more in-depth polling of Republican voters. "Why is Arnold Schwarzenegger so darned popular? ... How do they hang onto Bush but move away from him significantly enough to win?"

Iowa may represent the last chance for second-tier candidates. The polling reality there is different from the national momentum. It's a long way to the caucuses, and there's enough time left to see what effect the secondary candidates can have. In that sense, national polling runs risks of overshadowing the actual mechanics of winning a nomination.

Online polls get bashed, and deservedly so. You can't control for the sample, you can't control for the exposure, and you can't control for abuses. Other than that, they're great. We see this quite a bit in the blogosphere, where straw polls routinely return results that clearly conflict with the reality of the political races. Robopolls also get a lot of criticism, but also some defense for quick rollout response polling.

Which polls should be routinely ignored? I think the questioner wanted a brand name, but the panel's answer was any poll that refuses to release its sampling and methodology. If it can't be peer-reviewed, it should be disregarded -- just like scientific research. Good point, and one bloggers should consider and enforce.


TrackBack URL for this entry:

Comments (3)

Posted by Lod Whorfin | May 17, 2007 12:23 PM

Captain Ed:

Any idea what's up with
When I come to your page, it opens itself up in a new window. Thanks

Posted by unclesmrgol | May 17, 2007 12:41 PM

Why is Arnold so popular?

Well, in a blue state, a blue Republican can be popular. Arnold is pro-illegal immigration, pro-abortion (except partial birth and non-notification of a minor child's attempt to seek abortion), gun control, gay adoptions, believes that humans cause global warming, and supports universal health care (his position would obligate every citizen to purchase a state-defined health care contract). Certainly these positions are influenced by the views of his left-wing Democratic wife, Maria Shriver of Kennedy fame.

Posted by Carol Herman | May 17, 2007 4:47 PM

In 1966, Ronald Reagan was elected governor of California. Something of a Blue State; but down South from Anaheim towards Mexico, a very healthy conservative environment.

And, Schwartzenegger, to win, at first, had to bounce Grey Davis out of office.

Who put the money up for that? A very conservative guy. His name escapes me, now. But it was his $3-million dollars that bankrolled the recall effort.

ONLY after the recall was successful, did Schwartzenegger jump in.

And, I've already told ya. On that election day, when the networks were interviewing people leaving their polling station; I remember one interview that made the cut. And, was shown. The man was asked who he voted for. But he prefaced his answer, "that he preferred the conservative guy, whose name was also up on the ballot." However, he voted for Schwartzenegger because he wanted a republican to win.

California, Ohio, Pennsyvania, and probably New York, are behind Guiliani, now. Those are the big states. While the primaries are changing as well. I don't htink Iowa and New Hampshire will be worth chump change.

It's also possible, though there are no polls telling you so, that as Bush's popularity tanks, he's lost a lot of republican voters. Along with the Independents, and the "blue collar democrats" who all came along to make Ronald Reagan's runs for the White House such a success. (1980. 1984.)

While the GOP, itself, in 1976 decided to go with Gerald Ford. He didn't make it.

Sometimes, in politics, the best lessons come with a kick in the pants. The GOP has experienced many of them. It was, for instance, a shame they picked John Dewey to run in 1948. He got creamed in 1944. Still, one newspaper did run with the headline "Dewey won." And, Truman gleefully held it up for all to see.

When Ike came down the pike, Truman ran home.

The process of picking candidates was once the perview of the "back room boys." Sucking on cigarettes and cigars. As out front the balloting would take place, and not a single candidate reached the number needed to be selected.

I have no idea who the "record holder" was. I seem to think it took 23 ballots to get to a winner, once. But I don't remember name, Place. Or year.

Just that getting to the nomination is a tough fight.

Does our media, today, look to make this look easy?

If you're so smart, that Schwartzenegger only holds "blue cards," I think you're mistaken. I think he won because he knew how to approach elections with victories in mind. Others failed. Since he couldn't walk into the governorship, alone, without voters giving him the edge.

It's not color specific. But sure. For a newcomer to politics, Schwartzenegger's family provides lots of behind the scenes experiences; where knowing the people to hire, is left to the "professionals." Again, not something with a blue label on it.

Then? Ya gotta stay popular. Bush, here, did not.

And, because of Bush, ahead, the water gets riskier for others within the GOP. If you think in politics you only have to worry about your oponents, I think, not.

I think just like backstage on Broadway, the show out front is one thing. And, the bouncing ego's of the stars a whole other ball game.