Sep 24 2008
Now That Is One Skewed Poll!
The Washington Post/ABC News poll out showing Obama with a 9 point lead (52-43) illustrates my point about bad sampling techniques. There straight party sample breaks down 38% Dem, 28% Rep and 29% Ind. When Indies are asked which way they lean the sample goes 54% Dem, 38% Rep and 7% Ind.
Is anyone surprised with a 54(D) – 38(R) sample the result is 52(D) – 43(R) race? Not me.
The 9/7/08 Poll had a sample with a party + lean of 50(D) – 41(R), and the bottom line result for that poll was 47(D) – 46(R). I seriously doubt the Dem voters will be 54-38 over the GOP voters. Especially in battleground states. Maybe these pollsters should trying calling outside NY, DC and LA?
I was wondering about that one. I can believe that Obama might have a 3-5 point advantage today (he won’t have it after the debates), but NINE?
PPP just released a poll in Colorado which puts Obama up +7
Now what the issue is here is that the party split was
Dem 40
Rep 36
Ind 24
However August voter registration number per the spreadsheet available from the Secretary of State show the registration breakdown is
Dem 30.6
Rep 34.8
Ind 34.5
But in general AJ follow any of the polls and for the vast majority of the states the independents are near split with a few exceptions of heavy leaning states to either camp.
Only a couple of points are the normal leans for the independents and the ones in your post are way out of whack for that.
What looks to have happened here is PPP applied a model for national registration ratios to get their poll number oriented to that scale and overloaded the dem picture for how it plays within the state.
It is absolutely vital that this $700 billion “bailout” (it’s much more complicated than the word bailout indicates) is passed very quickly. No matter how bad it is.
Of people I have watched, talked to, and read for years, the more experienced and financially involved they are the more frightened they are about where we are. Right now, the total collapse of the American financial system is on the table. This is beyond politics, but on that level, we really don’t want to go into this election 4 weeks after a repeat of the Crash of ’29.
You may scoff – we came within 500 trades of exactly that scenario last Wednesday. We are within 500 trades of it every day until the government takes drastic action.
Poll numbers and the fine print…
One of the first things I do every morning after logging on is to check the Real Clear Politics polling page to read about the latest poll numbers that have been released, whether they be state polls or national. I rely more on state polls than natio…
Today’s gallup poll has the same result.
Then today’s “gallup poll” probably has the same kind of slanted sampling.
[…] “How the hell can people trust Democrats to get us out of a mess they enabled?†Then I read AJ Strata’s post where he took a look at the party breakdown, and it became clear: [Their] straight party sample […]
Can we all say: orchestrated poll numbers to 1) suppress republican votes, and 2)cover up the cheating and malfeasance in voter poll stations?
Remember, Kerry was ahead in the Gallup numbers as well. Do not be fooled.
[…] A few more thoughts on poll models. Â One of my commentors, MerlinOS2, noticed a risky disconnect between one poll’s party affiliation and party […]
[…] have a perfect example of this in two Colorado Polls out recently. Â The first poll was commented on by our Reader MerlinOS2: PPP just released a poll in Colorado which puts Obama up […]