Again, Scott Rasmussen apparently thinks that party ID changes only infinitesimally during the course of an election season. This is false. And with the Dems currently having something like an 8.4 percentage point edge in party ID, it’ll hardly be a disaster if the GOP gets a narrow plurality among independents.
Analysis of this issue by a libertarian leaning blog whose readership tends to be liberal (no, I have no idea how this happened either).
Would it surprise anyone that the segment of the Republican party I tend to call “honest conservatives” swelled the ranks of “independent” voters by a vote of disgust? Without actually changing their political opinions in any meaningful way? Wouldn’t that mean that, demographically, the “independent” voters were significantly more conservative in outlook than they had been before?
That’s a nicely written bit. Quite clear.
There’s an article currently on HuffPo that reports polling indicating Romney is scoring lower favorability ratings than Shrub.Credit to one of the commenters, who notes: These polls were skewed by sane people.
He’s up to 2% with Black Protestants now!
Black Protestants? What if they riot?
Elucidator has pointed out the relevance of the 2004 election results in Ohio to this thread. I find it very instructive. In Ohio, there was a HUGE (double digit) discrepancy between what the exit polls showed the voting going as, and what the vote actually went as. (In fact, there were exit poll discrepancies in three battleground states, all showing more Democratic votes than were actually cast.Here’s a link to pollster Steven Freeman’s analysis of the discrepancies.
In both the 2004 vote and the current polling issues, the reaction among Republicans and Democrats were the same: in both cases, the Democrats maintained that the pollling was accurate, indicating vote fraud in the 2004 election, and Romney’s doom in the 2012 election. Meanwhile, Repubicans maintain that Democrats were oversampled in both the 2004 election and in the 2012 election, negating vote fraud in the 2004 election and maintaining Romney’s hopes in 2012.
I think the fact that this is a recurring theme is very interesting. Is there in fact some hidden factor that makes pollsters oversample Democrats? Or did Republicans steal the 2004 election and are they ignoring the evidence of doom for Romney in 2012? Exit polling date in this election may be very illuminating, I hope the Democrats are watching the polls carefully, especially in Ohio.
Eventually we will probably just get to the point where election officials start openly “unskewing” the votes themselves.
“Oh hey, we had 53% of the vote go to democrats this year? Well that doesn’t match up with our registration database! Obviously these votes were miscounted. Let’s just adjust that down to the 42% we know are registered truly as Democrats!”
Well, some Dems maintained that the exit polls were accurate, and that there was vote fraud. Unlike with 2000, most Dems threw in the towel pretty quickly on claims that fraud reversed the 2004 election, to the extent that they gave the idea even brief credence in the first place.
Also, we were talking about exit polls in 2004, as opposed to your more standard polls. They’re different animals.
And even as far as that goes, I certainly missed the moment right after the Ohio polls closed, when the nets called Ohio for Kerry on the basis of the exits. My recollection was that they called Ohio for Bush not immediately but still pretty quickly, ending what little suspense there was in that depressing evening.
Now there were claims in 2004 by Democrats that Republicans were overrepresented by a few points in a number of garden-variety pre-election polls. Been there, done that, got the ratty old t-shirt in the bottom of a drawer somewhere. It didn’t turn out to be true then, which is one reason why I’m pretty adamant that it isn’t true now.
I’d still like to know - ASSUME for the sake of discussion that there was no conspiracy of Fox with CNN etc but that there was some OTHER factor that was resulting in Democratic “oversampling” - what would that factor be? It is NOT that the pollster adjust the turnout numbers, that hypothesis is already widely debunked. What is the hypothesis?
(We know why Rasmussen numbers show what they show - they assume the D/R/I a priori and adjust to fit.)
I would just like to say that as someone who knows nothing about stats, I’ve been following this thread and finding it very informative and helpful. Thank you to all who have participated.
Honestly, at this point the hypothesis seems to be: Democrats are horrible, so anything that shows people like them must be a lie.
So I flip past Fox News in between football games today, and what do I see? A breathless reporter talking about how the drop in Democratic voter registration in the swing states is SO MUCH MORE than the drop in Republican voter registrations. And that with all these volatile, unreliable independent voters out there, pollsters have to be very careful about predicting how they might vote.
Of course, I think pollsters get around that by, ya know, ASKING those voters who they might vote for. But I guess that’s all ouija-board voodoo made-up mumbo jumbo.
It’s starting to get really funny, this Republican obsession with the poll numbers. You don’t like 'em? Have Romney do something to change them. Don’t just stick your fingers in your ears,say “neener neener, your numbers are wrong” and claim bias. It’s laughable, I tells ya.
I read the entire thread and found it very informative- you have a great team here. Painful to watch adaher though; who is obviously quite passionate about his guy Mitt and having a hard time with the cold hard numbers. Hope he decides to take a break from the soft stuff that he is citing (e.g. Breitbrt, pundits) as there is no way to reconcile that emotionally charged rhetoric with polls data; he will twist himself inside out over it.
Welcome to the SDMB, Catballou!
Yes, there’s a lot of knowledgeable people who post here - hope you stick around!
It basically comes down to this- when pollsters call X number of people, a certain percentage of them self-identify as Democrats (and Republicans and independents). The ratio of D to R in the raw data for most of the recent polls is higher than the Republican supporters are expecting- so they suspect something is wrong. It doesn’t have anything to do with weighting- it’s what the respondents are saying. The data suggests that many people who recently may have identified as R now identify as I, and perhaps that some who used to identify as I now identify as D. That’s it.
Problem is, no one is expecting it. Republicans are just the only ones willing to point it out.
Again, if Democrats really think they’ll exceed 2008, that’s a big story that should be reported. No one is reporting it because no one actually believes it.
Here’s what I think should be reported: “Obama is leading in x polls by y amount.” That’s the story. I have no idea why you think something else, which only matters because of the impact on the number of votes, would be a bigger story than that.
That is what’s being reported. The problem is, no one is willing to stand by the internals, or even report the internals, because no one actually believes the internals.