The basic idea of polling is very simple. Say you have a coin that comes up heads 50% of the time. If you flip it a thousand times, it should come up heads 500 times, with a standard deviation of just under 16. That means there’s a two-thirds chance that the number of times it will come up heads is between 484 and 516, and it means there’s about a ten percent chance that it comes up heads 520 times or more.
Now say you have a coin and you don’t know that percentage of time it comes up heads. Say you flip it a thousand times and it comes up heads 520 or more times. Then it probably isn’t a coin that comes up heads 50% of the time in general.
If you view taking a random sample of 1000 voters as flipping a coin 1000 times, then it means that if you’re up 52-48 in a poll, you probably really are ahead. Now, being 52-48 is only barely outside the “margin of error” (which is 1.6% for each candidate, for a total of 3.2% on the difference), so some might scream STATISTICAL DEAD HEAT. But it’s not, it’s probably not a fair coin. In fact, there’s just as a good a chance that’s is a 54-46 coin as a 50-50 coin.
Actual polling is more imprecise because of issues with getting a good sample. The so-called “margin of error” is assuming that your sample is a perfect sample of the people who end up voting. There may be a bias within your sample — maybe Democrats are more likely to take your call or something — and some pollsters adjust their sample to fit what they think the population of people who vote will look like (certain percentage of olds, of youngs, of Democrats, of Republicans, etc.) and that’s hardly a perfect science.
But if you take a whole punch bunch of different pollsters with different methodologies and look at their numbers in aggregate, you can probably estimate the outcome pretty accurately if you use the right models or simulations. Maybe not in theory in the strictest sense but in practice.
When people say THE POLLS WERE WRONG, DEWEY BEATS TRUMAN, BRADLEY EFFECT, they usually mean 2 or 3 polls, not a hundred. If a hundred polls by several different pollsters all point in one direction, it’s just not that likely the election goes in the other.
And it shouldn’t be surprising that with some knowledge of statistics and election history, a person like Sam Wang or Nate Silver can come up with a reasonably accurate model for forecasting elections.
No one at a bank says “is it worth $1000 or $1200, I don’t know, let’s give up”. They come up with models to price things. And the argument that their models caused the financial crisis so we shouldn’t trust Nate Silver’s models is a nonsensical argument. Markets have bubbles because the only measure of price is what people are willing to pay. There’s no good analog of polling in that world.
Reader MK sends along an article about how old-time baseball people hated Nate Silver too. Does this remind you of anything?
PECOTA wasn’t any sort of dark magic. PECOTA was in essence a smart analyst who knows his history, draws defensible conclusions, chips away at uncertainty, and never claims to know more than he actually does. A good pundit, in other words. And that’s what made it and its ilk such a threat to the sports world’s pundit class. Statistics can’t replace good writing, but it can expose the bad, and sabermetrics represented a direct threat to the bad writers who had gotten away with being bad for far too long. These were the writers who used the same old false narratives to reach the same old misguided conclusions. (They used stats, too, incidentally, just the wrong stats—the noisy old metrics like RBIs and batting average.) Sportswriting isn’t a monolith, and many writers like Joe Posnanski combined their experience and access with the new methodology. But a lot of those pundits made their money on that margin of uncertainty in sports, yammering about heart and grit and all that ineffable crap that was never so ineffable that a hack couldn’t write 500 words about it for the early edition. And so they remained in the dark, stubbornly entrenched, missing out on a new way to analyze the game they were paid to follow.
HEY YOU
The only poll that matters is the one on Nov.6 2012.
gbear
This is exactly the kind of info that I don’t want to read any more. I’m not paying attention to stories about polls, just want to read GOTV stories and watch Obama rallys. I’m completely sick of pundits and polls and stories about math.
I’ll be back for Kay’s next posting.
jayackroyd
Central Limit Theorem, Baby!
dmsilev
It should be noted, by the way, that the big difference in the Obama victory probabilities being quoted by Wang and Silver come from differing estimates of the margin of error of the polls. Sam Wang is looking purely at the random-sampling errors discussed in this post. Nate Silver is, on top of that, adding an additional chunk of uncertainty to account for the possibility of a systematic error in the polling (Nate discussed this yesterday). If you take Nate’s model and you remove that systematic bias term, the Obama win odds will go up to something similar to what Sam Wang is calculating.
cyntax
You and Sam Wang have your work cut out for you; well, Sam more than you because his site attracts more people who aren’t happy with the Black Metrosexual Abe Lincoln being in the lead.
Also, there’s this typo:”But if you take a whole punch of different pollsters…”
Bunch not punch, right?
Unless that’s a Freudian slip about wanting to punch a bunch of the pollsters, which I could see too.
Jewish Steel
Thanks, Metromathmatical DougJ!
Punditing looks like such a cushy sinecure. I still can’t wrap my head around not making the barest of efforts and egaging with the new style. What kind of idiot rushes headlong into their own extinction?
Violet
@Jewish Steel:
Even stranger is someone like Chuck Todd, who in 2008 actually followed and reported accurate polling data, but since then has turned into a horse-race-at-all-costs Villager, only content if Dems are either tied or losing. That’s not refusing to engage. That’s having engaged previously and then purposely turning his back on it.
Both-side-do-it punditry must pay really well.
Spaghetti Lee
Good article, Doug.
Personally, I welcome the rise of statistical analysis in politics, but I’m kind of weary of it in baseball. Politics is serious, but baseball is entertainment, or supposed to be. I mean, I guess it’s useful for GMs, but I don’t see why I should care. Randall Munroe once referred to pro sports as a “weighted random-number generator”, but if that’s all I needed to have fun, there are plenty of random number generators on the internet. I watch sports for the storylines, the civic pride, the entertainment, etc., first.
Also, how big a divide is there really between the old-school and new school? The all-time leaders in Wins Above Replacement (the consummate new-school stat) include Ruth, Cy Young, Walter Johnson, Mays, Cobb, etc. You know, the same group of guys the old school would say were the best ever. It’s not like they’re coming to radically different conclusions.
dmsilev
@Violet: Chuck Todd is a perfect example of the Peter Principle. And looking back on it, his reporting during the 2008 primary campaign mainly amounted to being able to add up delegate totals correctly (which, granted, was more than most of the people following the campaign could do).
Davis X. Machina
@Jewish Steel:
Short-timer syndrome, for the older ones…
They won’t be on the boat when it hits the iceberg, having gotten off with the mail in Cobh.
Alison
@cyntax: Maybe that’s the term for a group of pollsters? You know, a pack of dogs, a herd of cattle, a punch of pollsters. I kinda like it…
Metrosexual Manichean Monster DougJ
@Alison:
Ha ha!
currants
@Jewish Steel: LEMMINGS….
Cermet
Or as the pro-thug polling agency – Rasmussen – just change your numbers to slowly agree with reality a few days before the election and presto – you are taken seriously by the media talking heads after feeding the thugs and their stupid sheep the story they want to believe (helps to steal the election in close races since you can point to the thug being ahead shortly before the real election.)
dead existentialist
Pft. I’m rooting for the Redskins today because when they win their last game before a presidential election, the Dem wins; when they have lost, the Repub has won. Something like 17 times now. Also, too, when the National League wins teh World Series, the Dems have a 67% win record for the presidency.
So GORGIII!
Democrat Partisan Asshole
Those who figure out how to expose bullshit are usually hated by those whose bullshit is exposed.
That Nate has managed to do it in two fields is extraordinary.
Edo
@Alison: If we need a term for a group of pollsters, I’d nominate “a deviation.”
cyntax
@Alison:
I do kind of like it in that context. I refer to my wife and my four month old as a “fuss.” My wife puts up with things like that.
Violet
@dmsilev: Yeah, it wasn’t like Chuck Todd was doing anything that should have been particularly remarkable, but since no one else was reporting actual data, he seemed like a breath of fresh air. Since then he’s turned his back on actual data and instead is pushing the horse race, even if he has to skew the polls or tell untruths (Sandy stopped Romney’s momentum!) to do it.
kd bart
I judge polls the way they do figure skating and diving scores. Throw out the high and low scores and average the ones in the middle.
PreservedKillick
@Violet:
Why is this strange at all? He cannot play in the mathematical world. He needs a horserace.
Can you imagine this year’s election without the horserace stuff? It’d be basically unchanging except for the bump after the DNC and the un-bump after the VP debate.
And there are literally BILLIONS of dollars sloshing around out there.
That’s the real story here, BTW. BILLIONS OF DOLLARS…for what?
NotMax
Oh my.
TV on in background. Just ran a clip from the files of the announcement of JFK being the winner.
Host came back and said (not making this up, really): “And that was circa 1960.”
No, idiot, that was 1960. Period.
Comrade Scrutinizer
@gbear: And yet, here you are…
Alison
@cyntax: She’s a good woman :P
Rick Massimo
If the people who make $5 million a year covering national politics want to throw their hands in the air and say “I can’t understand the math, and I can’t tell who’s going to win,” that’s fine with me. They can fill their pages and airtime with stories that cover the serious issues, explain (not just recite) both sides of the disputes and point out when one side or the other is lying.
You know – like they keep saying someone oughta do.
Spaghetti Lee
I was wondering where Paul Ryan was hiding a few days ago. Well, I think I found him: http://thechive.files.wordpress.com/2012/11/wasted-24.jpg?w=500
Violet
@PreservedKillick: Yeah, if the billions of dollars were spent on something helpful, like, oh, infrastructure or schools or even paying down the national debt, we might not be in such bad shape as a country. Such a waste, all that money we spend on politics.
chopper
one thing that drives me nuts is when people point to a poll showing obama ahead of romney by 8 and say ‘but the margin of error is plus/minus 4, so it’s really a tie!’.
the MoE is not some event horizon that obscures completely the reality of the poll. likewise, the probability of reality being a tie is not flat across the MoE; that is to say, the chance of O+8 being an inaccurate depiction of what is in actuality a tie is not nearly as much as if the results were O+1 over the same MoE.
the chance that a well-done state poll, showing O+8 with MoE +/-4%, is in reality a tie is pretty slim.
lol
@Cermet:
I don’t know why people don’t remember just how badly Rasmussen missed in 2010.
chopper
@PreservedKillick:
billions spent on the same thing that keeps guys like chuck todd employed.
advertising.
r€nato
@dead existentialist: I’d say that’s a perfect example of how correlation does not necessarily imply causation.
Anoniminous
@cyntax:
What about: Population of pollsters.
dmbeaster
In typical wingnut fashion, we hear that Sandy stopped Romney’s momentum while it is allegedly Obama’s Katrina. I am willing to bet that the best piano tuners are all rethugs, since they make their money by happily working with dissonance.
arguingwithsignposts
@dmsilev:
I think it would actually be interesting to do some sort of comparison between CT v 20.08 and CT v. 20.12 to see how much of that variation exists. i know we say it here, but data would be compelling.
Mark S.
@Spaghetti Lee:
I really like just looking at OPS to figure out how good a batter is. Baseball is really the perfect sport for stat nerds. Football has too many positions that don’t have any meaningful stats to look at, and it is pretty much impossible to judge how good a player is at defense simply looking at stats.
I don’t think Sabermetrics would work well with soccer, and I don’t know enough about hockey to judge on that sport.
r€nato
that post about how all the old-school sportswriters can’t stand Nate Silver, just underscores what a racket for lazy, no-talent hacks sportswriting is.
Speculum Spatula
Test
Beauzeaux
The problem with DEWEY BEATS TRUMAN is that the polls were taken by phone. Lots of people — poor people — didn’t have phones.
I was living in Missouri in 1948 and we did have a phone. The number was three longs and two shorts. We were also part of a large party line.
Pollsters didn’t call people like us. And they couldn’t call all the people without phones.
chopper
@dmbeaster:
i’d say that that excuse means they’re admitting that god didn’t want romney to win.
Anoniminous
@Mark S.:
At the moment 2012 hockey stats are zeros all the way down.
dmsilev
@NotMax: It was 1960 +/- 2 years, so tied within the margin of error.
Democrat Partisan Asshole
Basic statistics should be required to get a high school diploma in this country. That I have to explain what a margin of error is to a person with a college degree – something I’ve had to do over ten times this week – is utterly inexcusable.
AA+ Bonds
Bahrain detains rights activist amid defiance of protest ban
Bahrain extends detention of rights activist held over march
chopper
@Democrat Partisan Asshole:
what do you expect from a country that bought in to the “1% doctrine”?
RJ
The Ombudsman of the NYT does not seem to understand that the NYT was buying Silver’s credibility. The publisher of Douthat, Brooks, and Friedman has no credibility to offer Silver.
If Silver is correct in most of the races outside of OH, PA, and FL, one could use his numbers to judge the success of the Republican plans to suppress the vote of their opponents.
Cermet
@kd bart: You don’t appear to understand the idea behind polling when thousands of people are asked?
NotMax
@Beauzeaux
Also too, that in 1948 some major polling firms stopped doing election polls sometime in September.
burnspbesq
@Mark S.:
The things you can measure in soccer generally don’t seem to correlate well with outcomes. For example, percentage of passes completed is influenced by what kind of passes you attempt, and a back-pass from a defender to a goalkeeper under no pressure counts for exactly as much as a through-ball that rips open a defense and leads to a goal.
Democrat Partisan Asshole
@chopper: A tax system that favors the rich at the expense of the poor. Oh wait. We already have that.
arguingwithsignposts
@RJ:
Actually, that’s not entirely true. The NYT has credibility with some people, even if people within the NYT have an overactive imagination WRT its actual credibility outside their bubble.
quannlace
Wondering what kind of CGI graphical monstrosities CNN reporters will be interacting with, this year?
Bill E Pilgrim
So despite hard evidence that no such thing happened, the story from now on is going to be that Romney had “the momentum” and was going to win until Hurricane Sandy happened.
It’s like living in a pre-scientific society. Or post, I guess, in our case.
shortstop
Finally have a few moments to myself this week, so I just used three of them to visit The Corner for the first time in, I dunno, years? They’re falling all over each other insisting that polling shows a strong Romney win. The lack of dignity is stunning. Even the Sopranos character-looking poli sci adjunct from Suffolk University has backed off and is now equivocating, for heaven’s sake.
Mark S.
@r€nato:
It certainly explains Buzz Bissinger.
Democrat Partisan Asshole
@kd bart: No statistician who expects to remain employed, or who gives a shit about accuracy, would do that. Why would you?
catclub
“There’s no good analog of polling in that world.”
Actually, there are some pretty good analogs to polling in that world, but since they gave numbers that the banks did not like, the banks asked for and got a change in the rules of how they priced those objects. _THEN_, they collapsed when they ran into the actual world.
Collapse of the wavefunction indeed.
NotMax
@quannlace</a.
Cute anime-style girls in short-skirted school uniforms?
Hello Kitty?
Captain America?
Hologrammatic Founding Fathers?
dslak
DougJ, we’ve found your soccer mom replacement:
Women don’t matter; white guys matter. There’s still time for this to go mainstream!
Bill E Pilgrim
@quannlace: John King, AKA The Machine That Goes “Bing”
I finally realized what his mannerisms remind me of: a butler. With his dark suit and hands folded in front of him, leaning forward, after and before he manipulates the map.
Cermet
@r€nato: And why most sport writers are/were thug voters. Supidity is a requirement.
Rainey Day
To carry the sabrematics analogy one step further:
Mitt Romney = Juan Pierre
Schlemizel
@cyntax:
I thought maybe a “punch of pollsters” was like a ‘pride of lions’ or a ‘GOP of assholes’, just a way to describe a group of something.
EDIT: carp! looks like another great original idea I had that someone had before me 8-{D
Cermet
@Democrat Partisan Asshole: It is offered but rarely required.
Spaghetti Lee
@NotMax:
Yes please!
@Bill E Pilgrim:
Let ’em mope and whine. A win’s a win, and we’re gonna get a win.
Hoodie
It’s the need for a causal narrative to keep social status. Chuck Todd did not need that when he was simply playing the resident geek that they bring out of the closet for election night. Pundits are courtiers by nature, they making their living by subsuming their lives to those that they cast as great, important, newsworthy, etc, and their own status derives from access to that power. They’ve come up in a system that rewards obedience to power, not talent (that’s why Jon Stewart is a better “newsman” than any of them). If the horserace is just numbers, they’re worthless. There’s plenty to write about, they’re just not capable of doing it, so they do what they know how to do. They’re drunks looking for their car keys under the street lamp because the light’s better there.
Bill E Pilgrim
@Spaghetti Lee: Oh I prefer that they’re deluded, the only outcome it could possibly have is avoiding complacency from Democrats and increasing turnout a little. And the Republicans’ refusing to see why it happened even afterwards just chases them further into the dark for next time. It’s still amazing though.
Svensker
@shortstop:
Heh. Me too. Alternate Reality R Us.
That being said, I’m knocking on wood, throwing salt over my shoulder and keeping fingers crossed. No hubris here.
Craig
@lol: Rasmussen was awful in 2010, but the everybody was awful in 2010. That’s what happens when you adjust your final polling to match the aggregate. Ever since 2000 when he whiffed by 10pts, Scott Rasmussen makes sure his last polls for any given race don’t stand out from the crowd.
Robin G.
@Bill E Pilgrim: Yep.
The interesting thing is that, even were the narrative true, it’s absurd to claim it delegitimizes Obama’s (hopeful) reelection. Why shouldn’t Florida, North Carolina, and Virginia take into consideration how good a president is with dealing with a hurricane? It’s kind of an issue in those states.
Cermet
@Bill E Pilgrim:
Edited for somewhat improved accuracy.
arguingwithsignposts
@shortstop:
This is how you treat yourself?
dr. bloor
@cyntax:
I dunno–that could work. Murder of crows, riot of pussies, punch of polls?
As a Bill James fan since the early 80’s, I can attest to the fact that old-timey baseball writers are essentially a bunch of holier-than-thou, get-off-my-lawn types. Of course, it’s mostly comedy gold since guys like JoeScar are trying to drive the narrative for a federal election, while baseball writers are preoccupied with, well, baseball.
catclub
@Cermet: Of course, Charles Pierce seems to have left the field of sports writing at about the same time as the onset of the sabrmetricians, but we can at argue it was for some other reason. Even the statisticians still have to admit that some coincidences are just that.
ericblair
@catclub:
True, and it extended to the models as well: management did not understand and/or did not like the fact that certain numbers were not fixed in stone for eternity, were difficult to measure (and therefore had a margin of error), and could easily change over time. So they decided that reality wasn’t going to be the boss of them and did what they wanted anyway. Sounds familiar.
jeffreyw
Voting by fax and email in New Jersey?
catclub
@dr. bloor: In general, sportswriters are still better writers than political writers, aren’t they? Stats blind or not.
Bill E Pilgrim
@dr. bloor: Murder of crows, riot of pussies, punch of polls?
I think it’s “an annoyance of pollsters”. Especially if it’s dinner time when they call.
koolearl
@Bill E Pilgrim: This reminds me: Todays Philadelphia Sunday Inquirer has an article about the race that brought up the falsehood that Ronald Reagan was trailing Jimmy Carter until he surged late in the 1980 campaign.
Bill E Pilgrim
@Cermet: Better.
trollhattan
“November surprise” may be this op-ed the president has penned for America’s Finest News Source.
eemom
@Bill E Pilgrim:
So true, and so drearily exhausting to know with a million percent certainty, zero margin of error, that that is EXACTLY what is going to happen.
Next up: NOAA receives death threats
Ben Franklin
Jeralyn has some encouraging maps.
http://www.talkleft.com/story/2012/11/4/3546/47709
RSA
@Violet:
I wish that we did more to teach primary school children and high schoolers at least the basics of counting and combinatorics, measurement, error, and statistics. That’s pie-in-the-sky thinking, I suppose, but it would really help in the “Here are the numbers” versus “My gut feeling about voter enthusiasm tells me…” debates.
Craig
@Beauzeaux: That’s incorrect. Polling was performed in-person from its inception to about the 1970s-1980s. Gallup (and Roper) missed badly because they ceased polling in the last two weeks, but mostly because they used an inferior method called quota sampling.
The sampling issue has faded from public consciousness because the average person doesn’t know what quota sampling is and is thus deprived of an opportunity to feel smug and superior to those stupid, stupid, pollsters.
MikeJ
@burnspbesq:
The real question is, is it the things that *can* be measured that don’t correlate, or the things that *are* measured.
Before Bill James anybody could have come along and measured OPS, but nobody thought to. Right now nobody measure “minutes a striker spends in the last third of the field”, but it probably has a high correlation with winning (since if you spend all your time on offence the other team has less chance to score. Or maybe they simply need less time. I suppose you’d really have to measure it to know for sure.)
Craig
@MikeJ: Actually Brach Rickey and his team of statisticians discovered OPS when Bill James was in diapers.
shortstop
@arguingwithsignposts: I enjoyed it. I’ve got a great cup of coffee, I’m playing good music and they’re making complete jackasses of themselves. The trick is not to stick around longer than three minutes lest you join them in the ranks of the pathetic.
Now I’m going out to lunch and away from the computer like a non-obsessed person.
cyntax
@Anoniminous:
Population is good. Or a “mean” of pollsters?
@dr. bloor:
Punch is a solid choice; someone suggested “deviation” which I also like.
dead existentialist
@dead existentialist: Gah! I must correct my correlation. I f’ed that all up.
When the ‘skins win, the incumbent party wins; when they lose the challenger wins. This has been going on since 1940. Circa, also, too.
Chris
@AA+ Bonds:
Ah, the completely forgotten part of the Arab Spring…
Liberty60
ruh-roh:
Jim Cramer is predicting a landslide 440 EV victory for Obama.
Given his track record, I am seriously worried about now.
Business Insider shows a 440vote win. It includes Texas going for Obama.
I think this is just Karl Rove fucking with us.
kay
I’m interested in Florida, because Silver had Romney as the favorite and then he adjusted as new info came in, and now he has it as as O 46% ( or something).
But the punditry called FL for Romney long ago.
I would just love it so much if FL voters beat Romney and pundits. Those are some long lines down there. They look pretty enthusiastic to me.
It would be so like Republicans to tell the dumbasses that Florida was a Romney lock to distract from what looks like some really egregious voter suppression. 9 hour lines? Come on.
Go FL voters! I hope you beat your governor, Mitt Romney AND political media. I love an upset.
trollhattan
@Liberty60:
I’m with him on the “I’m taking the over on Obama. Everybody else is too close” and think the gaudy numbers are just Cramer gathering the sweet attention he loves with some arithmetic overstatement–shouting with numbers.
Now, who the hell is Andrew Beyer?
CW in LA
@dead existentialist: Didn’t Green Bay beat the Foreskins the weekend before the election in ’04? Or is that further evidence the election was stolen?
trollhattan
@kay:
Wang has Obama +0.5 @24.8% in Fla. Also, too.
trollhattan
Even the liberal BBC.
Mark S.
@kay:
I really want to win Florida badly as well. It would look so much better to win 333-205. Go Betty!
MikeJ
@Liberty60: While I wish that were true, I don’t believe it. Which is what separates us from the Republicans.
eemom
@trollhattan:
Yeah, wtf is up with that? I’ve given up clicking on Guardian or BBC at all in the last month because ALL I ever see is dumb horserace shit. They’re usually at least somewhat marginally better than that.
Tino De Angelis
Now say you have a coin and you don’t know that percentage of time it comes up heads. Say you flip it a thousand times and it comes up heads 520 or more times. Then it probably isn’t a coin that comes up heads 50% of the time in general.
Wrong.
Now say you have a coin and you don’t know that percentage of time it comes up heads. Say you flip it a thousand times and it comes up heads 520 or more times. The first time this is observed you remember “there’s about a ten percent chance that it comes up heads 520 times or more”.
Then it probably isn’t a coin that comes up heads 50% of the time in general. If you are familiar with statistics you can pick a desired confidence interval and then determine how many repetitions of this trial (1000 coin flips) are necessary to be able to draw a conclusion about whether or not the coin in question displays random behavior.Otherwise, some networks want to know what you think about Romney’s “chances”.
kay
@Mark S.:
I would just love it if all the clowns were clustered in Ohio (where Mitt Romney told them to go) and Florida turns out to be the interesting story.
Voters went ahead DESPITE Romney + media calling it for Romney.
I just enjoy these pundit- humiliation scenarios :)
FlipYrWhig
The way pundits seem to be treating “margin of error” is even stupider than you’d think. I saw John King, major idiot, reporting from Ohio once over the last few days. They had just shown the top-line number with Obama up by 4 or something, which, of course, was A STATISTICAL TIE! They then proceeded to talk about the Cincinnati area. Romney needs to win it, you see. So they flashed the poll results: Romney up by 5 there. So, King concluded, Romney is doing what he needs to do. In the upper right of the screen, margin of error was displayed: +/- 7. This detail, of course, was not at all part of the story. It was not a sign of a statistical tie. It just floated away as though it weren’t even there.
The pundit set seems to have gotten no further in their thinking about these things than to treat differences of 1, 2, 3, and 4 points as “close, practically tied.”
dead existentialist
@CW in LA: You are correct. The Redskin Rule has been revised.
Definitely gonna vote on Tuesday.
Anoniminous
@kay:
OFA is saying they are concentrating on getting Obama supporters, but not likely to vote, voters banked so they can then turn to ensuing their more likely voters do vote on election day, thus increasing total Dem vote. OFA is also saying the GOP is only maintaining their vote.
No idea if all that is accurate. Makes a nice story, tho’
Anoniminous
@cyntax:
Split the difference: a mean population of pollsters
waratah
@kay: Kay I have been watching Florida closely and hoping for the upset. I felt a little down when Nate and Sam did not think so but I am like you it takes determination to stand in those long lines, and I think there will be long lines again on Tuesday.
David Koch
It’s “Moneyball” all over again, only this time it isn’t scouts drafting players based on how they look and whether they have an good looking girlfriend, it’s Beltway pundits picking candidates based on whether Sally Quinn invites them to parties and whether Bobo swoons over them.
Nate is the political version of Bill James.
SRW1
The funny aspect is that our present day Majikthises and Vroomfondels don’t even get how their real enemy is not so much Nate Silver, who still puts in a bit of magic sauce in his simulations, but Sam Wang, who doesn’t.
I suspect that at a not too distant point in time (possibly next week Wednesday) the conclusion that the predictive power of the models is pretty good will become hard to dispute. From that point on, the discussion will move to the quality of the input for the models, i.e. polls.
There will be a bunch of casualties once that happens: pundits who are statistophobes, and pollsters who massage their stuff for partisan purposes.
Joel
@Spaghetti Lee: Well, it’s not like pundits and poll aggregators would come to different conclusions on Nixon-McGovern, either. The tough calls are the ones that require nuance. Like the old school guys claiming that Shea Hillenbrand was a decent baseball player, using selective numbers (i.e. batting average) in their case.
Joel
@chopper: back-of-the envelope, but I think a lead that is 2x the MOE is the same as 2 confidence intervals (or 1-.05^2) or 99.8%.
Joel
@MikeJ: Well, Branch Rickey measured OBP and it worked out pretty well for him… The issue was that the general public was behind the curve. The general managers had no incentive to share their insights.
burnspbesq
@MikeJ:
Dunno. In the game I know best, lacrosse, I’m a pretty determined holdout against the stat guys, because I don’t think the available metrics tell you very much about what actually contributes to wins and losses.
Take two of the best attackmen of recent years, Max Quinzani and Steele Stanwick, as cases in point. They’re both great players and were incredibly important to their teams’ success, but they did very different things, and they did things that are hard to quantify using existing metrics.
Quinzani’s job in the Duke offense was to work tirelessly in front of the net to get open, with the knowledge that when he got open, Ned Crotty would deliver the ball. Max could score four goals in a game and only have the ball in his stick for ten seconds. His shooting percentage is great, but it’s misleading because it doesn’t take into account the fact that he’s always taking high-percentage shots because he’s always one-on-one with the goaltender at point-blank range. What metric captures the work he did to get open?
In contrast, everything Virginia did on offense ran through Stanwick. He had multiple touches on every possession. There were games in which if you broke down the video I’m sure you would find that the ball was in his stick for 20-25 percent of all the time Virginia spent on offense. If he beats his man behind the net, forces the defense to slide, and makes the pass that forces the second slide, and that guy makes the pass that leads to a goal because the third slide doesn’t happen, what metric captures Stanwick’s contribution?
Rick Taylor
Putting on my stats hat now, this is not true, and is a common error people make interpreting statistics. All we can say is that if the coin was a fair coin (if the null hypothesis was true), then there’s a roughly 5% chance we would have gotten a total number of heads as far or further away from 500 heads as the one we actually got. We can’t say anything more, unless we have prior information about how likely the coin was to be fair before the experiment.
Robert Waldmann
“some pollsters adjust their sample to fit what they think the population of people who vote will look like (certain percentage of olds, of youngs, of Democrats, of Republicans, etc.) and that’s hardly a perfect science.” and they work for Repulicans.
Other pollsters adjust the sample so that the sample of adult respondents corresponds to the adult population according the Bureau of the Census, that is the Census plus extrapolation. Then those weights are used to weight those respondentes who say they are registered voters, those who say they are very likely to vote or those who pass some other likely voter filter.
This means that the pollsters “think”ing is about the adult population which is fairly well measured and not about who will vote which is a guess.
The whole unskewed polls nonsense is based on the unfounded assertion that independent pollsters make guesses about turnout and impose them. All who have explained their methodology insist they weight the sample of adult respondents to match the adult population (on age race gender and maybe other things definitely not including party affiliation (except for Rasmussen)).
In contrast, journalists have un named Republican polling sources who say that the Republican commissioned polls use “a different turnout model”. I think they mean that the Republican commissioned polls use a turnout model, that is impose assumptions about who will vote. I know of no reputable pollster which does this (no links but quite a few have said they use the approach I described).
Comrade Mary
@trollhattan: Oh, man. That is about to hit Facebook HARD, courtesy of reading-impaired, irony-deficient, and melanin-deprived right wingers given blessed, blessed confirmation of their deepest fear and greatest wish.
I bought lots of deadline snacks today. Why, oh why did I not pick up some popcorn?
grandpa john
@lol: Probably because it is not consistent with the narrative they are flogging. in 2010, Nate and Sam were kings of the polling world according to the Repubs. Now not so much, ahh fame is but a fleeting thing in the wing nut world. Here today, eaten alive tomorrow
grandpa john
@lol: Probably because it is not consistent with the narrative they are flogging. in 2010, Nate and Sam were kings of the polling world according to the Repubs. Now not so much, ahh fame is but a fleeting thing in the wing nut world. Here today, eaten alive tomorrow
grandpa john
@arguingwithsignposts: Well a lot of us who followed Nate before he went to the Times were worried about what his joining the NYT would do to his credibility, as most of us realized that Nate had more Credibility than the Times did.