More Early Voting
Every now and then, I write a long post to explain something complicated, and end up tying myself in knots. And then, just after it's published, I notice a telling example that neatly explains everything.
This afternoon, Nate Silver posted the early voting numbers from a seventh national poll. He'd taken to task a number of pollsters for using suspect likely voter models; Berwood Yost of Franklin & Marshall took the trouble to reply to his criticism:
...[i]n 2008 we are all concerned about how all of the new voters, who will likely be disproportionately young and minority voters favoring Senator Obama, will affect those models....The piece of data we looked at closely in our survey to corroborate our model was a question about whether the respondent had already voted. Seven percent of our sample reported they had already cast a ballot and Obama's lead over McCain was 46 percent to 41 percent among these early voters--five points. Those who reported voting early in our sample were more likely to be over 55 years of age (53%), male (53%), and non-Hispanic whites (75%). The same proportion of Democrats and Republicans (40% each) reported early voting in our sample.
Kudos to Nate for raising the issue, and to Yost for offering helpful transparency. In response, Nate wrote:
See, this is all I'm asking for -- evidence that the pollster has thought about what they are doing, rather than simply acting reflexively. We might disagree about certain things (in this case, the predictive value of a very small sample of early voters), but that's all in the spirit of a good discussion.
Sounds good, right? Yost was worried that the composition of the electorate might be shifting, so he took a look at those who had actually voted. And he found that they're white and old, and that they support Obama by precisely the same margin his likely voter screen had produced. Meet the new electorate; same as the old electorate.
Well, not exactly. The "very small sample," in this case, consisted of 89 voters. That's not a typo. Yost heard from 41 people who voted from Obama and 36 or 37 who voted for McCain (actually, it's not mathematically possible that McCain garnered 41% support, but let's leave that aside), and on the basis of that "five point" difference, validated his likely voter screen.
But wait, it gets better. The survey was taken from October 13-19, before most states even opened the polls for early voting. That's probably why just 7% of the sample reported voting early, lower than the 10-11% being reported in other surveys a few days later. Of the 89 "early voters," 66 voted by absentee ballot, and just 23 voted in person. We have, as I wrote in my last post, excellent data to suggest that absentee voting remains a bastion of GOP strength. Absentee voters are exactly what Yost describes - old, white, and relatively Republican. That, presumably, is why he found an even 40-40% partisan split. So even if his results were accurate, despite the tiny sample size, they were also irrelevant. He wasn't looking at early voting, but at absentee voting. No one expected to see an enormous increase in absentee balloting among youth and minority voters. And sure enough, he didn't find it.
Let me repeat the gist of my earlier post. We now have very good data on who's showing up to vote early in a number of closely contested states, and state-level polling data on that question in a number of others. In a word: Democrats. And by large margins - larger, in fact, than Obama's overall lead in any reputable poll. So polls that predicted that early voters would resemble the general electorate have a problem. And pollsters like Yost who used one set of polling data to validate another are suddenly confronted with the circularity of their logic.
Advertisement
Thanks Fly. I wonder why Nate Silver didn't flesh this out.
October 26, 2008 10:38 PM | Reply | Permalink
Great analysis.
What we don't yet know is whether Obama is expanding the universe of Democratic voters, or just shifting Democrats' votes forward in time.
But fortunately, it's not a make-or-break question for us this time around.
October 27, 2008 12:08 AM | Reply | Permalink
Excellent point. Given the possibility of congestion at the polls on election day it's still a smart strategy to get your votes in the bank early.
It's totally unclear to me if the early voters are those more or less familiar with the whole process of voting and are therefore less or more vulnerable to election day shenanigans. So are the early voters the new ones or the wizened veterans?
October 27, 2008 2:23 PM | Reply | Permalink
Thank you, fly, for fulfilling my wish for daily fotw analysis. Please keep up the brilliant work.
October 27, 2008 12:50 PM | Reply | Permalink
And when do I get more political satire in return?
October 27, 2008 12:53 PM | Reply | Permalink
I've been handing out satire like a Jehovah's Witness on speed all year long. Where have you been, mister?
Just hang on, I've got one in the pipe.
October 27, 2008 12:58 PM | Reply | Permalink
Psssst.... Wouldn't this be an appropriate moment to direct Mr. Fly to your new and exciting blog-home?
October 27, 2008 1:23 PM | Reply | Permalink
Thanks, Paige. I've been informed that it violates TPM policy to flog the blog in the comment threads, which is why I didn't mention it. I think that I should ask you to refrain as well, though of course I very much appreciate the thought and would humbly welcome a visit from said fly.
October 27, 2008 1:54 PM | Reply | Permalink
Ah. I see. Sorta. Well, glad to have been naively of service.
October 27, 2008 2:28 PM | Reply | Permalink
I don't think that's true - there is aright way to whore a blog post and a wrong way. The wrong way is to not ever be part of the conversation, but just show up, post an entire blog post with a link and leave again.
The right way is to say outfront you're whoring a post, and give the link and don't post the whole damn thing. Nobody really minds blogwhoring by regulars, Genghis. That's not the same thing as spamming.
October 27, 2008 3:29 PM | Reply | Permalink
"Regular poster" kinda brings to mind that old gem, "There's no whore like an old...." Ok, I better stop.
But it could lead to one HECK of a post. Let people vote - "Genghis: More Jehovah Witness on Speed... or Old Whore?" FlyOnTheWall could handle the polling analysis.
Betcha the campaign would be bitter. "Granted, he's a whore. BUT. Define 'old.' " ;-)
October 27, 2008 3:44 PM | Reply | Permalink
It's kind of gross, the way Genghis keeps flogging his blog in plain sight.
Tena, thanks for the tips. Genghis has been pimping out the blog in kind of a weird way, and he needed to be reined in on the blogwhoring.
October 27, 2008 5:57 PM | Reply | Permalink
Blogwhoring. Good one, Tena
October 27, 2008 6:16 PM | Reply | Permalink
Tena, I'll withhold my opinion on what constitutes spam, as I have a conflict of interest here, but the limitation is TPM policy, not a question of what other members think.
PS Ignore Articleman; he's just jealous and wants to promote his own blog.
PPS Also ignore Quinn; he's not a citizen of this country.
PPPS FotW, I'm sorry that we hijacked your thread.
October 27, 2008 6:23 PM | Reply | Permalink
No you're not. ;-)
October 27, 2008 6:28 PM | Reply | Permalink
Actually, I am. I shouldn't have replied publicly to Paige, as the consequences for the thread should have been obvious to me.
October 27, 2008 7:28 PM | Reply | Permalink
Oh, I'm just funnin' with you, Genghis... no worries from my end, hope there are none on yours.
October 27, 2008 7:37 PM | Reply | Permalink
What country? Manhattan?!?
October 27, 2008 6:38 PM | Reply | Permalink
He meant you're from the northern suburbs.
October 27, 2008 10:04 PM | Reply | Permalink