[Kerrigan]

Hmm. The latest Gallup poll indicates that Obama is ahead by only two points, 49 to 47, within the poll's margin of error.

This comes after a terrible month for McCain. This ought to remind people that nothing is certain, and that Obama shouldn't be measuring the drapes just yet . . . .

[Eug]

Originally Posted by Kerrigan 

Hmm. The latest Gallup poll indicates that Obama is ahead by only two points, 49 to 47, within the poll's margin of error.

This comes after a terrible month for McCain. This ought to remind people that nothing is certain, and that Obama shouldn't be measuring the drapes just yet . . . .

Front-runner Obama cautions against overconfidence - CNN.com

"For those of you who are feeling giddy or cocky and think this is all set, I just [have] two words for you: New Hampshire," the Democratic presidential nominee said during a fundraiser breakfast in New York. "You know I've been in these positions before where we were favored and the press starts getting carried away and we end up getting spanked. And so that's another good lesson that Hillary Clinton taught me."

[Dakar V]

I guess this means McCain's aggressive performance last night rallied the troops.

[Kerrigan]

I saw that earlier today, Eug. Obama is certainly smart about lowering expectations. The media wants to paint the election as already wrapped up, but Obama realizes that this could generate a counter-productive sense of complacency that could affect turnout.

Smart money is still on Obama, but it's becoming evident that McCain is actually within striking distance. He does indeed have a shot at winning by pressing the tax-and-spend issue, and the unfolding Acorn scandal is certainly not going to help Obama.

[stupendousman]

Smart money isn't on Obama. Especially given the number of undecideds, the "bradley effect" and the fact that whatever lead he did have was enhanced by unrealistic weighting of Democrats in the sample. If it's a statistical tie now in the polls, that pretty much means that Obama is down in reality. He needs to be up at least 6 just to be even. He's NEVER going to get double the record turnout that some pollsters are giving him in the polls.

[Sayf-Allah]

http://news.yahoo.com/s/nm/20081016/...aker_obama_odd

[Sayf-Allah]

http://www.fivethirtyeight.com/

[Kerrigan]

I hit the trifecta once in Kentucky on a $1 bet, it paid off pretty well. I wouldn't bet the house on McCain though. Or either of the candidates for that matter, it's just too close.

[BRussell]

Another Kerrigan poll thread! If ever there's an outlier poll, Kerrigan starts a new thread. Maybe someone should start a thread about yesterday's poll showing Obama leading McCain 53-39.

[ebuddy]

Originally Posted by Kerrigan 

I hit the trifecta once in Kentucky on a $1 bet, it paid off pretty well. I wouldn't bet the house on McCain though. Or either of the candidates for that matter, it's just too close.

You mean there's a chance neither of them will be elected?

[hyteckit]

Originally Posted by stupendousman 

Smart money isn't on Obama. Especially given the number of undecideds, the "bradley effect" and the fact that whatever lead he did have was enhanced by unrealistic weighting of Democrats in the sample. If it's a statistical tie now in the polls, that pretty much means that Obama is down in reality. He needs to be up at least 6 just to be even. He's NEVER going to get double the record turnout that some pollsters are giving him in the polls.

Bradley effect? You mean racist people who won't vote for Obama because he is Black?

[Chongo]

Originally Posted by hyteckit 

Bradley effect? You mean racist people who won't vote for Obama because he is Black?

No, they tell pollsters and exit polls they will/did vote for for an AA candidate to avoid being called a racist.
http://en.wikipedia.org/wiki/Bradley_effect

The Bradley effect, less commonly called the Wilder effect,[1][2] is a proposed explanation for an alleged discrepancy between voter opinion polls and election outcomes in American political campaigns when a white candidate and a non-white candidate run against each other.[3][4][5] Named for Tom Bradley, an African-American who lost the 1982 California governor's race despite being ahead in some voter polls, the Bradley effect refers to an alleged tendency on the part of some voters to tell pollsters that they are undecided or likely to vote for a black candidate, and yet, on election day, vote for his/her white opponent.

The theory of the Bradley effect is that the inaccurate polls have been skewed by the phenomenon of social desirability bias.[6][7] Specifically, some white voters give inaccurate polling responses for fear that, by stating their true preference, they will open themselves to criticism of racial motivation. The reluctance to give accurate polling answers has sometimes extended to post-election exit polls as well. The race of the pollster conducting the interview may factor in to voters' answers.

[hyteckit]

I think Kerrigan must be reading Drudge Report again, a site known to make up stories. First it's the fake Ophra story of her being a "Reverse Racist". Now, a made up Gallup Poll. How about going to Gallup directly instead?

http://www.gallup.com/poll/111211/Ga...McCain-43.aspx

[stupendousman]

Originally Posted by hyteckit 

I think Kerrigan must be reading Drudge Report again, a site known to make up stories. First it's the fake Ophra story of her being a "Reverse Racist". Now, a made up Gallup Poll. How about going to Gallup directly instead?

http://www.gallup.com/poll/111211/Ga...McCain-43.aspx

I went to your link. Here's what I found:

"Today's results show Obama with a two-point advantage over McCain using this likely voter model, 49% to 47%, this is within the poll's margin of error."

You're confusing the 4 different versions of the poll they have out. The one you are likely looking at is the new "expanded" weighting formula, that apparently counts new ACORN registrants "Mickey Mouse" and "Seymour Butts" as people who'll vote for Obama. See my "fantasy poll" thread for an outline of how these new methods, designed to give point bumps to Democrats, are bogus. That is unless you really believe that the Democrats will likely double their all time turnout lead record in a year where Republicans are energized and the "bradley effect" is likely in full force.

GOOD LUCK!

(ps. You owe Kerrigan an apology.)

[Timothy Leary's brain]

Originally Posted by stupendousman 

I went to your link. Here's what I found:

"Today's results show Obama with a two-point advantage over McCain using this likely voter model, 49% to 47%, this is within the poll's margin of error."

You're confusing the 4 different versions of the poll they have out. The one you are likely looking at is the new "expanded" weighting formula, that apparently counts new ACORN registrants "Mickey Mouse" and "Seymour Butts" as people who'll vote for Obama. See my "fantasy poll" thread for an outline of how these new methods, designed to give point bumps to Democrats, are bogus. That is unless you really believe that the Democrats will likely double their all time turnout lead record in a year where Republicans are energized and the "bradley effect" is likely in full force.

GOOD LUCK!

(ps. You owe Kerrigan an apology.)

That's funny. I went to the link and found this.

This estimate would take into account higher turnout among groups of voters traditionally less likely to vote, such as young adults and minorities. That model has generally produced results that closely match the registered voter figures, but with a lower undecided percentage, and show Obama up by six percentage points today, 51% to 45%.

There's also a new phenom out there call the 'reverse Bradley effect'.
http://www.google.com/search?client=...UTF-8&oe=UTF-8

GOOD LUCK, YOU LOSE!

I think he owes Kerrigan buttskwat. But you owe everyone an apology.

[hyteckit]

Originally Posted by stupendousman 

I went to your link. Here's what I found:

"Today's results show Obama with a two-point advantage over McCain using this likely voter model, 49% to 47%, this is within the poll's margin of error."

You're confusing the 4 different versions of the poll they have out. The one you are likely looking at is the new "expanded" weighting formula, that apparently counts new ACORN registrants "Mickey Mouse" and "Seymour Butts" as people who'll vote for Obama. See my "fantasy poll" thread for an outline of how these new methods, designed to give point bumps to Democrats, are bogus. That is unless you really believe that the Democrats will likely double their all time turnout lead record in a year where Republicans are energized and the "bradley effect" is likely in full force.

GOOD LUCK!

(ps. You owe Kerrigan an apology.)

Read the headlines, it's 49% to 43% according to the Gallup Poll.

Gallup Daily: Obama 49%, McCain 43%

The 49% to 47% was to account for the Bradley Effect.

Funny how right-wing conservatives argue Drudge Report has no bias.

[stupendousman]

Originally Posted by hyteckit 

Read the headlines, it's 49% to 43% according to the Gallup Poll.

The "new coke" formula. The one that uses unrealistic democrat turnout figures. The same crazy "quadruple their record" sort of rational we've discussed in this thread:

http://forums.macnn.com/95/political...ntasy-polling/

The one that uses the same tried and true methods they've always used shows a tie.

[besson3c]

There is no point in arguing either way over a one day snapshot produced by Gallup on the national vote. http://pollster.com is where it's at.

[besson3c]

Also, Obama's lead should be no surprise. It is more of a surprise why it isn't even larger. Historically speaking an incumbent party has never won when the economy is failing, and this particular incumbent is now the most unpopular president in the history of polling.

[stupendousman]

Originally Posted by besson3c 

Also, Obama's lead should be no surprise. It is more of a surprise why it isn't even larger. Historically speaking an incumbent party has never won when the economy is failing, and this particular incumbent is now the most unpopular president in the history of polling.

What does history show when the opponent is a member of the most unpopular Congress in history? Even more unpopular than the President? Oh..yeah, I don't think that's ever happened before, which is why all the assurances that Obama should rule are gross exagerations.

There is no past precedent for what's going on now, and if you do look at past trends there's no real reason for anyone to believe that Democrat turnout will quadruple their record when they did their absolute best, and the Republican base was discouraged. That's the sort of thing you have to believe if you want to believe that the polls are much more than a statistical tie. Even to polls showing a close race ignores the fact that there are likely as many people who will make a point of going out to vote against Obama than there are those who never voted before who'll go and vote for him. That's why at best you really can't give him more than the 3 or 4 point record they've had occasionally in the past and why the polls that give him 12-16 points are jokes.

[besson3c]

Originally Posted by stupendousman 

What does history show when the opponent is a member of the most unpopular Congress in history? Even more unpopular than the President? Oh..yeah, I don't think that's ever happened before, which is why all the assurances that Obama should rule are gross exagerations.

There is no past precedent for what's going on now, and if you do look at past trends there's no real reason for anyone to believe that Democrat turnout will quadruple their record when they did their absolute best, and the Republican base was discouraged. That's the sort of thing you have to believe if you want to believe that the polls are much more than a statistical tie. Even to polls showing a close race ignores the fact that there are likely as many people who will make a point of going out to vote against Obama than there are those who never voted before who'll go and vote for him. That's why at best you really can't give him more than the 3 or 4 point record they've had occasionally in the past and why the polls that give him 12-16 points are jokes.

There is plenty of past precedent. The voting public doesn't even know who their congress people are. Why is it that all polls show Democratic gains in congress despite being so unpopular? The popularity of congress really doesn't yield very much useful information, historically speaking.

[Powerbook]

Originally Posted by Kerrigan 

I saw that earlier today, Eug. Obama is certainly smart about lowering expectations. The media wants to paint the election as already wrapped up, but Obama realizes that this could generate a counter-productive sense of complacency that could affect turnout.

Well I guess the Democratic campaign still remembers how the lying cvnts of the Bush "CRP" swift-boated Kerry as a unpatriotic traitor, whereas Bush was a war hero and accomplished missions. [ ]

PB.

[stupendousman]

Originally Posted by besson3c 

There is plenty of past precedent. The voting public doesn't even know who their congress people are. Why is it that all polls show Democratic gains in congress despite being so unpopular?

Quite possibly the polls giving democrats an extra 10-16 points in weight, when they've NEVER done better than 4?

[besson3c]

Originally Posted by stupendousman 

Quite possibly the polls giving democrats an extra 10-16 points in weight, when they've NEVER done better than 4?

Huh? This is a different argument. Mine is that there isn't a relationship between the popularity of Congress and who people actually vote for at the congressional level. Since Congress is polling so incredibly low, the polls should show blowout gains for Republicans in Congress, right?

[TheMosco]

Originally Posted by stupendousman 

Quite possibly the polls giving democrats an extra 10-16 points in weight, when they've NEVER done better than 4?

Not all the polls are giving such a heavy weight btw.

[SpaceMonkey]

Originally Posted by stupendousman 

The "new coke" formula. The one that uses unrealistic democrat turnout figures. The same crazy "quadruple their record" sort of rational we've discussed in this thread:

http://forums.macnn.com/95/political...ntasy-polling/

The one that uses the same tried and true methods they've always used shows a tie.

So you're argument is that we should be polling the tried-and-true voters of the 2000 and 2004 elections then? Let us know when you've invented that time machine.

Given all of the Great Depression analogies that have been thrown around, perhaps we should be polling the 1932 public.

[stupendousman]

Originally Posted by besson3c 

Huh? This is a different argument. Mine is that there isn't a relationship between the popularity of Congress and who people actually vote for at the congressional level. v Since Congress is polling so incredibly low, the polls should show blowout gains for Republicans in Congress, right?

Bush isn't up for reelection regardless of what Obama tries to claim, so it's not like they are voting against an incumbent. On the other hand, the guys in Congress ARE the incumbents - all of them. You aren't going to get an accurate representation from the polls, because they are using weighting which overestimates the number of Democrats voting.

ACORN gets 200,000 phony voter registrations and the polls give Democrats extra weighting and say that despite NEVER getting more than 4 points over Republicans in turn-out, they'll get quadruple that this time. Sure...pull my leg.

[stupendousman]

Originally Posted by SpaceMonkey 

So you're argument is that we should be polling the tried-and-true voters of the 2000 and 2004 elections then? Let us know when you've invented that time machine.

Uh..no. Just using the same methods and formulas. Even with those, they still managed to distort Democrat turn-out.

[stupendousman]

Originally Posted by TheMosco 

Not all the polls are giving such a heavy weight btw.

True. The ones that are at least reasonable (they give Dems 4-6 points) usually show it between a 2-6 point race.

[SpaceMonkey]

Originally Posted by stupendousman 

Uh..no. Just using the same methods and formulas. Even with those, they still managed to distort Democrat turn-out.

Well I noticed that in the other thread, in the explanation by zerostar (who seems to be writing out of personal experience) the polling his firm had conducted in his area using 2004 voter registration still showed Obama up by 9, so perhaps you are yet again complaining about something that you don't have any special knowledge of?

Not sure why you care so much about the polls, anyway.

[besson3c]

Originally Posted by stupendousman 

Bush isn't up for reelection regardless of what Obama tries to claim, so it's not like they are voting against an incumbent. On the other hand, the guys in Congress ARE the incumbents - all of them. You aren't going to get an accurate representation from the polls, because they are using weighting which overestimates the number of Democrats voting.

ACORN gets 200,000 phony voter registrations and the polls give Democrats extra weighting and say that despite NEVER getting more than 4 points over Republicans in turn-out, they'll get quadruple that this time. Sure...pull my leg.

The amount of skew from ACORN, if any and applicable to the state is pure speculation, and wishful thinking on your part. I made a simple point based on historical precedence, whether you wish to acknowledge my point as being factual or not is up to you, I'm not interested in trying to challenge your speculation.

[stupendousman]

Originally Posted by besson3c 

The amount of skew from ACORN, if any and applicable to the state is pure speculation, and wishful thinking on your part.

The ENTIRE effort made in "weighting" is based on pure speculation. No one knows what turnout is going to be. We have past precedent, but a lot of pollsters are totally ignoring that and using fantastic numbers that make little logical sense. If you are asking about polls, you are asking about something whose results are based on a good part of speculation.

[besson3c]

Originally Posted by stupendousman 

The ENTIRE effort made in "weighting" is based on pure speculation. No one knows what turnout is going to be. We have past precedent, but a lot of pollsters are totally ignoring that and using fantastic numbers that make little logical sense. If you are asking about polls, you are asking about something whose results are based on a good part of speculation.

My original point was not about polls, but historical precedence.

[hyteckit]

Originally Posted by stupendousman 

The "new coke" formula. The one that uses unrealistic democrat turnout figures. The same crazy "quadruple their record" sort of rational we've discussed in this thread:

http://forums.macnn.com/95/political...ntasy-polling/

The one that uses the same tried and true methods they've always used shows a tie.

I hope you are not accusing Gallup of fixing polls to be weighted in favor of Democrats.

[hyteckit]

Oh look, Obama is up another point today based on the Gallup Poll.

http://www.gallup.com/poll/107674/Ga...tion-2008.aspx

It's now:

Obama - 50%
McCain - 43%

[stupendousman]

Originally Posted by hyteckit 

I hope you are not accusing Gallup of fixing polls to be weighted in favor of Democrats.

WHICH Gallup poll? They've been releasing 4 different versions of their poll, often times all with the same time frame and and 5 points between them.

Their normal, "traditional" poll that doesn't invent new criteria to boost Obama's chances this year by giving outrageous turn-out estimates still shows a statistical tie.

[hyteckit]

Originally Posted by stupendousman 

WHICH Gallup poll? They've been releasing 4 different versions of their poll, often times all with the same time frame and and 5 points between them.

Their normal, "traditional" poll that doesn't invent new criteria to boost Obama's chances this year by giving outrageous turn-out estimates still shows a statistical tie.

What? The "traditional" poll is the weighted poll that is based on voting behavior, which is weighted in favor of McCain. You are against weighted polls right? So why go with the "traditional" poll, which is weighted?

Why not go with actual poll data from Gallup. The poll below, or the daily poll is not weighted, and is based on daily poll numbers of registered voters, not potential turn-outs.

http://www.gallup.com/poll/107674/Ga...tion-2008.aspx

[hyteckit]

The OP somehow thinks the Gallup Poll shows a big swing in favor of McCain in the last few days. It hasn't. The OP is simply wrong.

Gallup's own data and title: Little Impact From Debate So Far

http://www.gallup.com/poll/111241/Ga...ebate-Far.aspx

From Sept. 28 to Oct. 17:

Obama's numbers fluctuate between 48% to 52%. Median: 50%
McCain's numbers fluctuate between 41% to 44%. Median: 43%

Not much have change since the past 3 weeks. So, the OP is wrong.

[stupendousman]

Originally Posted by hyteckit 

What? The "traditional" poll is the weighted poll that is based on voting behavior, which is weighted in favor of McCain.

Actually, it's weighted based on past precedent. Something that can actually be proven. It's not based on an estimate but rather how likely the people being questioned are to vote. If someone has NEVER voted before, chances are he isn't going to start just because Obama is on the ballot. That's the sort of thing the polls giving up to 16 points are doing.

Why not go with actual poll data from Gallup. The poll below, or the daily poll is not weighted, and is based on daily poll numbers of registered voters, not potential turn-outs.

Because there is never going to be 100% voter turnout. In general, Republicans are more likely to vote than Democrats. When the Republicans are demoralized and the Democrats put on their "A Game", the best they do is a 4 point advantage. You can't look at that, and simply guess that the Democrats are going to quadruple that. At least not with a straight face.

[stupendousman]

Originally Posted by hyteckit 

The OP somehow thinks the Gallup Poll shows a big swing in favor of McCain in the last few days. It hasn't. The OP is simply wrong.

Gallup's own data and title: Little Impact From Debate So Far

http://www.gallup.com/poll/111241/Ga...ebate-Far.aspx

From Sept. 28 to Oct. 17:

Obama's numbers fluctuate between 48% to 52%. Median: 50%
McCain's numbers fluctuate between 41% to 44%. Median: 43%

Not much have change since the past 3 weeks. So, the OP is wrong.

Currently, it's a statistical tie, unless you're using the "new" weighting method which uses a bunch of guessing and estimates that aren't based on anything really real.

[hyteckit]

stupendousman, make up your mind. You are either for weighted polls or against them.

You complain about weighted polls:

http://forums.macnn.com/95/political...ntasy-polling/

Now, this thread you argue for weighted polls.


But is still doesn't change the fact there hasn't been much of a change in the polling numbers in the past 3 weeks. The Gallup "traditional" poll hasn't change much. The actual Gallup poll hasn't change much. So why even make up fake stories on how the Gallup poll now shows a big upswing in favor in McCain when it hasn't.

Even Gallup themselves have said: Little Impact From Debate So Far

[stupendousman]

Originally Posted by hyteckit 

stupendousman, make up your mind. You are either for weighted polls or against them.

I am for weighted polls when they are using the traditional methods to determine "likely voters", though even those tended to end up overly representing Democrat chances. I'm opposed to weighted polls that invent reasons to give Democrats a 16 point advantage when they've NEVER gotten above a 4 point advantage.

But is still doesn't change the fact there hasn't been much of a change in the polling numbers in the past 3 weeks. The Gallup "traditional" poll hasn't change much.

Great, then it looks like it's been a tie for weeks! Bad news for Obama. Experts say he needs to be at least 6 points in the polls to actually break even.

[besson3c]

Why obsess over the results of one daily tracker when there are now so many of them? Use the Pollster averages:

[stupendousman]

Originally Posted by besson3c 

Why obsess over the results of one daily tracker when there are now so many of them? Use the Pollster averages:

The aggregate sites don't drop the far outliers and those negatively effect the averages (likely the very reason they exist as outliers) in ways that likely misrepresent the current trends.

Here's the Stupendousman Super Poll of Poll average that takes all the non-partisan polls for the most recent polling period (gathered from Realclearpolitics.com), drops outliers that don't come close to being within the margin of error of the rest of the results, and gives preference over polls of likely voters who use traditional methods instead of some crazy forumula that assumes a 6-16 point advantage for Democrats which goes against everything we know from past precedent. This keeps things current, and fair for both parties. It really makes no sense to use polls using data for about a week or so ago either.

Here's the latest polls that fit that category and how they average:

Reuters/C-SPAN/Zogby 10/15 - 10/17 Obama +4
Rasmussen Reports 10/14 - 10/16 Obama +4
Gallup (Traditional)* 10/14 - 10/16 2155 Obama +2
GWU/Battleground 10/12 - 10/16 Obama +4
IBD/TIPP 10/12 - 10/16 Obama +5

That comes to 3.8, just shy of the 3 points the usual "margin of error" is.

Now, using this method, I'd also drop a poll that showed McCain up say, 2 points or more, and any poll put together by just Republicans for Republicans. GIven the fact that the polls are ALL OVER THE PLACE as far as the race goes (and you know that it can't be both a 2 and a 14 point race) this method gives added weight to polls obviously used the same tried and true methods of determining who'll show up on Nov. 4th.

Again, I took the middle of the pack as far as polls go, and started pruning from the top and bottom only the polls that were well beyond the margin of error for the polls nearest to them on either ends. This punishes polls which are (likely deliberately) exaggerating turn-out for either party. The fact is, there was only one poll that was a point or two away that was dropped, and that was Hotline at +10, which is still 8 points away from the lowest poll on the scale which still has results within the margin of error of the poll nearest it. If you include them in the equation, it puts the percentage at 4.8. If you include all the polls but the partisan ones that have results within about the last week or so, you've got a 6.6 average

As you can see, if you include crazy outliers, and old polls it really effects the averages in ways that don't reflect the most current trends.

Based on a reasonable analysis of the polls, it would seem we've got about a 4 point race. McCain has to convince 2 percent to switch sides, and/or just get one percent and just half the undecided (which Obama hasn't closed the deal on). This doesn't even take into account the "Bradley Effect" that pollsters witnessed during the primaries that experts say would mean Obama would need over 6 points to overcome.

So, I think it's wishful thinking that a reasonable analysis of the polls show a dead heat right now, even if a few respected polls have shown that of late. But, it's not unreasonable to point out that if you just look at the polls which seem to be doing things the way they've always done it with acceptable results (that still usually exaggerated Democrat turn-out) it's as McCain's people have said "within striking distance" and even more likely if you include variables that the polls are totally ignoring.

If you look at the polls today, the only way Obama would get the turnout the 8-14 pointers claim is via a miracle from God or massive voter fraud. I don't doubt the latter is possible after seeing the current events in the news, but I'm pretty sure God isn't going to help Rev. Wright's man this time.

[besson3c]

Heh... you need to read the Pollster methodology to learn how they come up with their numbers. If you did you'd realize that they account for outliers. Then, you need to look at the historical accuracy of pollers like Zogby to learn that they suck.

You'll excuse me if I don't put too much stock into your methodology of essentially massaging the results to work in favor of your argument, but I'll hold you to this when or if these same pollsters show bigger leads to see how your methodology shifts

[besson3c]

Also, not that it matters, but you ought to take these trackers over a 3 day period, not just a single day. Ras over a 3 day period has Obama up by 5:

http://www.pollster.com/blogs/us_oba...rasmusse_7.php

[stupendousman]

Originally Posted by besson3c 

Also, not that it matters, but you ought to take these trackers over a 3 day period, not just a single day. Ras over a 3 day period has Obama up by 5:

http://www.pollster.com/blogs/us_oba...rasmusse_7.php

I believe it was 4 this morning and I was using the 3 day period.

I'll stick by my method. I'd bet it's a lot more likely to reflect the final totals in the end.

[besson3c]

Gallup doesn't update until 1 PM. I guess the Traditional view ought to be weighted to account for the following, huh:

Gallup's "traditional" likely voter model, which Gallup has employed for past elections, factors in prior voting behavior as well as current voting intention. This has generally shown a closer contest, reflecting the fact that Republicans have typically been more likely to vote than Democrats in previous elections.

http://www.gallup.com/poll/111124/Ga...aditional.aspx



The problem is, when you get to cherry picking weighting, which model to use (LV/RV), etc. your results do not take into account these sorts of variables. This is why CNN, Pollster, Realclearpolitics, Politico, etc. use the averages of all major pollsters, not the stupendousman average.

[stupendousman]

Originally Posted by besson3c 

Gallup doesn't update until 1 PM.

Okay. So here at 2:30, here's how the polls go based on my criteria:

Rasmussen Reports 10/15 - 10/17 Obama +5
Reuters/C-SPAN/Zogby 10/15 - 10/17 Obama +4
Gallup (Traditional)* 10/15 - 10/17 Obama +2
GWU/Battleground 10/12 - 10/16 Obama +4
IBD/TIPP 10/12 - 10/16 1020 LV Obama +5

That puts the race at 4. Even Gallup's "expanded" polling shows it a 4 point race. Pollster has it at about 6.

The problem is, when you get to cherry picking weighting, which model to use (LV/RV), etc. your results do not take into account these sorts of variables. This is why CNN, Pollster, Realclearpolitics, Politico, etc. use the averages of all major pollsters, not the stupendousman average.

It doesn't take into account variables that people are just guessing about. When someone guesses that the Democrats will get up to 16 percent greater turn-out than the Republicans, there's really nothing in any past election cycle that would support that as being realistic. Time and time again polls have overestimated Democrat turn-out. Kerry was sure that he'd win in 2004 based on polling estimates of turn-out. That sort of thing simply isn't accurate. Even using traditional methods have resulted in extreme over-estimation of Democrat turn-out.

Over the course of thirty years, Republicans have turned out in greater percentage than Democrats. That's not something that has to be guessed at or estimated. It's set past precedent. Democrats have to have an "A game" just to break even and when Republicans aren't energized that will get them at best 4 points. Suggesting they'll get 16 points or even 8 points up is fantasy, based on any reasonable analysis of trends and past precedent. There's no reason to take a poll seriously that engages in that kind of lack of professionalism.

I don't see anything wrong though with suggesting that Democrats are on their "A game" and even doing better than normal, but with Palin and the fierce opposition to Obama during a presidential election year (and a real potential for a "Bradley Effect"), there's no reason to believe that the Republican turnout won't be near the same as 2004. At best, you're going to have another 4 point advantage for Democrats. I think if you can find the internals for the polls that have it at a close race, those are probably the numbers they are using, and even those are likely the best case scenarios for Democrats.

[kobi]

Originally Posted by stupendousman 

Over the course of thirty years, Republicans have turned out in greater percentage than Democrats.

Unless you count the early-voting numbers in many of the swing states.

North Carolina Early Vote

In North Carolina on the first day of early voting:
64% Democrats
21% Republicans
15% Unaffiliateds

Illinois Record Turnouts

Illinois is having a record number of early voting. More than 187,000 have voted so far.