Zogby and Bismarck's Sausage
By streiff Posted in User Blogs — Comments (1) / Email this page » / Leave a comment »
Otto v. Bismarck famously opined, "If you like laws and sausages, you should never watch either one being made."
His observation on the process of drafting laws seems equally appropriate to the Zogby polling operation. As an interesting development in the SD Senate race we get to look inside the Zogby' method. The sight is not a pretty one. It is something between a dowsing rod and reading goat entrails.
According to the Washington Post panic set in on the Daschle campaign because of a 3-point lead for Thune:
For starters, they say, the Zogby poll published in the Rapid City Journal is flawed. It showed Republican Thune leading Daschle, 48.5 percent to 45.5 percent, just within the margin of error. At first, however, the poll had shown an even larger Thune lead, which seemed so improbable that the pollsters adjusted their voter turnout estimates and arrived at the narrower gap.
So what were the original numbers that put Daschle's knickers in a knot? The Rapid City Journal clears up the mystery.
Poll results can differ significantly depending on how pollsters define "likely voters." Zogby's initial report, for example, showed Thune leading Daschle by 49.6 percent to 43.9 percent, a result Pfeiffer called "absurd," adding, "No sane person in America believes this is a six-point race."
Diedrich led Herseth 48.8 percent to 44.5 percent in those first results, and Zogby said that gave even him pause.
Zogby said he arrived at those results using the ratio of Republicans to Democrats in exit polls conducted during the 2000 presidential election. Those results were 49 percent Republican and 33 percent Democrats.
Democratic turnout could have been low in that election for a number of reasons. Thune overwhelmed his Democratic opponent for the House that year, there was no Senate race and there was no significant Democratic get-out-the-vote effort.
So with a wave of his magic wand John Zogby takes a Thune lead that was outside the margin of error and created the story of a race that was too tight to call.
However one looks at it, it is not science, it is not mathematics and clearly demonstrates how little we know about the polls that drive the news.
Although I wouldn't count out the possibility that Mr. Zogby's adjustments went the wrong way entirely, I don't think he's wrong in principle. There might very well be higher Democratic turnout in South Dakota on Tuesday than there was four years ago.
I wouldn't mind a Thune victory at all, but I've grown to be wary of good news, especially unexpectedly good news.
I think the Daschle/Thune race will, like the national election, be a squeaker. I'm not certain Thune will win (I'd give him better than even odds, but I'm not certain), but I think it will be close.
'Course, I don't live out there and am not keeping up with the race, so there could be a welcome turn against Mr. Daschle. I won't complain if I turn out to be wrong, and Mr. Thune finds himself with a comfortable margin of victory on Nov. 3rd. But I come by my cynicism honestly, so I hope you'll forgive me a little skepticism.