How Many White People Will Vote for Romney Because He's White?

Bradley's Revenge

If Obama loses, the left will blame imaginary racists.

By JAMES TARANTO

If Barack Obama's re-election campaign ends in failure, you can expect to hear a lot about a man named Tom Bradley.

Bradley, who died in 1998, was a five-term mayor of Los Angeles, serving from 1973 until 1993. In 1982 he ran to succeed California's then-Gov. (also now-Gov.) Jerry Brown. Polls showed Bradley leading, but he lost narrowly to the Republican nominee, Attorney General George Deukmejian. Bradley lost again, this time by a landslide, in a 1986 rematch.

Thirty years later, the first Deukmejian-Bradley contest is remembered for having produced the idea of the "Bradley Effect." This was a hypothesis, or a legend, that aimed to explain why the polls erred in predicting a Bradley victory. It proposed that for racial reasons a significant number of voters told pollsters they were for Bradley (who was black) but voted for Deukmejian (who is white).

The Bradley Effect has been invoked in several other elections since, most notably the 1989 contests for governor of Virginia and mayor of New York. Black Democrats Douglas Wilder and David Dinkins, respectively, won those races, but they underperformed relative to late polls.

Whether the Bradley Effect hypothesis is true is a matter of considerable dispute. The secret ballot makes it untestable, and the polls might have been wrong because of bad methodology (bias in sampling or weighting) or just ordinary statistical error. In the case of pre-election polls, it may be that enough voters simply changed their minds before Election Day.

Whether or not there was a Bradley Effect in 1982, no one argues there was one in 2008, when black Democrat Barack Obama handily defeated white Republican John McCain. But this time around, with Obama leading Romney only narrowly in most polls, and trailing in others, the idea has been revived.

Here's contrarian Californian Mickey Kaus, a grudging Obama supporter:

Voters who were genuinely enthusiastic about Obama in 2008–and therefore told pollsters they were voting for him and voted for him–might have second thoughts in 2012, but be reluctant to express them for fear of either seeming cruel (to a pol they tend to like personally) or racist (last hired, first fired!). . . .You'd think this would be a possibility pundits would take seriously right around . . . now. At the least it will give them something to talk about.

We'd note that this column raised the possibility that something like the Bradley Effect was inflating Obama's "likability" ratings, but that was last December so it's ancient history. Now, it is lefties, preparing to apportion blame in the event of an unfavorable outcome tomorrow, who are alighting on the Bradley Effect.

Among them is former Enron adviser Paul Krugman, in an especially weird post on his New York Times blog. Krugman opens by inveighing against the Financial Times for publishing a headline describing the presidential contest as being "on a knife's edge." That is "deeply misleading," Krugman contends, because accredited scientist Nate Silver has declared that Obama is the prohibitive favorite.

But Krugman spends roughly half his post vindicating the FT by noting arguments against that proposition:

As Nate says, it's definitely possible that the polls are systematically wrong. The obvious ways they could go wrong, cell phones and Latinos, favor Obama rather than Romney; but maybe pollsters are overcompensating for these factors, or maybe there's a large Bradley effect distorting poll responses. Reporting about these possibilities would be interesting.

Reporting on a Bradley Effect would present quite a challenge, since it would entail finding voters who tell reporters that they deceived pollsters and confirming their honesty. One can, however, do a little analytical musing about the possibility of a Bradley Effect.

What's important to understand is that for the left, the Bradley Effect is not only an analytical construct but a moral one. Stephen Kaus, a doctrinaire San Francisco Democrat, made this clear the other night, tweeting to brother Mickey and yours truly (quoting verbatim):

"Can you braniacs explain why the Bradley effect isn't out and out racism. Bubba can't pull the lever for blk guy."

Thus belief in the Bradley Effect is a consolation prize for the left: If Obama loses, it reinforces their sense that racism is alive and well in America--and that they, who reject it, are better than most Americans.

But even if the Bradley Effect exists, construing it as indicating racism makes no sense. Stephen Kaus's formulation--"Bubba can't pull the lever for blk guy"--begs the question. To a doctrinaire lefty, it is axiomatic that many Republican voters are motivated by racism. It's safe to surmise Kaus believes that is true both of Romney voters who tell pollsters their true intent and of those who tell pollsters they are voting Democratic. Is there any reason to think it is more likely to be true of the latter group than of the former?

If anything, intuition suggests the opposite. Imagine a focus group of white Mississippians in 1946 in which participants are asked their opinions of Sen. Theodore Bilbo. None of them will vote against Bilbo, since he is running unopposed, but they express varying degrees of enthusiasm about his racist ideas. Wouldn't you assume that those who openly and strongly support them are the most racist, while those who are reticent or express reservations are the least?

Of course, America in 2012 is not Mississippi in 1946. White supremacy was prevalent there and then and is stigmatized here and now. The hypothetical Bradley Effect voter is operating within the context of that stigma. He falsely tells pollsters that he is voting for the black Democrat because he feels either guilty or ashamed to be voting for the white Republican.

The guilty Bradley Effect voter is one who has internalized the axiom that it is racist to vote Republican. That is, he agrees with Stephen Kaus.

The ashamed Bradley Effect voter does not think it is racist to vote Republican but is averse to the harsh judgment of those who do. That is, he is intimidated by people who think like Stephen Kaus. (Another term for this phenomenon is "social acceptability bias").

This column is a Bradley Effect skeptic. We doubt that such guilt and fear are a major source of bias in election polls. But if they are, that wouldn't demonstrate that liberal suppositions about race in America are true, only that they are influential.

Here's one conservative list of excuses for an Obama victory.

racism is alive and well in America--and that they, who reject it, are better than most Americans.

I shouldn't, because the evidence is manifest for all to see and only those too stupid or ignorant can't see it, but I feel I need to say for the record: Voter suppression is real, and it is being perpetrated entirely by one side - the Republicans.

I say this not because I don't think Obama is going to win - I believe he's going to win reelection and I believe we will see that result of this election before midnight tonight.

I feel I need to say it because of the twisted, lowdown scumbags who keep saying its just some figment of imagination, or some attempt at a preemptive excuse.

Voter suppression is real; it's being perpetrated openly by The governor of florida and the secretary of state of Ohio, and not so openly by rightwing crazies like True the Vote and other paid lackies of the rabid rightwing. And maybe the only thing that's lower than those actually doing it are those who live and promote the lie that it doesn't exist.