Sam Spade
SamSpade
Atlas Star
Posts: 27,547
|
|
« Reply #3 on: August 15, 2008, 10:36:18 AM » |
|
POLLSTER FIGHT!
(it's kinda like a cripple fight, but less amusing)
A general point:
The main problem with PPP that I have as an issue is that they have worked for Democrats in the past in North Carolina and were considered solely a Democratic firm before their first perusals into "independent" polling in 2006, which solely consisted of polling NC-11 and maybe something else I missed.
Therefore, we have no track record for them and no way to judge what their biases or non-biases are.
Considering the situation to be the opposite to Strategic Vision in 2004. There we had a firm that was considered solely a Republican firm before their first perusals into "independent" polling.
What did that election show us? Their samples leaned too Republican, except in their home base area of Florida/Georgia. But they were no less wrong to the Republican side than R2000 was to the Democratic side, which makes sense considering who runs R2000.
But should we consider them an "independent" pollster now? Yes, unless the sample says otherwise. Just merely keep in mind that their samples typically (but not always) lean too Republican with the exceptions as above (they were also ok there in 2006 as well).
As for Rasmussen, he may be a Republican, but there's no history that he's ever done any polling for any Republican ever. The only polling he did for any politician were a couple of polls for Badnarik in 2004, amusingly.
Rasmussen obviously gets lambasted for his 2000 national result, as well he should. But the interesting thing about that election was that the state polls he ran were pretty damn accurate. And in 2004 and 2006, his state polls were pretty much on the money - I think only M-D performed better and it was close.
Now, I have issues with his method, especially in the summer. One day polling in the summer is especially problematic for, oddly enough, the reasons that PPP mentions. Theoretically, the method, though, should cause the same problems that most summer polling creates - jumping around and a Democratic lean.
One last thing - additional callbacks of people don't necessarily make the polls more accurate. Look at what M-D does.
|