Atman
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:11 PM
Original message |
My theory about why the new polls always seem to be so far off. |
|
Okay, I'm going to toss aside the notion, for the sake of this post, that we're simply being lied to. I believe it to a certain extent, but I don't think it tells the whole story, or even the largest part of it.
I've been noticing an trend among most of the polling data that has come out during the last election cycles...it seems that where once pollsters sought the opinions of about 1200 people -- this number always seemed to be about "normal" -- in most of the current polls, the seem content to base their assumptions on the opinions of just 400 or so people. And the MOE's are often 5 or 6%, instead of the 2 or 3% which used to be considered normal. They're polling 1/3 the number of people and extrapolating trends said to represent the opinions of a nation of 300,000,000.
Of course, as a percentage of the population, that's probably statistically insignificant in and of itself. I'm sure the many Stat Wizards on DU can help me out on this one. I admit this is purely anecdotal, based upon nothing but my own observations. But doesn't it seem like such a small sampling, 400 people -- is just more prone to being unrepresentative of the larger population?
.
|
question everything
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:15 PM
Response to Original message |
1. Polls are conducted using land line phones |
|
as more and more people switch to cell phones, they are left out.
One anomaly, though, is that it is the younger crowed that rely exclusively on cell phones, the Obama crowd, yet New Hampshire results were in favor of Clinton.
I no longer answer my phone directly, at least, not in the evening and in the weekend when everyone is at home. While I am on the Do Not Call list, all the calls are from Democrats wanting money or seeking my opinion and I am no longer interested.
|
havocmom
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:16 PM
Response to Original message |
|
They are manipulating the formula and samples to get the results they want to promote.
|
alfredo
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:17 PM
Response to Original message |
3. Large numbers of undecided voters and soft support adds to the problem. |
DJ13
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:20 PM
Response to Original message |
4. Theres one discrepancy you need to factor in ......... |
|
While the polls for Democrats seem to be too far off to explain with anything resembling rational thought, the polls of the GOP side seems to be well within the statistical error rate that polls have long had.
Why?
Who's fudging the Democratic poll results?
|
patrice
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:25 PM
Response to Original message |
5. I'm not a stats wizard, but I do know a little bit about research in the Social Sciences. |
|
Edited on Sat Jan-26-08 01:29 PM by patrice
I think you're right and the reason I think you're right is that we may in fact be in the middle of a true "paradigm shift" amongst the general population of the U.S.
The Invasion and Occupation of Iraq and everything about it, especially since it is our sons, daughters, husbands, wives, friends etc. who are being sent and feeding back concrete information one way or another has opened our eyes to who and what The Machine, in all of its manifestations, is and what it actually does. No, this isn't absolutely everyone in the U.S. who is waking up, but it IS a critical mass.
ALSO: The Machine runs on numbers, which are referred to as quantitative data. People are more complex, more variable, than numbers can show. Social Science research has been talking about this for a few decades. A simple example of the issue is that you can get people to poll one way and then when you interview them and take their actual words as data, they more often than not will contradict however they responded to poll questions. Interviews, naturalistic observation and such are referred to as Qualitative Data in Social Science research. And Social Scientists do not as yet have valid reliable procedures for extracting "conclusions" from masses of qualitative data. BTW, one thing qualitative data is being used for is to define populations from which samples are extracted for quantitative polling purposes. i.e. when you're looking for a certain range of results, you can increase the likelihood of getting those results from a poll, by using a population with certain qualities, which said qualities are accessible now by means of the absolutely enormous databases that are being acquired by means of The Machine right at this minute. This is one of the reasons that Right to Privacy and this whole FISA battle is one of the BIGGEST issues of all times.
|
EmperorHasNoClothes
(1000+ posts)
Send PM |
Profile |
Ignore
|
Sat Jan-26-08 01:36 PM
Response to Original message |
6. Since the polls are run by independent companies |
|
you would think it would be in their favor to get them as accurate as they can. The media will stop paying for them if they continue to be so far off.
|
DU
AdBot (1000+ posts) |
Sun May 12th 2024, 07:59 PM
Response to Original message |