Nov 02 2008
Obi-wan Kenobi Joins Forces With AJStrata
Jim Geraghty’s mentor Obi-wan Kenobi agrees with my assessment that either there is a historically high Democrat advantage this year, or historic voting patterns will dominate and McCain-Palin could win:
So here is the McCain supporter’s homework assignment. First, polls like Gallup, Washington Post/ABC, and some of the others with voting models that give enormous advantages to Democratic turnout  will be right or wrong Tuesday night. *
…
We will find out whether they suffered from the same inner culture that skewed the exit polls in 2004. Ignore’em for now. Instead, tomorrow morning look at Battleground which has a good record and, again, has it around four now. If Obama’s numbers are starting to decline then go to Rasmussen. His turnout model is more favorable towards Obama but he will probably catch a fall-off in Obama’s numbers if it is happening. At midday watch for IBD-TIPP, the most accurate poll in 04. It has got it around two now and is actually saying for the record that Obama may have picked up strength over the weekend only because  trick-or-treating Republican parents weren’t answering polling calls. If Obama’s strength shows a fall in that poll, consult this space and find out what you are saying about Pennsylvania.
Obi-wan is being coy – AJStrata has no such weakness. Look, one of these models is right, but all the recent polling shows McCain-Palin support equaling or surpassing Obama-mania in many states, VA and PA are in play. If McCain can grab these two states Obama’s path to victory is seriously limited.
Here is a wonderful essay on what the Obamatrons have been doing. It is pretty long, but worth the read.
I can’t help but wonder. Since these pollsters have used a skewed number of democrats/republicans/independents all along could it possibly be that the pollsters are in collusion with the Obama campaign? Did the Obama campaign tell the pollsters that ACORN was going to register a huge number of democrats? As you see, I don’t put anything past Obama. He has shown he is willing to steal this election any way he can. Besides, the media is in the tank with him the pollsters might be too. They are, after all, part of the media.
I have never believed the polls. They are part of the “discourage and rattle the republicans so they won’t vote team”. They have never put a republican ahead ever. They use too many democrats. They are definitely not taking into consideration that some democrats are voting for McCain this year. They don’t take into considertion that a majority of reputlicans don’t want to talk to them. I know I am not answering my phone unless I recognize the number lately. There are too many election phone calls and I have already voted. Not that these phone calls would sway me even if I hadn’t.
I believe it was SBD who did a major posting on verification of the databases for profiles and the costs involved.
My guess is that a lot of the smaller polls simply can’t afford those kind of bucks and they just take the registration numbers and weight for that.
Well ACORN has been padding the books since 1970, not just this year.
We know places have well over 120% of their census estimates voting age population registered.
The Obots spammed and skewed the exit polls and were over rated by 7%.
All these factors just lean in the same direction to oversample Dems as a form on normal business and then if the polls oversample beyond that or weight them higher it gets even worse bias built in.
Crosspatch, that Zombietime article’s conclusion captures perfectly what I felt in ’04:
“Yet it may very well be that an army of glum, dispirited and pessimistic conservatives will reluctantly trudge to the polls on November 4, each one imagining they are the only remaining person in the entire country voting for McCain, and lo and behold — they’ll turn out to be a silent majority after all. “
I went to bed thinking Bush had lost. All I wanted to do was sleep and forget. I wasn’t even going to turn on Fox News in the morning, but I did. And boy was I surprised!
…
AJ,
I had the following E-mail exchange with that ex-political pollster I mentioned before. You will find it extremely illuminating:
Trent to ex-pollster:
He (Kevin “Coach†Collins of the collinsreport.net) seems to have hit on the idea of hostile field polling questions without being a pollster.
His observation about TIPPS poll asking “Do you display the American Flag” being damned interesting from the point of view of people lying to pollsters about everything.
>TIPPs asked “Do you display an American flag?†:
>Obama 38/ McCain 52/ undecided 9.
>
>People who wave a flag are undecided? Does
>anyone really believe this?
Ex-Pollster:
Yeah, that would be the start of it; if I were getting paid to do it and I thought of the flag question, I’d use it. Typically for a sample size big enough to cover 67 stratifications (the current minimum number you need to adequately cover the American public — you’d need more to do it state by state, maybe as much as 3x as many) you need around 1200 respondents. To do a good hostiles protocol you need about 15 stand in questions for the one that you think they’re lying about (if you think they’re lying about more than one, you can overlap some questions, but obviously that produces a spurious correlation in the results. Or at least that’s obvious to me and the people who know what they’re doing; I’ve seen a lot of researchers screw that one up). The 15 stand in questions all need to correlate with each other at about 60% or better.
And now you see why no network even tried it this time. To get those 15 questions you probably would need to test around 100 good ideas, in about 25 nationwide surveys (since they won’t all start out in the same survey and you have to see how they interact, so you’d test say 7 surveys in your first round … maybe 3/4 of questions would look promising in the second round, so you’d test 5 …. then 3 … then start doing mix-and-matches that would easily use up another 10 surveys). You’d have to pay for 25 polls to make your polls be accurate again. Forget all ideology; Fox wasn’t going to pay for anything like that any more than MSNBC.
By 2012, though, they may all have to.
Another Friend on the list Comments:
The well-poisoning effect on public willingness to respond telephone polls is not due to commercial telephone solicitations.
The media and celebrities have done their level best to delegitimize their cultural opposition, which IMO has resulted in their cultural opposition, i.e., normal people, refusing to answer campaign and even political questions by pollsters. Their feelings haven’t changed, but their willingness to divulge those to strangers has.
This goes double for union members in Pennsylvania who have apparently learned that their union hierarchy is using pollsters to ferret out unbelievers and bad-thinkers.
Ex-Pollster:
I’m sure there’s some of that going on. But straight up, used to be if I hired 10 college kids to do interviewing, trained them properly, etc. they’d get a hangup rate of maybe 15% and a useful result rate of about 30% (that’s with a well-designed poll that would make someone feel good about answering, which after all is the whole idea). Nowadays with the best trained kids and the beat designed poll in the world, you’re lucky to keep hangups under 40% and the useful rate is more like 10%. (Things that are not useful and are not hangups are, e.g., people who are not quite all there mentally (a surprising number, until you ask yourself “Who’s home all the time?”), super-hostiles (those numbers have ballooned — people who don’t hang up because they’re too busy ranting; various others). This is of course for commercial stuff, which is what I do; but my colleagues over in political work tell me it’s this way cubed and squared over there.
Whether it was people selling timeshare condos, celebutards making fun of them, or just a different climate, I don’t have evidence for. Probably all of that and many more I haven’t thought of. But nowadays regular people sure don’t like pollsters, and online surveys are more or less begging to be lied to and screwed over.
AJ,
The bottom line on getting accurate and trust worthy polling today is that you need the following:
1) You need a minimum of 1200 respondents to cover the 67 demographic stratifications that accurately cover the American population.
2) Getting those accurate stratifications costs millions. A subscription with a reputable demographic data firm costs millions per and requires a multi-year contract. The minimum term is two years and has the highest per year cost. Most polling firms used by the main stream media are using 10 plus year old demographic data.
3) To do a hostile field poll, that deals with how people are now lying to pollsters, you need 15 questions with a 60% correlation between them to get one statistically valid and accurate answer.
4) To test enough questions (100) to get that “golden 15,” you must conduct _25 polls of 1200 respondents_ to test those 100 questions.
All of the above is for a single question on a national level. AKA “Who will you vote for, Obama or McCain.”
Additional questions require the same process of vetting the “Golden 15” and may require a separate polling team to do all of the above.
I’ve added your post to BlogWatch: Final hours to 2008 Election
I also agree with your assessment, which is why I’ve added / linked to your post.
[…] in almost all of his primary contests. There are enough undecided votes in each of these states to deliver a McCain victory. The vast majority of undecided voters are white suburbanites. On balance, it looks like Obama has […]