« UnityWatch: Hillary Student Chair Endorses McCain | Main | Circulating In Republican Oppo Precincts...Obama On Iraq From 2004 » Obama And A Racial Premium12 Jun 2008 05:06 pm Comments (24)
I don't think the Bradley effect is necessarily gone for good. However, in this particular election, those who usually vote straight ticket Democrat but aren't comfortable voting for a black man have enough excuses and rationalizations that they'll feel comfortable admitting they won't vote for him in a poll. I think the Bradley effect shows up when low info racist white voters who normally vote Democratic (or Republican) don't have a "reason" not to vote for the black guy, so just tell pollsters they'll vote for him. But Obama is a Muslim, doesn't say the Pledge, is an elitist, is out of touch with working class voters, whatever. The racists who are voting against Obama simply because he is black have their talking points and don't feel the need to lie about who they are going to vote for. 
I want to think we've progressed as a society since elections where the Bradley effect supposedly had a hand in a candidate's defeat (i.e. - Tom Bradley) or surprisingly narrow victory (i.e. - Doug Wilder). I'm not sure though. I certainly don't think the citizens of Kentucky and West Virginia by and large voted against Obama in their primaries because of his race though. Hillary Clinton had a much more natural connection with rural white voters in many states for a number of reasons, particularly her decision to stress economic issues in these parts of the country. I do think you will find voters who vote against Obama based on his race haven't been supportive of the modern Democratic party in some time.
Glancing at the polling data, I think that many of the large amount of undecideds that currently exist should be considered to be open to Obama on the issues but resistant because of his perceived culture, values, religion, race, etc. So right now, they don't feel comfortable with either candidate. They aren't destined to go McCain, but they are going to be very receptive to attacks on Obama (like the ones you see spread through e-mail and by the less savory GOP surrogates).
Your poll is not worthy of my attention.
All of the above but especially "We don't know what will happen yet."
There is a little bit of political science research on this issue that runs a bit contrary to the media coverage of the topic. The media suggests racially (or gender) biased voters lie to pollsters and say they support Obama, for instance, even though they plan on voting for McCain. The research suggests that voters motivated to vote against someone because of race are most likely to say "undecided" or "don't know" when answering pollster questions (or that they don't intend to vote, when in fact they do). This can lead to a couple or a few percentage points away from the minority candidate if there are quite a few 'don't knows' or 'undecideds' in the survey (there can of course be true undecideds as well, not racially biased). Thus, any poll where Obama is ahead but there are large %s of undecideds, we should look at the cross-tabs and see: (1) are most undecideds white?; and (2) assume at least 1/3 to 2/3 of voters not choosing a candiate are likely to vote for the white candidate.
Chris, This political research falls under the name "The Bradley Effect", from a 1982 campaign. It's not new, and large parts of the assumptions have been if not denied, then certainly questioned. It's not an accepted doctrine at all, and your methology is pretty off as well. At least 1/3 of undecided white voters will vote for McCain, unless it turns out to be one of the biggest landslides in history, and if your margin of error is from 1/3 to 2/3, then its essentially pointless.
James, no, you are wrong. The research I cite was published this decade (see work by A. Berinksy, a professor at MIT).
The research still falls under the term the Bradley effect, can you cite the specific paper - the only one i can find seems to relate only to racial policy, not electoral candidates, and uses 1992 data. Political Context and the Survey Response: The Dynamics of Racial Policy Opinion, Journal of Politics, 64:567-584 (2002) http://web.mit.edu/berinsky/www/context.pdf And also, you didn't actually cite specific figures save the 1/3 to 2/3 figure at the end, so I'm not sure what is being taken from his work save you generalizing and extrapolating. Apologies if there is such a paper which I have missed completely.
That's one of the papers I was thinking about. If I recall, the article doesn't just look at policy. It also looks at the mayoral race in NYC between David Dinkins and Rudy Giuliani. The thesis of the article, as I stated, is that people express racial bias through non-response of survey questions, not lying and saying you support the black candidate but actually vote for the white candidate (which is typically called the Bradley effect). Berinsky also has a book with greater detail, and there is a book by Keith Reeves too on this topic. There is a racial effect, and it is especially in non-response (e.g., those saying they are undecided --> more likely to vote for white candidate in contests w/ a white vs. black candidate). So this means, when a poll has McCain 40% and Obama 52%, that's great news for Obama. Even if the large bulk of the undecideds vote for McCain due to racial bias, Obama still wins. But if the poll is McCain 40%, Obama 44%, Obama has reason to be nervous about where the undecideds will go.
The article doesn't mention Dinkins or Giuliani, nor anything about electoral candidates. It's data is 1972 paired with data from the 90s. the Mayoral race is mentioned in the book (around 5-10 pages), but it's a stretch to think these figures for that specific race in 1989 can be ported over to Obama, especially since we now have a wealth of primary data. And I'm still unsure where the 1/3 to 2/3 figure comes from, unless we are literally going to take the New York mayoral election and transpose it on Obama and McCain. And specifically with Giuliani, it might be similar to the UK conservative party winning in 1992 - it may not have been race, but that many people were simply embarrassed to admit they were going to vote for him. http://en.wikipedia.org/wiki/Shy_Tory_Factor
Sorry, the article on Dinkins-Giuliani is "The Two Faces of Public Opinion." It shows Jewish voters, older voters, and women voters were less likely to respond to the survey and more likely to say "don't know." Liberals are more likely to respond to surveys. The results show the 'don't know/undecideds' led to about 4% points lower for Dinkins in the actual election compared to the pre-election polling. But you're right, it's just one election in New York City. I doubt the bias can be extrapolated to Virginia, Nevada, Florida, Ohio, or any other swing state (note sarcasm). It's certainly better evidence for an effect than a wikipedia page. which is all you offer. Last I heard, wikipedia is not peer-reviewed.
Well, if you're going to fall back on that peer-reviewed favorite, sarcasm, then no, it can't be extrapolated. Because its one event, twenty years ago, and New York has the biggest Italian-American community in America, and to ignore that when an Italian-American candidate is running and reduce to 'black' and 'white' implies a lack of subtlety to say the least. Or religion, when Giuliani in 1989 gained higher than expected figures in white Catholic (which of course overlap noticeably with Italian American) areas. It also completely ignores the fact of even then Giuliani was an extremely divisive figure. And that Jesse Jackson had called New York 'Hymietown'. And of course you have the Clymer disagreement, which Berinsky acknowledged - Clymer quoted from NYTimes: "Take the ''lying'' excuse. In almost all exit polls, voters drop a questionnaire in a box, so they do not seem likely to be fudging to please an interviewer." The exit polls were way off as well with the mayoral race. The point of the Shy Tory, which is backed with historical exit polls and actual election result, was a far bigger swing than Dinkins (almost 7%), was to show that there can be non-racial reasons for not admitting your election preference. The link was to show you the theory, and so you didn't have to hunt down everything like i have on your ever changing references. This was in the hope that perhaps you are interested in actually considering the factors with an open mind. But given virtually all your posts are asking for balance, by which you mean solely more critical posts on Obama, saying "we know about John McCain", and proclaiming that the Student Chair article IS news, that was a long shot. But you aren't really looking for actual debate on this and you still haven't answered as to where you get the 1/3 to 2/3 figure.
No, I'm not looking for a debate. I'm trying to teach you. Here's what can't be extrapolated about New York City--the people there are MORE tolerant than the white people in the rest of the country. Here's an idea. Maybe you can read the Reeves book I cited in the next 20 minutes and tell me how that is completely wrong as well. If you don't think race causes white people to engage in patterns of non-response, please learn how to read.
"No, I'm not looking for a debate. I'm trying to teach you". Glad we got that cleared up. As I never disagreed with non-response, I stated that using data a minimum of 15 years old, and a very specific race from 1989, which ignores the race of one half of its candidates, and ignoring any 'reverse Bradley effect', hundreds of polls in thiselection about this specific candidate or anything else which clashes with the two books you've read, may not be the most sophisticated way to judge what happens in this election. Now, since you are the only one of us who can read, you can tell me nice and slowly so I can understand, where...do...you...get...1/3...to....2/3...from?
Are white voters allowed to break away from Obama without being racially biased? Or is it CW that all white people who don't support Obama are biased. The latter. Well, as I recall, a particularly generous Slate writer allowed that we might just be spiteful. Marc, that's an incredibly insulting poll, unless you're looking for the views of the Obamabots.
James, since we don't have data on this presidential election that has yet to occur to know how behavior will occur AND we know that non-response and claiming 'don't know' or 'undecided' has led to racial bias in past elections (sorry there hasn't been research on other elections that fit better--but to my knowledge, no black candidate has ever run against a white candidate in a general presidential election, which is what it sounds like you are looking for). Anyway, what I was suggesting with my 1/3 to 2/3 comment, was an estimate of presidential polls. So for instance, if McCain is polling 45% and Obama is polling 45% in Pennsylvania (just hypothetically), and the remaining 10% claim 'don't know' or 'don't intend to vote' or 'undecided.' Given that we know some of these voters not answering may have racial bias, Obama needs to estimate that 1/3 to 2/3 of the 10% of non-responders is likely to vote for McCain. This totals somewhere between 3%-6 or 7% of the undecided breaking toward McCain. Very likely actually, but I am extrapolating for the future. Obama's campaign, it has been reported, is thinking the same way I am.
What swing state senator would be taking the DC to NY shuttle? Sherrod Brown makes sense, but I think there are direct flights from DC to Cleveland. I suppose it depends on your definition of swing state. Connecticut is polling closely, so maybe Joe Lieberman or Chris Dodd? Or maybe someone without direct DC flights to their state. Jeff Bingaman?
Not looking for a presidential, but you have for example Ken Blackwell, Deval Patrick, Lynn Swann, Erik Fleming, Harold Ford, Jr., Michael Steele who ran in 2006 - all of which combined should reveal more than the dataset mentioned, and could then negate the specifics of any one election. But thank you for the figures. I can see how you got to it, although I disagree with the final tally, but even if it is correct, the fact that it has a margin of error of 100% (33% versus 66%) still makes it problematic.
James, sure, these need to be studied. But academic studies don't work like the media. It takes years between an election and the data to be analyzed, the papers to be reviewed for problems and accuracy, and then finally published. Maybe you can do the study on these people, mostly elected in 2006.
No Chris, I don't need to do it because I don't think it bears real relevance to this election. Given the methodology, it wouldnt be too hard to collect data from the polls, isolate the undecided, then compare from actual result. It wouldn't take that long to get rough data that could falsify your theory or maybe show perhaps you have something. Since the burden of proof for your claims is on you, it kinda makes sense for you to try to verify. But you don't really have any interest in actually verifying the data, and since I'm about to go away for two weeks, I might as well stop wasting my time. This is one of the most unpredictable elections for years, no one is currently sure how it will pan out, there are apparently over 51 billion possible scenarios. Yet you have the answer to it, based on two books you read, with data no more recent than 1992. You sound like something from the Da Vinci Code. It's pathetic. the original study was on surveys on racial policy, in 1972 and 1992, which they then extrapolated onto the 1989 NY mayor election. So you are extrapolating an extrapolation of one single event with data at least 16 years old and you have no problem with that? one study versus fifty fucking primaries and caucusesthat have just been held, for which huge amounts of data are easily available, but you don't need to look at that data. You make no effort to take it further, to actually test it, yet you confidentially assert that Obama camp has the same figures. Comparing your workouts with theirs is a bit like Da Vinci and an Etch-a-Sketch. Seriously. You have two books, one idea and no fucking clue. An error range of 100%, and 2/3rds of undecideds will vote against Obama solely on racial basis? So if he's less than 7% ahead in the polls, he's going to lose? You're either stupid, or, more likely, you are anti-Obama for some reason and are trying to be subtle about it. I'm away for the next two weeks, which probably sounds like a gambit, but its not - if you see me post elsewhere, its not me. I'm on holiday, and I need a break from all this nonsense. No one else has commented on your post because I imagine you were regarded as not worth debating. You suggested I was unable to read, which reminded me of this:
Otto: Apes don't read philosophy. Wanda: Yes they do, Otto, they just don't understand it! Let me correct you on a few things; Aristotle was not Belgian! The central message of Buddhism is not "Every man for himself!" And the London Underground is not a political movement! Those are all mistakes. I looked them up. If you want to want to continue in two weeks time, ok. It's only six elections, you should have got them done by now. But I still may not be ready for more horseshit like before.
Talk about distorted answers due to a rigged poll. It's not possible to choose "will vote for McCain because they think he's a better candidate." According to your idiotic poll, you either vote for Obama (as a racist or non-racist) or vote for McCain (as a racist).
Correction: It's more accurately
|
Subscribe to this blog by Email Twitter || On Facebook || QIK Live Channel  Recent Posts
Monthly Archives
|
| 
Copyright © 2009 by The Atlantic Monthly Group. All rights reserved.
Are white voters allowed to break away from Obama without being racially biased? Or is it CW that all white people who don't support Obama are biased.
Posted by Paul | June 12, 2008 5:27 PM