Good news for Obama.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
PoliGAF 2012 |OT4|: Your job is not to worry about 47% of these posts.
- Thread starter Amir0x
- Start date
- Status
Again, you're describing a position that no one is taking. Saying "Gallup is an outlier" isn't the same as saying "ignore Gallup entirely."
Many simply do not seem to be understand what I'm saying. Just look at the wording you're using here.
You have decided that the other polls "accurately represent things" and that Gallup is therefore wrong and innacurate. I'm asking you to consider the possibility that the other polls are more wrong and Gallup is closer to a true picture of the current race.
What do I think the odds are that Gallup is more correct? It's less than 50%, given that Gallup is one poll among many. Looking at Silver's analysis, I'd say it's about a 29.6% chance right now.
PhoenixDark
Banned
I seem to remember you saying similar things before the first debate lol
Obama should do well given Romney's Libya fuck up, but he should still be preparing to knock this out of the park. It's time to close the deal. Btw if Romney seriously plays the "optimal" card on Monday, all bets should be off: Obama should finally call out Romney and the GOP for their borderline unpatriotic behavior
But historically, they're not good predictors of margin of victory.
Yeah, especially considering their early voting numbers are identical to the NBC/Marist poll.
I really dont know how someone can recover from that.
When people say "Obama is going to win," you realize that they're saying "Obama is probably going to win," right?
When people say "Gallup's poll is inaccurate," they're saying "Gallup is most probably inaccurate." No one is going to couch every post they make in probabilities; we can parse the difference.
AlteredBeast
Fork 'em, Sparky!
The fact of the matter, national polls only matter in an election that is won or lost by virtue of the popular vote. Gallup doesn't matter, etc.
Swing states, and specifically Florida, Ohio, NV, and Iowa matter.
Swing states, and specifically Florida, Ohio, NV, and Iowa matter.
I...literally don't understand how you of all people would take up the argument that, given five pieces of evidence that suggest one thing and one that suggests something wildly different, we should focus on the piece of evidence that suggests the wildly different thing.
Necessary, if we assume the average of the polls is an accurate picture of the situation, then each poll will be assumed to be "wrong" (because inaccurate) by an amount that represents their deviation from the mean. Gallup's deviation from the mean is by far the largest, so necessarily it's the most wrong. This does not mean we don't average it in! It just means that averaging it in makes it CLEAR that it is wildly at variance with what we have accepted to be a correct picture of reality.
edit: Is the problem that I'm not explicitly stating my constant assumption that the average of the polls should be accepted to be the correct analysis of the situation?
Is that poll just discounting everyone who's already voted because yeah, I don't see how Obama who has doubled Romney up on actual votes could be behind, especially when another poll just the other day had Obama up several points and corroborated those same early polling advantages.
DragonSworne
Banned
I wouldn't worry about NH not going to Obama.
Out of all the swing states, they know Romney best. NH and Mass are very codependent work and politics wise.
Out of all the swing states, they know Romney best. NH and Mass are very codependent work and politics wise.
RustyNails
Member
Opiate, even cnn labelled gallup poll as an aberration
Black Republican
Member
instead of the Dewey beats Truman newspaper, Obama will just use a Gallup spreadsheet on election day
For the record, the Iowa poll stresses me out a lot more, but on average Obama is still doing okay there (nor is it vital to his success if he continues to show up well in CO and NV). We'll see whether PPP shows a 20-point lead for Romney or just a much smaller figure for early voting.
People can keep voting for Romney until it pushes him over the top? It's certainly possible. If we're looking a ~3 point race nationally, anything can happen.
the interesting thing is the 66/32 split matches up with marist's 67/32 split. it's the 9-point swing in their overall results that's waaaaaaay off.
yeah i don't think we can trust that. was grove one of the lesser pollsters?
We should focus on all of them simultaneously. That is what I'm arguing.
Instead -- using your example -- people seem to be focusing on the five pieces of evidence they like, and dismissing or marginalizing the one they don't. I absolutely agree that the preponderance of current evidence shows Obama ahead. What I am not doing is marginalizing or ridiculing the evidence that he is not ahead. If I had 5 pieces of (valid) evidence to support a position, but then a new piece of (Valid) evidence contradicted that position, I would not say, "lolz that evidence is silly." I would take that evidence seriously and consider it when making conclusions.
That is precisely what Nate Silver is doing, for example. That 29.6% chance of Romney winning is not being pulled from thin air. If Gallup had Obama up by 7 instead of down by 7, I'm quite sure Silver's predictions would look considerably more favorable for Obama right now, perhaps eclipsing the 5:1 odds he peaked at before the first debate.
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
:lol
PhoenixDark
Banned
If Iowa is virtually tied on Election Day, Obama should win simply due to early voting. Romney could win by 2-4% points on November 6th voting and still lose the state.
I'm more concerned about NH, Virginia, and Florida.
I'm more concerned about NH, Virginia, and Florida.
I really hope NOVA delivers...
cartoon_soldier
Member
Damn, I was hoping for at least +2-3 O in IA
Obama defines Romnesia
So good
So good
Unfortunately I'm pretty convinced VA and FL are lost to Obama. NH seems to be the definition of toss-up.
metalslimer
Member
Like altered pointed out above, it does not matter who wins the popular vote so national poll are worthless in general. I honestly don't even understand why they even bother with them. Speaking of which, I'm sure someone has done a mathematical calculation of the extremes someone could lose the PV by and still win the presidency.
Its cool early voting is looking very good.
We have 3 Iowa polls now:
NBC Obama up 8
YouGov Obama up 4
PPP Romney up 1
Romney's 29.6% chance of winning is based on the Electoral College, which has very little to do with Gallup's national tracking numbers (hence why people are saying national polls like Gallup "don't matter").
Frank the Great
Banned
that ppp iowa poll sucks 
PhoenixDark
Banned
I think Florida is certainly lost but Virginia might stay blue due to Hispanics/youth vote.
When he was doing debate prep people complained he wasn't campaigning enough.
PoliGAF 2012 |OT5| OBAMA CANNOT PLEASE US
metalslimer
Member
I'm sure after that first one he won't ever take a debate lightly again. He has to stay campaigning though. Holing himself up in the room would be the worst thing for him to do this close.
Edit: That romnesia video is amazing lol
Yes, I think your last sentence is the crux of it. And there I think you may be misinterpreting when people dismiss a poll like Gallup as an outlier.
I want it included in the polling aggregates because I know that cherry picking polls in the aggregates creates a distorted view. The whole point of an aggregate is to capture everything.
But me, Ghaleon, I think Gallup's poll is quite obviously bogus. And I dismiss it entirely, based on how disconnected it is from the aggregates, and given the methodology, etc. So I toss Gallup out of my consideration of where the race is at. Thus (though I don't word it this way) "lolz Gallup," "fuck Gallup," etc. I think they're junk right now. But I still want them included in the aggregates.
hydragonwarrior
Member
So are the odds of Obama losing rising now? :/
Nate gives them a pretty good weight (4/5 bars), but a three point lead is suspect since every other poll is showing NC being red. Kind of like this latest Iowa PPP poll, except the opposite.
Like Nate says, never cherry pick polls. Look at them as a whole to see what direction the state is in.
i believe in harvey dent
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
Not weird at all. It just means that Obama's GOTV was very good.
Obama is in VA as well today.
djplaeskool
Member
Going to go pay the Ad Valorem on my car and toss my ballot in the toilet (read: Voting for Obama in Georgia) today. *sigh*
Nate's model adjusts state polling averages based on national trends, especially if the national polls are newer than state polls
No, his odds of winning are slightly decreasing...
Did they actually call it an "Aberration" or did they call it "an outlier?"
It is certainly an outlier, and all I'm asking is that people not ridicule outliers. As long as their methodology is sound and they appear to be following the scientific process, there is no reason not to take them seriously and factor them in to our conclusions.
The popular vote will probably be close. Your vote matters.
metalslimer
Member
Eh..maybe Obama losing the popular vote will give the necessary push to get the EC abolished anyway. I'm not going to be a hypocrite after complaining about 2000 for so long (although Gore should have won the EC too but that's another matter)
- Status
Similar threads
- Locked