The slight move toward McCain in the national polls appears to have stalled. In most polls, Obama has gained a bit in the last few days. The only exception among those outfits who polled through Saturday is the IBD/TIPP poll, which shows only a 2% lead for Obama.
The betting services are even more encouraging for Obama, particularly state by state. If McCain were to win on Tuesday, given the improvements in polling since 1948 it would be a much bigger upset for pollsters than Truman over Dewey.
BTW, I was polled by Rasmussen on Saturday. The automated operation was professionally handled, though I noted that the question about presidential approval for Bush preceded the question about who one favored in the 2008 election, which might suppress expressed McCain support very slightly. Other questions that might influence the 2008 presidential preference question were asked near the end of the survey, as they should be. It made me wonder if the outlier polls showing double digit leads for Obama have even more serious question-order problems than polls (like Rasmussen) showing 5-8% leads.
Buy a half-hour spot Monday night (when it's too late for Obama to counter):
1. Make strong substantive arguments that push conservatives to vote.
2. Conclude with: "The polls show a blowout landslide win for Senator Obama. It would be customary for me to say 'The polls are wrong!'. Sad to say, the polls are likely right. I don't want to lie to you, but I urge you to do the right thing."
This might have the effect of both stimulating the conservative get-out-the-vote efforts and simultaneously suppressing the Democratic efforts. Why vote for Obama when his opponent has already conceded a loss?
It's his only hope. It also has the benefit of his speaking the truth, always a novelty for a politician. Of course he won't do it.
What makes you think he has the money to do it? I don't believe the Republican IE arm could buy this time for him to do a personal pitch, and his own campaign's account is probably close to empty already, with almost all the money already either paid in advance for television advertising or earmarked for GOTV operations. Any further reduction in the latter is likely to simply concede outright the states in which Obama is best-organized, which includes pretty much every battleground state.
You are interpreting my suggestion on a more literal level than it was meant--I was speaking abstract strategy and you are speaking tactics.
Nonetheless, to respond seriously, I understand Ms. McCain has a dollar or two tucked away.
I think your strategy would be too mavericky for even McCain. When he announced he was giving up on Michigan, that decision cause Obama's numbers to spike upward in the state. People want to vote for winners, not long shots.
The 'Polls' called Gore and Kerry as big winners, even in EXIT polling on Election Day !!!!!
The polls this year are WORSE than ever ! When you see 10 + point swings overnight from the same pollster - you KNOW it's crap ! And you can't just 'average them' - if you take 2 wrong numbers and average them, you come up with ANOTHER WRONG NUMBER, not the RIGHT number !
I suspect Brobama may well win ( and we're all F'd ), but to put ANY reliance in polls this year is lunacy !
I think question order could influence the pick, but I doubt it is enough to change the likely outcome. I just think McCain is in a place where there really isn't anything he can do to win. I do think the race is going to be tighter than polls seem to indicate popular vote wise, but I think Obama will likely have a very clear and decisive win electorally.
But short of Obama being found with a live boy or dead girl or the political equivalent over the next 24 hours, I just don't see how he is going to lose the election.
The likeliest explanation for outliers in either direction is that with all the polling going on there have to be some results far from the others. In this sense they are not outliers at all, but just the more extreme draws from the distribution.
Of course the polls differ somewhat in their methods of identifying likley voters and other things, so some will lean one way or another just as a result of that.
As I understand it, because the McCains file separately, Mrs. McCain is limited to same contribution limits any other private citizen would be. She can't just pour X million into the campaign--she's probably capped at around $20k, and I'd be surprised if she's not already maxed out. (I have no particular expertise in election law, so it's possible I'm mistaken about this--I would welcome a correction from someone who's familiar with the relevant statutes.)
Paul Milligan:
Well, it's just objective fact. In 1948, the overall volume of polling was very low, there was less understanding of the need to weight for demographic subgroups, the margin of error was of course known, but generally not actually released with the results of the poll, question ordering and wording was not seen as significant, etc. etc. There have been many improvements in polling methodology since 1948.
The polls projected Gore as the winner of the popular vote, which was accurate. There was less state polling, but in the closest states it was definitely within the margin of error. Kerry was never, ever favored to win in the late polling (mid-October till the end)--not in the popular vote, not in the battlegrounds. Democrats convinced themselves that he had a chance in that the undecideds could break in his direction--not unreasonable, but it did not happen. At this point, even if undecided voters break to McCain, he can't make up his deficit. That's particularly true in states like CO and NM, where it looks like more than half the ballots have already been cast.
We have not seen that. Unless you're talking about a one-day sample from a rolling three- or five-day tracking poll, in which case you don't understand polling very well. There's a reason they're averaged over X number of days, you know.
That would only be true if polls were off by completely random amounts, due to factors we can't possibly foresee. That could conceivably be the case, but it wasn't in the primaries, and what seems more likely is that knowing the different factors out there that polls are bad at measuring, it is possible to aggregate and adjust polling data to produce a more accurate picture of the state of public opinion.
To deny the reality of all polling because it doesn't show an outcome you want is just wishful thinking, I'm afraid.
Are you sure?
You can make a killing on Intrade.
Only if the non-responders differ substantially from the rest of the population. They might, but I suspect irritation at pollster calls is a bipartisan phenomenon.
That it is a serious concern. My understanding is that pollsters are aware of this problem, and working to correct it (weighting by party ID and other relevant demographic characteristics being the most obvious method of accomplishing this), but nonetheless, it's a problem that is becoming increasingly severe with each election cycle. Nonetheless, there are countervailing factors, and on balance my guess is that, as with previous elections, the results will fall pretty close to an aggregated, weighted average of the various polls.
I think in the long run that the polls probably are over sampling or overweighting democrats, and that the vote totals will in the end be much closer than polls indicate, but I don't think McCain has a snowball's chance in hell of pulling out the victory.
So while I think a double digit lead prediction for Obama is probably way out of line, and may indicate something funny within the poll internals, my guess is that the result-an Obama win is the correct one, even if the margin is much smaller.
I would go on about the difference between sampling error and bias, but somehow I don't think you are attempting to make a subtle point about statistical analysis.
Do you think that refusal to take a poll is correlated with political leaning? I don't see the connection.
However, most of these pollsters normalize by various factors, including political leaning, which would eliminate the skewing effect.
Would it only be so.
In the campaigns of 1856 and 1860, local newspaper would publish polls of the passengers of steamboats docking in town on their way through, or of a passenger car on the railroad. Since there has been some interest in polls of some sort for at least the last 152 years, I doubt they will go away just by being wrong once.
You really don't see a likely correlation between refusal to take a poll and political leanings? If voter A says "I'd love to tell you about my political opinions! How much time do you have?" and voter B says "My political opinions are none of your damned business. Don't call back. [click]", you don't think voter A is more likely to be for Obama and voter B for McCain?
I'm a public-school teacher, and my principal sent out an e-mail on Friday telling all employees that we can say and do what we want off campus, but we must keep our political opinions to ourselves in the classroom and anywhere else on campus where students are. (Apparently there had been complaints.) As far as I can tell, all the McCain voters were already doing exactly that. I certainly refuse to tell my students how I vote (or about my religious beliefs) when they ask. I'm proud of the fact that students who have tried to guess how I vote over my 15+ years of teaching have a 50% success rate. This time around, some have concluded that my refusal to tell them means I'm voting for McCain, so even 17-year-olds think there's a correlation between reticence and Republicanism.
The only poster I've seen on the wall at school is for Obama. I haven't complained or put a McCain poster next to it, because it's in the teacher's lounge, which is forbidden territory for students.
Don't believe them- they are skewed toward favoring the Democrats from faulty samples- and tryng to discourage you from going to the polls.
Every vote COUNTS!
So bring as many like-minded friends as you can (and perhaps a lawn chair) and Vote McCain-Palin 2008
We have to show up in force- to defeat the voter fraud by ACORN!
And please, don't waste your precious vote on a write-in for Hillary, Huckabee, Nader, Paul, Barr, Romney or Baldwin. God Bless them ALL!
It's a 2 party race people- and it's going to be close.
It will either be McCain or Obama Nov. 4.
Don't let the polls, the media or long lines at the voting polls discourage you!
Get out the VOTE Nov. 5th
McCain-Palin 2008
God Bless John McCain and Sarah Palin
God Bless America
God Bless Our Troops
www.JohnMcCain.com
www.ObamaExposed.com
www.ObamaFiles.com
www.ObamaCrimes.com
www.ObamaWTF.com
IIRC, the non-response rate is very very large. If you ask enough questions to do the demographic balancing, it drives the non-response rate up. It also adds in the variation due to the inaccuracy of how to determine those demographic balances.
Perhaps a better way would be to go for a higher response rate with a one or two question poll.
No.
Nice opinion piece from the bloggers at Powerline: A Closing Argument for John McCain
Not all high school graduates, of course. Just the one's who couldn't manage to advance any further. McCain still leads among high school drop outs...
Answering a couple of other points raised above:
1) Polls in the PA primary did not show Obama and Clinton even. The majority of the late PA polls had Clinton up by 7-10 points, which pretty nicely brackets her 9-point margin. The RCP average is skewed by a single PPP poll, with Obama up by 3, which proved to be a serious outlier.
2) Pre-election polls did not "call Gore and Kerry as big winners" in 2000 and 2004. Keep in mind that pre-election polls are not exit polls. In 2000, most pre-election polls had Bush's share of the vote at about 47-48%, which was just about spot-on. Many polls seem to have gotten the distribution of votes between Gore and Nader somewhat wrong--they overstated Nader's vote a little, and therefore showed Gore slightly behind in the popular vote.
In 2004, the large majority of late polls had Bush narrowly ahead; only a couple of polls came out in the late days that showed Kerry up by 1-2 points. The general picture they give is of a close race in which Bush is the likely winner--again, pretty much how it turned out.
PollingReport.com has some good data on this subject.
Since coolpillow's pitch included, "Get out the VOTE Nov. 5th," I'm guessing s/he's either: (1)kidding, (2) trolling, or (3) even less rational than you think.
Several people have answered you, but here's some more data. What you're calling a 10-point error was a 3.1-point error. The RCP polling average was Clinton by 6.1. The final result was 9.2. Big deal.
RCP now shows Obama ahead in PA by 7.3. So more than twice the size of the earlier error.
But by all means, keep hope alive.
I hear they're putting pollsters in shower stalls and shrubbery to ambush the resisters.
[BTW, I'm a liberal and I decline telephone polls.]
Could your problem with the polls be that they haven't cooperated with your race-baiting fantasies?
BTW sampling survey theory is a special and isolated field within statistics. They use different methods and terminology from most mathematical statisticians. The whole thing can get pretty esoteric.
We’ll see how all the polling did on Tuesday.
"Could your problem with the polls be that they haven't cooperated with your race-baiting fantasies?"
Your remark is extremely unfair and rude. Do you think it contributes to the discussion here? Don't you think it's much better to point out a flaw in someone's data or reasoning than fling insults?
Boulder (Univ. of Colorado)- 78394 out of 121503 mail in ballots returned (64.5%)
Denver- 101510 out of 194631 mail in ballots returned (52.1%)
Arapahoe County (Denver Burbs)- 109756 out of 206494 returned (53.1%)
El Paso (colo springs)- 96570 out of 153357 returned (62.9%)
Jefferson- 150647 out of 228753 requested (65.8%)
Among mail in ballots Republicans have returned 64.9% of all the mail in's they requested
Democrats have returned 63.3% of the mail in ballots they requested
Unaffiliateds have returned 54.3% of the mail in ballots they requested.
Assuming 92% of Republicans support McCain, 86% of Dems Support Obama and Indies split evenly, the PROJECTED vote totals thus far look like this:
McCain - 665255 51.5%
Obama - 625698 48.5%
Can you tell where your numbers came from? The numbers here are different:
You and I have had many exchanges, and though I find some of your ideas objectionable (as you may well find mine), we've always addressed each other respectfully, i.e., we've disagreed without being disagreeable. I'm a big believer in that, and often a pain in the ass to others who aren't.
If you read both the comment I responded to and the one I linked, you noticed he went out of his way to be racially provocative and I'd say, though you may disagree, racially offensive. I believe "race baiting" is a conservative description. In his previous post he also made certain predictions about how the polls would bear out his demeaning, insulting assessment of Obama. Clearly he was wrong.
I don't think another commenter's blatant violation of any reasonable standard of civility gives me license to do the same. My comment was pointed, and frankly right at the edge of what I consider civil. But since it was: (1) directed only at his remarks, not at the person; and (2) factually accurate (or at least plausible -- I was after all speculating), I don't think I stepped over the line. But if you disagree I'd welcome your explanation. I'd rather maintain whatever civility survives on these boards than be right.
However, most of the polling data I've seen the last several weeks suggests that independents are going for Obama by a moderate but significant margin.
You probably already know it in (the original) Yiddish, but just in case, here it is. I make no promises about the quality of my transliteration:
Az di bube gehut beytsim vult zi geven mayn zeyde.
(pls. don't tell me to Google it:)
Obama 51.0%
McCain 47.9%
Of course, all of this is speculation. We'll know for sure on Tuesday.
"I don't think I stepped over the line. But if you disagree I'd welcome your explanation. I'd rather maintain whatever civility survives on these boards than be right."
I disagree. You don't need to go the extra mile to maintain civility at the cost of being right.
Hopefully civility will return after Wednesday!
And FWIW, you were not unfair in your response.
I read the original Paul Milligan post. While parts of it are provocative, inaccurate and poorly expressed, I don't find it so out of bounds as to warrant a personal attack on your part. He didn't insult you did he? If you think he's wrong then tell us why. If he thinks that race explains actions and events, then he should be free express his opinion or analysis without receiving a vituperative response crafted to cut off debate. We cannot ignore race as an explanatory variable in certain aspects of average human behavior. Here I'm expressing a statistical fact as opposed to a judgment about a particular individual.
In any case, even if he went over the line, we should make every effort to avoid doing the same-- two wrongs don't make a right.
BTW, I have never found your ideas "objectionable," merely inaccurate, and I hope I have never given any impression to the contrary. I tend to think in quantitative terms and this sometimes leads to my being misinterpreted. I try for clinical detachment, but don't often succeed.
"I have read on more than one site (sorry, can't remember which ones now) that the hang-up rate has been going up for years, and is now over 80%. That seems to put even the most 'scientific' polling technique in a class with chicken entrails and tarot cards."
Excellent point. If the hang up rate is really that high, then it argues for that population as being no different at least with respect to political opinions. The response to cold calls differ markedly by geography. Based on the testimony of cold callers, the South and the Midwest are the most polite parts of the country. California and New York are most rude. So this raises the possibility of an induced geographical bias.
Yes. I've been speculating in a similar manner. McCain's birthplace is not an issue, and I think it's obvious he won't win. But if he does, it could go like this.
McCain wins. McCain dies in office. Palin takes over. She proceeds to quickly do a bunch of really dumb and corrupt things, a la Troopergate. Her approval ratings dive below Dubya's. Ds have a huge majority in Congress. They vote overwhelmingly to impeach her. Nancy Pelosi becomes our second female president.
The Troopergate stuff would not lead to her being impeached, either before or after she's sworn in. But she would be likely to do other stuff like that, ultimately leading to her impeachment.
But this is all very speculative, because McCain is so far behind.
The exit polls in 2004 were wrong, but that has zero to do with the with actual preelection polls in question here. I think it is important that people not say false things, regardless of their ideology.
She was not tried by a court, but an investigator hired by the Alaska Legislature found that she "abused her power as governor [in] attempting to get Trooper Wooten fired."
That was a hypothetical scenario kind of post. If you note that I started with "if McCain wins ... "
Someone has already answered the question; she has not been tried. Whatever investigation the Alaska legislature carried out, stated that she has abused power (as read in the title of the news story, and then the body of the news report went on to tell us she has not broken any laws or something to that effect)
no dice. if McCain dies in office and Palin gets promoted to Pres, there will be a new VP, and if Palin gets impeached in office, the new VP will become the Pres.
the only fantasy resulting in Pelosi being the president has to be a quickie affair without a lot of time between mccain's event (disqualification/incapacitation/death) and palin's event (impeachment).
Best thing to do is to go vote, and wait for Tuesday night.
Mark Blumenthal had a National Journal column on this issue that goes into the ins and outs of it pretty well. I've never had success putting links up here, but it shouldn't be too hard to Google for it.
In other news, Jeffrey Hart endorses the secret Muslim commie.
"The good people will turn out in droves."
And statements like this will turn out the other good people in droves as well to vote for your opponent.
"You can make a killing on Intrade."
Exactly. Intrade = signal. Polls = noise.
;-)
Either the news report is wrong, or you're wrong. She did break the law, and then lied about it. See here.
Of course you're absolutely right. It was silly of me to not think of that.
Where did you get your degree in mind-reading? My silence could mean many things things other than what you think it means.
And while I don't think there's any point in turning this into a thread about Pelosi, I think it's worth pointing out that no investigation has found that she has abused the power of her office.
His prediction was comically wrong, and pointing it out constituted much of my comment. But that's not what I objected to.
Intentions are frequently opaque. One person's quantitative analysis is another's cover for bigotry. I don't assume bad intentions without evidence. For example, I don't ascribe racial animus to your arguments about Obama's LSAT's, his IQ, etc., even though racists parrot similar lines as socially defensible alternatives to their real feelings. And yes, it's simple logic that just because a racist say "XYZ" doesn't mean saying "XYZ" makes one a racist.
But there's no ambiguity here. PM's personal antagonism toward Obama is obvious in statements like "he will be wheeled into the General election, drooling about 'change'." In that context referring to him as "Brobama" and accusing him of wanting to blame his failures on "Whitey" might or might not signal racial hostility, but at the very least they constitute race baiting -- i.e., they're intended to offend, possibly in the hope of provoking an accusation of racism which can be turned against the offended.
So when you say you "don't find [PM's comments] so out of bounds as to warrant a personal attack on [my] part," I agree insofar as, like you said, "two wrongs don't make a right." But I find them intentionally obnoxious and beyond the pale.
But apart from being snarky, I still don't know what you thought was over the line about my response. Here's what I said: "Could your problem with the polls be that they haven't cooperated with your race-baiting fantasies?" If I replaced the last three words with "predictions," would you agree it's an inoffensive question that offers a speculative answer, which on the evidence may be correct? If so, it comes down to the last three words. I think "fantasy" is an accurate description of PM's prediction in May that come October Obama would never see the north side of down 10% in the polls. And "race-baiting" I've already covered. So please be more specific about where you think I went too far.
You haven't. That said, I wouldn't presume to tell you what you should or shouldn't find objectionable. What pushes our buttons is something we have limited control over. How we express those objections is another matter, and whatever you may have felt, you've been unfailingly civil.
My emotions get the better of me all the time, which is one reason I believe in ground rules for discourse.
Skepticism about and noncooperation with the media is evenly distributed among both parties? On what planet?
As someone who often disagrees with zarkov, I want to chime in and say that this has been my experience too.
The story says this:
In the process of explaining that Palin told a brazen, transparent lie, the reporter throws Palin a bone by saying that Palin's claim about the report being "a partisan smear job is debatable." Trouble is, it's not. The Legislative Council which voted unanimously on 7/28 to hire Branchflower is composed of 8 R and 4 D. And Palin was obviously not a candidate at that time. And in the month of August, Palin repeatedly expressed her support for the investigation, and promised to cooperate with Branchflower. It was only after she got the call from McCain that she decided to smear Branchflower by accusing him of partisanship. Interesting coincidence.
And the LC voted unanimously on 10/10 to release the report.
The facts in the report speak for themselves. Show us where partisan bias is embodied in the report. You can't. Palin hasn't even tried. At the very least, the report proves beyond any doubt that Palin was telling another brazen, transparent lie when she said this: "no pressure was ever put on anybody to fire anybody."
Palin is a serial fabricator. There's no room for any doubt about that. As documented above, she told "the reverse of the truth." Why is this OK with you? I guess for the same reason it's OK with you that Bush and McCain are also serial fabricators.
i am not sure how to post the link, but here is the actual report of the investigation, as opposed to what some political blog selectively quotes to suit their point.
download2 dot legis dot state dot ak dot us/DOWNLOAD.pdf
on page 8,
Finding 1.
...Palin abused her power by violating Alaska statute 39.52.110(a) of the Alaska Exec Branch Ethics Act.
Finding 2.
Walt Monegan's refusal to fire Wooten was not the sole reason he was fired by Gov Palin, it was likely a contributing factor ... Inspite of that Gov Palin's firing of Monegan was a proper and lawful exercise of her constitutional and statutory authority to hire and fire exec branch dept heads.
What I noted was Palin's opponents cling to Finding 1, while her supporters to Finding 2.
Here is the pdf.
Many other primary documents, including the police interviews of Sarah Palin et al from 2005, can be found via the wiki article. google troopergate.
I think you did a nice job of summarizing the report.
Not so unanimous as jukeboxgrad pretends, is it?
The fact that all 12 members of the Legislative Council are appointed by just two people makes their collective opinion (such as it is) far less persuasive. Perhaps Alaskan readers can tell us whether the President of the Senate and Speaker of the House are enemies of Palin and whether they put other enemies on the committee. We know she shook up the corrupt Republican power structure of the state and made quite a few enemies in the process.
Further on in the AP story:
That last sentence seems to me to be quite sufficient to discredit the report utterly. Why should we believe any of it?
All you actually have is one R who said "I think there are some problems in this report." Yet he voted to release it, despite those alleged "problems." If he though it was a "partisan smear job," then why vote to release it? If he thought the "problems" were serious, then why vote to release it? And let us know if you can find Stevens anywhere telling us what those "problems" are.
The vote speaks for itself. Any lawmaker who didn't approve of the report was free to vote to not release it. This many did so: zero. A vote to release the report is a vote showing approval of the report.
You are completely failing to address the fact that Palin repeatedly expressed support for the Branchflower investigation. Until she got the call from McCain. Why would she do so if the report was instigated by "enemies of Palin?"
And if all Rs and Ds are her enemies, then who has standing to investigate her? No one? Only people she certifies as her friends? Only people appointed by the governor (like the Personnel Board)? How can you have a democratic system where an executive claims to be immune from any investigation that isn't run by their friends?
Nice try. Palin was already trying to obstruct Branchflower, even before French made that statement. And there's nothing wrong with the statement he made. French didn't say it would happen. He said it could happen. Those words are materially different. And there's absolutely nothing wrong with him saying that. As Judge Michalski said:
Further, French didn't do the investigation. Branchflower did. Aside from that, the facts in the Branchflower report speak for themselves. Which of those facts do you dispute or deny? It's now abundantly clear that Palin was lying when she said "no pressure was ever put on anybody to fire anybody." Why is it OK with you that she said that?
Because Branchflower took statements under oath from credible witnesses, and you can read the statements in the report, and they prove that Palin broke the ethics law. And they prove that Palin is a liar. And Palin herself has not disputed the facts described by the witnesses. Instead, she simply told a lie about the conclusion of the report. She brazenly stated that the report said something different from what it actually said.