August 30, 2006
Behold, the Poll of Polls
Observers of the 2004 Election may recall the Real Clear Politics "poll of polls" strategy of taking available polls, and averaging them together in a running assessment of where a given race is at. Such a tool has now been released for the race between Mike McGavick and Maria Cantwell. It's definitely worth checking in on as the election unfolds.
The current poll average is skewed by a recent Survey USA poll, a firm I've noted my preference for here. This poll does raise an eyebrow given that it's out of the range of other major pollsters, and since Survey USA is kind enough to provide cross tabs, I gave it a closer look. One challenge is that until after the primary, Survey USA is using only "registered voters," rather than the more reliable "likely voters" used currently by Rasmussen and Strategic Vision for example. But, the real tale of this poll is that it says Cantwell is leading 51%-45% with rural voters, and 54%-42% in Eastern Washington. Anyone who wants to believe that is a good sample can be my guest.
Another smaller note on the Zogby poll. It uses an interactive Internet methodology that has its critics. That criticism is interesting, though even at that, one can still capture trends from long-term Zogby data so it's still of some value. However, this tidbit in Zogby's description of the race caught my eye: "Maria Cantwell remains in a strong position...thanks to a campaign strategy making much of gay marriage." Um, ok.
Every poll will have its warts - for example no pollster is yet including 3rd party candidates - but the "poll of polls" approach is always an interesting tool. It did fabulously well showing the actual trends and results going into the 2004 election, both for the presidential race as a whole, and the "battleground states" as well. Keep an eye on it in this race.
UPDATE: Also from Real Clear Politics is this blog post on Cantwell. I essentially agree with their summation of the race:
Right now, Cantwell's lead over Republican Mike McGavick stands at 9 points. That number is inflated a bit by a new SurveyUSA poll showing her with a 17-point lead. Three other polls released in the last two weeks show the race between 5 and 8 points. The first batch of polls after Labor Day will show exactly where this race stands. If Cantwell busts out to a double digit lead in the RCP average, the race is probably over. If McGavick can keep it close - meaning 5-7 points or better - with the debates and a bit of luck, he'll make it to election day with at least a chance of scoring an upset.
Posted by Eric Earling at August 30, 2006
08:10 AM | Email This
1. Again, I say, (unfortunately, I guess) it's Cantwell by 4 in November.
2. polls are for lemmings; do i vote blindly? no; i read, try to remember who did what to me in legislation and courts and then decide; left and right nuts like polls; the lazy like polls; sensible people are busy making up their own minds;
3. It sounds like Mike McGavik's got his work cut out for him. But then that's not really news anyway. Maria Cantwell has held that post for too long. I doubt that the poll results have anything to do with gay marriage, especially in rural and eastern Washington.
4. Polls...Schmolls....It's not over, til it's over.
Remembering Truman vs Dewey(1948). Also, the exit polls had Kerry winning over President Bush.
Cantwell by 4 King County mail in ballots? Sounds about right.
I think that Mr. McGavick has a pretty effective strategy. Focusing on his likability and buisness experience while talking about "changing the tone" in Washington in an abstract way without offering specifics play to his strengths.
I'm sure Ms. Cantwell will try to tie McGavick to President Bush. It may erode some of McGavick's support, but I think that more people respect the president than polls that focus on "job approval" may indicate. However, if McGavick is tied to the congressional Republican leadership he's toast.
I suspect that most statewide contests will be very close for the forseeable future. Which is why it would be nice to have a voting process everyone could trust.
6. I suspect that most statewide contests will be very close for the forseeable future.
Therein lies the problem. A close race in SayWA means the 'Rat "wins". If the McGavick-Cantwell race is close, and it looks like it might be, you can count on King Co. "just happening" to "find" so-called "lost" ballots, just enough times and in just enough numbers to give the "victory" to NeerDoWell.
7. How much more $$ will McGavick loan his campaign? $5 million? $7 million? 10 million?
8. #7 Just like Cant-vote-well did. LOL
9. RealClearPolitics? What a joke! They predicted a Bush landslide when the result was, in fact, the closest election in US history. These guys are a bunch of jackasses.
Ah, critical stinker's out from under its rock again....irrational and irrelevant as usual.
Why would anyone want to place credibility with a joint like RealClearPolitics when you could go to oh, say DUmmicrat.Underwear.com...right stinker? Thanks for sharing your delirium with us.
Critical Thinker -
Uh, is that why Real Clear Politics final poll of poll average (a link I provided in the original post) had Bush winning by 1.5%, when the actual margin was 3% (51%-48%)? And is that why the final Real Clear Politics electoral projection (http://www.realclearpolitics.com/Presidential_04/RCP_Electoral_Count_Final.html) had every state correct except Wisconsin, where the final margin was a Kerry victory by only 11,000 votes out of 3 million cast?
By gum, you’re right, they were totally wrong. Brilliant analysis!
12. Can we realy trust the polls put out by CNN/TIME or AP/UPI or from the usial liberal talking heads?