GSO/Guilford Pols

September 2014

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30        

« The hurricane agenda | Main | Back it up »

Nov 01, 2012

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341cc33e53ef017ee4a26d74970d

Listed below are links to weblogs that reference They can't both be right:

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

bubba

"Look state by state, says FiveThirtyEight.*"

Look elsewhere:

"We can’t know until Election Day who is right. I stand by my view that Obama is losing independent voters decisively, because the national and state polls both support that thesis. I stand by my view that Republican turnout will be up significantly from recent-historic lows in 2008 in the key swing states (Ohio, Wisconsin, Colorado) and nationally, because the post-2008 elections, the party registration data, the early-voting and absentee-ballot numbers, and the Rasmussen and Gallup national party-ID surveys (both of which have solid track records) all point to this conclusion. I stand by my view that no countervailing evidence outside of poll samples shows a similar surge above 2008 levels in Democratic voter turnout, as would be needed to offset Romney’s advantage with independents and increased GOP voter turnout. And I stand by the view that a mechanical reading of polling averages is an inadequate basis to project an event unprecedented in American history: the re-election of a sitting president without a clear-cut victory in the national popular vote.

Perhaps, despite the paucity of evidence to the contrary, these assumptions are wrong. But if they are correct, no mathematical model can provide a convincing explanation of how Obama is going to win re-election. He remains toast."

Ed Cone

Here's another take on the independents.

I'm ready for Thanksgiving.

formerly gt

Someone said they see this as either a narrow win for Obama, a narrow win for Romney, or a landslide for Romney. I agree. I don't see an Obama landslide.

I respect statistics and modeling. But, I hope Nate's models are built on faulty data. He's either going to be elevated to god status or be written off next Wednesday.

Thanksgiving? Miss the kids?

(Being an NC State fan, I'm ready for ACC basketball this year. Daughter is still at home)

Andrew Brod

Nate's models are built on polling data. If the polls are systematically wrong (as opposed to randomly wrong), his model will be wrong.

Hugh

I've intentionally given robo-polsters the wrong info this time around. I suspect there are many others who do the same.

Ed Cone

FGT, I do miss the kids, but really meant that I'm more than tired of this endless campaign season.

Kim

I'm going with the Redskin rule. We'll know Sunday afternoon.

Thomas

Only five days until the start of the 2016 campaign!

Andrew Brod

One more on 538: "Re: Nate Silver, most amusing thing about this election is watching political pundits make sports fans look like PhD mathematicians."

Ged

Nerds Rush To Nate Silver's Defense - The Atlantic

Tim

Get ready for Coneheads to develop the mental disease to be known as RDS.

Great comment Mr. Bubba.

(-:

Tim

Bill Yaner

We've struggled all the way down from ancient Greek and Roman civilizations to understand and predict the future. Pollsters and political pundits are only the latest manifestation of this compulsion to "know" what's going to happen before it does.

Then comes roaring down the highway this massive, dark force of nature we try to humanize along the way by naming it Sandy. How dangerous could a gal named Sandy be? Come on Sandy, cut us a break.

Then sayeth the raven, "Nevermore".

Andrew Brod

Hey, kids! Think you can do better than Nate Silver or Karl Rove in predicting the electoral vote? Then enter this contest and make your own prediction!

But tell us how you did. I'm sure all the usual commenters here will want to congratulate you.

Roch

Obama: 305
Romney: 233

http://on.wsj.com/Qrt3ak

MojoNixon

Freakin' pessimist:

Not Even Close!

Andrew Brod

Okay, so Roch predicts 305 electoral votes for Obama, while Mojo says 358 (same as the 2008 result) and Frog says 268.

For context, Nate Silver's now-cast stands at 304 today, nearly identical to Roch's prediction, while Drew Linzer (Votamatic) says 332 and Sam Wang (Princeton Election Consortium) says 318.

Roch

Obama won 365 electoral votes in 2008.

In 2008, I predicted Obama would win North Carolina by 19,200 votes. He won by about 14,000. This time I think Obama will maintain NC with a 40,000 vote margin.

Andrew Brod

So he did. I was going by the WSJ map, which gives the contest-taker the option to start from the 2008 electoral map, and its 2008 map shows Obama with 358.

It's amusing that it's the Journal that "took away" those 7 votes from the Democratic candidate.

Predicting NC for Obama takes you out on a bit of a limb, though I guess the same was true in '08. Nate Silver colors NC pink, but Linzer and Wang call it a toss-up. So does Real Clear Politics, but that's less revealing given that RCP puts a total of 146 electoral votes in that category.

Account Deleted

That WSJ map has Arizona blue and I just don't think that's going to happen.

I'll go with 290-248 for the president.

I could see Va. going for Romney but enough independents mixing ballots to give Kaine the senate seat over Allen.

Roch

Andy, I haven't checked, but could the difference be caused by re-apportionment of electors since 2008? Same states yielding fewer electors because of population shifts?

Roch

Jeff, Arizona or New Mexico?

Account Deleted

Their base map has Arizona blue.

Also, you are probably right about the changes in the EC. Here is a map with the state losses highlighted for easy comparison to your CNN map from 2008.

Andrew Brod

That's probably what it is, Roch. Good catch.

Spag

Romney 295, Obama 243

Obama loses NC by 5% or more.

Spag

Map

Worst person on the internet

Betting pool, anyone?

polifrog

Agreed, Spag. Obama loses NC by 5% or more.

That Roch believes there will be an increase in Obama's support in NC throws into question his other state calls. Not that I believe it, but PPP a couple of weeks ago found only 70% support for Obama among black voters. At least some of that will result in a loss of black support in NC, much it due to gay marriage.

270 Romney is my low end for Romney after trowing Obama some bones like Arizona and New Hampshire

Obama's is an eroded base. Modest swings in the electorate translate to massive swings in the EC and the swing is away from Obama. Relative to '08 Obama has lost independent support, black support, female support, Jewish support, and most importantly enthusiasm. I keep expecting to see at least a few Obama signs and bumper stickers materialize as we close in on Tuesday. I haven't. Those swing voters who are swinging between, not Romney and Obama, but voting or not, are unlikely to turn out for Obama this time around. Doubly so for first time voters which Obama was successful in drawing to the voting booth in '08.

Conversely Romney has made gains relative to McCain in '08 among those choosing between voting or not and has likely picked up a few Obama supporters as well.

Worst person on the internet

Does anyone else find this intriguing? What 3 things do we all have in common here?
1) Strong political views,2) Never being wrong and 3) Winning. (Some are also obsessed with their idols' and villains' predictive powers as well.) Look how many have stuck their necks out already here in front of God and the world . We really need to have a pool here. Winner supplants Silver/Krugman as God for one year. Last place banished or condemned to Kudlow/Cramer status. We could do split pot or winner take all. Who's in?

Spag

CP, some people won't participate because they fear the Fonzie Factor. Being demonstrably wrong is something they can't cop to.

polifrog

Is the news-observer the source of Roch's expectations?


Roch

There is something wrong with the WSJ map's display of what it says are the Real Clear Politics averages for some states. Although Arizona is red on my screen, click on Arizona and it shows Obama: 49, Romney: 44, but the RCP average for Arizona is actually reversed, with Romney at 49. Mistake with Georgia too.

Gerry Alfano

I voted yesterday at Weatherspoon. I was happy to see a really long line of young voters and didn't mind the wait to vote. What I heard from the young voters was that they really appreciated the convenience of having an early voting site on campus. Republicans should really be ashamed of the fact that they tried to restrict voting. We need more participation in the democratic process, not less.

Ed Cone

I don't really get the looking-for-comforting-numbers thing. Not saying it doesn't happen, obviously it does, just that it's alien to me. I'm not looking to the projections to feel happy, I'm looking for a sense of what's actually happening on the ground, good or bad (from my perspective).

Maybe it's a personality thing, but I'm more likely to fret over the numbers that predict an outcome I don't want than to celebrate the friendly projections (or, to be more accurate on Silver's methodology, probabilities).

If you are spinning for a campaign, you want to keep your people energized, but if you're just following at home and actually want a sense of what's happening out there, why bother? There's going to be an election next week, so the mood management seems moot.

The predictors themselves will see their reputations rise and fall on the results. For the poll crunchers, including Silver, those results will speak not just to their models, but to the accuracy of the polls themselves. It's not magic, it's data crunching, and the idea is to improve both the data and the crunching over time.

As far as personal predictions and electoral counts go, it's a game, like an office pool for the NCAA tourney, but one in which I have even less information and intuition. As such, it's fine by me, but I have no insight that would give me some sense of pride in getting the exact number right, or make me feel foolish for getting it wrong. So without looking at the WSJ map, I'll say Obama 283, Romney 255.

Re our home state, as I've said before the conditions seem right to me for a Romney victory, with the possible difference coming from the strength of Obama's ground game. I'm a lot more interested in the analysis of what ends up working and what didn't, and its application to future campaigns, than I am in guessing how things will turn out.

polifrog

Ed Cone:

For the poll crunchers, including Silver, those results will speak not just to their models, but to the accuracy of the polls themselves.

I see. Silver won't be wrong because he was unable see through the noise but because the noise was too noisy.

Otherwise much to agree with in your comment, Ed. Even the NCAA tourney comparison. I have no interest in the NCAA but my wife does. She accused me of playing the brackets (if I said that right) yesterday when she saw me fiddling with the WSJ state by state above.

Ed Cone

Silver and the others who crunch polls are by definition dependent on the underlying poll data. The post I linked in my comment speaks in some detail to that issue. GIGO, and all that.

Different poll crunchers use different methodologies, from simple averages to much more complex weighting. The ability to tell signal from noise is what they're selling. They should be judged, after the fact and repeatedly over the years, by their performance versus the individual polls, and each other.

If Silver is wronger than rival poll crunchers, then his methodology will be shown lacking relative to theirs, and vice versa. But all of them start with the same polling data, so the accuracy of that data is highly relevant to their outcomes.

Worst person on the internet

Obama 290 Romney 248

Ed Cone

Worst, my dad told me bet against the Tar Heels -- that way I'm happy either way.

Not the first time you've reminded me of my dad.

Worst person on the internet

you nailed me

polifrog

Some people argue that intrade represents, through the pursuit of wealth, a form of popular opinion on the election stripped of clouding desire. Although the thinking is appealing it is not my personal experience.

I find that wagers cloud my judgement.

Good call on Worst, Ed.

Andrew Brod

I used to think that political markets were the wave of the future. If traders have a monetary incentive to be right, they're going to try to be right. However, the problem is information. A commodities trader might have information that goes beyond what he reads in the newspaper; for example, he might be a producer or a broker.

But the vast majority of traders in political markets appear to have no more information than what the polls say. As a result prices in those markets mostly follow the polls. And when they don't, they indicate a great degree of credulity, e.g. prices can zoom up for a day thanks to one piece of news. Sure, arbitrage and other traders often bring the price down, but so do additional waves of polling.

I could be wrong, but I don't think the political markets (Intrade, U. of Iowa) have done systematically better than the polls.

P.S. If it's reasonable to consider it significant that traders in political markets have a monetary incentive to be right, it's only fair to do the same with analysts like Nate Silver and Drew Linzer, whose business models need them to be right or at least close to it.

Andrew Brod

So you say you don't like polls? Well, here's an interesting indicator. It might even be right.

polifrog

'Nother Prediction:

It will become clear tomorrow night that for having vastly overdrawn his store of political capital, Obama will be kneecapped by America and in the process, will have destroyed the Democrat party in ways that those who believe in liberty had no hope in achieving without Obama continuing to draw upon nonexistent political capital after the midterms.

53%(R) - 47%(O)

"47%" will haunt Democrats far longer than it ever haunted Romney.

Polling party identification assumptions are off. With the exception of the bluest of the blue states there is no +D party identification this election.

Andrew Brod

Actually, Obama can't destroy the "Democrat Party" because there is no party by that name.

Andrew Brod

In what I presume is 538's final forecast, Nate Silver is predicting 315 electoral votes and a 51-48 popular-vote win for Obama. Frog's predicting 47-53. We'll see who comes closer.

Roch

Revised:

Obama: 318
Romney: 220

http://on.wsj.com/RyMjUM

polifrog

No call from Brod...

Roch

More:

Presidential popular vote:
Obama: 50.15%
Romney: 48.75%

Senate:
Republicans +1

House:
Democrats: +4

North Carolina presidential
Obama by 40,000 votes.

Marshall

Obama helped carry Perdue in NC last time. (sorry excuse for a governor)
McCrory helps carry Romney in NC this time.

Account Deleted

Looks like Roch was closest without going over assuming Fla for Obama.

I'm surprised Romney lost Virginia and Florida. The demographics have shifted toward the future.

The comments to this entry are closed.