When it comes to pundits, why are we selective about the ones we listen to?

[johnny9fingers.livejournal.com]

Over the past few months here, we have had a number of posts, many of them tagged obama's uphill battle, which have attempted to explain the course of the US Presidential election in relation to collated public perception.

Therein we have been shown the wonderfully incisive political analysis of various pollster agencies based on specifically-framed questions which are asked of demographic slices of the general public, the answers to which are then subject to some less-than-rigorous processes, and which finally give whatever grouping within the political sphere either comfort or despair.

But, come the election results, we have a measure to judge the relative accuracy, or otherwise, of the pollsters and their analysis. Have they asked, er leading questions? Is there a pre-existing bias that prevents them from being able to accurately assess the information resulting from their questions? Are the pollsters trying to influence the results of the polls, and thereby influence the results of the election?

So...

Which organisations/individual analysts/pundits do the panel think called this election correctly?

IMHO Nate Silver was the best, and Rasmussen the worst.

Comments

[badlydrawnjeff.livejournal.com]

The good news, I guess, is that the pollsters generally nailed it down. We'll be able to sift through how well they did so later, but it's good to know that scientific polling in fact ruled the day, and the concerns about party identifications and such were completely unwarranted.

[johnny9fingers.livejournal.com]

Yesterday Rasmussen was calling 50-49 in Romney's favour.

Some of them didn't nail it down at all.

[kylinrouge.livejournal.com]

Rasmussen and Gallup were way off the entire time. Confirmation bias made people believe in them despite the fact virtually every other poll disagreed with them.

[badlydrawnjeff.livejournal.com]

And yet 50-48 is well within the margin of error.

[dwer.livejournal.com]

Well. Nate Silver nailed it down.

[badlydrawnjeff.livejournal.com]

And Nate Silver was able to nail it down because the polling was good.

[fizzyland.livejournal.com]

I made this for you - I know you never took my wager but if you want a color update, it's all yours.

Just tell me you're going to stop making predictions. Because you're really bad at it. Nate Silver should be your new guru otherwise.

[luzribeiro.livejournal.com]

Any poll question other than "Who will you vote for, A or B?" is useless.

[johnny9fingers.livejournal.com]

Well, this time the popular vote and the electoral college seem to agree: but there has been a recent election where that wasn't the case. Even such simplified questions when asked of voters in non-crucial districts will not necessarily give an accurate result.

[politikitty.livejournal.com]

Actually, new research from Justin Wolfers is showing that non-response rates are hurting that sort of polling.

By instead asking "who do you think will win?" you're able to tap into the collective social network and get a sense of the entire population, not just those who are likely to pick up the vote. Quick overview here (http://www.bloomberg.com/news/2012-11-06/obama-will-win-say-expectation-polls.html)

All really fascinating stuff.

[devil-ad-vocate.livejournal.com]

Polls are good for predicting trends; they are not - and never have been soothsayers.

[jonathankorman.livejournal.com]

If our media sphere worked the way it should, all of the pundits who said that Nate Silver was a propagandist who was making things up and didn't know what he was talking about should be completely discredited. But of course that won't happen.

(And speaking of propagandists, that clip reminds me to say that Yes, Minister was a deliberate work of conservative political propaganda. (http://reviewsindepth.com/2010/03/yes-prime-minister-the-most-cunning-political-propaganda-ever-conceived/) Really.)

[johnny9fingers.livejournal.com]

Jay was one of Maggie's chaps, true. I'm from rather a different conservative grouping: Macmillanite (http://en.wikipedia.org/wiki/Harold_Macmillan) One-Nation-Tories (http://en.wikipedia.org/wiki/One_nation_conservatism).

Essentially Keynesian mixed-economy advocates. Paternalistic, patrician, and emphasising personal duty and social responsibility and cohesion. Pretty much to the left of today's Labour Party in the UK, and the Dems in the US, actually. But still conservative. Conserving of good manners, compulsory Latin and classical Greek in schools; the idea of High Culture, Intellectualism, and learning: all of which are now unfashionable in the Anglo-Saxon cultures. (Though I'm glad to see that folk are now beginning to regard racism, sexism, and homophobia as extremely bad manners.)

I always regarded Maggie as a trumped-up grocer's daughter with C19th economic-liberalist tendencies. Jay was a different matter, being rather cleverer and better educated. I disagree with many of his opinions but still think he can write a damn fine comedy, much of which still has pertinence, even now.

[underlankers.livejournal.com]

We're selective in listening to them because ultimately pundits exist to make our reality validated, not to provide any kind of genuine substance on the issues. In general people do not accept the prospect of being wrong a sixteenth as well as what they tell themselves they do.

[a-new-machine.livejournal.com]

Basically this.

[musicpsych.livejournal.com]

I honestly didn't pay much attention to polls or pundits this time around. I paid attention to the presidential race and the debates, but my blue state vote wasn't going to swing the election, so why bother worrying about it? I knew it was out of my hands, and I had to move forward no matter who was elected. I had to Google "Nate Silver" to figure out why people were talking about him. (At first I confused his name with "Nate Fisher," the character from Six Feet Under.)

We're selective about pundits we listen to because we want to feel right, so we listen to ones we are likely to agree with. I think there are leading questions in polls, and pundits no doubt pick and choose answers that support their claims. I also think people are less honest when answering questions on polls, since they are telling someone else their beliefs, and may be judged on that. At the ballot box, it's just them and the ballot.

It's also important to keep in mind that polls may not include a reliably representative sample; the pool of people they survey may include people not voting, and people who are voting might hang up or not take the time to answer. I suspect there are theories about how big a representative sample would need to be to be accurate, and I think that probably works most of the time, but not all of the time. The only perfect poll is the election. (I also think TV ratings are inaccurate, but that's another story...)

[musicpsych.livejournal.com]

Another question: Do you think it's worthwhile to break down polls based on demographic info - race, age, etc.? Why or why not?

[chron-job.livejournal.com]

It is useful, if it has predictive value. If it doesn't predict anything, it can't be tested, and isn't useful.

[telemann.livejournal.com]

Nate Silver does not consider himself a pundit. He's a statistician and psephologist. His statistical analysis of punditry (saying they are worthless) is why they hate him so much.

Even yesterday, MSNBC and Howard Fineman were dissing Mr. Silver pretty hard.

[fizzyland.livejournal.com]

I agree.

[dwer.livejournal.com]

and they were all. wrong.

[johnny9fingers.livejournal.com]

Well, he seems to have come out of this best of all. Good for him.

[sophia-sadek.livejournal.com]

The kind of pollster that is most insidious is the one that starts with a survey and ends with an appeal for financial support.

[johnny9fingers.livejournal.com]

Well...how much can you afford?

[anfalicious.livejournal.com]

IMHO Nate Silver was the best, and Rasmussen the worst.


I thought that was common knowledge up until about 6 months ago. Being out of the loop (live in Oz and took some time off news this year) I missed the start of that info war, but could remember the 2010 and 2008 elections where Nate Silver was pretty much spot on and Rasmussen was lulzworthy.