20 (!) count'em 20 new state polls - Updated in polling section

[The Vorlon]

http://www.uselectionatlas.org/USPRESIDENT/GENERAL/CAMPAIGN/2004/IMAGES/POS_RNC_BG_June2004.pdf

Not sure how to read these results.

Just about all of them seem "ok" with some of the margins for Both Kerry and Bush seeming a bit big in places.

Public Opinion Strategies (R)is a DEEP blue GOP firm - To describe these guys as "partisan" just doesn't seem strong enough some how.

Damn good firm, just super super partisan also.



Here, for what they are worth.

Washington State - Kerry +6.5%

   Bush 42.9
   Kerry 49.4

Iowa - Kerry + 6.9%

   Bush 41.9  
   Kerry 48.8

Arkansas - Bush +5.3% -

   Bush 49.5  
   Kerry 43.8

Arizona - Bush +3.0%

   Bush 48.3  
   Kerry 45.3

Florida - Bush +4.9%

   Bush 49.2
   Kerry 44.3


Michigan - Kerry + 3.2

   Bush 43.3  
   Kerry 46.5

Minnesota - Kerry + 4.3%

   Bush 42.2  
   Kerry 46.5

Missouri - Bush + 4.9%

   Bush 48.9  
   Kerry 44.0

New Hampshire - Kerry + 0.6%

   Bush 44.7
   Kerry 45.3

Nevada - Bush + 0.7%

   Bush 46.1
   Kerry 45.4

New Mexico - Kerry + 3.4%

   Bush 43.7
   Kerry 47.1

New Jersey - Kerry + 7.1%

   Bush 41.7
   Kerry 48.8

North Carolina - Bush + 5.9%

   Bush 49.1
   Kerry 43.2

Ohio - Bush + 3.4%

   Bush 47.1
   Kerry 43.7

Oregon - Kerry + 3.5%

   Bush 42.3
   Kerry 45.8

Pennsylvania - Kerry + 1.5%

   Bush 45.6
   Kerry 47.1

Tennesse - Bush + 10.7%

   Bush 52.3
   Kerry 41.6

Virginia - Bush + 8.4%

   Bush 50.3
   Kerry 41.9

West Virginia - Bush + 2.9%

   Bush 47.1
   Kerry 44.2

Wisconsin - Kerry + 2.8%

   Bush 44.6
   Kerry 47.2

[The Vorlon]

Like I've said before, if PA stays this close, Bush's ground game will carry him here.

It was this close last time around, and his "ground game" didn't carry him.

In Pennsylvania, the "ground game" argument may have some validity actually.

In 2000, Gore/Brazille did a very very good job in Pennsylvania, while the Bush folks did BRUTAL at turnout.  

The so called "ex-urbs" - the little cities in a ring around Philadelphia about 50ish miles out the Bush folks frankly did very badly in 2000 at getting people out.

Bush lost by 4.14% in 2000 in Pennslyvania, and most of it was turnout.

[The Vorlon]

Kerry starts with a 4% edge man.  I'm pretty sure he takes it in a close electin

Kerry's "edge" is artificial, as it is based entirely on very high turn-out from the Gore team in 2000.  Gore's people had set up shop and were fully staffed in PA by May of 2000.  Bush's team was never even fully staffed.  There were huge pockets in the "ex-urbs" (as Vorlon said) and in the "T" were there was no Bush effort at all.

This year, Bush has had his HQ established in Harrisburg since February or March.  The HQ alone in perminatly staffed with hundreds of devoted workers.  There are several hundred subs stations throughout PA.  Each is fully staffed.  There is one Bush foot soldier for every 400 people in PA.  Regional offices are all over staffed.

Kerry has an office in Harrisburg that is staffed with about 10 people.  No regional offices.

Bush made the mistake of highly over estimating devoted GOP support in this state in 2000.  He won't make that mistake again.

I expect Kerry will get his act together in Pa, and they have "outsourced" a lot of stuff to groups like ACT, but on balance, I think I agree a "little" more with supersoulty than I do the Boss on this one.

[The Vorlon]

When I add these polls to pollbooth, they will push Bush above 270 by moving 5 of 7 states out of the toss-up column.

Oregon and New Mexico move to Lean Kerry, while West Virginia, Florida, and Ohio (by the vary barest of margins) move to lean Bush.  Wisconsin and New Hampshire stay toss-ups.

The Tennessee result looked a bit high for Bush to me, and Washington and Iowa results looked a bit high for Kerry, but other than that, these mesh very well with the other stuff I have seen.

PS - I see in Pollbooth you are using that LA Times piece of %^%^%^%!...

I emailed the LA Times polling folks, the actual partisan breakout was:

Democrats 38%
Republicans 24%
Other: 38%

They are defending this as a representative sample!

[The Vorlon]

Yeah, I know it's a terrible poll, but I have no objective reason to disqualify it....so far, I haven't been rejecting polls just because the party ID breakdown is off, so I figure I shouldn't start because of this poll, even though it is the worst I've seen.

Fred's All Beef Double Decker Burger Poll has Bush +9 in the "deluxe with cheese" subsample.  Can I get that in Pollbooth too...?



The LA Times is a member of the National Council on Public Polls and that last poll broke at least 5 or 6 or the required disclosure guidelines of the organization.  For what (little) it is worth NCPP has started an investigation.

They "may" have to publish some token "clarification" or something, but in reality, nothing will come of it.

[The Vorlon]

does their partisanship skew their numbers to the right by a point or 2?

The honest answer is I don't know.

Historically speaking these guys are deadly accurate - These guys are a good as Zogby claims to be - These guys poll for about a dozen Senators, and dozens of Congessman.  They are hired because they are good.

POS is a combat polling firm - and when you run a campaign, you want your polls to be accurate, not biased.  You need to know the truth, whatever that truth is. - No point putting time and money into a state you can't win.

Speaking solely for myself only, If anything, when polling internally for a campaign you actually try to be just a tad pessimistic, maybe a point or 2, if anything.

A lot of these states, Washington, Oregon, Iowa, Nevada, maybe New Hampshire, Arizona, New Mexico, etc actually look a tad optimistic to the Kerry side IMHO.

[The Vorlon]

Maybe, but out of the 12 or so polls that their has been in Florida this is the 2nd best for the Bush team so that could be a reason it was released. Not saying that its the reason, but you never know

On the other hand, it is the most favorable to Kerry polls I have seen for Washington and New Mexico and Nevada...

Who knows.. ?

Remember too... 20 polls, we can expect 1 to be wrong outside the MOE.

Off the top of my head I'd vote Tennessee on this batch...

The suspicion of bias in polls by partisan firms comes not so much from the belief that individual polls are biased, but that only a small percentage of these polls ever become public.  Candidates generally only release polls that show them in a favorable light.  And even if the poll "leaked out", it is possible that the leak was intentional.  

I think the best policy on Partisan polls is just put a big "D" or a big "R" next to them, and let folks make up thier own minds.

For what it is worth:

Gore won the 20 states polled by 0.01% in 2000

In these polls Bush leads these 20 states by 0.78%

If you take out New Jersey, 2000 versus 2004 is almost identical over the 19 states.

Certainly not crazy by any means, but I guess you never really know.

Maybe the "vast right wing conspiricy" is just most subtle than they used to be...

[The Vorlon]

I don't trust partisan firms. Even ones that would otherwise be good firms.
Too many times watching Yes Minister...

Most people think Yes Minister is a comedy.

If you have ever worked in government, you know it is actually a documentary.....