BenGrimm
Formally known as burntorangeVOLffle
- Joined
- Jul 22, 2008
- Messages
- 16,093
- Likes
- 10,138
For polling to be accurate the sample must be a mirror representation of the eventual electorate. Oversampling makes sense if the eventual electorate is in fact unevenly split among D and R voters.
Rather than oversample based on 2008; it would make more sense to oversample based on registration records and changes in registration records in a given state.
The problem with following the trend of the polls is that some of these polls have be using different oversampling schemes at different times - so changes in the poll could reflect a change in sentiment or a change in sample with no change in sentiment.
Which would be a good argument for paying attention more to trends in the polls than snapshots in time.
Granted I don't watch any Of them but there was a time when Fox always injected opinion into their news coverage. Has that changed?
all the other cable networks are so far to the left, they make Fox look hardcore right-wing. Fox is more moderate than any other network.
no, what he just said indicates you can't even trust a trend from week to week. over a much longer time period, even the trends aren't helpful in this type of race.
With the caveat of my last paragraph - if the oversampling is not consistent from polling event to polling event (within the same poll - e.g. Gallup tracking poll) then any change could be the result of sampling issues OR true changes in
sentiment OR a combination of both. We do see multiple swings in the oversampling rate from poll to poll so that raises a flag.
Likewise, using likely voters has proven more reliable than using registered voters (particularly when we see turn out rates in the mid 50s).
With the caveat of my last paragraph - if the oversampling is not consistent from polling event to polling event (within the same poll - e.g. Gallup tracking poll) then any change could be the result of sampling issues OR true changes in
sentiment OR a combination of both. We do see multiple swings in the oversampling rate from poll to poll so that raises a flag.
Likewise, using likely voters has proven more reliable than using registered voters (particularly when we see turn out rates in the mid 50s).
Is that really saying that "virtually every taxpayer earning under $100,000 pays an average rate of no more than 8 percent of their income in taxes." I feel special, and most of my friends must because we all definitely pay more than 8%.
Bump. Lawgator must have missed this, I just want to make sure he has accurate information when he decides his vote.
He has managed to do all of this without having to seriously and substantively defend his first-term failed promises or shortcomings, and without having to say much, if anything. about what, if anything, he might do substantially differently if he is fortunate enough to win again.
Unless I missed it, the president has yet to give a detailed answer to why he has failed to meet or even come close to his promises about reducing the unemployment rate. Saying that the task was harder than he initially thought isn't (or shouldn't be) a convincing explanation.
He hasn't given a detailed answer as to why he and his top advisers, led by Treasury Secretary Timothy Geithner, failed to focus sufficiently on reviving the housing market, rather than just bailing out banks.
He hasn't explained why his own administration is now saying that at least 6 million Americans, most of them in the middle class, will indeed face a tax increase (penalty) in 2014 if they do not buy health insurance -- a new estimate substantially higher than earlier ones.
He hasn't explained whether he shares any blame for the failure of budget talks on a grand compromise. And if the art of presidential leadership is to cajole your foes into doing deals they don't want to do, what are we to make of his famous charming effectiveness?
He hasn't given a detailed defense of the vast expansion of the security state under his watch -- a policy that, in effect, has doubled down on the global war on terror-based approaches that his predecessor, President George W. Bush, initiated.
He hasn't given a detailed explanation for why he didn't close Guantanamo, as he had promised he would.
He hasn't said how, even with a Simpson-Bowles-style budget deal, the country is going to seriously grapple with long-term unfunded liabilities in the tens of trillions.
I could go on.
I don't think many people disagree with you on this point.
I don't disagree that a given poll could make the same mistaken sampling assumption over and over, but multiple polls using different methodologies, all showing the same trend, should give greater confidence that the consensus result is accurate. Not to say it is impossible that Rasmussen alone is right and the other ten or eleven are wrong. Its just extremely unlikely.