Forum Home | Log in | Signup | My Page | Rules | FAQ | Boards | Members | Search
« ALEA Home | Email msg. | Reply to msg. | Post new | Board info. Previous | Home | Next

Re: Poll: Obama opens substantial leads in key swing states

By: DigSpace in ALEA | Recommend this post (0)
Wed, 26 Sep 12 6:22 PM | 53 view(s)
Boardmark this board | The Trust Matrix 		Msg. 10330 of 54959
(This msg. is a reply to 10329 by Cactus Flower)
Jump:
Jump to board:
Jump to msg. #

no question Nate puts it on robocalls and landlines, the right puts it on sample weighting. The two narratives both segregate the rolls similarly.

I just don't know if they actually do sample weighting, or if the relative weights are simply a consequence of the method (type of call/phone). Are the deliberately pre-determining the party mix they choose to use as a representation of what will happen? Or are they simply reporting the party mix that their polling discovers and that different methods discover different mixes?

Most of that data is buried at the premium subscription level it seems.

I still watch Ras pretty closely, Dems turn-out poorly and generally over-poll (at least over my lifetime) 2008 being the exception.


 - - - - -
View Replies (1) »



» You can also:
- - - - -
The above is a reply to the following message:
Re: Poll: Obama opens substantial leads in key swing states
By: Cactus Flower in ALEA
Wed, 26 Sep 12 6:13 PM 		Msg. 10329 of 54959

this is just desperation. when they were doing well, there were no complaints.

but silver has addressed this and it seems mostly to be explicable by polling medium. the republican leaning polls mostly result from calling landlines. no coincidence that older people tend to have landlines and preponderantly vote republican.

http://fivethirtyeight.blogs.nytimes.com/2012/09/24/the-statistical-state-of-the-presidential-race/?gwh=85AF5D4AF46293028BE4E83D603298EC

if anything, the issue is with the landline polls.

« ALEA Home | Email msg. | Reply to msg. | Post new | Board info. Previous | Home | Next