Election links
November 3, 2012, 2:36 pm
Filed under: Uncategorized

PostLibertarian points out something I didn’t know, something very retarded, about how poll results are being reported. (Nate Silver uses correct poll numbers; another point in favor of his ‘model’!)

Michael Barone predicts a 315-EV Romney win. Why? “Fundamentals.” Which is swell and all, but howcome nobody can explain why these mysterious “fundamentals” aren’t really showing up in poll numbers? Do the “fundamentals” only kick in post-polling? I’m afraid actual votes trump “fundamentals” and polls, however flawed, are at least trying to be an estimate of the former. I don’t even know what “fundamentals” is trying to be.

Look, all I’m asking as a Romney, Inc.-preferrer is some quasi-believable fantasy scenario to hang my hopes on. Something rooted in actual observable fact. If it’s, ‘all the polls are systematically overcounting (D)s’, I’m prepared to believe that! But I need a workable theory as to how on earth that is happening. Not even necessarily a great theory (I believe in all sorts of wacky theories!), but at least something. Is that really too much to ask?

It’s like, there’s a reason that porn scenarios (I’VE HEARD) involve guys who are, like, cable technicians and pizza delivery guys, instead of attaches to the UN undersecretary or world-renowned neurosurgeons. You want to be able to believe in the fantasy without straining yourself too hard. “Yeah, I could totally be that guy.” But Barone yammering about some mysterious ‘fundamentals’ that (far as I can tell) exist mostly in his head, that’s giving me nothin’. Bad (R) porn. And really, what’s worse than bad porn?

Looks like Radley Balko’s prediction is the same as mine, Obama with 290. But, he loses me with his ‘advice’ to (R)s, that what they need to do is to “Stop hating on immigrants.” The oft-repeated notion that (R)s don’t get support of immigrants (and, of racial groups like blacks) simply because they ‘hate on’ or don’t ‘reach out to’ those groups strikes me as tremendously patronizing and insulting – to those groups. It can’t actually be that those groups have actual, deep-seated, sincere pro-lefty ideas that they see as being in genuine, honest conflict with those of the (R)s? They only vote for (D)s because the (R)s talk mean to them and don’t ask for their vote? How did the great Bush/Rove-era push to embrace the perennially ‘culturally conservative’ Hispanics and lastingly convert them to (R)s, work out in 2008 I wonder? Why, let’s ask President McCain! This, like nominating McCain in the first place, is the sort of concern-trolling ‘advice for (R)s’ I have come to expect from people who don’t actually want (R)s to win. The sad thing is how often the (R)s follow such advice.

RELATED RE: IMMIGRATION: Jeb Bush thinks Texas will ‘go blue’ by 2016 due to demographics. Oh yes, the (R)s being nicey-nice to immigrants is what will save them….

20 Comments so far
Leave a comment

Historically, no incumbent has taken less vote share than their first pass and won.

Comment by A Lady

I assume you’ve seen this?

Comment by Sonic Charmer

Yes, i have, but how is my comment any different than ‘historically, state polls have been more predictive of who will win the Presidency’ in that light?

Comment by A Lady

1/2 of a fair point. Yes, people do point to state polls’ historical record of (relative) accuracy as a point in their favor. But, independent of all that, state polls are statistical attempts to estimate voters’ intent. They are either unbiased or have some bias.

If they have some bias, that would be very interesting, and I would love an analysis of the mechanism for it….

Comment by Sonic Charmer

well, one argument is simply that many of the pollsters weren’t around ‘historically’ (which is true, there are a ton of new polling firms this cycle, some of whom weren’t around in 08 either) and that this means using a historical argument to privilege state-level polling over national polling is not really kosher.

another argument is that this cycle the people paying for the polls have incentive to preserve a horserace narrative until the last minute and thus the D+N shifts back and forth without reflecting the on-the-ground state of affairs accurately. this one of course implies that on-the-ground is not a neck and neck horserace, though.

and still another argument is Lying Democrats. they don’t just cheat, they LIE ON POLLS MAN.

and the most common argument i think i’ve seen is that indies don’t favor the incumbent in most of the polls (this is in fact the case), state or national and it’s pretty uphill to win if they don’t.

Comment by A Lady

re: new firms, well for the record, my main reason for privileging state polls is not historical, it is because states are what matter when it comes to the electoral college.

if all polling firms en masse are fudging numbers to ‘preserving the horserace’ then I reckon it’s about time for them to flip the switch back to accuracy, don’t you? time’s running out, election’s on Tuesday.

I’m not sure there’s comfort in a poll showing indies don’t favor the incumbent, that also shows the incumbent winning. Do I believe the poll? Then Obama’s ahead. Do I disbelieve the poll? Then I lose my evidence that indies don’t favor the incumbent..

From where I’m sitting, Lying Democrat Pollees is just about the best theory anyone’s come forth with. Which, is not encouraging.

Comment by Sonic Charmer

well, all the arguments require at least a splash of lying democrats to work out.

Comment by A Lady

Assume that polls measure “likely voters”. Assume that this is ascertained by the pollster by receiving an unqualified “yes” to the question: Have you already voted, will you vote next Tuesday, or will you send an absentee ballot by the deadline? Is there are D-R bias in the reliability in the answer to that question? It seems likely, only insofar as there is a D-R bias in general reliability. Do pollsters automatically factor in this reliability factor? I dunno. It would be terribly impolite of them to ever say so.

Comment by Nick B Steves

Understood, so there could be a bias in stated intent: perhaps Ds are more likely to be lying when they say Yes I Will Vote.

If this were the case, then some actual tangible evidence (however flimsy) for it would be what would help the fantasy.

And P.S. you also need an explanation for why the dynamic didn’t show up in previous years’ polls.

Comment by Sonic Charmer

the early voting numbers. state level polls for some reason draw samples that have higher early voting reportage than in previous cycles and higher than national polling of early voters. the national results are pretty close to the actual reported ballots received, while the state results are not. there could be any number of reasons for the divergence, but it’s not typical and appears to be current to this cycle.

Comment by A Lady

Just so I understand, you’re saying that state polls are oversampling voters who are likely to vote early? and that they disproportionately lean (D)?

Because SkepticalCynical just kinda convinced me in the ‘Enthusiasm’ thread that (D) leaners’ lack of enthusiasm meant they were *less* likely to be early voters this cycle…

Comment by Sonic Charmer

but that’s just it, they aren’t likely early voters, they are claiming to have already voted. either the polls are somehow capturing a lot of D early voters and not so many R early voters or well, lying democrats again. it’s pretty easy to say you’ve already voted whether you did or not.

it’s a simple explanation. the only enthusiasm D leaners might have left is sitting around in their own homes, cheerfully saying or pressing a button to indicate they voted when that isn’t the case and nobody will ever really know they didn’t vote.

what many people are using as signs of enthusiasm is also stuff like crowd turnouts, which have been poor for obama but pretty solid for romney.

i mean, sure, maybe nobody lies to a pollster ever and for some weird reason they end up with 30-40% early voters and also the same people who won’t show up as they did in 2008 crowds will all show up plus some extra on election day and obama wins. it’s certainly possible.

but at this point lying democrats looks more occam friendly is all.

Comment by A Lady

Ah ok I misunderstood, you’re saying that state polls (and only state polls?) are getting responses claiming an early-voter rate that doesn’t line up with the actual early voters we see? That would be interesting.

I guess I do have to hang my hat on the lying-(D)s theory. Anyway, mericfully I’ll find out soon enough, hopefully before I go insane :)

Comment by Sonic Charmer

It is mostly state polls, reporting 30-40% early voting rates (and they report it like that ‘30% of voters in $state have already voted’) when the states that have records of received ballots have rates that are 10-15% of registered voters (and 15-20% of actual voters, which are fewer.)

Comment by A Lady

Sonic Charmer is, I believe, a mathematician, but he is having trouble understanding the idea that the polls are weighted +7 for Dems, when they are unlikely to enjoy a +7 advantage this time because of the fact, not the supposition, that there is less enthusiasm for Obama now, than four years ago.

Comment by Pastorius

They aren’t ‘weighted +7 for Dems’. They are constructed via some attempt to sample demographics, and the sample they end up with *comes out* +7 for Dems (or something along those lines).

One interpretation of this is that it’s actually evidence that the country really just is D+7 at this time. I don’t know otherwise, not really. You say it’s ‘unlikely’ but why?

‘Less enthusiasm for Obama than 2008′ sounds somewhat plausible, but as we discussed, there are demographic changes that counterbalance it. More immigrants, young people who have become first-time voters – lean Obama. Old white people who have died off – leaned McCain….

Comment by Sonic Charmer

“Fundamentals” in this context means “the economy sucks”. I realize that Smart People only understand elections by means of stochastic modeling of statistical margins of error in polling instruments, which are themselves an ideal and incorruptible beacon illuminating the America political soul.

But there’s an obscure branch of academia called Political Science where they study the voluminous records of 200+ years of voting in the United States and attempt to generalize from it. And one of those generalizations is that economic performance explains a large fraction of election outcomes. I know, absurd! But Barone totally is one of those guys.

Comment by SkepticalCynical

If (a) the economy sucks and (b) this makes people want to not re-elect Obama then (c) why are they telling pollsters, at this time, that (in effect) they plan to re-elect Obama?

I understand that people like to look at these sort of ‘fundamentals’ and whatnot, but to do so *to the exclusion of what actual poll data is telling us* doesn’t make a whole lot of sense to me.

Comment by Sonic Charmer

If this is suspicious, the only explanation I can think of is my (D)s-answer-pollsters’-phone-calls theory I made up a while ago.

It would be nice not to have to make up theories…

Comment by Sonic Charmer

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Get every new post delivered to your Inbox.

Join 501 other followers

%d bloggers like this: