I honestly don’t understand what the kerfuffle is over what Nate Silver has been doing at his 568 blog at the NYT. Here’s one side of the story:
On MSNBC’s “Morning Joe” today, Joe Scarborough took a more direct shot, effectively calling Silver an ideologue and “a joke.”
“Nate Silver says this is a 73.6 percent chance that the president is going to win? Nobody in that campaign thinks they have a 73 percent chance — they think they have a 50.1 percent chance of winning. And you talk to the Romney people, it’s the same thing,” Scarborough said. “Both sides understand that it is close, and it could go either way. And anybody that thinks that this race is anything but a tossup right now is such an ideologue, they should be kept away from typewriters, computers, laptops and microphones for the next 10 days, because they’re jokes.”
and here’s the other, from Paul Krugman:
For those new to this, Nate is a sports statistician turned political statistician, who has been maintaining a model that takes lots and lots of polling data — most of it at the state level, which is where the presidency gets decided — and converts it into election odds. Like others doing similar exercises — Drew Linzer, Sam Wang, and Pollster — Nate’s model continued to show an Obama edge even after Denver, and has shown that edge widening over the past couple of weeks.
This could be wrong, obviously. And we’ll find out on Election Day. But the methodology has been very clear, and all the election modelers have been faithful to their models, letting the numbers fall where they may.
Yet the right — and we’re not talking about the fringe here, we’re talking about mainstream commentators and publications — has been screaming “biasâ€! They know, just know, that Nate must be cooking the books. How do they know this? Well, his results look good for Obama, so it must be a cheat. Never mind the fact that Nate tells us all exactly how he does it, and that he hasn’t changed the formula at all.
The definition of “objective” is that something is not based on feelings or opinion. Are Nate Silver’s results objective? I honestly don’t know. To the best of my knowledge he hasn’t shared the details of his model with us, just the results. That he expresses those results in numbers does not make them objective. That’s what’s been referred to as the “illusion of accuracy”.
Perhaps someone can explain it all to me. I really don’t understand.
There’s one thing I think we can have very high confidence in: one candidate or the other will be elected. There’s a 100% expectation of whoever wins winning and the other candidate losing. The result will be the result.
I’ve done enough modeling in my life to know that frequently the models don’t do as well as astute observers in making predictions.
…
As for this election, I stand by my statements of the past: Romney’s to lose. I think it is too late for him to lose it at this point.
One doesn’t have to look too hard to find articles like this: The Florida Voters Who Can’t Forgive Obama. This is not the sign of a winning re-election effort.
Looking at the behavior of the campaigns also paints a story: Romney and his campaign are upbeat, Obama and his campaign are downbeat, surly, grasping at straws and name-calling.
And as I keep re-iterating, Presidents that win second terms do better the second time around. There have been two exceptions in 220 years, and both involved rather extraordinary circumstances not evident this time around.
Nate probably has more invested in being right than he has in getting Obama elected.
As I think I’ve written before, I developed my first mathematical model more than 40 years ago. I think it’s rare that you get a lot out of them that you didn’t put into them.
No model is objective. What variables one includes is a subjective decision. What data to use is a subjective decision. How to treat possible outliers, is a subjective decision.
The idea that models, both statistical and theoretical, are objective is horse crap.
As for predictive modelling and evaluating those models, use a hold out sample and see how well the model predicts in that case. Look at the MAE and/or the MAPE. And the more data you can get the better since more data points means you can have more sophisticated models (read more explanatory variables). If you can’t do that, then the model really isn’t that useful, especially for events that happen only a few times over a number of years. That is, can Nate use his model to predict the poll results?
Suppose he has data for 2010, 2011, and so far into 2012 (i.e. up to September 2012). Now using that data, can he fit his model to the data for 2010 and 2011 and then predict January 2012? What is the MAE and the MAPE? 2hen append January 2012 data to the historical data and predict February 2012 and again look at the MAE and MAPE. How good are his models at predicting the various polls?
As somebody who does forecasting as his day job, that is what I do. I don’t give a crap about statistics on goodness of fit, t-tests, and all the rest. Does the model predict well for the hold out sample? If no, useless even if the various statistics for that model are excellent.
I find the individual state entries that Nate Silver reports to be useful (such as, Florida is projected to go 50.3 to 49.1 to Romney, with a margin of error of plus/minus 3.2%) I think some weighing of the polls as a part of aggregation makes sense since all of the polls are not created equally. But all this tells me is that Romney is probably leading in Florida, but its close enough that the result may differ based upon any number of inherent limitations in polling and random events.
I don’t believe the bottom line (Obama has a 72.9% chance of winning today) is very useful. Partly because I don’t understand his approach to converting state race predictions that fall within his own margins of error, to an overall probability of the outcome. But also because its hard to determine whether any failings in the polls will only relate to a given state or will reflect systemic problems that exist in most states.
Reminds me of Zogby. He got the 2000 election right and for awhile I kept hearing alot about his polls and when queried the inevitable response was, “Well he got 2000 right!!” Do we hear about Zogby much these days? Didn’t think so.
There now it is in the right thread….
One of the primary methods of testing the accuracy of a predictive model is via hindcasting. Has Nate Silver or anyone else in the polling business ever discussed this?
Ray Fair certainly does but his econometric model is gauged to predict the popular vote rather than the electoral vote. His latest update is that, essentially, the election is too close to call.
Until Silver actually shows us the inner workings of his “model” I think people can say whatever they like about him. He certainly cannot offer an “objectivity” defense. In social science we do not presume objectivity. In fact, the opposite is assumed until someone demonstrates the fairness of their methodology.
@Icepick
You do not read the OTB comments. So, I will bring them to you.
You and @Steve Verdon are right wingers and probably racists. You are anti-science, and you definitely do not accept evolution. You believe that math = magic.
You would also get at least 10 “mean girl” thumbs down votes.
Here are two facts that may be somewhat illuminating
First, our friend Mr. Silver started out as a blogger for the far left whack job site Daily Kos. That should pretty much tell you where he comes from ideologically.
Second, he gets paid by the New York Times, which you might just say has a teensy weensy bit of bias in the same direction, Just maybe.
If Nate Silver likes his job, he’s going to come up with what his employers want. Oh, he can use statistical mumbo jumbo and computer projections only he has the exact details of to make it seem objective, but that’s pretty much the bottom line. He needs to sing the songs the people that pay him want him to sing.
Glenn Kessler, a ‘fact checker’ at the WAPO found out the same thing when he critically dissed a fallacious but major story his own paper had written on Mitt Romney’s career with Bain Capital which really should have been reported to the FEC as a contribution in kind by the WAPO.
That’s how it goes.
Regards,
Rob
You and @Steve Verdon are right wingers and probably racists. You are anti-science, and you definitely do not accept evolution. You believe that math = magic.
This is one of the reasons I’m considering saying “Fuck it” and voting for Romney. Partly it’s personal pique, and partly that such behavior (not only its supporters but also those closely tied to the Administration regularly demonize everyone that disagrees with them) should be punished. Steve V., my vote would be social signaling, even if irrelevant.
Also, the longer this Administration goes the more incompetent it looks in every way. I’m sure that they will manage to fuck up the recovery efforts for the hurricane. The President saying with a straight face the other day that we don’t leave anyone behind, after the Administration just left four behind to die in Benghazi, just speaks to either utter disingenuousness or utter incompetence. Or both, most likely.
Shorter version: Obama’s supporters plus the mess in Benghazi have just about convinced me to vote for Romney.
@Icepick
The OTB crowd and others like them are a subset of all of President Obama’s supports. They are very vocal and vicious, but they are not representative of most Obama supporters. The Right has it’s share of these also – Sean Hannity, Ann Coulter, Laura Ingraham, etc. I disagree with @steve on many things Obama, but he has not delved into crazyland.
Unfortunately, TastyBits, a lot of people, more than a majority, seem to have bought a ticket to Crazyland. They may not be there yet but they’ve got on the bus. Check the post at the top of the page.
TB, not saying most have. But a lot of prominent ones have, and Obama seems to encourage it.
I have to admit, voting for Romney just to see the reaction of others who have bought a ticket to Crazyland is a bit tempting.
The President saying with a straight face the other day that we don’t leave anyone behind, after the Administration just left four behind to die in Benghazi, just speaks to either utter disingenuousness or utter incompetence. Or both, most likely.
Ahhh, but he brought home the bodies, so they weren’t really left behind. Left to die, sure, but left behind, no.
TB,
Sigh….
Yes, I’m a right winger….
I’d be completely in favor of decriminalizing marijuana, even most other drugs. I wouldn’t encourage anyone to do some of the really dangerous drugs, but some people are going to do them no matter what, and as long as it doesn’t pose a danger to me not really my business. I’d also favor decriminalizing prostitution. I don’t mind if a woman wants to get an abortion. It is unfortunate and there are other options, but I don’t feel I can tell another person what to do with their body. Don’t even care about gay marriage….if you found somebody you want to get married to (you silly fool :P) and they are the same sex….I just don’t care* (so legalizing gay marriage is fine with me).
As for science and evolution, I argued tirelessly for sometime over at OTB for evolution and I’m sure my posts are still there. Needless to say the few right wing commenters there were peeved.
So, given that many over there do think I’m a right winger it shows the complete intellectual vacuousness of said commenters, or as I like to refer to them: the drool cup brigade.
Heh…a true mark of distinction.
*Unless it is two hot chicks who are also into exhibitionism, then I care very deeply.
@Dave Schuler & @Icepick
Within the partisans, there is always a large number of crazy people, but they are always there. Sometimes, they can affect the larger general population, but usually, they are contained. The internet, cable, and talk radio have allowed them to be more vocal, but they are not much the pamphleteers of the past. VP Cheney peppering a fellow hunter with birdshot is the closest we have gotten to the Burr-Hamilton duel.
I still remember the Reagan, Clinton, and Bush the younger Presidencies. These all had the crazies. I do not recall much craziness for Ford and Bush the elder, and I do not recall much about Carter. Nixon was a scumbag, and he should have been tarred, feathered, and run out of town.
@Steve Verdon
I was trying to give @Icepick a little taste of the OTB nonsense. Any deviation from the party position is treated as an attack. Both sides do it, but over here, there is cross pollination with the OTB site. It goes beyond simple name calling, and it makes intelligent discourse difficult.
As you have done here, you need to attempt to establish that their stereotype is wrong. The fact that you and @Icepick may know a little more about mathematics than they do makes no difference. (NOTE: I do not think you all are right wingers.)
All sites have trolls, and if that is a problem, do not feed the trolls.
Ahhh, but he brought home the bodies, so they weren’t really left behind. Left to die, sure, but left behind, no.
Okay, fair point, I take it back.
VP Cheney peppering a fellow hunter with birdshot is the closest we have gotten to the Burr-Hamilton duel.
God how I miss Dick Cheney.
TB, I’m a total right winger. I think the John Birch Society* are a bunch of faggotty communists. To Hell with Reagan, I want Attila the Hun for President. Strong on national “defense” and a zero tax policy – we’ll just steal the money we need from our enemies instead. I could SELL those policies in an election and have fun doing it!
* Are those guys still around?
And economics….
I have my political leanings (biases if you prefer), but I can look at policies and economic situations using different models (although I tend to prefer a micro approach). For example, I understand the logic of a cap-and-trade program. I could go into nauseating detail that might put some into a coma. My objection to them isn’t the economics–i.e. the public finance aspect of the program but another area of economics: public choice. Gaming the set up of a cap-and-trade program is a very serious problem and was apparently an issue when Europe tried to implement such a policy (the initial disbursement of credits was overly generous as to provide a nice windfall profit to participants. I’d argue we probably have an even bigger problem here than Europe does. It is one of the reasons I support ideas like Greg Mankiw’s Pigou Club….which again puts me at odds with many conservatives.