Poll Denial
Let me say that if all the polling, and the analysis of people whose work I respect, like Nate Silver’s, turns out to be wrong on election night, I will be forced to re-evaluate what numbers I take seriously going forward. At a minimum, I will expect an explanation for what went wrong.
After all, evidence is evidence.
Having said that, I am astonished at the degree to which many who are rooting for Romney seem to be in total denial about the polling. For example, the following from Katrina Trinko at NRO:
But regardless of partisan breakdown, Republicans should be wary of taking any polls as completely accurate.
“Part of the reason the Democrats won in 2008 was that when it looked as if McCain was going to lose, some Republicans stayed home,” argues McLaughlin. “So if President Obama is in a dead-even race with Mitt Romney in so many swing states, if the Democrats can convince enough Republicans they’re going to lose, it could take a one-point loss for the Democrats to a one-point win.”
Emphasis mine.
The thing that is remarkable about the above is that it not only in based in an approach that privileges preference over reality, it comes with a built-in fairy tale to explain any non-preferred results! Using the logic above the polling can be wrong whilst predicting the actual outcome and, better yet, the wrong polling (that was actually right) wasn’t just wrong, but it caused the wrong outcome to occur!
Life would be better for all of us if we were all, regardless of partisan preferences, a tad more grounded in empirics.
They’re standing on the bow of the Titanic trying to convince the rats not to jump off. Panic. Pure, sweet panic.
Do they know that they are a minority party?
Its over. Probably Dems keep Senate. The big surprise may be the House. Possibly a bloodbath.
This must be the liberal psych-ops Jan was blathering about yesterday. Just get enough of the MSM to feed Republicans discouraging poll information and they’ll be so demoralized that they’ll sit out election day. Because, in the right wing bubble, it couldn’t be possible that Obama is actually leading in the polls.
@Steven Taylor:
Townhall has this:
“An indication of the media running with a narrative opposite of what the country actually believes”…
If this is an indication of anything it’s an indication of their desperation and/or them being unable to understand polling.
It is one thing to say polling doesn’t matter in June, but in September?
@Jay_Dubbs:
I still believe that GOP SuperPACs will drop huge amounts of money on a number of House elections. Question is if Obama’s coattails will be enough.
Steven: I’m not sure you’re reading the comments very charitably. At least the portion you quoted makes two non-controversial statements: 1. Republicans (like anyone else) should be wary of taking any polls as being completely accurate (emphasis mine) and 2. just because the polls say your side is losing doesn’t mean you shouldn’t go and vote anyway.
We know polls are subject to substantial measurement error, and while averaging (like RCP or Nate Silver’s approaches) does offset some of that, you can’t get rid of all the factors that could make a poll inaccurate, particularly when both the true population parameter and the population of interest are not known with certainty (who will vote or not, and how they will vote) until after the election. For example, a 2008 turnout model will probably overstate Obama’s lead (youth enthusiasm for Obama in 2012 is substantially lower than in 2008, which means youth voting is likely to regress toward historical averages), while another turnout model (say from 1996 and 2004, a less close election) might understate it.
And I don’t see anything in the unquoted section either that seems to argue for your interpretation; it reads to me as restating what public opinion researchers know, which is that PID is dynamic over the course of a campaign and thus it’s difficult to judge if a sample is overweighting one party over another.
@PJ:
You’re kidding right? A website, side-banner poll? You may want to scroll down and read the disclaimer that’s always at the bottom of those…
well, there’s still the actual “vote” that takes place- that’s the final poll. and seeing as obama said he can’t/won’t work with congress i guess we’ll have 4 more years of nothingness.
@Jeremy R.:
Thing is, they have. But they still argue that it’s an indication…
@PJ:
Oops, I misread that post of yours I responded too. I see you were calling attention to their silly spinning. Sorry.
@Chris Lawrence:
Actually, I would argue that you are being overly charitable. If the admonition was a general one aimed at cautioning readers of polls that they are not perfect, I could live with that. But that parenthetical is from you, not the author.
I absolutely agree with this. However, that is not what the quote is saying or, at least, I think it takes an overly charitable interpretation to see it that way. I think it is looking for an excuse for the results. However, even worse, I think it is building in an excuse for a Romney loss this November for those who don’t want to believe the polling now.
Again, I think you are being too charitable. I don’t see it as a basic cautionary tale about polling. I see it as an attempt to assuage the concerns of Republicans in a way that does not educate about the potential foibles of polling, but rather to undermine polling in the minds of Republicans.
I will look at it again tomorrow, as perhaps I am being unfair.
@Jeremy R.:
That’s ok 🙂
@Chris Lawrence:
This article wasn’t written in a vacuum, though. There is a whole context here.
If you really believed that everyone wanted to get rid of Obama, and all they needed was a (Republican) alternative, then it could take some time to adjust to polls heading the other way.
Most pollsters get paid to do polls. When they blow the results then they stop getting hired to do polls. Now, not all pollsters are like this, you’ll see universities do polls from time to time, but for the most part, pollsters don’t want to be wrong because it’s bad advertising for them. That’s why it’s a bit presumptuous to pick apart their methodology (with only partial information), and it’s definitely crazy to think there’s a conspiracy. Especially with something as pure as straight up election polling. Polling on issues and questions though… more art than science.
Now, I’m not saying that pollsters are mathematical gods that us mere men should never challenge. But at some point, people have to realize a lot of these pollsters have track records that extend beyond this particular election.
I even tend to stay away from Rasmussen bashing. Sure, they tend historically to lean republican. But their precision still makes them useful.
Matt Tabbi had an article a ways back called “The most boring election ever” (or thereabouts, I be drawing from memory here) that suggested the “Media” has a vested interest in this being a close horse race.
Theory being: The TV stations are where the money goes, mostly, and the pollsters are also businesses which don’t do as well if people are not heavily focused on it race. Add to that the PAC’s. Those boys know how the big donor$ act when they detect the scent of “Lost Cause”, or “Eau de Faile”, as it is sometimes known, certainly.
He predicted that all the above would spin it as a close race as long as they could, the limit being at the cost of their credibility. I would guess he must have to be referring primarily to the pollsters there… ;).
We’re talking about a party where a scary proportion of their members still aren’t sure the President was born in the US, think the Earth is under 10k years old, are completely against the idea of biological evolution, and still think global warming is a myth. What’s disbelief in a couple polls compared to that?
@Chris Lawrence: I just re-read the piece. I still see its main purpose as a means to give NRO readers (i.e., Republicans) reasons to doubt polling. Further, I don’t think that the purpose is to doubt polling in a scientifically skeptical way (I am all for that), but rather, starting with the title, to suggest bias.
I consider it a possibility, although not a probability, that all these polls are wrong for any number of reasons. Because, as you note, “it’s difficult to judge if a sample is overweighting one party over another.”
However, I don’t think that the purpose of the piece, and certainly not the way it will be received by most voters, is a clinic of sampling methodology. I think it plays into the notion that pollsters are all part of the dreaded MSM and are, therefore, always biased against Republicans.
@Jeremy R.: In other news, Ron Paul will be our next President.
In other words, epic fail for Townhall.
@Jay_Dubbs:
Very unlikely. Far too many uncompetitive districts via gerrymandering.
@Steven L. Taylor:
It’s funny how the first lie of defense of the party of personal responsibility — see Mitt’s quote about how all democrats cannot take responsibility for their actions — is everyone is biased against us.
(btw, that wasn’t a typo in the previous sentence).
@Console: “Sure, they tend historically to lean republican. But their precision still makes them useful. ”
No, unless you know their bias very well.
@PJ: @Fiona: No, because he is the worst president EVAH. And a Marxist to boot.
“Let me say that if all the polling, and the analysis of people whose work I respect, like Nate Silver’s, turns out to be wrong on election night, I will be forced to re-evaluate what numbers I take seriously going forward. At a minimum, I will expect an explanation for what went wrong.”
Your problem Dr Taylor is that YOU are in denial, duped by all these rigged polls which show Dems turning out to vote by margins of 10% verus Repubs on election day. That may happen on Pluto & Mars but not in the USA this Nov. You, Doug & the other far left posters on here just cannot see how the CNNs, Pews, CBSs, NY Times, PPP are oversampling Dems to push an agenda.
I notice you didn’t mention the Daily Tracking polls surveying thousands of people that show the race is a tie with Obama Job Approval in the mid 40%. In fact, I hardly ever see the huge daily trackers mentioned on here – only rigged polls from the likes of PPP & CBS showing gazillion more Dems than Repubs voting this Nov. Keep deluding yourselves that this race is over. Btw, Nate Silver, the guy you respect so much, is a far left hack who cannot be trusted because he works for one of the most left wing organs in the world.
@Smooth Jazz: This will be easily settled in just over a month.
I will graciously acknowledge the error of my ways should it prove that the polling ends up to be wildly out of sync with reality.
Will you do the same if your rejection of the polling ends up to have been incorrect?