Mostly I have been saying I don’t want to say anything more about the election, but a good, old friend I haven’t seen in a long while emailed to ask my last thoughts on the election before the canvass. So I wrote the below:
I cringe to think I’m thought to know anything much. But here’s what I see.
EARLY VOTING: It appears that more than half the voters in some critical states—NM and CO?—have already voted—and O has been polling well ahead there. The election is already under way; we can no longer talk about it in the future tense—and this is where Obama’s having built up a lead early may—should—pay off. Many people—and more than half in some swing states—now vote early.
DO WE TRUST POLLS? You say you don’t trust polls: I can say I don’t trust any given poll, but the deadweight of the polls has all leaned in one direction. Cumulatively, there’s a lot of data there.
BRADLEY EFFECT: Is there a Bradley effect? My hunch has always been “yes, there is.” But all the political scientists say no. This fellow Charles Franklin actually designed a poll question intended to ferret one out and he derived perhaps one percentage point for McCain out of the data.
THE PATH TO VICTORY: If you play the electoral college game, it’s very hard to see McCain’s route to victory. As you know he’s been making a big play for Pennsylvania, and I bet he does have some unmined potential there. But even that might not be enough—suppose he gets PA; he needs not to lose some very large number of Nevada, New Mexico, Colorado, Montana, North Dakota, Indiana, North Carolina and Virginia, all of which are wobbly or lean Obama.
MY BIG CAVEAT: Ground game. Obama’s got the big ground game for registration and turnout. But there’s one last move McCain can make: vote suppression. Are there interestingly fewer voting booths in some districts than others? Are there more stringent ID checks in some places or for some kinds of people than others? What’s happening with all these early-cast ballots? Are they being safeguarded?
The thing is, for that really to work, it takes a devoted—and I mean seriously, willing-to-go-to-prison devoted cooperative set of state and local parties. Early signs are, they’re not *all* that devoted. Some may be—Ohio? wouldn’t care to make the accusation—but Florida under Crist appears as though it is not.
But that’s the last ditch McCain and the GOP could die in, if they care to.
As I say, I don’t think I’m adding too much value there, but that’s what I have to say.
10 comments
November 2, 2008 at 12:41 pm
Ben Alpers
Not to beat the deadest of dead horses, but why do you have a hunch that there’s a Bradley effect?
As you note, political scientists suggest that there’s no data at all to suggest a Bradley effect in any election in the last fifteen years.
And the data that suggests that there was a Bradley effect in the 1980s and early 1990s is itself highly contested. Indeed, people involved in polling the 1982 California gubernatorial race that gave the effect its name deny that the Bradley effect explains Tom Bradley’s loss to George Deukmejian. WNYC’s On the Media had an excellent segment about this last week.
The executive summary: much more mundane polling mistakes, most notably underestimating turnout in the Central Valley and undercounting absentee voters, probably explain the polls that had Bradley winning on the eve of the election. The Bradley effect was more or less invented as an explanation by Mervin Field, the most prominent California pollster, on the basis of no evidence whatsoever. It had the virtue of blaming lying voters rather than polling mistakes for his own failure to accurately poll the Bradley-Deukmejian race.
And, in fact, the Bradley effect makes precious little psychological sense. Why should people lie to pollsters about not voting for a Black candidate in a close race in which there are plenty of socially acceptable reasons to vote for his opponent? And why should race and race alone, of all possible socially unacceptable reasons for not voting for a candidate, produce such a measurable and regular spate of voter lying, an effect otherwise unseen in polling?
There’s no question that many voters are racists and that many of them will vote for McCain because they are racists. But why should we expect these people to lie to pollsters in any greater numbers than anyone else? It makes no sense…and it never has.
November 2, 2008 at 1:29 pm
eric
Personal experience. I know, I posted that anecdote is not the singular of data. But of such are hunches born.
November 2, 2008 at 1:35 pm
Ben Alpers
Care to elaborate on what sort of personal experience you’re talking about?
November 2, 2008 at 1:38 pm
eric
I don’t want to impugn people I’ve spoken to. But haven’t you got the impression that a large number of people with conservative feelings of one sort or another feel a put upon by p.c., they have the idea that they’re not allowed to express their biases, but will cheerfully do so in the voting booth?
November 2, 2008 at 1:46 pm
bitchphd
Hm. But don’t most people who feel that way actually express their biases quite freely, as if they’re striking a blow for freedom?
November 2, 2008 at 2:34 pm
wren
In Ohio’s defense, they have a Democratic governor and Sec of State this time around. I doubt many county Republican parties are going to risk trouble for the off-chance the self-described air pirate and black criminal John McCain can pull out wins in VA, FL, CO, etc. Forktime.
November 2, 2008 at 3:10 pm
Ben Alpers
don’t want to impugn people I’ve spoken to. But haven’t you got the impression that a large number of people with conservative feelings of one sort or another feel a put upon by p.c., they have the idea that they’re not allowed to express their biases, but will cheerfully do so in the voting booth?
Absolutely, eric. And this sort of thing takes place in other kinds of settings, too. Back in 2004, I would have cast my (entirely symbolic) Oklahoma presidential vote for David Cobb had he been on the ballot here. But he wasn’t and Oklahoma doesn’t allow write-ins. The proper Green thing to do would have been leaving my ballot blank. But I voted for Kerry because I felt very strongly about casting a ballot against Bush. However, I would not have been forthcoming about this choice in my Green circles.
However, for the Bradley effect to be a factor, you need to assume two things:
1) That people treat pollsters the same way they treat their (putatively) “pc” friends and associates.
2) That, so far as pollsters are concerned, this effect somehow uniquely takes place in cases involving Black candidates. That is, though certain conservatives lie to their “pc” friends about liberal candidates in general (at least that’s my anecdotal experience), they lie to pollsters only about Black candidates.
I think both these assumptions are problematic.
November 2, 2008 at 4:16 pm
ignobility
I’m inclined to think we might see the opposite of the Bradley effect, at least in my part of the country (rural Appalachia). I think many a redneck would be afraid to admit to his confederate flag-waving fellows that he is voting for the black guy.
November 2, 2008 at 4:25 pm
eric
I’ll be happy to see my hunch not borne out by the evidence.
November 2, 2008 at 9:18 pm
Urk
I tend to think, with just as little hard data to back it up, that ignobility is on to something here, as some folks, rednecks and/or cultural republicans (those folks who aren’t evangelical and really don’t make enough money to vote republican but do so anyway) will, safely out of sight of their Limbaugh quoting &/or stars’n'bars waving buddies, pull the lever for Obama. Mostly because they figure he’s going to win anyway, so they may as well get on the right side of history and make a vote they can tell their grandkids about. I base this entirely on close observations of my kith and kin over 4 decades or so in the Arkansas Ozarks.