.

[Sbane]

Bawlexus, can you show us proof that most pollsters sampled the number of Democrats and Republicans the same as the exit polls( not sure why these are considered to be accurate, they are also polls) in 2008. The best way to interpret polling is to look at the average, and nationally it shows Obama to be up by 2 and an average of NC would likely show Romney up. That is the most reliable way to look at polls, not trying to decipher individual polls subsamples.

[Wiz in Wis]

I think I read somewhere that SurveyUSA was pretty unique in that they don't weight by party at all, just report the results of their overall sample in crosstabs. The likely voter screen is probably generated by a question or two towards the beginning of the survey asking participants about their registration, past voting, or polling location.

Is the partisan spread favorable to Obama... maybe.
Is the likely voter screen to tight, or too loose... maybe.
Are we spending too much time arguing about what this means for Obama/Romney 6 months out... definitely.

[Sbane]

Bawlexus, can you show us proof that most pollsters sampled the number of Democrats and Republicans the same as the exit polls

Not sure why I would need to proove this to you, since I never made that claim.

Then why do you feel the need to go through every poll and examine its partisan breakdown, especially when it's something not asked by many pollsters, and something that can change based on the question asked?

[Sbane]

Ok yeah go ahead and do your thing. I'll go ahead and keep tabs of the polling averages.

[Queen Mum Inks.LWC]

Database entry: https://uselectionatlas.org/POLLS/PRESIDENT/2012/polls.php?action=indpoll&id=3720120521019

[Tender Branson]

Let's compare this:

76-18 among Democrats

with this:



...

Obama's probably underestimated by about 4% right now in NC, because at least 10% more Democrats will back him until November (40 x 0.1)