Here is a chart, and please check my data, but I have been trying to understand the logic in oversampling of one party or another. I remember the democrats back in 2004 complaining that there was a 2-3 point republican oversampling, and same in 2008 when they were giving a 7 point democratic over sampling. Well for the most part they were right, well except for PEW who historically overstates Democrats, while Gallup and Rasmussen have been right on.
So whats going on here is very interesting as looks like all the polling agencies are following PEW and they seem to be the one's oversampling. Only thing, is they adjust for the over sampling, not sure the rest of the polling companies are adjusting for it... Here is my guess, with the currently assumption that Obama leads with independents by 4%. Rasmussen's numbers are approximately Rep+2 minus Independents -4 giving todays numbers of Obama +2% and those polsters that just quite don't get it are doing first grade math Pews Dem+8 plus Independents +4 coming up with Obama +12%.. Everyone else is interpolating somewhere in between.... It's all a guess between Obama up 2% and 12%....
If I were to guess today, i would say that youth voters will be down 30%, Hispanics down about 10% the rest of the voters about the same percentage wise. Overall number of voters will be closer to to 2004, and the oversampling should be closer to 2010 where it was Dem+2, that would give a polling number of Obama +6.... Hope this helps
| Democrat Oversampling | 2004 | 2008 | 2010 | 2012 | |
Pew - historical |
-5% |
-15% |
-8% |
-8% |
|
| Fox | -8% | ||||
| Cnn | -8% | ||||
| Reason rupe | -8% | ||||
| NBC | -7% | ||||
| Monmouth | -6% | ||||
| Gallup - historical | 3% | -9% | -2% | -5% | |
| NY Times | -5% | ||||
| Actual | 3% | -7% | 2% | ||
| Rasmussen - historical | 2% | -7% | 2% | 2% | |
| Avg | 1% | -10% | -2% | -6% | |
