Holy shit, that was great. John McCain! :lol
-
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.
PolliGaf 2012 |OT5| Big Bird, Binders, Bayonets, Bad News and Benghazi
- Thread starter RustyNails
- Start date
- Status
Holy shit, that was great. John McCain! :lol
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
dafuq?
I'd much rather get a corndoggin' triple double than farm, to be honest. My friend lived down there for the bulk of 2012 for a job. He ate a lot of corn dogs when he lived there. Like a true outsider!
speculawyer
Member
What the fuck is wrong with them? When your opponent is out there with Bruce Springsteen, you don't counter with Meatloaf. Jeez. Just ignore that aspect and do something else.
speculawyer
Member
Now I understand why they are so paranoid about Shariah law coming to the USA . . . they don't want the competition.
RiccochetJ
Gold Member
Ever heard of sample size and margin of error?
I really like this Sherrod Brown ad in OH. I'd almost like to see Obama in it. But I guess there would be trouble figuring out who'd be in the driver seat, haha.
http://www.youtube.com/watch?v=ettjF5iOc4E
http://www.youtube.com/watch?v=ettjF5iOc4E
speculawyer
Member
It's a tie!
Joe Shlabotnik
Banned
If they rounded it could be juuust within the MOE. As pointed out before though, the actual numbers would still suggest a 90% certainty of a real lead.
His point is that four points is outside the MOE. It could still be technically within the MOE with rounding, but being within it doesn't mean there's a 50/50 chance of a Romney/Obama lead either. In fact with this poll it's very unlikely there's a real Romney lead.
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
50-46 with 3.5% MOE means Obama could be at 47.5 AND Romney could be at 49.5 at the same time. If there's a 3.5% MOE, then the leader could have 3.5% more OR less than his showing in the poll, and the trailing candidate can have 3.5% more OR less than his showing in the poll... so Obama could be at 53.5% and Romney could be at 43.5%.
In other words up to 7 points spread is technically in the margin of error.
Meatloaf ain't metal.
Just saw that video, wow. Romney had that 'WTF is this guy doing' look.
speculawyer
Member
damn
That's what this campaign does though. Obama has large crowds? Let's photoshop more geriatrics in. Obama sings Al Green? Quick, let Mitt sing something! Trickle down economics doesn't work? Well, let me tell you bout that trickle down government.
el retorno
Member
sucks. hopefully it wasn't serious
ClovingWestbrook
Banned
Ugh
PhoenixDark
Banned
Obama should go visit him in the hospital. He'll owe Reid a lot if he wins
Aaron Strife
Banned
Seconded. Jeez...
I won't let my sorry-ass quips be dictated by fact checkers.
The Lamonster
Member
he's not mature enough.
MasterShotgun
brazen editing lynx
Damn. Stay strong Reid. Same to everyone else involved.
Half and half
Member
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
This is true because the sampling goes both ways - in other words, Romney could theoretically be up by 3.5% more than what they're showing and Obama could be down by 3.5% from what they're showing. The only way its out of the margin of error with a 3.5% MOE is if there's a 7% spread.
Aaron Strife
Banned
Margin of error is greatly exaggerated by the media. If they're going to claim a 3.5% MOE really makes a 4 pt lead a tie, they may as well concede that the polls are useless.
So they're tied.
Right, I get it now- a margin of error per result. I always thought it referred to the margin of error of the spread itself; I.e.: could add or subtract 3.5% to/from either Obama or Romney's result.
I am disturbed because I've been following polls for many years and have taken statistics classes yet missed this crucial detail. Maybe I was distracted by a redhead that day.
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
That assumes that there's a very large error ratio on the large number of samples. There's never been a poll showing that kind of deviation. That only describes the margin of error on that particular sampling.
There's also a chance that they're understating Obama's lead, and under stating Romney's.
It could be 54-43 O+11. But neither are likely.
Amen. The GLBT citizens of this country will look back on Barack and Uncle Joe very fondly.
Goes to http://www.barackobama.com/video/id/Tb60nFeJsNc?source=forward-doodle
I have to say, though the road to equal rights for homosexuals will be excruciating in this country, this adminsitration has taken a far more ballsy stance as an advocate than I ever expected.
If most of the marriage equality amentment votes have favorable results, and if Baldwin wins in Wisconsin, and if one or two more good Supreme Court justices are nominated, I'd say these years will have been a really good era for this ongoing civil rights story. And I think it'd be near-undeniable that this administration has set a better tone, a better environment for progress.
Obama could also be up 3.5% and Romney could be down
He knows that, he was demonstrating for me how it could be within the margin of error (I was like, "3.5 is clearly less than 4...")
Don't worry. We have lots of other Ohio polls to compare to, most of which say about the same thing. If this poll were an outlier, it'd be significantly different from the average, and it really isn't. So it adds another chunk to the pile of evidence suggesting than Obama has a small lead in Ohio.
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
In all likelihood, the poll is relatively close to accurate - there's been too many samples to say that the poll's methodology is consistently flawed. If anything, I thought polls from 2008 underestimated Obama's support, although the political climate is far different now, which probably explains why they've messed around with the likely voter model so much over various polls.
It depends. This is how it was explained to me by a pollster.
Black Mamba
Member
Nooooooo.
You do not double the MoE unless the sample is 100% with no undecideds. Without that, the MoE between differences is lower.
And again, MoE simply refers to the 95% confidence interval. When you say something is 55 with a MoE of 3% it does not mean 52 is still likely. 52 or less happens one out of 40 times is what that means.
If you have a 4 point lead with a 3% MoE you are likely to win roughly 90% of the time. THIS is what happens. Yes, 10% of the time you could lose.
If you are up 7 with 3% MoE, you still could lose 1% of the time. Does anyone say "oh well, that 1% means is possible therefore TIED." no,
95% is used because it's 2 times the standard deviation (or 2 times the average distance from the mean). But it's simply a plot point in the end. It is still highly unlikely to be at the margins.
A 4 point lead with 3% MoE is a HUGE lead, statistically.
edit: in my haste I forgot the opposite is true, Obama could be up another 6 or so. That is also part of it.
It's only a one out of 40 chance Mitt and Obama can move 3% or so towards each other.
You do not double the MoE unless the sample is 100% with no undecideds. Without that, the MoE between differences is lower.
And again, MoE simply refers to the 95% confidence interval. When you say something is 55 with a MoE of 3% it does not mean 52 is still likely. 52 or less happens one out of 40 times is what that means.
If you have a 4 point lead with a 3% MoE you are likely to win roughly 90% of the time. THIS is what happens. Yes, 10% of the time you could lose.
If you are up 7 with 3% MoE, you still could lose 1% of the time. Does anyone say "oh well, that 1% means is possible therefore TIED." no,
95% is used because it's 2 times the standard deviation (or 2 times the average distance from the mean). But it's simply a plot point in the end. It is still highly unlikely to be at the margins.
A 4 point lead with 3% MoE is a HUGE lead, statistically.
edit: in my haste I forgot the opposite is true, Obama could be up another 6 or so. That is also part of it.
It's only a one out of 40 chance Mitt and Obama can move 3% or so towards each other.
I didn't mean the poll itself was disheartening, I meant learning that my own manner of interpreting them has been flawed was disheartening. :lol
I remain confident in my understanding of SD and syllogism's graph makes perfect sense to me so I'm just gonna try and shake it off and move on with my newfound knowledge.
Black Mamba: 3.5% moe on a 4-point lead, but otherwise very helpful, thanks.
I remain confident in my understanding of SD and syllogism's graph makes perfect sense to me so I'm just gonna try and shake it off and move on with my newfound knowledge.
Black Mamba: 3.5% moe on a 4-point lead, but otherwise very helpful, thanks.
heliosRAzi
Banned
He who fights monsters...
PhoenixDark
Banned
Josh Marshall (TPM editor) keeps harping on a very important point: Ohio pollsters who include cell phone users constantly show larger Obama leads. Very important point, especially in relation to Ras polls which do not contact cell users
Wonder if their poll also excludes people who drink water
PPP polls don't either
Angry Grimace
Two cannibals are eating a clown. One turns to the other and says "does something taste funny to you?"
Ras polls are all robot polls that don't call cells. In fact, I think a lot of pollsters don't call cells.
- Status
Similar threads
- Locked
- Locked
- Locked
- Locked