Looking at the big picture, with all these bad, odd, and partisan polls
TKS
Jim Geraghty Reporting
10/24
Day by day, more readers point to details of these polls that are just flat-out odd. John notes:
<<< On Election Day, perhaps 36% of the voters will be Democrats and 30% Republican. [Some polls have shown an even larger margin — Jim] And perhaps the first solidly red state incumbent to lose since Bush became President will occur in Montana. >>>
According to the Newsweek poll - again, of adults, not likely voters - 20 percent of Republicans support a Congressional resolution of censure against Bush for the wiretapping program.
Listen - I don't know what's going to happen on Election Day. The Republicans could end up with huge losses; if that comes to pass, I don't want readers out there saying, "Jim, you promised!"
And you'll notice that this blog usually spotlights polling data that doesn't look too bad for the GOP. This is because I think these numbers are often overlooked, underreported, and undermine the conventional wisdom; I presume you get the gloom everywhere else. If you're not getting the gloom, seek it out just as a reality check.
Having said that, my aim is to never do what our friend Kos just did, touting Quinnipiac's results in Florida to be reason for optimism while declaring the same firm's polling in Connecticut to be garbage. (Sure, somehow the company completely forgot everything it knew how to do right from one state to the next.) Similarly, feel free to blow raspberries if you ever find me calling a lead growing from 2 to 12 points a "precarious lead."
The only pollster who I out-and-out mock is Zogby, and even with him there is a school of thought that after two bad cycles, he knows his reputation is on the line this year and he won't be using as much "secret sauce."
However, day by day, I'm increasingly agreeing with TKS reader Scott:
<<< The GOP is either going to get shellacked, or stage one of the most stunning upsets in electoral history. I don't really see a scenario where the GOP loses "only" 16 or 19 seats. Either these generic polls are right, and the GOP will lose 30+, or these generic polls are totally misleading (like 2004's exit polls) and the GOP is destined to lose a small number of House seats and retain the majority.
My money is on the latter...
Outside the generic polls, I see *no* evidence *whatever* of a big night for the Dems. I see spin, spin, and spin. Nothing but spin. Meanwhile, countervailing that spin are some objective facts: lackluster performance of Lamont in "blue" CT, where anti-war sentiment ostensibly runs high; failure of Dems to leap to breakaway leads in more than a handful of House races, and those ones where they have gotten big leads are districts with idiosyncratic properties this time around; proven superiority of GOP's GOTV machine; money advantage; the apparent strength so far of GOP's advance voting turnout; etc. On top of these things you have the slowly-improving fortunes of GOP Senate candidates in TN, MT, OH, and MD.
I believe the Lamont race was the canary-in-the cave-test of the idea that the Iraq War would cause a Big Democratic Wave Over Iraq. Lamont got 51 percent of the Democratic vote in a blue-state primary, and he's going to down to defeat to an independent who supported the invasion and still wants the US to win. If there was no Iraq-centric surge among Dems in CT, how is it logical to think there will be one against the GOP among moderates and independents in TN, IN, FL, or MO?
The "facts" in support of a Dem rout boil down to polls that, if we had believed them on the afternoon of November 2, 2004, would have caused us to start saying "President Kerry." The polls were wrong in 2000, 2002, 2004, and they've been getting "wronger" with each passing election, especially when they try to measure pro-GOP sentiment. >>>
When I see things like two Florida governor's polls come out the same day, one showing a 2 point lead, one showing a 16 point lead, I'm almost ready to say, "okay, the polls are all messed up, nobody knows anything."
Or you see stories like this that John McIntyre of RealClearPolitics spotlighted:
<<< Democrat Jill Derby and Republican Dean Heller are tied in the race for the 2nd Congressional District seat, according to a poll by the Mellman Group of Washington, D.C.
Each drew 40 percent of the votes in a telephone poll last week of 400 voters in the district. Minor party candidates received 5 percent and 15 percent were undecided.
In Mellman's last poll, conducted in early September, Heller led Derby 44 percent to 35 percent.
It is not until you get down to the 8th paragraph when the reader is told that the Democrat Derby's campaign paid for the poll. >>>
Yet this year we've seen partisan polls showing their own candidates down while the non-partisan polls show them up. (One school of thought is that a partisan poll offers a more optimistic assessment for their client; they're depicting how the campaign is going when the ball bounces the way the campaign wishes; the other school of thought suggests that a partisan poll may be more cautious, because they'll lose their clients next cycle if they let their candidate get caught unaware that he's trailing.)
If all of the polls are loused up, and they're oversampling Democrats, and the pollsters aren't taking the Republican GOTV effort into account... then Election Day could turn out significantly different than the MSM expects. If I were feeling gutsier about predictions, I might stick my neck out and make a prediction in line with Barron's, or even more optimistic for the GOP. But then again, I thought the dominant issue of the campaigns would be terrorism and national security, not a 24-7 discussion of Mark Foley's pants and which Republican is secretly gay.
tks.nationalreview.com
msnbc.msn.com
dailykos.com
dailykos.com
dailykos.com
time-blog.com
users2.barrons.com |
