The Question of the Hour, Or, Why Are the Polls So Screwy This Year?

Today on RealClearPolitics, Sean Trende:

Given what we know about how individual states typically lean with respect to the popular vote, a Republican enjoying a one-point lead nationally should expect a three-to-four-point lead in Florida, a two-to-three-point lead in Ohio, and a tie in Iowa. Instead we see Romney ahead by roughly one point in Florida, and down by two in Ohio and Iowa.

Why are the polls so screwy?  Trende’s own analysis is worth reading in full, and Rob, James, Troy and I spent much of the hour on the podcast we recorded this morning–with the able assistance of Ricochet members, and political professionals, Adam Schaeffer and Rick Wilson–trying to figure it out.

But here’s the short answer:

Nobody knows.

  1. Carver

    Why were mortgage loan applications or appraisals  so screwy in 2007?

  2. Keith Preston

    Never in the field of human hysteria have so many been so misled by so few.

  3. Valiuth

    Sampling bias. That is the best explanation. Pollsters for some reason can not sample a representative slice of America. They then try to correct for this by making up constants from thin air…

  4. Schrodinger

    Jay Cost also has an article on polling at RCP.

    My take from these articles is that, this year, pollsters are divided into two groups. They differ dramatically in their assumptions about the demographics and party ID of the voters that will actually vote. Thus, the results differ significantly. Using an average, like RCP, is no help. One group will be right and the other wrong. Splitting the difference is meaningless. This is a “place your bets” election. There is no fence to sit on.

    The other point is that the state polls are often run by local, less established polling organizations as compared to the national polls. Again, the inconsistency between the two may be a matter of methodology.

    So, who do you trust?

    Personnally, I will be watching VA, FL and NH on Tuesday night as indicators as to which set assumptions is correct.

  5. Arizona Conservative

    Rush suggested this morning that the left is still skewing polls to prepare and bolster their accusations of fraud in the case of a Romney win. They think close polling numbers and, in some cases, polls showing O with a solid lead will give the left a chance to claim wide-spread GOP voter fraud (voter ID fraud).

  6. Jim Chase

    If polls are meant to be the barometer of the people, and the polls are screwy, it stands to reason that the answer to your query is simply: People are screwy.

    Now, if only a way to measure the psyche of poll watching …

  7. DocJay

    Since when have we had an election like this?  1860 maybe but this one is more serious.

  8. Redneck Desi

    In medical clinical trials it is called selection bias….any poll which has an equal or higher democratic weight to 2008 is absolutely ludicrous.

  9. Matt Blankenship

    Has anyone heard anything reliable on what the campaign internal polls are showing?  I know this is guarded info, but it seems like someone must know something.  I am desperate to feel a little better before Tuesday. 

  10. ConservativeWanderer
    Matt Blankenship: Has anyone heard anything reliable on what the campaign internal polls are showing?  I know this is guarded info, but it seems like someone must know something.  I am desperate to feel a little better before Tuesday.  · 0 minutes ago

    I think you can pick that up from the demeanor of the two campaigns.

    Who looks confident, and who looks desperate?

  11. Yeah...ok.

    So where is the podcast already? – Thank you.

  12. Dave

    It nothing more complicated than this: the polls are attempting to predict a future event and predicting the future is really hard, if not impossible.

    We invent all kinds of techniques to do so. Some are silly, such as psychics and crystal balls, some are much better, such as polling.  But as much as polling aspires to the status of a social science, it is still trying to document something that hasn’t happened yet.

    Every pollster will admit this; no one pretends they are doing anything other than attempting to approximate what will happen on election day.

    Another way to put this is: it’s screwy because each event is unique. Although there have been plenty of previous elections, this particular one has never happened. And it won’t happen until next Tuesday.

    Neither the past nor the present predicts the future.

  13. tabula rasa

    If you look at the RCP polls for today you’ll see CBS/NYT/Quinnipiac polls that show Obama ahead in Florida (+1), Virginia (+2), and Ohio (+5).

    Thank heaven for analysts like Battleground Watch, who dig into the numbers.  According to this article, the way these pollsters got to these numbers was to over-sample Democrats.  In Florida it’s Dems +7 (Dem 37, Rep 30), in Virginia it’s Dems +8 (Dem 35, Rep 27), and in Ohio it’s Dems +8 (Dem 37, Rep 29).

    With continual reports of a positive enthusiasm gap for the Republicans, the only way the pollsters can get Obama ahead is to oversample Democrats in egregious ways.

  14. ConservativeWanderer
    tabula rasa: With continual reports of a positive enthusiasm gap for the Republicans, the only way the pollsters can get Obama ahead is to oversample Democrats in egregious ways. · 1 minute ago

    Edited 0 minutes ago

    Tab, that’s been the pattern for months now.

    In order to get Obama a lead, they gotta oversample Democrats. If you remove the oversample, Obama’s lead goes away.

  15. Matt Blankenship
    ConservativeWanderer

    I think you can pick that up from the demeanor of the two campaigns.

    Who looks confident, and who looks desperate? · 22 minutes ago

    I hope you are right.  I just play around with that electoral map on RCP, and even when I give Romney every state he should/must win (VA, NC, FL, CO, IA, even NH, etc) I am still stuck at 267.  That means he’s got to have one of WI, OH, MI, PA, or NV.  And I just can’t wrap my mind around a Republican winning any of those except Ohio, and Ohio is way too close for comfort. 

    The whole thing is close, and it shouldn’t be–as others above have said.  That’s what’s so vexing.  Look, I couldn’t stand Clinton, Gore, Kerry,  or the Obama of 2008.  But I could understand how the casual apolitical centrist undecided moderate could vote for them.  This Obama, with this record…I just don’t understand it.  The 30% of the country that is hard left/liberal, I get.  They’re always going to vote for the Democrat.  But this should be a 57/43 election.

  16. David Williamson
    Peter Robinson: 

    Nobody knows. 

    Other than Doug Mataconis.

  17. Duane Oyen

    I sure hope the optimism about “extending the field” to some purple states is better founded than it is for Minnesota.  I’ll drink a cocktail on recorded video (I hate booze) or a cup of coffee (I hate coffee) if Romney wins Minnesota.  There is no chance at all in this vale of wackos, Strib poll notwithstanding.  My guess is they are just wanting to pick up a few conservative subscriptions and will try anything. 

    The electorate in this state is certifiably nuts.  Wellstone (still treated as a saint), Franken, Klobuchar, Mark Dayton, etc.

  18. BrentB67

    All this consternation over polls is wasted heartache by a lot of smart people, especially here at Ricochet.

    Let me be presumptive that the majority of folks reading this post support Gov. Romney. If that is the case, why do any of us care about the polls?

    If a poll shows Gov. Romney leading are we not going to make any more donations, pull up the yard sign, or not vote? – I think the answer to all is No.

    If a poll shows Gov. Romney behind are we going to double our donations, put a sign in our neighbor’s yard, offer to drive someone to the polls? – Maybe.

    If we support Gov. Romney we should be doing all the things within our power to effect the change irrespective of the polls. Log off CNN, Politico, RCP, etc. and pick up a phone, put out a sign, and for darn sure get off the couch and vote.

    Time to stop spectating and get in the game.

  19. ConservativeWanderer
    BrentB67: All this consternation over polls is wasted heartache by a lot of smart people, especially here at Ricochet.

    Let me be presumptive that the majority of folks reading this post support Gov. Romney. If that is the case, why do any of us care about the polls?

    If a poll shows Gov. Romney leading are we not going to make any more donations, pull up the yard sign, or not vote? – I think the answer to all is No.

    If a poll shows Gov. Romney behind are we going to double our donations, put a sign in our neighbor’s yard, offer to drive someone to the polls? – Maybe.

    If we support Gov. Romney we should be doing all the things within our power to effect the change irrespective of the polls. Log off CNN, Politico, RCP, etc. and pick up a phone, put out a sign, and for darn sure get off the couch and vote.

    Time to stop spectating and get in the game. · 11 minutes ago

    It’s the same reason people like to yell at the umpires at baseball games.

  20. Pig Man
    Schrodinger’s Cat: 

    Personnally, I will be watching VA, FL and NH on Tuesday night as indicators as to which set assumptions is correct. · 4 hours ago

    Nice analysis Cat.   If the national polls giving Romney +2 or more are correct, Obama is toast.  Why did you pick VA, FL and NH?  Because these are the earliest one to close?