Obama/Biden vs. McCain/Palin

[jim]

Statistically most polls have a margin of error. As to if they really represent reality, we will find out tomorrow.

While N<1,000 or so seems like a small sample, statistically it's not. If you have a coin that can only have two outcomes, you don't need to flip it that often to determine that it is fair.

[Popeye_Card]

Has anyone ever thought that these polls might not be right?

I mean, I understand that if all of them are saying the same thing, then it's probably somewhat truthful, but aren't these polls are rather small sample sizes?

Also, I've heard multiple times now that the majority of early voters have been for Obama, does that mean that Obama has all these new voters for him, or just energized democrats wanting to get out and vote early? There's no way to tell that

I fully understand that the polls may not be right. And I do fear that some people look at polls, see their candidate comfortably ahead, so they don't bother to vote, letting the underdog close the gap.

But the states that will likely swing the election have been polled, and polled, and polled. When there's that many polls, you start to get a reasonably good sample size. Any of the states in toss-up territory could still flip, of course. But the way it's playing out so far, Obama has a 238-118 lead among "strong" states (10%+ margins). McCain needs virtually everything to go his way to have a chance. If you give Obama just half of the states he has a current ~5% lead in, he wins.

[JG1982]

Statistically most polls have a margin of error. As to if they really represent reality, we will find out tomorrow.

While N<1,000 or so seems like a small sample, statistically it's not. If you have a coin that can only have two outcomes, you don't need to flip it that often to determine that it is fair.

Meh, I'm not sure the coin argument holds much weight, because while I understand there's two choices, most people only see one choice, and if you happen to poll 750 registered democrats then your poll is going to be skewed that way..

That made sense to me..maybe not to anyone else

[jim]

Statistically most polls have a margin of error. As to if they really represent reality, we will find out tomorrow.

While N<1,000 or so seems like a small sample, statistically it's not. If you have a coin that can only have two outcomes, you don't need to flip it that often to determine that it is fair.

Meh, I'm not sure the coin argument holds much weight, because while I understand there's two choices, most people only see one choice, and if you happen to poll 750 registered democrats then your poll is going to be skewed that way..

That made sense to me..maybe not to anyone else

There is an entire science to polling. I wouldn't say "meh" to it, and while I do more or less believe the results I also believe that there is enough random variation where you could get results that differ from the polls.

[Fat Strat]

Has anyone ever thought that these polls might not be right?

I mean, I understand that if all of them are saying the same thing, then it's probably somewhat truthful, but aren't these polls are rather small sample sizes?

Also, I've heard multiple times now that the majority of early voters have been for Obama, does that mean that Obama has all these new voters for him, or just energized democrats wanting to get out and vote early? There's no way to tell that

I think that's always a concern. I would think you'd look at states where the results have been erratic for places where McCain might win. And the national poll is probably meaningless as far as the outcome.

But if one is outpolling the other in a state day after day and poll after poll, I think it becomes harder to dismiss. The sample size gets bigger and bigger as the polling goes on.

I agree, and would also like to add that Obama brings in aspects that pollsters both haven't faced before AND have always had trouble adjusting to. Turnout is very difficult to project and an atypical election (a black man facing a white man with a women VP) along with a candidate with a strong youth following makes for a difficult polling situation. I think that the polls are going to be further off than they normally are (and they're normally off by several points). What's yet to be seen is which way are the polls going to be wrong and which methods are going to end up being more accurate?

We'll find out tomorrow. I still think it will be closer than many think (popular vote nationally and popular vote in some swing states), but I think the end result will be as expected.

[clement]

These pollsters are a lot more smart and sophisticated than that. They all have models to make sure that the 750 or however many people they poll represent a reasonable cross-section of voters to sample from. That's why when you look at 50 polls from one state, they all more or less say the same thing. Different pollsters will use slightly different methodologies, which might explain why one poll might consistently show a particular candidate up by an average of 8 points and another poll might consistently show a particular candidate up by an average of 3 points. But aside from these kinds of discrepancies, if say, Wisconsin were to flip for McCain, the answer would not be because the polls were wrong.

[Fat Strat]

These pollsters are a lot more smart and sophisticated than that. They all have models to make sure that the 750 or however many people they poll represent a reasonable cross-section of voters to sample from. That's why when you look at 50 polls from one state, they all more or less say the same thing. Different pollsters will use slightly different methodologies, which might explain why one poll might consistently show a particular candidate up by an average of 8 points and another poll might consistently show a particular candidate up by an average of 3 points. But aside from these kinds of discrepancies, if say, Wisconsin were to flip for McCain, the answer would not be because the polls were wrong.

Well, yes and no, but only because of the methodologies involved. It might very well be that the polls rightly measured the wrong information, but only because the voting demographics were different that projected.

Of course, we're talking about percentage points here as well. When a poll is "wrong" it's almost always "wrong" within or near their own stated margin of error. Which means it wasn't really wrong at all.

[BenNX74205]

Has anyone ever thought that these polls might not be right?

I mean, I understand that if all of them are saying the same thing, then it's probably somewhat truthful, but aren't these polls are rather small sample sizes?

Also, I've heard multiple times now that the majority of early voters have been for Obama, does that mean that Obama has all these new voters for him, or just energized democrats wanting to get out and vote early? There's no way to tell that

I think the polls are way off. Because the polls are almost always way off. People lie to pollsters all the time, for various reasons. I think the "coin flip" analogy is overly simplistic here. It's true there are only two outcomes in both scenarios, but when testing a coin, you're flipping THE SAME COIN over and over and over. Each poll is asking different people. Yeah, 1000 is probably statistically significant, but it presumes that the answers will be reflected on election day, when they almost never are.

I think PA is going to flip. Obama doesn't resonate with PA Democrats the way Kerry and Clinton do. McCain does resonate with the large veteran population. Palin energizes part of the base as well. Considering the narrow margins PA has gone Dem the last two elections, I think this is enough to flip it red tomorrow. That means based on the 2004 electoral map, Obama needs to gain 35 electoral votes instead of just 14--that's a pretty big swing.

And to everyone saying McCain is wasting money in the Northeast, I understand you are saying McCain has a poorly run campaign and is wasting money by campaigning in these true-blue states where polls have him down double digits and he has no chance of winning. What is your reasoning as to why Obama is also increasing spending in these states? Is his campaign run poorly? Does he just have too much money? Is there any other possible reason why he would ramp up campaigning in Maine besides the notion that it will be a close race in that state?

If it's stupid for McCain to spend there when it's hopelessly lost, isn't it even stupider for Obama to spend there when it's already won?

[Popeye_Card]

Statistically most polls have a margin of error. As to if they really represent reality, we will find out tomorrow.

While N<1,000 or so seems like a small sample, statistically it's not. If you have a coin that can only have two outcomes, you don't need to flip it that often to determine that it is fair.

Meh, I'm not sure the coin argument holds much weight, because while I understand there's two choices, most people only see one choice, and if you happen to poll 750 registered democrats then your poll is going to be skewed that way..

That made sense to me..maybe not to anyone else

Well, as has been stated on here, this country is made up of ~30% hardcore Democrats, ~30% hardcore Republicans, and the remaining 40% is usually up for grabs. Virtually every measure has Obama winning over the independants at a greater rate than McCain. Additionally, Obama stands to gain more from traditional non-voters who might go to the polls this time than McCain.

IMO, there's just a few things Obama should fear. 1.) The elderly in Florida supporting the old white guy rather than the younger black guy (not saying it's racist--just who they think is "presidential"). 2.) The newly recruited base not showing up at the polls when it comes to crunch time. 3.) The Bradley Effect.

I know two of those deal with race, and you want to dismiss that. I don't think there is going to be hardcore racism show up in this election. But I do fear that the more subtle racism that's still quite rampant in this country could determine the race. I do think this country is ready for a black president (or a female president). But there's still that mental block that we view the old white guy as "presidential", while we look for every reason not to trust the one that seems different. Let's be honest, has there ever been another major candidate who has had his experience and associations questioned more than Obama has? People probe around for reasons not to trust him. Just look at the list of emails flying around for Obama vs. McCain.

[thrill]

OMG PAGE 666