Who is modelling the correct polling sample?
There seem to be two very distinct threads of Obama/McCain polling out there. One like the AP and Rasmussen tracking seems to show the race at Obama +4/+6. The other like Newsweek last week (Obama +15) or today's LATimes/Bloomberg (Obama +12/+15, Bob Barr takes a couple points off McCain.)
The only sense I can make of this is that they are modeling their samples pretty drastically differently. Are the low polls using what we should assume will be an inaccurate 2004 turnout model, or are the Newsweek/LATimes estimations really out there? (How do you predict turnout this election?)
(It's not just Dem/Repub/Independent. It's also the age, income, evangelical, black, female demographics that are likely to turnout fairly differently than previous elections.)
These polls are so far out from the election that they're pretty meaningless. However, come fall, a narrative of a close race versus the narrative of a blowout creates a very different dynamic within the electorate. This will matter.
The only sense I can make of this is that they are modeling their samples pretty drastically differently. Are the low polls using what we should assume will be an inaccurate 2004 turnout model, or are the Newsweek/LATimes estimations really out there? (How do you predict turnout this election?)
(It's not just Dem/Repub/Independent. It's also the age, income, evangelical, black, female demographics that are likely to turnout fairly differently than previous elections.)
These polls are so far out from the election that they're pretty meaningless. However, come fall, a narrative of a close race versus the narrative of a blowout creates a very different dynamic within the electorate. This will matter.
posted by mikevotes at
10:11 PM
2 Comments:
I tracked last years Aussie poll from about a year out. The disparities ran the whole distance and none of the pollsters got it right.
Voting is compulsory here suggesting better poling outcomes, you don't need to guess turnout.
It seems the big problem was that voter attitudes were all over the place. Either more voters were more informed or simply more confused.
By
Cartledge, at 1:41 AM
And if you look at the US Dem primaries, there were polls all over the place, too. Part of the reason for that is the different polling agencies had trouble projecting the voter mix.
By
mikevotes, at 7:09 AM
Post a Comment
<< Home