For a lot of this campaign, I've been worried - sort of closet-worried, actually, because I feel weird talking about it - about the Bradley effect - basically the idea that lots of white people are closet-racists and, when they get the polling booth, will flash back to 9/11, see how much the name on their ballot resembles "Osama," and vote McCain. Or just decide they can't handle a black president.
But this NYTimes article - which, fair warning, rambles a bit - tells me the issue of race is both more complex and less worrisome than I thought. It's a good read.
P.S. Now that the Cubs have choked, I'm even more neurotically worried about Obama, despite the polls. I really thought that this was the year for both of them.
Subscribe to:
Post Comments (Atom)
Posts
I like this question. I like it a lot. And I have comfort to offer.
ReplyDeleteFor one thing, the Bradley effect and Obama's turnout are both hard to guess and work in opposite directions. So starting off, you've already got 50-50 when it comes to things we don't know about.
You should check out 538's dicussions of the Bradley effect. He leans left but I like most of what he says.
One key idea is that even if there is a Bradley effect it's not limited to white people. In the primaries, many black people told pollsters they were undecided, but when it came time to cast their ballots, did so disproportionately for Obama. Separately, I'm not convinced that the Bradley Effect is more than an isolated incident, or that its relevance hasn't decreased with time. Finally, I'll say that Nate's model actually uses primary data to guess how undecideds will vote in the general election based on how they did in the primaries. So he assumes that states like Ohio will show Obama underperform, and states like North Carolina will show him overperform.
Finally I'll say this: those who would be "subject to" the Bradley effect are largely people who would be voting Democratic this year for sure if it wasn't a black guy. And largely, undecideds' main issue is whether they can trust Obama. If I had to point to one reason Obama has been surging recently, it is that those voters are just now tuning in. They've seen Obama during the financial crisis. They saw him in the debate. And he's acted in a trustworthy manner. Those people who know they don't like what's going on now are becoming OK with the idea of Obama as president (all else aside, I thought he looked damn presidential in that debate) and we are now seeing the resolution of that discomforting question: why is this election close? Undecideds know they don't like the status quo but weren't sure they were comfortable with Obama. That's changing.
The cherry on top: I can tell you first-hand that Obama's ground game is the stuff of legend, and will be what history takes away from his victory in November.
Let me begin by stating my hope that rambling is not too off-putting for those of you reading this blog. I feel as if my posts and comments tend toward verbosity and confusion rather than brevity and coherence, and to those offended I apologize.
ReplyDeleteThat said, let’s ramble about racism! I would like to (facetiously) dub all unstated racism in non-white communities The Kanye West Effect, in honor of a man who has dedicated much time and energy to racializing every aspect of his life, from his appearance to his occupation to his statements (e.g. “George Bush hates black people” at the MTV awards).
On to the real topic. From everything I have read, most of it required for The Politics of Minority Representation in the United States (Yay for topicality at NYU!), my understanding of the Bradley effect is the same as Chris’: the Bradley effect is a narrowly-defined, largely irrelevant idea. Troubling, though, is the amount of attention the Bradley effect gets in the media whenever a black candidate is near an election. This disproportionate focus on an irrelevant conclusion worries me, as I view it as yet another indicator that none of the mainstream media outlets actually care about the factuality of that which they are stating, but rather only about the ratings they receive and the controversy they generate (as these both lead to increased revenues at little cost).
This statement may be a bit too general. Allow me to rephrase: many news outlets seek to maximize their relevance/impact while limiting the risk to which they are subject, which can often lead to the use of disclaimers (e.g. “Some have said…” or “Researchers say…”) to create a space in which anything goes, where outdated research can be presented in a supremely polarizing way.
Rather than rely on the Bradley effect, I think we need a more general rule. Perhaps something along the lines of: When there is enough going wrong in the country/state/county/etc. to create believable excuses for voters who might otherwise lie about their preferences to hide biases, the pre-election polling will more accurately reflect the actual electoral vote. Stated more concisely: when people have less reason to lie, fewer people will lie. This second version is a blunter construction of the idea, but it nicely highlights the limitations of such a general rule: of course fewer people lie when they have less reason to do so. This conclusion seems only natural and antiquates the very rule that spawned it, putting us back at square one: no Bradley effect and people lie when they have to do so.
In dismissing my own general rule above, I was assuming the following: lying leads to risk of the lie being discovered; being discovered to be a liar uses up social capital (i.e. good will); social capital is limited for each person. All this implies that to lie is costly (in terms of the energy it takes to execute and maintain the lie and also in terms of the costs of damage control if the lie is ever found out), so it is only natural that fewer people will lie when there are fewer incentives to lie.
Well, I am at about a page of rambling, so I will stop myself now. Please tell me if any of this was unclear or needs elaboration. As Alan Greenspan once said, "If I seem unduly clear to you, you must have misunderstood what I said."
I kind of see where you're going. Let me say first, though, that no one has the right to apologize for rambling as long as I'm still contributing to the site.
ReplyDeleteBut moving on. You say the following:
When there is enough going wrong in the country/state/county/etc. to create believable excuses for voters who might otherwise lie about their preferences to hide biases, the pre-election polling will more accurately reflect the actual electoral vote.
I think I agree with your idea, but I've got to say that in any election there will something so wrong with either the candidate or the country to allow for this. I mean, seriously, it's November, and the best they've got on Obama is Rev. Wright. Think about that. That is nothing. What did they do to Kerry, and to Bill Clinton, or would have done to Hillary? They've got nothing on Obama, but what they do have still counts just as much.
(Side rant: If I hear "possibly the most important election in history" one more time I'll go postal. I strongly maintain that this election is nowhere near as important as in 2004, and even calling that the most important election in history is tough. I mean, seriously? You think John McCain could even possibly be as bad as G. Dubbs? I don't think so.)
Also, with regard to when and why people lie, some researchers may have said (did you see what I did there) that people will lie about issues like race even when there is no risk of getting caught, and (to a lesser degree) even when they talk to an automated message instead of a person. So I don't know if you can get by without taking that into account.
Oh, I just saw that I misread the statement I quoted earlier. I don't feel like editing what I said, so I'll ramble some more.
I don't know if stuff going wrong in the country matters and, if it does, which way. If you're a "maybe racist," and the country is in crisis, are you more likely to vote on emotion/your gut and vote against Obama or more likely to really go the extra mile to think it out? I think it's hard to say either way.
I gotta quick talking now. Love the Greenspan quote, and I wouldn't mind seeing you clarify your hypothesis.
Ok, you guys have made me feel better. And even if you hadn't, Daniel Okrent over at HuffPo did:
ReplyDeletehttp://www.huffingtonpost.com/daniel-okrent/bradley-who-here-comes-th_b_132987.html
very clear, very logical, very reassuring