Nate Silver isn’t a pollster… Nor is he affiliated with PPP. Not sure why you bring him up.

And the point is that the pollsters apply their secret sauce to account for sampling bias ?

Surely if the poll is shown to be accurate in retrospect that is what matters ?

I am confused (perhaps it is still the shock of this election hitting me square in the face), are you suggesting that PPP simply shaped the race via a Field of Dreams type approach. If you build it they will come? If you wish it, the votes will follow.

But Silver’s calculations are all based on the rigged polls. He guessed right because they guessed right. There was no objective information that turnout would be what it was. They all got lucky. That is definitely not science.

It is impossible to know how much it shaped the race. Did it make a difference in the outcome? It was close in the three swing states we lost, FL, VA, and OH. Who knows?

They were EV…. bingo. No real sense trying to tell a different story. Many thought things were one way. They just weren’t.

Isn’t that the gist of the controversy this year with all the pollsters, each pollster had to estimate, or guess, what the turnout would be for each individual party. In this case PPP guessed correctly, in 2008 Rasmuessen guessed correctly. That is the secret sauce. Otherwise a pollster would randomly grab a 1000 people off the street, ask them who they are going to vote for, and from only that information make a prediction of who will win and who would lose.

> What excuse is there for him?

Other polling analysts did try this cycle, and had mixed results.

The main reason is, that, uh, constructive redistricting makes it hard to correlate voter patterns to a win. Florida’s topline totals are very closed to tied – 50/50 between Pres. Obama and Gov. Romney.

Yet, the Congressional delegation is something like 19-7 in favor of the GOP. Based on the concept of the House of Representatives, being a fairly democratic body, you would expect it to be divided roughly equally.

The same is in true in Democratic states – California, New York, Illinois, Massachusetts.

The name of the game is gerrymandering – you cut up your opponents districts in a way to concentrate their voting bloc into the smallest number of districts possible, and then you dilute your voting blocks just enough to safely create districts that are 1.5-to-1 in your favor. That way you don’t have to really run much of a re-election campaign, and your opponent’s can’t really do anything about it. 10 years of easy re-elections.

Trying to capture this in a polling model is very hard, or impossible. Also, polling individual Congressional seats is pretty difficult.

Huh? Is that news? PPP was ‘rigging’ the polls to the 2008 model? I thought that was known, every time we cracked open polls to see D +8. Or is it that they were doing weightings in there.

We conservatives argued against this being a 2008 repeat, a lot of evidence mitigated against it … and … oops, they were right. First, you cant claim PPP was ‘rigging’ to be dishonest since, well, it gave a good result. And this was a Demographic and not political aligned weighting (although the 3 groups mentioned it boils down to much the same thing).

BUT YOU MISS THE REAL POINT. The REAL STORY. How did Romney get fewer votes than McCain?!? How did that happen? If Romney had achieved the McCain level, it would be a telling the whole 2008 turnout model might be thrown out.

So perhaps the question is less “How did youth/black/hispanic vote stay up?” It is “How did all the other turnouts go down like it did?”

The difference between weighting and rigging is the process. Read the posts. Study science instead of lapdogging the lefty activists here.

You cannot determine accuracy by one data point.

HA… you can’t retroactively call a poll rigged… Now it’s ppp (whom I thought would be colossally wrong). Whoever was right, you’d call it rigged? This isn’t “science” this is math. PPP guessed, and I do say GUESSED, right. To call it rigged is very bizarre. Every campaign cherry picks the polls to sway public opinion. I really think this is barking up the wrong tree.

You get a lot MORE info from EV, as you can tell who voted.
It changes the whole GOTV game. Its silly to say EV spells the end of getting election day info, it tells you a wealth of information!

Had you had a local person look into it, they could have told you if EV trends were better than 2008 or not or more like 2010. One thing that frustrated me was the GOP side talking about “Hey, we are doing better than 2008 on this or that.” when comparing against a 7 point win was the wrong metric! Yes, we cut from 7 pts to 2pts, but it didnt cut it did it.

The more I think about it, the more I conclude that this is more an issue with GOTV then about polls. Any given poll has a 3% MOE. Any given GOTV effort can get you 3% delta in turnout. WE GOT THE 2008 TURNOUT MODEL BECAUSE THEY RAN AS GOOD A TURNOUT / GOTV OPERATION AS IN 2008 AND WE RAN AS BAD A TURNOUT / GOTV OPERATION IN 2008!

Gerrymandering is a nice excuse that both parties like to draw on, but when you have a turnout disparity at the level that we’re talking about it’s insufficient. Take a look at the swing districts; take a look at places like Wisconsin, on the state level-races; and take a look at the swing state results.

Kyle, that’s weighting.(2008) Unless you’re claiming they made up their results. If so, they were very accurate in fake data. The death of a pollster would be to be wrong. Why would they try to fake it?

No its not. It is not reasonable and objective. You can do a back test on known data to show that it is not. For example, take the turnout from 2004 and use it to adjust the polls in 2008. It will not give you the correct answer. You can do the same with 2000 and 2004 and get the same incorrect answer. The only time it works is between 2008 and 2012. That is not accurate.

Seriously? Let’s blame the polls? I’ve been trying to argue my point below, but this is laughable. Come on Neil, it’s not the MSM polls you should be upset about — or any polls. Romney failed to convince the American public about his vision for the future. Period.

I have my own diary on this now.

http://www.redstate.com/wosg/2012/11/09/it-was-the-gotv-all-along/

This article is just flat out wrong. PPP didn’t “rig” their polls. Like every other poll, they weighted the results to reflect a certain demographic breakdown. There is nothing dishonest about that.

PPP projected turnout based on the information gathered during the calling. I really don’t understand this diary. PPP was accurate. Can’t fix a problem without defining it first.

The objective data was created by how the respondents self defined. This was D+6. Romney/Rove pushed back on this data because they could not convince the media or the money backers they had a fighting chance if they bought into reality. This was a sad case of intentional self dellusion.

Maybe we will realize climate science is accurate and well defined as well? Evolution next? The anti-science, anti-intellectual movement in the GOP is ridiculous. There is nothing snobbish about an education. Nothing.

No.

Obama wasn’t on the ballot.

Blacks and young voters are much more likely to turn out for a black President than for some white Congressman.

So did every pundit that our side trusted: Michael Barone, for example.

Why aren’t we having diaries criticizing him???
Nearly every conservative pundit said Romney was going to win. They were obviously telling us what we wanted to hear, because conservative journals pay them to tell us that.

The next election that comes up, I will listen to Nate Silver, not to Michael Barone.

I care only about final results. I don’t care if they involve some guesswork.

I think the conservative echo chamber, which includes the blogosphere, has succumbed to the same measure of success as reality stars as to what is relevant. The more audacious the claim (without pushing too much the limits of absurdity) the more publicity it garners. Publicity, in reality television, my friends is its own reward.

Cases in point:

Dick Morris -Romney will win in landslide. Any self-respecting numbers guy and more importantly, one who is living in the real world, would know that such claim is preposterous for either one of the following two reasons: there has not been a landslide election since 1964 (Johnson/Goldwater). Since the 1980s, the electorate has become much divided and the candidates less distinguishable (all running to the right of center). Fertile ground for a non-landslide.

Donald Trump -Need I say more? No I shouldn’t. I’d feel more cheapened than I already do talking about Dick Morris.

Fox News -Need I say more?

John Sununu -What he said about Colin Powell was red meat to the conservative base, but effective in alienating the minorities and the youth whose sense of race is a bit more evolved than us old timers.

Neil Stevens -Blaming those who got it right (PPP, Nate Silver, etc.) for making a more plausible assumption garners him one of his most discussed diaries. To many of his readers’ credit, they see the foolishness in his non-approach.

The accountability of reality stars is only to themselves. They are as good as the publicity they attain, good or bad. This, non-surprisingly, fits the conservative dogmas of you’re on your own, and the end justifies the means. If you attain success, regardless of the damage you cause to your environment and fellow humans, you have nothing to apologize for.

If we were such individual beings, we don’t need society and in extension we don’t need politics. After all, politics is the game of who gets what and how much. We have finite resources and we use politics to duke it out as to who gets what and how much. As such, we’re inherently interconnected. The less greedy and self-serving we’re the healthier the societies in which we spend our short lives.

How is a target party ID mix any more objective that a target demographic mix? Seriously, you are being ridiculous – they were both weighting it based on subjective opinions of turnout and it turns that averaging everyone’s subjective views produces a better prediction – everything else produces denial.

Nate Silver is engineering his own obsolescence. As Sean Davis pointed out a week or two ago at The Daily Caller, setting aside the whimsical secret sauce of the way Silver “weights” polls according to his personal judgments (or biases), what Silver does can be done by anybody with the time to feed poll toplines into Microsoft Excel with a Monte Carlo plug-in. You need to be a better mathematician (or at least have more patience) than a humanities major like me to do this stuff, but that’s hardly saying much.

Even more fundamentally though, Silver’s premise is: polls are trustworthy, and the guy who is ahead in the most state polls in key states is going to win. Where would we be without Nate Silver to deliver us this insight? I watched him have a little Twitter-spat with Matt Drudge about the polls in 2008. The ’08 polls were “dead-on,” Silver insisted. Fortunate for Silver, since he built his reputation on interpreting them. If polls are dead-on and a reliable predictor of outcomes, what, exactly, do we need Silver for?

I won’t give you a percentage figure on the probability, but this moment is likely the peak of Silver’s popularity and influence. Conservatives like Cost raised his profile through their back-and-forth with him, and now the election results have given him vindication. His book is flying out of Amazon warehouses.

But where does he go from here? Obviously the first election where his predictions happen to be substantially contradicted by the actual results, his reputation as a soothsayer will take a substantial hit. He will have explanations or excuses, and the rest of the media will cover for him, and he will survive, but he will be damaged nonetheless. Others have pointed out that he didn’t even make an attempt to predict house races this year, and this was a shrewd move on his part; his methods clearly don’t work nearly as well in that arena.

But even assuming the best-case scenario for Mr. Silver, he has nowhere to go but down. The novelty of his methods will wear off. Other people will copy them, with minor variations and modifications, and start producing similar statistical probability forecasts every election cycle. The pedestrian simplicity of what Silver actually does, the lack of any truly revelatory innovation or insight in it which I alluded to earlier, will start to become more and more apparent.

The career forecast for Nate Silver is to become either useless or superfluous.

Now that said, it won’t stop him from enjoying a long career as a talking head and newspaper and magazine columnist alongside the host of other poll wonks and prognosticators, most of whom are useless or superfluous themselves. And aside from the exasperation caused by the smug, casually-informed liberal internet commentators who treat Silver like a guru, I don’t begrudge that kind of career to a baseball starts nerd made good. But I hope he enjoys this peak while it lasts, because the product he sells is one that tends to lose a lot of its shine, sooner or later, one way or another.

“Where was our GOTV?”

RIGHT. That is what I concluded after looking at the exit polls and the results. There were millions of soft Republican voters who didnt show up.

http://www.redstate.com/wosg/2012/11/09/it-was-the-gotv-all-along/

Had Romney done better GOTV he might have confounded those polls.