Following up on this, it appears that the racial demographics of voter turnout are relatively stable.
If you modeled the turnout on the 2004 figures and ignored the fact that the population is less white than in 2004, the results still wouldn’t change nearly enough to cancel out the Obama bounce being found.
Today’s polls have put Obama’s chances into uncharted territory. 538 has Obama’s chances at 82%, a new high. The “Now-Cast” i.e. what would happen if the election were held today, has Obama’s chances at 98%. He also has Obama’s share of the popular vote at 52.1%.
A bit of sanity from a big-league conservative blogger (though he still uses the weighting and demographic arguments to theorize that Obama’s lead isn’t as big as it appears).
Earlier **bengangmo **asked a question about the new voter ID laws and how that might skew the polling. He kind of got blown off, but I think it’s worth considering here.
The voter ID laws will certainly disenfranchise certain people, and it’s generally assumed that those people will fall into certain demographic groups that typically vote Democratic. If that’s true, then the definition of “likely voter” will need to be refined. A likely voter will have to meet whatever criteria were set in the past plus have a valid photo ID.
I’ve not studied this as much as many of you have, but as far as I know, none of the polling firms are taking this into account. I suppose if *none *of them are, then it wouldn’t explain the “oversampling” issue. However if some of the pollsters who are showing better numbers for Romney are tossing out some percentage of their sample who fall into the “targeted” demographic groups, then I would think that that would push the numbers towards Romney.
This should be pretty simple to resolve. The media should just say straight out that they expect Democratic turnout to be even greater than 2008. A lot greater, actually. Because that’s what their polls are showing. Why are they reporting the Obama lead, but not this tremendous news that Democratic turnout is going to set new records? You’d think that would be newsworthy.
Or, are they trying to give the false impression that independents are swinging Obama’s way?
Fivethirtyeight.com isn’t a polling firm, but it’s forecasts do take the voter suppression efforts into account.
I check in with 538 regularly, but I don’t recall seeing this covered. Did Nate address this in a particular post or is it just scattered throughout?
Could you give a quick summary of your argument that that’s what ‘their’ polls are showing? (And clarify who ‘they’ are.) I’m missing that here.
Not that even enormously increased turnout is always a big story. John Kerry got enough votes in 2004 to win the 2000 election in a landslide. I don’t remember hearing much about it.
It was that whole exit polling mess, where the liberal media just couldn’t grasp that so many people voted for Bush in Ohio but refused to admit it. Hell, if I had voted for Bush, I’d lie about it.
The polls are showing more Democratic turnout than in 2008. That’s a really big deal given that 2008 was a huge year based on a once in a generation candidacy.
Here’s the CBS/NY Times internals that the media made a big deal out of today. Obama leads big in major swing states, but the real story is that Democrats are set to turn out in their largest numbers ever:
See the crosstabs of the PA poll OMG linked to, specifically question 47d. Of those claiming Democratic voter ID 99% have photo ID that (they believe) would suffice. Including 100% of Black voters surveyed.
A bigger deal for the intent and marketing for both sides than for the actual effect methinks. Nonzero but not huge. Nates’ take concurrs. (Yes, he did cover it.)
The polls are showing more Democratic voter ID among all registered voters. If turnout is the same percent of all those voters and the GOP lean to more turnout is no greater than usual then of course the Democratic turnout among likely voters is going to be higher as well. Some of the move in the LV polls may indeed be that the greater than typical GOP turnout that the firms had been factoring in has been dialed back some to more historical norms as Democratic enthusiasm has increased and GOP enthusiasm has decreased some (conventions may have actually mattered this time) … but the spread was already in the RV numbers (other than Rasmussen’s and Gallup’s) before the turnout (likely voter) adjustments were applied.
The likely voter models are showing more Democrats voting than in 2008. If that’s true, it’s a phenomenon that should be reported on. I can only believe that it’s not being reported that way because no one actually believes it.
Do you believe that Democrats will turn out in greater numbers than 2008? Is the country more Democratic in 2012 than it was in 2008? I have yet to see a party ID poll that shows that.
Gosh. Even Gallup’s daily tracking is now giving Obama a 6 point lead, 50 to 44. That’s as big as the post convention bounce peak.
Rasmussen alone is putting it tied for popular vote and even they are stating that Obama has a narrow lead “in each of the Key Three States in this year’s election, Ohio, Florida and Virginia.”
Interesting bit about Rasmussen’s adjustment of likely voter status based on voter party ID (Per freepers no less):
So no matter what their sampling shows they “correct it” to match what they think it should be based on their believe that it does not change much. Rasmussen’s numbers show no significant change in voter party ID? Well no shit! They correct those pesky shifts out of their data!
And in response to your question adaher, I believe that significantly fewer of those who are going to be voting (who are mostly the same people who voted last time) are willing to call themselves Republican at this particular point in time, more are willing to call themselves Independent, and maybe a smidgen more willing to label themselves Democrats, the result being a significantly higher relative percent of total voters self-identifying as D compared to R … more as a result of the loss from the R side than from gain or greater turnout on the D side.
As for the possibility that Democratic turnout may be unusually high, especially in swing states, Gallup does address that, and does find that Democratic enthusiasm is surging. Since June up 20 points to 73% as very or extremely enthused, compared to GOP up about 9 to 64%. Yes that portends for record Democratic turnout. Gallup nicely understates it:
But is that higher than 2008? If it is, it’s a pretty big story. It’s THE story of the campaign. The media is reporting as if Romney is losing the race because he’s not appealing to independents, yet Romney is still doing quite fine among independent voters.
If Romney is losing because of a surge of enthusiasm among Obama supporters, that’s the story, not Romney’s missteps, which should be costing him among independents.
Adaher, the total population of the USA per the 2000 census was 285,230,516.
By the 2010 census the total population was 308,745,538, for a total increase over 10 years of 23,515,022.
That’s an average increase of 2.35 million per year.
Why should you be surprised if voter totals are higher than 4 years ago?
In early October 2008 Gallup had voter enthusiasm 71 D to 51 R, a 20 point spread, more than the current numbers. But the question was phrased differently: “Compared to previous elections, are you more enthusiastic than usual, or less enthusiastic than normal” instead of the current “How enthusiastic would you say you are …” - so it’s not quite an apples apple comparison. (Why they changed it I don’t know.) Still you have a pretty high Dem enthusiasm number, especially in the key swing states, and more than usual who are willing to label themselves Democrats compared to those willing to say they are Republicans (at this moment). Do that math. A high percent times a high percent equals what?
Higher as a proportion.
Anyway, Hugh Hewitt asked the same question I did :
HH: Do you expect Democrats, this is a different question, do you, Peter Brown, expect Democrats to have a nine point registration advantage when the polls close on November 6th in Florida?
PB: Well, first, you don’t mean registration.
HH: I mean, yeah, turnout.
PB: Do I think…I think it is probably unlikely.
That’s the guy who did the poll! Something is going on when pollsters won’t stand by what their polls are saying when asked.
We may get some better sense of probable turnout over the next few weeks as early voting starts … high numbers there should portend for high turnout overall.
A good pollster recognizes that their poll is just one data point with its own possible blips, that their house is just one house with its own unavoidable house effect. 9 points is, at this point, an outlier number, and he knows it, as he should. Others are coming out more like Obama by 3. 538’s model weighting the polls for house effects (and so on) gives it an Obama 2 point lead in FL and a 67% probability of a win. Gallup perhaps should learn to have the same lack of hubris.