Democracy Corps Poll

[millwx]

Not sure if this was posted here.  The poll is not TERRIBLY recent (June 14-17th), but from what little I can tell (based on when it was posted on PollingReport) it's "new" in terms of being just released.  Apologies if I'm wrong and duplicating an old thread.

The poll was done by Democracy Corps (actually conducted by Greenberq, Quinlan and Rosner Research).  They polled 1004 LVs.

Head-to-head: Kerry 49%, Bush 48%, other 2%, unsure 1%
W/ Nader: Kerry 45%, Bush 45%, Nader 6%, other 1%, unsure 3%

This represents very little change from their previous poll.  Bush has improved (or Kerry declined) by 1% in both the head-to-head and with-Nader results.

[The Vorlon]

Never, ever apologize for posting a poll

[millwx]

Never, ever apologize for posting a poll

Vorlon, assuming the Fox poll results (in another thread) are correct (and I have no reason to doubt Tedrick), what is your take?  I suppose if the Fox results are three-way, their head-to-head could be closer.  We've got several widely varied poll results from the past week.  Most show it reasonably close, but a few lean heavily one way or another (ABC/WP leans Kerry - though their sample is SLIGHTLY disputable - while Harris and Fox lean Bush, and the remaining few are close - within a few points).

I know things are fluid, and the polls have MOEs, and could even occasionally get a bad sample, but, seriously, this seems like more spread than we've typically had this year.

[The Vorlon]

Never, ever apologize for posting a poll

Vorlon, assuming the Fox poll results (in another thread) are correct (and I have no reason to doubt Tedrick), what is your take?  I suppose if the Fox results are three-way, their head-to-head could be closer.  We've got several widely varied poll results from the past week.  Most show it reasonably close, but a few lean heavily one way or another (ABC/WP leans Kerry - though their sample is SLIGHTLY disputable - while Harris and Fox lean Bush, and the remaining few are close - within a few points).

I know things are fluid, and the polls have MOEs, and could even occasionally get a bad sample, but, seriously, this seems like more spread than we've typically had this year.

We have actually had fairly few really crazy samples this year.

Step Number 1 - Compare "Apples to Apples"

The Harris Bush +10 was LIKELY voters...

Hmmm... A likely voter polls screening for level of interest right dead in the heart of wall to wall Reagan Funeral coverage... would that skew to the GOP... ?

Hmmm.. let me think on that... I'll get back to you...

Using the 3 way race, REGISTERED voters.. these are the last few polls:

Harris  Bush 48 - Kerry 42 - Nader 7%
Pew    Bush 46 - Kerry 42 - Nader 6%
Tipp    Bush 43 - Kerry 40 - Nader 5%

These numbers are all, well, well with in MOE, in basic agreement.
TIPP doesn't slam the undecided they way the other two do, so their lead and total numbers being a tad lower is fully understandable.  

The Bush lead in PEW is also a bit low, as PEW noted a sharp Bump for Bush after Reagan dies (The first two days of PEW were pre-Reagan Death)

Lets toss in Fox, assuming the +7 to Bush is correct, it will likely be something like Bush 46 - Kerry 39 - Nader 5 (Opinion Dynamics doesn't slam the undecided too hard either)

Giving us:

Fox: (?)   Bush 46 - Kerry 39 - Nader 5%
Harris     Bush 48 - Kerry 42 - Nader 7%
Pew       Bush 46 - Kerry 42 - Nader 6%
Tipp       Bush 43 - Kerry 40 - Nader 5%

Again, well within MOE, they all agree.....

Now add in Wash Post/ABC.....

Fox: (?)   Bush 46 - Kerry 39 - Nader 5%
Harris     Bush 48 - Kerry 42 - Nader 7%
Pew       Bush 46 - Kerry 42 - Nader 6%
Tipp       Bush 43 - Kerry 40 - Nader 5%
Wash/Post Bush 44 - Kerry 48 - Nader 6

Hmmmm....

How does that song go....?

"One of these things... is not like the others....
One of these things... doesn't belong...."

The ABC sample at +8 to the Dem side in party ID (while not super crazy) is a bit off and USUALLY if one area of a poll is "non-random" other areas you may not be abee to test are also non-random.

When you crank through the raw data of a poll, the more your weightings change things, the more scared you should be.

If you take your raw data, weight it, and the results don't change, chances are you have "nailed one"

As always... let's wait for the next round of polls

[© tweed]

This last series of polls have varied by 15%.  I don't know who to believe...so I can find a reason to change my map.