If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Salon)   How does Gallup explain their discrepancy in the election polling? By blaming Nate Silver, of course   (salon.com) divider line 100
    More: Fail, Gallup, FiveThirtyEight, rational decision, statistical model, DailyKos  
•       •       •

6617 clicks; posted to Politics » on 13 Nov 2012 at 9:28 PM (1 year ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



100 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest

Archived thread

First | « | 1 | 2 | » | Last | Show all
 
2012-11-13 04:27:58 PM
Yeah, good luck with that.
 
2012-11-13 04:31:13 PM
He turned them into a naught?
 
2012-11-13 04:32:16 PM
It's not Nate Silver's fault your methodology sucks
 
2012-11-13 04:35:08 PM
The gist of the argument is this- accurate aggregators will limit the number of people doing polling, thus eroding the accuracy of aggregators.

Based on where Gallup was polling, I think Nate was chucking them as an outlier anyways. Nobody will miss you, Gallup.
 
2012-11-13 04:36:08 PM
Business Week had an article on this. They basically said this was a win for Gallop as it put them in the headlines for a few weeks.
 
2012-11-13 04:49:05 PM
Well, Gallup shouldn't have sucked out loud this year.
 
2012-11-13 05:00:21 PM
That message from Newport sounded frankly unhinged.
 
2012-11-13 05:16:03 PM

ginandbacon: That message from Newport sounded frankly unhinged.


If my company, which makes its money on its credibility, had lost all credibility, I'd probably be unhinged too.
 
2012-11-13 05:20:55 PM
I posted a comment on that blog entry last Friday, but they wouldn't post it because it has to be approved by moderators first.

Obviously, it was not a complimentary comment.
 
2012-11-13 05:26:47 PM
Look at him all thin gaying up the place.
 
2012-11-13 06:38:21 PM
You know Gallup could do three or four types of polls, produce a wider range or data and aggregate their own internal numbers and release them as a top line poll. They'd get better results and not look like morons after the fact.

Although honestly, I think they were just carrying water for the Romney camp. There were plenty of criticisms of Gallup going back years that said their likely voter model was deeply flawed wrt measuring youth and minority voters but did they address those issues? No. They took a gamble that they were right and that the youth vote etc. would stay home but they were wrong, now they're biatching about Nate Silver because he happened to be the waiter who delivered the shiat sandwich they, themselves, ordered.
 
2012-11-13 07:24:57 PM

dameron: You know Gallup could do three or four types of polls, produce a wider range or data and aggregate their own internal numbers and release them as a top line poll. They'd get better results and not look like morons after the fact.

Although honestly, I think they were just carrying water for the Romney camp. There were plenty of criticisms of Gallup going back years that said their likely voter model was deeply flawed wrt measuring youth and minority voters but did they address those issues? No. They took a gamble that they were right and that the youth vote etc. would stay home but they were wrong, now they're biatching about Nate Silver because he happened to be the waiter who delivered the shiat sandwich they, themselves, ordered.


This. Exactly this. Gallup certainly didn't suffer from a lack of polling, just a diversity of polling. They should be splitting up their shiat and trying different stuff and aggregating their own shiat.
 
2012-11-13 08:14:33 PM
I heard this guy interviewed on Marketplace on NPR. He's been saying the same thing since before the election, and he's sounded very defensive. Yes, if there are no polls to aggregate, aggregators will have no job. But I heard him interviewed both before and after the election, and he still has not offered any explanation of why Gallup was so far off the mark. He seems incapable of simply saying "We made some assumptions that turned out wrong. We'll do our best to fix our model and do better next time."
 
2012-11-13 08:15:20 PM
Does anyone know if Gallup had a change of leadership, recently? A change in leadership that might have been more...receptive to a Romney win?

/not arguing conspiracy
//just asking a question
///seriously and honestly, just asking a question
 
2012-11-13 08:23:23 PM
I honestly dont think they were carrying water for Romney - they were really off in 2008 as well, but it was in Obama's favor. I think they just stink.
 
2012-11-13 08:56:51 PM
It's not easy nor cheap to conduct traditional random sample polls. It's much easier, cheaper, and mostly less risky to focus on aggregating and analyzing others' polls.

It's not easy nor cheap to run a scientific lab. It's much easier, cheaper, and mostly less risky to publish textbooks that bring together various widely accepted papers to show the consensus in the scientific community to make the information palatable to the public and present a general picture of what a plurality of labs report so labs that have suspect procedures, when the results of their procedures don't fit with other experiments, are ruled out as the most likely and accepted explanation. 

Individual polls can be inaccurate. Aggregates of polls allow the polls themselves to be analyzed like any data set so the most likely trends can be seen.
 
2012-11-13 09:09:59 PM
They would have still sucked whether Silver pointed it out or not.
 
2012-11-13 09:17:08 PM
It's not easy nor cheap to conduct traditional random sample polls. It's much easier, cheaper, and mostly less risky to focus on aggregating and analyzing others' polls. Organizations that traditionally go to the expense and effort to conduct individual polls could, in theory, decide to put their efforts into aggregation and statistical analyses of other people's polls in the next election cycle and cut out their own polling.

My, how passive aggressive.

"WELL FINE MAYBE WE WON'T DO POLLS ANYMORE!"
 
2012-11-13 09:30:33 PM
...Gallup sounds Republican.
 
2012-11-13 09:30:58 PM
This is the exact same argument we heard 10 years ago that aggregators like Fark and Drudge would kill real journalism...

...uh oh.
 
2012-11-13 09:32:19 PM
If you stopped sucking at your job, then you wouldn't be so mad at nate for doing it correctly.

Now nate is going to be on every right-winger's election watch, waiting for him to be off by 0.000005% in his analysis so they can rip him apart as being stupid or a fraud. Of course, these same right wongers don't give a fark about their sources being wrong 100% of the time.
 
2012-11-13 09:32:24 PM

Quasar: "WELL FINE MAYBE WE WON'T DO POLLS ANYMORE!"


We're taking our polls and going home!
 
2012-11-13 09:32:42 PM

impaler: They would have still sucked whether Silver pointed it out or not.


If a poll sucks in the forest and an aggregater isn't there to feel it did it really smell like balls?
 
2012-11-13 09:34:05 PM
wait wait let me post this again :

'Is that some kind of Math You Do As A Republican To Make Yourself Feel Better?'

Boy that just never gets old....
 
2012-11-13 09:35:08 PM

Triumph: This is the exact same argument we heard 10 years ago that aggregators like Fark and Drudge would kill real journalism...

...uh oh.


Correlation does not equal causation.

Fox, and the media's reaction to Fox, had probably the greatest impact on the downfall of real journalism.
 
2012-11-13 09:36:49 PM
Because polls that aren't close (nail bitters) don't drive news/web traffic.
 
2012-11-13 09:37:35 PM
Take your ball and go home, Gallup. Nothing of value will be lost.
 
2012-11-13 09:40:37 PM

eraser8: Does anyone know if Gallup had a change of leadership, recently? A change in leadership that might have been more...receptive to a Romney win?

/not arguing conspiracy
//just asking a question
///seriously and honestly, just asking a question


I was curious about this too. Just taking a quick look at the gallup site, they've had Jim Clifton as CEO since 1988. But, maybe there have been other executive changes that might explain it.

Or, they just really farked up in assessing likely voters... by weighing for the fact that liberals are less likely to have land lines, and stuff like that. A series of small mistakes could result in projections like theirs.
 
2012-11-13 09:42:50 PM
i1.kym-cdn.com

/sorry
 
2012-11-13 09:43:12 PM
i.imgur.com
 
2012-11-13 09:44:31 PM

Triumph: This is the exact same argument we heard 10 years ago that aggregators like Fark and Drudge would kill real journalism...

...uh oh.


There is a difference in that it does take money and skill to conduct a poll. I suppose 2 dudes in a room could conceivably call the proper sample size in a couple of days and generate the data needed to give a poll. Of course, they'd have to have a way to sell said data to make their 20 hour days worth while.

But in your world that's apparently the same as writing code, having users generate a majority of your content and have a ready made platform to sell ad's.

/you should go to your local community college and audit statistics and marketing (specifically marketing analysis)
//because polling is easy
///that's why everyone does it
 
2012-11-13 09:44:42 PM
I was polled by Gallup via my cell phone (only phone I have) about six months ago.

They never called again. I guess they didn't like my answers.
 
2012-11-13 09:45:30 PM
Just like fark stopped news from existing!

Also, Gallop might as well stop, they couldn't be anymore wrong as is
 
2012-11-13 09:45:35 PM
Skew'em.
 
2012-11-13 09:47:37 PM
Gallup is butthurt over getting its ass kicked by a bunch of small robocall firms and having Nate Silver call them on it.
 
2012-11-13 09:48:07 PM
If too many people aggregate the polls, there won't be enough numbers to go around!
 
2012-11-13 09:50:55 PM

Heraclitus: wait wait let me post this again :

'Is that some kind of Math You Do As A Republican To Make Yourself Feel Better?'

Boy that just never gets old....


Heh.

/and they would've gotten away with it too, if it weren't for that meddling [Nate Silver] [increasingly diverse electorate] [sabotage of Orca] [ [fill in your own] ]
 
2012-11-13 09:52:35 PM
I know Gallup isn't exactly a "pundit", but...

imgs.xkcd.com
 
2012-11-13 09:53:47 PM
i1125.photobucket.com 

/It wasn't a horse race, dammit.
 
2012-11-13 09:56:22 PM
"How does Gallup explain their discrepancy in the election polling? By blaming Nate Silver, of course"

I don't see where they blamed their discrepancy on anyone. Whined that the pollsters do all the dirty work, and guys like Silver sit back and get the credit? Sure. But they sure didn't blame Silver for their faults.
 
2012-11-13 09:56:56 PM

Nadie_AZ: Business Week had an article on this. They basically said this was a win for Gallop as it put them in the headlines for a few weeks.


Well infamy is a type of fame, I suppose.
 
2012-11-13 09:57:45 PM

skipjack: Triumph: This is the exact same argument we heard 10 years ago that aggregators like Fark and Drudge would kill real journalism...

...uh oh.

There is a difference in that it does take money and skill to conduct a poll. I suppose 2 dudes in a room could conceivably call the proper sample size in a couple of days and generate the data needed to give a poll. Of course, they'd have to have a way to sell said data to make their 20 hour days worth while.

But in your world that's apparently the same as writing code, having users generate a majority of your content and have a ready made platform to sell ad's.

/you should go to your local community college and audit statistics and marketing (specifically marketing analysis)
//because polling is easy
///that's why everyone does it


Actually, yes. virtually every large newspaper does polls. "Dudes in a room" are cheap, scripts are easy to make, and the math is not very difficult.

This is why Gallup was so shocking. Nate Silver has almost fifty firms listed who did polling in the last election. 90% of them were damn close. Gallup was a gigantic outlier. Even internet polls got this one right.

It should not have required a Nate Silver for Gallup to see that they were consistently an outlier and try to figure out what was going on. They didn't. I could use two dudes in a room and I'd get better results than that.
 
2012-11-13 10:03:24 PM
Dear Gallup,
As Maine goes, so goes Vermont.
Love,
1936
 
2012-11-13 10:03:39 PM

unyon: The gist of the argument is this- accurate aggregators will limit the number of people doing polling, thus eroding the accuracy of aggregators.

Based on where Gallup was polling, I think Nate was chucking them as an outlier anyways. Nobody will miss you, Gallup.


They certainly won't after Nate gives them the summary downgrade in updating the model for the 2014 season.
 
2012-11-13 10:07:01 PM
And they would have gotten away with it if it hadn't been for those urban voters.
 
2012-11-13 10:11:11 PM

StopLurkListen: ...Gallup sounds Republican.


PUH-lease! Gallup is owned by Selection Research which "provides to clients verbal feedback and a formal written report with information about employee job suitability and developmental areas, information about the spouse / partner's internal resources for relocating and acculturating, and family factors and conditions." Does that sound nice?

Let's have a look at some of their clients:

ARCO Coal Company
AT&T
Boeing
Dow Chemical Company
Ernst & Young Ltd.
GE Capital
General Dynamics
General Mills International
General Motors International

Koch Industries

Lockheed Martin
MasterCard International
Monsanto
Procter & Gamble
Reuters
Rio Tinto Alcan
Rockwell
Siemens
Verizon
Western Digital

/I'm done
 
2012-11-13 10:13:21 PM
Whether or not Nate Silver even exists doesn't change the fact that they failed to predict the winner.
 
2012-11-13 10:13:22 PM
We have a reverse law of the commons with polls. It's not easy nor cheap to conduct traditional random sample polls. It's much easier, cheaper, and mostly less risky to focus on aggregating and analyzing others' polls. Organizations that traditionally go to the expense and effort to conduct individual polls could, in theory, decide to put their efforts into aggregation and statistical analyses of other people's polls in the next election cycle and cut out their own polling. If many organizations make this seemingly rational decision, we could quickly be in a situation in which there are fewer and fewer polls left to aggregate and put into statistical models. Many individual rational decisions could result in a loss for the collective interest of those interested in public opinion.


Well, bye ... the world needs ditch-diggers too, sonny.

Oh, you say you want to keep your job? OK, but here's the problem: While you may in fact conduct polls, you bloody farking FAIL at using the resulting data.

Instead you conduct the poll and then vend your fat face to any broadcaster dumb enough to pay the fee for your utterly perfunctory intuitive blathering.
 
2012-11-13 10:21:27 PM

Britney Spear's Speculum: Whether or not Nate Silver even exists doesn't change the fact that they failed to predict the winner.


^^^^^^^^^^^^

Unless their model was predicated on selecting whoever Nate Silver didn't pick--like putting your money on black because some idiot swears his system says the wheel will come up red--their poll model should have been independent of his system, and whatever he did or didn't do is not relevant to their results.
 
2012-11-13 10:25:27 PM
i45.tinypic.com
 
Displayed 50 of 100 comments

First | « | 1 | 2 | » | Last | Show all

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »






Report