All of Nate Silver's State-Level Polling Predictions Proved True 576
kkleiner writes "For the last few months, the political pundit class has been at war with NYT/FiveThirtyEight blogger Nate Silver. Joe Scarborough of MSNBC called him a "joke," while an op-ed in the LA Times accused him of running a "numbers racket." But last night, Silver triumphed: every one of his state-level presidential predictions proved true. "
Re:Not how statistics works (Score:5, Informative)
Re:Good for him (Score:5, Informative)
While the possibility of that is certainly true, it is false for this recent presidential election. Not all districts have reported in, but the most recent numbers show that Obama is ahead in the popular vote by a hair under 2.9 million votes.
http://www.google.com/elections/ed/us/results [google.com]
Re:Seriously... (Score:5, Informative)
Re:All? (Score:4, Informative)
Re:Why Nate? (Score:2, Informative)
Re:But when? (Score:2, Informative)
Except that most of the pundits claimed that the race would either go in Romney's favour (Fox News), or would be a squeaker (pretty much every other pundit). In reality, Obama won pretty handily, and only those people who actually looked at the numbers (instead of their "guts") got it right. The National polls showed, in aggregate, by the end of the campaign, a 1-point advantage for Obama (which was pretty much right on the money). The State polls, except for Florida, were also very close to the end result. But the pundits decided that polls didn't matter, and what mattered most was their "narrative".
Well, facts matter. Numbers matter. And the post-modern view of most of the political class that only "narratives" matter is hokum.
Re:Math (Score:5, Informative)
All things considered he was fairly accurate on the 2008 election as well. I think he got one state wrong if that as well.
So as far as I know he is 2 for 2.
Re:Good for him (Score:3, Informative)
You're just rationalizing. The reason only landowners could vote was an artifact of the historical evolution of political power in England. When the Normans invaded England, they handed out perpetual land tenures to their supporters. This directly led to the Magna Carta, which locked in the landed gentry's rights when the dynasty tried too hard to centralize power. This resulted in a devolution of political power, somewhat unique in Europe.
This notion that only landed white men could hold political power wasn't something that was thought through. It just _was_, like the idea that only opposite-sex couples could marry. Only when society changed and challenged the convention did people _invent_ rationales for keeping or changing the status quo.
Also, women were able to hold and own property in their own right long before the American Revolution. Likewise for blacks. As far as I can tell, the idea that women were only fit ruling the household, while men ruled in the community can at least be traced back to Athens. The Athenians used to make fun of the Spartans, who let their women cavort outside the household, and participate in all sorts of "manly" things.
Re:Good for him (Score:5, Informative)
David Brooks gave an interesting response, he said that the electoral system forces the candidates to make an effort to play somewhat to the middle.
David Brooks is completely wrong about this. For instance, I'm in Ohio, the swing state that everyone was focused on for months. Here are some of the appeals I got in my mailbox and on billboards:
- Obama is actually the son of a convicted drug dealer and a porn actress, not a Kenyan. And he's been lying about his name the whole time.
- Voter fraud is a felony (but only in neighborhoods that are mostly poor and black).
- Any kind of increase in taxes will cause the economy to collapse.
- Social Security should be abolished.
Tell me exactly how that forces candidates to appeal to the center.
Re:Math (Score:4, Informative)
The fantastic, insightful website http://natesilverwrong/ [natesilverwrong] website was very helpful in showing me how Silver would be entirely proven incorrect, with quotes from lots of people about why he would be shown as an idiot, and his methods were skewed.
It seems to be down now. Not really sure why. :-)
Because you mistyped the url. :)
Actually, the site really is down [natesilverwrong.com]. Guess someone was embarrassed. The cache [googleusercontent.com] still exists, though the site seemed to be completely devoid of content anyhow.
Re:Math (Score:5, Informative)
I've always found the best way to find great news sources is to hold them accountable and stop using them when they screw up the big stories. For example, when the media was shocked by the 2008 crash, I wasn't. I had predicted it 5 years earlier (not necessarily when, but the fact that it would happen). How? I took a look at the small handful of pundits and bloggers that accurately predicted the demise of the tech bubble and looked at what they said would be the next bubble. If people actually started paying attention to the sources that get it right, vs the ones with the largest reach, places like FOX wouldn't exist. What I have found over the past decade is that far left independent news sources get it right far more often than mainstream (or far right) new sources.
The election is another great example. Some people weren't surprised, and those are the ones that we should look to next time, unless we enjoy being a bunch of dumbfounded idiots all the time.
Re:Math (Score:0, Informative)
But it actually WAS a somewhat close race:
National Vote: decided by 3,000,000.
Virginia: 100,000
Ohio: 100,000
Florida: STILL too close to call
Re:Not how statistics works (Score:3, Informative)
What I gathered from actually reading his blog though, these were treated as related events.
Obama losing Ohio was to mean he lost VA, and FL, and the Presidency and maybe CO.
He posted on (Sunday I believe) that the only way Obama was going to lose the election was if all polls had made underlying assumptions that skewed them to Obama, he added that historically this wasn't the case.
Essentially his 10% loss for Obama was a prediction that the polls were skewed by enough for him to lose (not margin of error, but actual skew in underlying assumptions that every pollster made).
The fact that this was not the case leads all of his predictions to be accurate (even the ones approaching 50/50).
The numbers come from runs of the model he uses with various tweaks, but they assume things such as a failure to predict one midwest state flows into others, or even into the nation. His models DO NOT assume independence of the results when coming up with numbers.
The fact that his essentially 50/50 (FL 50.3% Obama) turned out to be right is promising, though we don't have a tight call on the other-side to measure if there was skew (next closest state is NC with 75% chance going Romney).
his popular vote was quite close too.
This is why he calls the "tipping point states", his model pretty much assumed if Obama lost OH, he would also lost other safer states such as NV and CO, he would therefor of lost the election if losing OH (though technically he could of won VA (which he did) and NV to fill back up, and that was somewhat accounted for in the model (see the chances of NV and VA as tipping point states), as different geographic regions were given some independence in the models.
Re:Math (Score:5, Informative)
but for instance, we had no way of knowing if a "Bradley Effect" would have been in play
Sure we do. Nate Silver has looked at this effect a number of times. If it exists at all, it's tiny.
http://www.fivethirtyeight.com/search/label/bradley%20effect [fivethirtyeight.com]
I'm not saying that the probability of systematic error is large, just unknowable.
It is knowable, and that's exactly why Nate Silvers forecasts are so much more accurate than anyone else's. He does the donkey work to minimise these systematic errors.
It was a perfectly reasonable and scientific position for a Republican to say "Romney's chances are equal to the probability of error in the polls, and I hope that probability is large."
No, it was really, really dumb. Not only can systematic errors be minimised, but margin of error is not going to go in the same direction on all polls.
Re:Math (Score:5, Informative)
People need to learn that statistics and polling are sciences. Like all sciences they are inexact, with a margin for error; but the chances of the poll averages being wrong in this case were incredibly small.
No, actually the poll averages were not correct. Read up a bit more on what he actually did.
The point is actually that the poll averages are reasonably likely to be wrong, because some polls are designed much better than others. Most news outlets just average the polls. Nate Silver weights them in an attempt to give more weight to accurate ones. So, the simple averages of polls are right in most cases, but in a handful of states are sufficiently skewed by biased polls to give an incorrect prediction. Nate Silvers' weighting of polls, on the other hand, got all 50 states correct--and in many so-called "contested" states actually nailed the Romney v Obama share perfectly to 0.1%!
Re:Math (Score:5, Informative)
No, not the walk, the fact that she actually went to get some facts. In reality, I bet she was just really pissed off at Rove for being a hyper-partisan dick. You're correct that it was difficult to find pundit-free air time on Fox, but it was there and not badly done.
Of course, any (marginal) good will Fox earned was immediately blown when the usual ass-hats (Hannity, O'Reilly, Palin, etc) got on the air and asserted that Romney lost because the people who voted for Obama are ignorant parasites who "want stuff and things" from the government - as opposed to all the old, rich, white, straight men that voted for Romney (you know, the "job creators") that just want money and power ...
Re:Math (Score:5, Informative)
It was completely staged. Did you not hear her say 'the sound cut out here when we practiced this'.
Fully agree on your second point. But then that too was staged. And predictable.
Which is why I only turned to Fox for the lols. Funny to watch them stroke out when things started turning blue.
Re:Good for him (Score:4, Informative)
Side A wins 51% of the (states/districts/electoral votes) by 1 percent.
Side B wins 49% of the same category by a landslide.
Now do you see the problem? This kind of bullshit is why Gerrymandering exists; you district your populations so that the districts you can't possibly win (less than, say, 45% support in the last election) lose all of your supporters, and move those supporters over to other (adjacent) districts that your party actually has a chance at (say, 49% support before). Now, assume everybody votes the same way they did before. The first district still ends up with the same winner-take-all result, but by 80% instead of 55%. The second through nth districts change from one side winning by a little to the other side winning by a little... resuling in a landslide victory (by winner-take-all district) for your party, even though the popular vote is *still* against you.
The only differences at the presidential levels are A) you can't Gerrymander states (not practically), and B) low-population states receive disproportionate votes.
A) is dealt with by focusing on appealing to one swing state (very close to an even split) at a time, until you're polling at just over 50% there, and then moving to the next. You don't have to win them all, and you don't have to win them by much at all.
B) is the only reason the candidates bother to compaign in the "flyover" states at all, not that which way Colorado or Montana or Idaho were going to break actually impacted anything in this particular election. It sure did in 2000, though. I understand small states have a disproportionate vote, but I still think it's stupid.
Re:Math (Score:5, Informative)
This always sounds like a great argument, except that the evidence we have doesn't really bear it out.
Even in 'two-party system' countries, the two parties seem to change far more than that theory would allow for. Where are the Whigs and the Federalists now? In the U.K., the Whigs transformed into the rather different Liberals, who were decimated by the rise of Labour yet persist as the smaller third party, the Liberal Democrats, after a merger with the Social Democrats.
Often the 'two-party system' theorists excuse these changes by constructing narratives where there are periods of stability followed by some kind of 'exceptional event' which causes a 'realignment', but to me, this is really just retrospectively imposing a narrative on much messier events, to fit your convenient belief.
Most strikingly, consider this country, Canada. We have an identical electoral system to the U.K., which is often argued to be a two-party system (notwithstanding the changes I noted above), just as much as the U.S. system - the same arguments are made, that people believe only two parties can possibly attain power, so they only vote for those two parties, and the electoral system reinforces this.
Yet here, at the last federal election, the Liberals - who had been one of the main parties for over a century, and were considered to be part of a two-party system along with the Conservatives, the only two parties realistically capable of attaining power - were virtually destroyed at the polls. They now have just 34 seats in a 308 seat chamber, and got 18.9% of the popular vote; they'd never previously in their 140+ year history polled under 20%. The strangest thing is there isn't even the possibility to construct any kind of narrative of an 'exceptional event' leading to this - people just flat out got sick of the Liberals and felt they ran a terrible campaign with a terrible leader, and so they just up and voted for other parties. Even though according to the two party theorists, they shouldn't have done, because they shouldn't have believed those parties could possibly win. But they did, and now the NDP - a party which had never previously gained more than 43 seats, or just over 20% of the vote - is the second party, with 103 seats and over 30% of the popular vote. That's only the most recent example; Canada has had a much more turbulent party history than the U.S. or the U.K., despite sharing the same system as the U.K., with all the arguments about it encouraging two-party stability.
I don't have all the answers as to how things _do_ work, but I think the two-party theory is pretty weak and not really borne out by a close reading of the history of mature democracies, even ones that are commonly considered to be two-party systems.