Posted on 10/28/2008 9:50:19 AM PDT by Zakeet
The validity of all of the MSM / Liberal University poll results depends on the critical assumptions that: (1) Democratic voters are so enthused over Barry they will turn out in record numbers and (2) Republican voters are so disgusted with Bush/McCain they will refuse to vote in record numbers.
Unfortunately for the pollsters, that premise does not appear to be correct in the case of our second largest state, Texas. As noted HERE, actual data from the Texas Secretary of State indicates both registrations and early voting numbers are higher in Republican strongholds than Democratic areas. This is the opposite of what we should see if the pollsters' underlying assumptions were correct.This brings up the next question: how can one adjust for Democrat oversampling bias?It remains to be seen if this pattern holds for other states; perhaps other Freepers might care to search for objective data from other areas, and perform a similar analysis.
Actually, it's not difficult using Zak's Rule of Thumb. Here's how it works:
Assume:By extension, this results in the following:
- Democratic voters are oversampled by two percentage points;
- The biased sampling will be split more or less equally between Republicans and Independents [or in other words, the actual turnout will be about 1/3 Democrat, 1/3 Republican and 1/3 Independent];
- Democratic and Republican voters will favor their party's candidate by roughly 90/10 margin; and
- Independents will more or less break evenly between the two major parties
In this case:
Which results in the following bottom line for a two percent Democratic over-sample:
- Two percentage points will be deducted from the Democratic vote;
- One of the two percentage points will go to the Republicans – and it will result in a 0.9 percent increase for the GOP and a 0.1 percent increase for the Democrats;
- One of the two percentage points will go to the Independents – and it will result in a 0.5 percent increase for both the Republicans and Democrats.
- The Republicans will net out with a gain of 1.4 percentage points – 0.9 percent from their own troops and 0.5 percent from Independents
- The Democrats will net out with a loss of 1.4 percentage points – 2 percent due to bias, but 0.1 percent added back due to GOP defectors and 0.5 percent added back from Independents.
| Bias Amt | Add to Rep | Sub from Dem |
|---|---|---|
| 1% | 0.7% | 0.7% |
| 2% | 1.4% | 1.4% |
| 3% | 2.1% | 2.1% |
| 4% | 2.8% | 2.8% |
| 5% | 3.5% | 3.5% |
| 6% | 4.2% | 4.2% |
| 7% | 4.9% | 4.9% |
| 8% | 5.6% | 5.6% |
| 9% | 6.3% | 6.3% |
| 10% | 7.0% | 7.0% |
| 11% | 7.7% | 7.7% |
| 12% | 8.4% | 8.4% |
| 13% | 9.1% | 9.1% |
| 14% | 9.8% | 9.8% |
| 15% | 10.5% | 10.5% |
Despite what the MSM may report, this election doesn't appear to be lost – by a long shot!
I'm not sorry for shamelessly naming the algorithm after myself.
Great post!
When I look at polls this gives me a chance to really evaluate them.
"Show me just what Mohammed brought that was new, and there you will find things only evil and inhuman, such as his command to spread by the sword the faith he preached." - Manuel II Palelologus
And how about the so-called Bradley/Wilder effect?
Unless Republicans have seriously embraced early voting (which is a possibility), the closeness of much of the early voting could indicate that we will turn out in larger numbers than the Dems.
With what’s at stake here, it wouldn’t surprise me.
Reagan and Mondale were more comparable than the two sets of candidates in this race. Mondale took only his home state of WI. Obama might not even win IL. He will probably win DC for 3 EV.
Here's also another hope and prayer that the downstream House and Senate races are also major turn-abouts.
Hey, maybe this explains why the media's leftist screamin'-demons are lying their butts off, and thrashing about like vampires with holy water in their eyes.
You forgot a key demographic!
The “Voting McCain, but too lukewarm to talk with pollsters” group which equals about 5%!!!!
“Democratic voters are oversampled by two percentage points;”??????????????????????????
Here’s my take on a couple of recent polls [part taken from another poster yesterday]:
SurveyUSA Virginia: O-52%, M-42%
http://www.surveyusa.com/clien...15-9928-119c5590638e
But, look at the fine print:
They surveyed: R-30%, D-38%, I-30%
2004 Exit Polls showed: R-39%, D-35%, I-26% for virginia.
Are there 8.5% more Democrats in Virginia than in 2004?
Are there 23% fewer republicans in Virginia compared to 2004?????????????
Another bogus Poll:
North Carolina OBama 49 McCain 48
http://www.publicpolicypolling...lease_NC_1027484.pdf
BUT who did they poll???
Democrat: 48
Republican: 35
Independent /other:17
Zakeet Zings Zogby!
(film at 11)
“The model of an Obama victory rests on the assumption the Dems are going to turn in numbers never seen before. And well if those numbers aren’t going to be there, that model is in trouble.”
I think his turn out in very Blue States will be unprecedented, NY, CA, MA, IL, NJ, etc.., but Repub turn out in contested and uncontested Red States will equal The One’s. I think it’s possible Bambi could win the general but lose the electoral. Oh yeah, McCain will take NH.
You forgot a key demographic! ... The “Voting McCain, but too lukewarm to talk with pollsters” group which equals about 5%!!!!
Both are simple adjustments.
Pick your own number for both effects. For example, assume 4 percent for the Bradley effect and 5 percent for other polling bias (e.g. won't talk, tells pollster what they think the pollster wants to hear, etc.) Finally assume a Democratic over-sample of, say, 5 percent.
Finally, add the numbers up (in this case, a total Democratic bias of 14 percent), look up the magic number (in this case, 9.8 percent), and then adjust the Rat vote down by that amount (9.8 percent) and the Pubbie vote up by that amount.
In your example, you get a 19.6 percent swing!
As you can see, it doesn't take much bias to totally trash an MSM / Liberal University poll.
They also do not account for the millions of mad Dem women and men at Obama stealing the nomination from Hillary with ACORN voter fraud along with Dean/Reid/Pelosi. They are mad and have been out helping McCain in key battleground states. A lot of the sourced negative info on Obama has come out of the PUMA’s.
There are a couple of things that have caught my attention. One is the absence of McCain bumper stickers, yard signs, etc. Part of that is probably explained by the “lukewarm” thing mention, but part of it is just as likely people not wanting to have their property vandalized by supporters of “The One”. As close as these polls always seem to be, I think you’d expect to see more open displays of support for McCain. Aother thing I notice is that here in my very liberal gentrified Nashville neighborhood I don’t see near as many Obama signs as I remember seeing for the candidates of other presidential years. All in all, I think support for McCain has been somewhat intimidated into silence and support for Obama overhyped and inflated.
I understood “of,” “and,” and “the.”
Probably the best question.
Makes a HUGE difference who they ask, and the questions.
I have thought for many years that the media must get their poling list from their subscriber list.
Nobody has ever called me for my opinion.
I remember way back when, before we decided our leaders by popular opinion polls, that we used to vote to actually elect them.
(funny, huh?)
Nowadays it seems as if anybody with a list of people can produce a poll, and help steer public opinion along with the big boys in the media.
Mondale is from Minnesota, not Wisconsin.
But, you are correct: he won the EV only from his home state (MN) and DC. However, I remember an interesting fact: Mondale won MN by a very small amount... a few thousand votes.
To put it in perspective, had one more person voted in the election for Reagan in each precinct (or one person voting for Mondale had stayed home), Reagan would have won MN.
If only one person in half of the precincts flipped their vote from Mondale to Reagan, Reagan would have won MN.
I see a similar thing here is Taxacussets. I live in Middlesex county north west of Boston and the signs in my town and surrounding towns is 3 to 1 in McCains favor. Its very strange.
When I speak to folks I get more M supporter than O supporters. The real kicker is I work a second job at night and the parking lot is about 50-50 bumper stickers, whats strange is they are mostly teamster but supporting M. Not saying M will take MASS(No way) but it just really strange. The support is not adding up for the annointed one.
You're right. Hope I didn't offend any of the ice fishing contingent.
In all honesty, I think a lot of people in MN wish Mondale was from WI.
However, I doubt many in WI would want him.
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.