Posted on 10/17/2004 2:05:36 PM PDT by West Coast Conservative
I understand, I think.
dvwjr, you’re saying that from a scientific perspective, dissimilar polls can’t be combined to reduce the MoE. And comebacknewt points out that because most polls over time are reporting very similar results, the results are probably true, just statistically ummeasurable.
So the real uncertainty of the results is smaller than the MoE, we just can’t measure it precisely? Wouldn’t that mean that the claim that “polls show the race in a dead heat” is only technically correct but misleading?
The MSM may in fact turn on kerry soon. They turned on Dukakis and they turned on Mondale. As soon as it appears out of reach they will lash out. We're not quite there yet, but if polls keep turning in bush's favor it could happen.
The problem with the 'dead-heat' within the MoE statements always made is that the individual making the statement always assumes that if the poll numbers seperating two candidates are between (1*MoE) and (2*MoE) that there is an equal probability that their is a statistical 'tie'. The problem with that statement is that there is a MoE 'centered around each reported poll number represented by a 'bell-curve' distribution.
If you were to graph the two intersecting MoEs of say Candidate A who leads Candidate B by the MoE of the poll, then you would find that the probability of a 'tie' would be the area under the intersecting curves, which means that while it is possible that lagging Candidate B is tied with leading Candidate A, it is MORE probable that Candidate A is actually ahead...
Hope this helps,
dvwjr
Because Republicans have jobs. (shameless quote lifted from PJ O'Rourke)
There are areas of error than the ones you are discussing. For example how about all those who only have a cellphone number. FCC rules say they can not be dialed by pollsters cause they have to pay for incoming calls.
And the new internet phones from companies like vonage and others can have area codes that don't always represent where the party actually lives.
I am an Iowan and believe that Bush will take Iowa, when all is said and done, no problem. I teach at an Iowa college and see no enthusiasm for Kerry at all, and I would normally expect to see some. Virtually no Kerry signs on cars at the college, but a few Bush signs on cars. Around town, I see about half and half Kerry and Bush signs; remember this is a college town.
In 2000, Bush was a far less well-known quantity, and the DUI thing had just surfaced, yet Iowa was still very close. Now Bush is a known incumbent, and Iowans ADORE incumbents of all parties and re-elect them over and over. The Liz Cheney comment will hurt Kerry in Iowa, as well; that simply isn't how Iowans think people should take it upon themselves to speak in public about someone else's child. Bush will take Iowa. I have never been in any doubt about this, regardless of earlier polls. Wait and see!
Should have said "both parties" and not "all parties".
Honestly, though I think this one is right on target. I think 4% is just about right: it's consistent with all the recent polls, including the 4% MOE ones showing a tie, the 4% MOE one showing Bush up by 8%, and the other 3% MOE poll which had him up 3%.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.