The “Bradley Effect” in Reverse

Much has been made of the Bradley Effect and its role in pre-election polling.  For those needing a refresher, the Bradley Effect is

Named for Tom Bradley, an African-American who lost the 1982 California governor’s race despite being ahead in voter polls, the Bradley effect refers to a tendency on the part of white voters to tell pollsters that they are undecided or likely to vote for a Black candidate, when, on election day, they vote for his/her white opponent.

The CW is that Obama’s likely results will be about 2 percentage points below what the advance polling suggests.

But what if the reverse is now true, 26 years later?

Putting on my psychologist hat, I can make what I believe is a pretty good case that things have changed quite dramatically in those 26 years, especially with regard to the issue of polling versus actual voting.

Think about someone responding to a phone poll.  Chances are, the person on the phone is not alone.  They may be surrounded by family members or friends.  Even if they are alone at the time, being included in a national poll is a noteworthy event for most people and likely to be mentioned to family members, coworkers, and/or friends when the opportunity arises.

So the poll-taker — a disembodied voice over the telephone — is not the only audience involved.  In fact, I might argue that the poll-taker is the least important person in the mix.

As views about race and race relations are in flux, it’s reasonable to infer that family groupings or workplace groupings or friend groupings do not change all at once.  So there’s always going to be some tension.  One member of the family might be colorblind, but don’t let Grandpa hear you speak highly of a black man or there’ll be hell to pay.  Or something like that.

Voting, on the other hand, is a truly private activity.  Asking someone who they voted for is akin to inquiring about the specifics of one’s sex life — it is only to be undertaken by true intimates.  (Plus, one can always lie. . .but I digress.)

This sets up the potential for a “Reverse Bradley Effect”, in which someone tells the pollster that he or she would not vote for Obama, yet actually pulls the lever (or mark the spot or, God forbid, touch the screen) for Obama come election day.  And no one would be the wiser.

As it turns out, the polling in advance of the Democratic primaries understated Obama’s support.  Kos was kind enough to crunch the numbers, and showed

what we saw time and time again was Obama exceeding expectations, oftentimes by significant margins — an average of 7.065 points, as a matter of fact. So there was no Bradley Effect. In fact, it’s quite clear that the opposite was really the case.

Just to be clear:  I’m not discounting the Bradley Effect when, um, Bradley was running for Governor in California.  Instead, I’m raising the possibility that the phenomenon has reversed itself in the ensuing 26 years.  After all, John McCain was first elected to the Senate in 1982, and look how he’s changed in the past 26 years!

6 Responses to “The “Bradley Effect” in Reverse”

  1. gcotharn says:

    I agree the Bradley Effect may have reversed itself. No one knows, and there’s no way to conject until after the election. You are correct that 26 years is a looong time. There has been much social shift in those years.


    Regarding 1982, I’ve wondered this about the Bradley Effect: could it have been partially documenting 1) hidden racism and 2) hidden fear of appearing racist?

    In the case of #2, I mean this: what if a voter is not racist, does not like Bradley, and has a weak spine? Isn’t it logical such a voter would claim they will vote for Bradley b/c they would not want to appear racist to a poller? Such is the path of least resistance. When such a poller gets into the privacy of a polling booth, they vote for who they really want. Their vote has nothing to do with racism. Their poll answer has everything to do with having a weak spine.

    I suspect the Bradley Effect was partially or largely about hidden racism. Yet, how much it was about hidden racism is conjecture. No one knows, with any certainty, how much. I suspect some significant portion of the Bradley Effect resulted from #2, i.e. weak spines. I’ve no feeling for how big that portion was, beyond an instinct that it was significant.

  2. DrGail says:

    Yes, cgotharn, I believe the Bradley Effect has everything to do with hidden racism, at least as the B.E. is defined. The reasoning goes that the people who were polled were too chickens**t to tell the pollster they wouldn’t vote for Bradley, but then they gave full vent to their racism once they were in the voting booth.

    My conjecture is that this effect may well have reversed itself — that people who are polled are more worried about letting their lack of racism show to their associates than they are about admitting racism to pollsters. If this is so, then the polls UNDER-report Obama support rather than OVER-report his support (as the CW regarding the Bradley Effect would have us believe).

    I agree with you that all of this is conjecture, but I believe that given the KOS numbers, there’s a good argument for a reverse Bradley Effect.

  3. Dynamic says:

    I think you’ve raised a possibility, but I’m afraid I don’t consider it especially likely. You’re optimistic about the change we’ve seen, and I think it’s a fair optimism, but while I think the Bradley effect may finally have been mostly nullified, I don’t know that a reversal is too likely. Those who are willing to vote for Obama are also those who are likely to be proud of voting for an African American and would make no bones about it. Certainly situations such as you described may occur (and probably already have, several times) but I don’t know that I believe they’ve occurred often enough to raise a serious difference.

    What I think is much more likely is what may come to be called the “Obama effect,” given that it may propel Obama to victory. This effect is a result of the fact that telephone polling can only be directed at landlines and NOT to cellphones. This limits the pool of available respondants – and it does so by cutting out precisely the most devout supporters of Obama. I could see an easy couple of percentage points swing on any given poll because of this oversight.

  4. DrGail says:

    I had shared your concern about the under-representation of likely enthusiastic Obama supporters because cellphone-only users were not included in the polls. But then I started seeing this very issue discussed in the blogosphere and the corporate media. According to this source, many of the major polling outfits have started including cellphones in their polls.

    I too am not very certain about how large an impact what I’ve dubbed the “Reverse Bradley Effect” will have: only time will tell. My intent really was to challenge the CW, which says the Bradley Effect still holds and that one can’t be too sanguine when the polls show Obama to be leading McCain.

    And of course none of this discussion about the Bradley Effect or its reverse has any bearing on the true partisans out there. As you say, “Those who are willing to vote for Obama are also those who are likely to be proud of voting for an African American. . .”

    My husband and I often discuss that the true racists out there, who would never vote for an African American under any circumstances, aren’t likely to vote for a Democratic candidate under any circumstances either. Hence, the effect of his race is only really relevant to independents. Or at least we hope that’s that case.

    I hesitated to go too far into the weeds in the original post, but there is also a social psychological phenomenon called the “self-perception theory” (pioneered by Daryl Bem). The basic idea is that people’s attitudes are formed and solidified as a result of observing their behavior. E.g., “I must like chocolate because I always seem to be eating it.” It is intensified when someone makes a statement of their attitude.

    Hence, someone who is polled and indicates a preference for McCain (when they might actually be more pro-Obama) would be more likely to indicate a preference for McCain in the future, simply because they have already said it once. This, of course, adds another huge layer of complexity to the issue and would tend to argue against the “Reverse Bradley Effect”.

  5. By Any Means Necessary says:

    There’s plenty of evidence to throw the original “Bradley Effect” out the window as a dated phenomena. Here’s the evidence “the Bradley Effect” is gone: (1) Hillary Clinton (2) Joe Biden (3) Bill Richardson (4) Chris Dodd (5) John Edwards (6) Mike Gravel (7) Dennis Kucinich. Once you recognize this fact – x 7 – you must question the spin abound in defining the “reverse” Bradley Effect.
    In the mid ’70’s I listened to civil rights activist Julian Bond speak on campus. During the Q & A time, to everyone who took issue with him, many posing intellectually developed questions, he immediately responded, “You’re a racist”. What we were witnessing was a precursor to the conditioning of a susceptible white America. Americans who were just learning the history of despicable acts of racism. Americans who were not personally victimized by black on white crime, and could rationalize the inherent racism of affirmative action. Americans who were distanced from isolated black Americans and bought into believing they couldn’t understand because, after all, “it’s a black thing”. They were racists. And they felt guilty. We felt guilty. We are guilty.
    It’s time we make amends for our ignorance. What a historic opportunity we have with Barack Obama.
    NOW you understand The Reverse Bradley Effect. We must bear the fruit of our pitiful racial conditioning. Nothing else matters! “By any means necessary.”
    Do you accept this truth – x 7, or mock, seeking the social and personal acceptance you’ve sought your whole life?

  6. Bob Haskell says:

    I wrote it down this way the other day, reflecting on this (and I agree with and appreciate the way this article breaks it down):

    “People are worried about the “Tom Bradley effect” this year, as even people who’ve said they’ll vote for Obama WON’T once they’re in the privacy of the ballot box, because they’re more racist than they’ll let on (Black candidate for governor in California Tom Bradley was ahead in the polls but lost the election, in 1982).

    But I have an alternative version of the way things might go. Maybe, just maybe, some people who’ve said they won’t vote for Obama because we just “aren’t ready” for a black president – or whatever other self-deceptive reason they dredge up to justify their unwillingness to vote for him – just might, once they’re alone, vote for Obama, as the only sane choice, overcoming their racism in the private moment of choice.

    The “reverse Tom Bradley effect”… Given the drastic consequences and importance of this year’s choice, it just might go that way.”


  1. The Obama Effect | Comments from Left Field - [...] there is the possibility that the Bradley effect (Black politicians underperforming in the polls) may come into play on…

Leave a Reply

Your email address will not be published. Required fields are marked *

Connect with Facebook