Correcting The Polls 2

The results you get from polling depend on who you ask, and how you weight each subgroup among those polled. In a poll of 1000 adults found randomly walking the corridors of the New York Times building, Obama is favored to win in a historic landslide. As mentioned earlier, most polls give a heavy bias to Dems when they add up the results and the interesting part is that they tend to call things a tie.

Rassmussen has data on how people self-identify which is not really subject to jiggering if the sample is widely enough based. Self-identification changes over time with more people being willing to identify with one party or the other depending on how the party is seen. This produced the following:

Based on extensive surveying, Rasmussen calculates that 35.4 percent are Republicans, 34.0 percent are Democrats, and Independents are 30.5 percent.

The above is from the methodology page at UnSkewed Polls which, interestingly enough is close to the same  basis for my initial prediction calling for a 5 point win by Romney based on data from a left-leaning polling company. I assumed an even 33% across the board split.

UnSkewed Polls has taken this data and applied it to 12 different polls and come up with an average of 7.8% in Romney's favor.

UnSkewed Polling Data Monday, September 24, 2012 7:58:05 PM
Poll	Date	Sample	MoE	Obama(D)	Romney(R)	Spread
UnSkewed Avg.	9/4 - 9/20	--	--	44.0	51.8	Romney +7.8
Reason/Rupe	9/13 - 9/17	787 LV	4.3	45.0	52.0	Romney +7
Reuters/Ipsos	9/12 - 9/20	1437 LV	2.9	44.0	54.0	Romney +10
NBC News/WSJ	9/12 - 9/16	736 LV	3.6	44.0	51.0	Romney +7
Monmouth Univ.	9/13 - 9/16	1344 LV	2.5	45.0	50.0	Romney +5
QStarNews	9/10 - 9/15	2075	3.0	44.0	55.0	Romney +11
NY Times/CBS News	9/8 - 9/12	1162 LV	3.0	44.0	51.0	Romney +7
Democracy Corps	9/8 - 9/12	1000 LV	3.1	43.0	52.0	Romney +8
Fox News	9/9 - 9/11	1056 LV	3.0	45.0	48.0	Romney +3
Wash. Post/ABC News	9/7 - 9/9	826 LV	4.0	45.0	52.0	Romney +7
CNN/ORC	9/7 - 9/9	875 RV	3.5	45.0	53.0	Romney +8
IBD/CSM/TIPP	9/4 - 9/9	808 RV	3.5	41.0	50.0	Romney +9
ARG	9/4 - 9/6	1200 LV	3.0	43.0	53.0	Romney +10

LV is likely voters, RV is registered Voters, and MoE is margin of error. Ju8st because the bro's name is on the ballot does not make it 2008 all over again. 

If this turns out to be correct, pollsters in the future will be forced to reveal a bit more of their methodology along with the results lest they be denounced as mere shills for one side or the other.