When writing about the Presidential preference polls, I often get friendly reminders along the lines of, “hey, meathead – - Obama still is the leader.” True enough, although the lead is small; but if the election were held tomorrow, McCain would win. There are several factors at play, including the re-emergence of the Bradley-Wilder Effect.
First, the easy factors, then a discussion of Bradley-Wilder.
Studies of opinion polling indicate that the margin of the leading candidate is usually overstated, even if only by a small amount. The average overstatement has been estimated to be about 2%. Even if we assume that because this race is already close, the overstatement is minimal, say 1%, that still brings the contest to a virtual tie.
Second, because the Republicans have a small Electoral College advantage, the Republican candidate can win a close race while running second in the popular vote, as Bush did in 2000. The difference very likely must be small, probably no more than one per cent in the nationwide popular vote; but this race is very close right now.
The Bradley-Wilder Effect was first observed in the 1982 California governor’s race, when popular Los Angeles mayor Tom Bradley was well ahead in the polls but lost the election by a narrow margin. The effect was observed in another prominent race when Douglas Wilder won the governorship of Virginia by an extremely narrow margin, despite polls showing Wilder with a comfortable lead. The Bradley-Wilder Effect was observed in a number of other elections during the period 1982-1996. In all cases, the election involved a black candidate who received far fewer votes than the public-opinion polls indicated.
Virginia Gov. Douglas Wilder
There seems to be a general consensus that the Bradley-Wilder Effect disappeared after 1996, although most observers believe it can re-appear in certain conditions. There is also a general consensus, with which I disagree, that the Bradley-Wilder Effect is a function of racism; see, for example, Susan Estrich’s article, Race and the Democratic Party.
In my view, while there may be a certain degree of racism involved, the Bradley-Wilder Effect results primarily from social pressures to conform to certain behaviors and the desire of voters not to be labeled as a “racist” or an “Uncle Tom.” Such labels represent a serious form of social ostracism in this country and can have negative impacts on personal, social, political, employment, and business relationships. There can be substantial pressure on someone who normally votes “Party X” to voice support for the “Party X” nominee; if the nominee is black, but someone has decided to vote for another candidate for any reason, there arises an incentive to claim support for the “Party X” candidate or to declare as “undecided.” This skews the pre-election polls so that the “Party X” candidate has a larger apparent share of the vote.
But if the Bradley-Wilder Effect disappeared after 1996, will it impact this election? I believe that the answer to this question is “yes” because the conditions which gave rise to the Bradley-Wilder Effect appear to be present today.
Obviously, the first condition is met: the leading Democratic candidate is black. Also, the Bradley-Wilder Effect came into play in well-publicized races. Tom Bradley was the first black mayor of Los Angeles; Doug Wilder was the first black governor of Virginia; David Dinkins was the first black mayor of New York City; Harold Washington was the first black mayor of Chicago; all of these elections are cited as Bradley-Wilder cases and all were well-publicized. The current Presidential contest certainly meets this second condition.
The third condition is also met: there are issues which are race-related, directly or indirectly. Ironically, the presence of those issues has been induced primarily by race-carding by the Obama campaign.
The Bradley-Wilder Effect is signaled by a relatively large number of “undecided” voters. This phenomenon is unquestionably present in the 2008 race. I have calculated that the standalone polls have been reporting “undecideds” in the range of 10% to 12% consistently since early May of 2008. Undecideds for the current period, early August to date, are about 10.6%. In contrast, at a similar stage in the 2004 Presidential election, “undecideds” were an average of 4.9% in the standalone polls reported by Real Clear Politics. The average of “undecideds” in June of 2004 was 7.4%; in early July, 5.5%; in late July, 7.5%. In the first half of August in the 2000 Presidential election, “undecideds” were about 4% to 5%, very close to the 2004 figure.
The most likely explanation for the high proportion of “undecideds” is the presence of a Bradley-Wilder Effect in this year’s election. The impact of Bradley-Wilder has been estimated at 3% to 10% of the actual vote. Even at the lower end, this difference defeats Obama.
The Rasmussen Tracking Poll as of this writing on August 9, 2008, has been flat for nine days, with a 1% Obama lead in every report. If the leader-overstatement factor is a minimal 1%, the race is tied; a minimal Bradley-Wilder Effect means a McCain +3 popular vote. If the actual vote is McCain +3, then virtually all of the battleground states become Red states: Colorado, Ohio, Florida, Virginia, Missouri, North Carolina, Nevada, New Hampshire, and possibly Michigan as well. The Bradley-Wilder Effect may be higher, since the difference between past “undecideds” and the “undecideds” in the 2008 race has been close to 5%.
Please keep in mind that voting decisions are often far more complex then we assume, and movements in opinion polls are similarly complex and may not mean what we think they mean. For example, there may be alternative reasons or a better explanation for the high proportion of “undecideds” in the 2008 polling results to date. Nonetheless, I have concluded that there is a very realistic probability that the Bradley-Wilder Effect is in play.
RELATED ARTICLES:
Polljunkie, August 6: New Polls Confirm Obama Has Problems
Polls: McCain, Obama In Close Race; Trends Are Against Obama
RECOMMENDED READING:
Trend Analysis – - Why Obama Can Lose
ENPR – - Race Talk Bogs Down Obama
The Election Is Not Held Today
JOIN US! Visit American Sentinel
Sign the petitions:
American Solutions: . . . . . . . . . . The Lexington Project (loads slow):
August 9, 2008 at 7:50 pm
From the primary polls, The undecided voters went to Clinton if you go by the numbers. I am hoping that same thing will happen with the GE polls with the undecided going to McCain. **==
August 9, 2008 at 8:30 pm
If the election were held tomorrow, IMO, the likley reason McCain wins is because his experience is finally being seen as valuable–and the petty charade that Obama has been promoting as him as already president has exposed how empty, inexperienced, arrogant and clueless he actually is.
August 9, 2008 at 9:54 pm
Honestly, I really don’t see how Senator John McCain (R-AZ) can win the presidential election given so many American voters are so dissatisfied with the current Republican administration of President George W. Bush (R) in the White House, ranging from the Iraq War to the domestic economy. Polls after polls to date have been showing that, and it’s simply undeniable! Besides, many conservatives, the Republican Party’s core base of electoral support, have vowed to stay home on Election Day in order to allow Senator Barack Obama (D-IL) to get elected because they hate McCain’s occasional display of his independence from his party when casting vote on important issues in the U.S. Senate. If Senator McCain were a Democratic presidential nominee instead of a Republican one, which he now is, then he would be assured of victory in such race for the White House. To be fair, Senator McCain is a good man, a decent and honorable one, too, but he is simply in the wrong party at the wrong time.
August 9, 2008 at 10:03 pm
Faceless: I don’t even know where to start . . . let me put it this way: given all you say, Obama should be cruising 10 points up or better; he’s not; why not?
August 9, 2008 at 11:31 pm
This assessment is pure speculation and not factual. Obama has a very solid lead in electoral votes….it’s all about the EC, baby. During the debates, McCain will pull a Nixon, although Richard Nixon was lightyears ahead of McCain in intellect, analysis and political acumen. Not to say I supported Richard Nixon. John McCain is an anachronism–let’s hope he catches up with the rotary phone after he finally masters Morse Code. Besides my quip, McCain lacks the ability to lead because of his narrow background, narrow thinking, narrow knowledge…perhaps he needs a weeklong vacation in Siam…and is too quick to pull the trigger on some lesser equipped country. Let him take on China or Russia…LOLOLOL. We are at the precipice of an American Dark Ages and McCain will finish stacking up the rocks closing the cave door. If you guys had any brains, you’d start learning Mandarin.
August 9, 2008 at 11:57 pm
RJK – “This assessment is pure speculation and not factual.” “During the debates, McCain will pull a Nixon . . . ”
QED
August 10, 2008 at 12:02 am
RE: RJ Kruger’s response: Now there’s an inescapably revealing presentation of exactly why Obama’s devotees mirror him: Lots of words and nothing worthwhile said.
Unless it’s malignant.
August 10, 2008 at 12:53 am
So McCain supporters are actually hoping that a significant percentage of the polled electorate are racist liars? Don’t get me wrong, it’s entirely plausible; but no matter your preferred candidate, shouldn’t it be viewed as a grim possibility instead of reason for wishful thinking?
August 10, 2008 at 1:24 am
In RE Bradley, from Los Angeles: true, and another commonality with Glibama – Bradley lost, then won. However he was much of Glibama, a complete cipher, a “listener” pumped up to Hymalayesque dimensions by the LA Times, nevertheless a crappy mayor who lasted in office saying nothin’, all roaches around very active. It’s during Bradley’s tenure that the LA seccession movement got traction (the Valley vs. metropolitan).
The only good thing one can say about Bradley is that he wasn’t Coleman Young – if Glibama’s elected, we might find solace in this comparison.
August 10, 2008 at 2:39 am
A few days ago, Mike Barnicle, substituting for Chris Matthews on MSNBC’s Hardball, asked Pulitzer Prize winning journalist, Chicago Tribune editor and consistent Barack Obama supporter Clarence Page about this very issue:
‘Hardball with Chris Matthews’ for Wednesday, August 6, 2008
http://www.msnbc.msn.com/id/26074322/
[[BARNICLE: What do you think the percentage factor is in these polls about people who don‘t tell the truth?
PAGE: Well, I‘m glad you asked that, Mike, because I called it the Clarence Page factor during the primaries. I learned from experience that if Barack Obama wasn‘t 6 points ahead going into a primary, chances were very good that Hillary Clinton was going to win it, like New Hampshire and a number of other states there.
It‘s not always that people lie to pollsters, but as you implied with your scenario at the factory gate, a lot of people just don‘t respond to pollsters. If they‘re not going to support the black guy, they just won‘t respond. And that‘s something that Andy Kohut at the Pew Center and other folks have found. So I think Barack Obama still has reason to be nervous if he‘s only 4 to 6 points ahead.]]
August 10, 2008 at 3:17 am
Yeah okay. If you put the math together, the average looks today of the electoral college Obama would win by 55 electoral votes. So let’s take facts and not assumptions.
August 10, 2008 at 6:59 am
nicolae carpathia: You missed a very important point. Try again.
August 10, 2008 at 7:01 am
Joey D: Your conclusion is based on assumptions, not facts. Where are you assuming that those “undecided” votes go?
August 10, 2008 at 11:44 am
Every body is ignoring the Hillary women who are still upset about the whole process.
If you add that to the Bradley effect- OBAMA is toast in this election.
This is from a Hillary supporter who will not support Obama.
August 10, 2008 at 11:53 am
Did I? Enlighten me.
August 10, 2008 at 11:58 am
nicolae: “In my view, while there may be a certain degree of racism involved, the Bradley-Wilder Effect results primarily from social pressures to conform to certain behaviors and the desire of voters not to be labeled as a “racist” or an “Uncle Tom.” Such labels represent a serious form of social ostracism in this country and can have negative impacts on personal, social, political, employment, and business relationships. There can be substantial pressure on someone who normally votes “Party X” to voice support for the “Party X” nominee; if the nominee is black, but someone has decided to vote for another candidate for any reason, there arises an incentive to claim support for the “Party X” candidate or to declare as “undecided.” This skews the pre-election polls so that the “Party X” candidate has a larger apparent share of the vote.”
By tagging people as “racist” without justification, Susan Estrich and you are part of the problem.
August 10, 2008 at 12:10 pm
That’s good. Barack can come back. Hillary can’t…She’s being buried by her own supporters. No matter how much she campaigns for Barack, the judgement has already come down. A Hillary loss will spell doom to her presidential. She and her husband have shown themselves to be exactly who their political enemies have said they were.
August 10, 2008 at 1:11 pm
this guy must be on drugs.
i am far right, but under no illusion, lets go to the numbers. pull out an electoral map and rasmussens state by sates. everything else does not matter. his polling has been the most accurate now for 3 consecutive presidential elections. now flip the states around to see what it would take for a MAC win. he has, as correctly stated, less than a 40% chance. which is no chance in vegas. without a meltdown, SNOBama is pres. duh. prepare for the lectures, the disarming, and high taxes. as an expat, i will likely renounce citizenship if cap gains taxes rise to 38%. the USA is long past the era where it is the land of freedom. what a joke it has become.
August 10, 2008 at 1:54 pm
I actually think that Charles Blow’s recent op-ed in the NYT:
http://www.nytimes.com/2008/08/09/opinion/09blow.html?em
is more accurate as to the reality of this situation, but now we’re getting into speculating on what’s going on inside people’s heads, which is kind of pointless. You yourself said there was a “certain degree of racism involved.” 5% of whites polled actually admit that they would not vote for a black guy, which means if what you’re saying about polls being skewed due to the “Bradley-Wilder Effect” is true, then the number is significantly higher than that. This means that if McCain wins in a close election, his margin of victory will be largely due to racism.
The tone of your article, and of many of the posts here, seems to be taking solace in the “Bradley-Wilder Effect” keeping Obama out of office. Although that doesn’t make you, or even a large percentage of people lying to pollsters out of some kind of PC pressure racist, in my view it’s being complacent toward racism in others which is rather cynical.
August 10, 2008 at 2:56 pm
“…the Republican candidate can win a close race while running second in the popular vote, as Bush did in 2000.”
The old canard. The democrats manufacture fraudulent votes in the big cities which they control. Take away 1 or 2% from Gore’s popular vote total leaves him well behind Bush. That’s the sort of bullshit the electoral college was designed to counterbalance.
August 10, 2008 at 3:00 pm
Look, let’s use 2004 as the baseline to determine what may have changed that gives BHO a shot. If McCain were to get the same 62 million votes Bush got, he wins. McCain is seen as more of an independent so he actually picks up some Kerry votes (maybe not much) which should expand the 2004 margin at first glance, adding to the 62 mill and for BHO, subtracting from the Kerry 59 mill. But BHO apparently has registered some new voters (although many of the young ones aren’t the most reliable voters) so perhaps that should offset the McCain gain and bring it back to about a 2.5 mill McCain lead.
BUT, the big BUT is the folks on McCain’s right. Many of the Bush voters of 2004 have threatened staying home, conceding the race to the Dems to teach their party a lesson. So in essence, to all these people claiming that Bush has hurt McCain, you have no data to back it up. McCain’s problem is NOT that the Republican brand is hurt but rather many Bush voters feel he is not Republican enough. Remember, BUSH WON (even after Iraq & Abu Gharib). Iraq since then has only gotten better (even if you think slightly). BUSH had about a 50 % approval at re-election time. He is down to about 33% now. (www.rasmussenreports.com) So who are those 17 % and where did they go? Well, if you follow thetrendlines, the most significant drop came during the immigration controversy last year. That’s what took him from the low 40’s permanenetly down to 35%. Those people are not about to register Dem or support BHO. They may stay home. McCain will have to work hard to draw them back but more they understand the disastrous consequences of electing BHO, the more they will come back. As for the 2 or 3 points Bush has dropped this year, it is directly a result of the economic conditions and the way both McCain and BHO have bashed Bush in the media for political points w/the Pres not able to defend himself. Those pts won’t affect McCain who is not connected to Bush on the economy as he is on foreign policy & immigration.
So remember, most of the Bush/Republican/McCain-haters would vote against McCain in any election and voted against Bush in 2004. They are factored in and not worth assessing for the purposes of projectingthis election. It is the staunch conservatives on the right who are relectantto get behind McCain mainly b/c of the immigration issue that will decide if McCain wins OR to concede this race to BHO.
August 10, 2008 at 4:41 pm
nicolas: I don’t subscribe to the New York Times, so I don’t know what Blow had to say. Trying to discern what opinion polling says about demographic phenomena is not “speculating on what’s going on inside people’s heads,” which would be pointless.
You say: “You yourself said there was a ‘certain degree of racism involved.’ 5% of whites polled actually admit that they would not vote for a black guy, which means if what you’re saying about polls being skewed due to the ‘Bradley-Wilder Effect’ is true, then the number is significantly higher than that. This means that if McCain wins in a close election, his margin of victory will be largely due to racism.” You make an assumption not supported by logic, i.e., that whites who will not vote for a black candidate either lie about their preference or claim to be undecided. That is not supported by any evidence in the polls themselves – -obviously, that 5% figure came from respondents who admit they won’t vote for a black candidate, and that is not a Bradley-Wilder Effect problem – - and it may well be that most avowed racists are like the ones I know, who aren’t the least bit shy about saying what they think. The polls are not skewed by racism, as such, but by social conditioning which may have nothing to do with racial attitudes.
August 10, 2008 at 4:46 pm
AB: For what it is worth, the polls so far indicate that Republicans intend to turn out and tend to vote for the Republican candidate in about the same proportions as usual. McCain has about 87% support within the Republican Party; Obama has about 80% support in the Democratic Party. Ultimately, I believe that differences concerning immigration policy will be secondary when compared to the issues of whether we continue to have an effective national defense and who gets to make the next four nominations to the Supreme Court.
August 10, 2008 at 10:38 pm
“In my view, while there may be a certain degree of racism involved, the Bradley-Wilder Effect results primarily from social pressures to conform to certain behaviors and the desire of voters not to be labeled as a “racist” or an “Uncle Tom.”
So, the results to a “Candidate Blackguy vs. Candidate Whiteguy” poll are skewed because people don’t want to labeled racist, but the results to a poll asking “Would you not vote for a black guy cause he’s black?” are not skewed for the exact same reason?
I don’t doubt that there are plenty of people who factor into “Bradley-Wilder” because they feel socially pressured into saying they’ll vote for the black guy, even if they won’t for reasons that have nothing to do with race. However, saying that there is no portion of the “Bradley-Wilder” percentage that is actually hiding some amount of racist feelings is illogical, and inconsistent with the opening of your above statement.
So how many people are absolutely not racist but are for the white guy but won’t admit it because they’re paranoid at being labeled “racist” vs. people who are for the white guy for reasons having something to do with some amount of racist feelings which they won’t admit to because they don’t want to be labeled “racist?”
Trying to answer that question requires speculating on what’s going on in people’s heads. However, since we are in agreement that the latter group does exist, or at least “may exist to a certain degree,” if JMC beats BHO by a narrow margin (like say, a couple football stadiums full of people in Ohio or a cafeteria full of people in Florida) Then racism will certainly have been a factor in his victory.
August 10, 2008 at 11:24 pm
nicolae carpathia: You seem to be intent on proving that John McCain will benefit by a racist vote. Most observers believe that Obama will receive better than 93% of the black vote and some unknown portion of those voters will support Obama because he is black; therefore, Obama is the beneficiary of a “racist” vote, mais non?
August 13, 2008 at 5:42 pm
[...] Why McCain Wins If The Election Is Tomorrow [...]
August 14, 2008 at 6:45 pm
[...] Why McCain Wins If The Election Is Tomorrow [...]
August 27, 2008 at 5:48 pm
[...] standards. The continued high proportion of “undecideds” affirms the possibility of a Bradley-Wilder Effect in this year’s contest. Accusations of racism by Obama supporters will make that prospect [...]
August 27, 2008 at 5:48 pm
[...] standards. The continued high proportion of “undecideds” affirms the possibility of a Bradley-Wilder Effect in this year’s contest. Accusations of racism by Obama supporters will make that prospect [...]