Having lived by the maxim that things are always worse than they seem rather than better, and taking the Democrats’ fortunes in the last two presidential elections fully into account, I can certainly relate to those Obama supporters who are suspicious of poll numbers showing their guy comfortably ahead in most national and many pivotal “battleground” state polls. In large part, these folks are achieving record levels of sphincter torque over the much-discussed “Bradley Effect,” wherein a black candidate’s poll numbers are inflated because whites who are unwilling to admit that they don’t want to vote for a black person simply lie to pollsters about their intentions.
This supposed phenomenon is so named because it is presumed to explain L.A. Mayor Tom Bradley’s loss to George Deukmejian in the 1982 California gubernatorial race after polls showed him well ahead. The same effect is supposed to explain why Doug Wilder won the Virginia governorship by such a tiny margin in 1989 after polls showed him with what appeared to be a strong lead. Ditto, I might add, for, David Dinkins’s unexpectedly close mayoral victory In New York City in ‘89 and Carol Mosely Braun’s narrower than anticipated win in the Illinois Senatorial race in ’92. More recently, the Bradley Effect was invoked to explain Barack Obama’s surprising loss to Hillary Clinton in this January’s New Hampshire Democratic Primary. I have to say that I bought into the Bradley Effect as at least a partial explanation for what happened in New Hampshire, and I have also been wary of making too much of Oby’s strong poll numbers for the same reason. However, NBC News political director Chuck Todd got me to thinking (Way to go, Chuck!! Nobody else though it remotely possible.) the other day, when he pointed out that neither Bradley or Wilder actually ran behind their respective polling percentages in ’82 and ’89, nor for that matter, did Congressman Harold Ford, Jr. in his unsuccessful 2006 Senate bid in Tennessee . More ammo for Chuck’s argument can be found in this chart from those mighty fine folks over at the Pew Foundation showing that compared to pre-election polls, the final vote tallies in major races involving black candidates actually appeared to reflect not so much desertion in the ranks of their declared supporters, as massive shifts of the “undecideds” to their opponents.
In Bradley’s case, for example, just a few days before the election, a Field poll showed him with 47 percent of the vote to Deukmejian’s 41 ,and another poll had the difference down to 45-44. Thus, on the eve of the balloting, save for the roughly 3 percent support spread among three other candidates of the “ no way in hell” variety, 8-9 percent of those polled ostensibly remained undecided. When the votes were in, Deukmejian had claimed just over 49.3 percent to Bradley’s 48.1, but since percentage-wise, Bradley had actually outperformed his final polling numbers, those shocked at his defeat should probably have been yelling “Liar! Liar! Pants on Fire!” not at the folks who told pollsters they were backing Bradley, but at a huge chunk of the self-described undecideds who wound up going for his opponent with suspicious uniformity.
Things were a little different with Obama in New Hampshire this year, but not much. The final pre-primary Real Clear Politics poll average showed Oby with 38.3 percent of the vote and a seemingly done-for Hillary at 30. Loverboy John Edwards and Bill, “I am, too, Hispanic!” Richardson accounted for another 18.3 and 5.7 percent respectively, leaving 7.7 percent in the undecided column. When tally-up time came, BO was in fact nearly 2 points below his polling percentages, while HC had gone nine points over hers. However, the two white boy also-rans had also lost more than 1 point each, suggesting that perhaps Oby had been shot down not by the Bradley Effect but the “Hillary Effect,” which reared its fearsome head when HC managed to pull a lot of folks off the fence and into her tent by transforming herself, in but the blinking of a single moist eye, from macho and formidable to feminine and vulnerable.
I may be wrong, of course—I think I may have been mistaken about something ten or fifteen years ago, although, for the life of me, I can’t remember what it was—but I’m guessing that in polls involving black candidates and white voters, the bulk of the real liars are to be found among the “undecided” folks who know damn well they aren’t going to vote for a black person but are telling pollsters they are still uncertain. This would make sense, after all, because it strikes me as a lot easier to believe that you aren’t really lying when you say you haven’t made up your mind even though you actually have than when you say you’re going to do something knowing full well you aren’t.
Still, what’s the difference? Call it “Bradley Effect” or not, doesn’t this mean that Oby could still lose in races where he seems to have a big lead right now?” Obviously, he could, but, all other things being equal, insofar as the reliability of the polls might actually rest less on the “truthiness” of people who have already declared themselves for Barry O than on the credibility of those who still claim to be undecided this late in the game, the key number is not the size of BO’s supposed lead so much as the percentage of the vote he projects to receive. A new Big Ten Poll l shows the following results for Pennsylvania:
Obama/Biden 51.9%
McCain/Palin 41.5%
Other 1.2%
Undecided 3.4%
Refused 2.0%
Let’s assume that the 1.2 % who intend to vote for the likes of Darth Nader don’t come to their senses or, God forbid, convince others to wallow with them in their miserable stupidity. Let’s also assume every single sovereign voter claiming to be undecided is lying through his or her rotten teeth or store-bought teeth and actually plans to vote for McCain and every one of the uncooperative SOB’s who refused to answer does the same. Even then, Angry Geezer and Stylish Airhead would still come up short unless they can convince an additional roughly 3 percent of the electorate to abandon the leading team in the waning seconds of the fourth quarter and come over to the losing team’s sidelines.
Obviously, similar conditions would apply in any state where the polls consistently show Obama-Biden above 50 percent. When they fall below that, however, after taking a gander at some of the wild swings in the Pew Foundation chart, I’m obliged to be a bit more skeptical of their prospects anywhere they can’t win simply by holding on to what they’ve got.
Clearly, unless somebody issues them a polygraph, pollsters can’t ever be any more accurate than their respondents are truthful. Still, if the oft-cited and equally oft-misread examples of Bradley , Wilder, et. al. are any guide, the most important question about their findings may not be just how many people are lying but what they are lying about.
Of course, a lot of liberal types seem to presume that any white vote against Obama (or any other liberal black candidate, for that matter) has to be racially motivated. A lot of them surely will be, but it isn’t as if there aren’t some legitimate reasons to have some doubts about Oby (some of which, even his supporters share) or that simple partisanship won’t explain a lot of votes either way. At any rate, the Bradley Effect, or at least the fear thereof, is clearly real enough to make us wary of projections about what will happen on Nov. 4, and I can only imagine how nervous the network nabobs will be on election night when trying to use exit polls to tell us who’s winning. Without denying or excusing the Bradley Effect’s racial component, it strikes me that some of the responsibility for its tyrannical grip on our perceptions of this election lies with those who have presumed to make the simple act of a white person voting against a black person seem like prima facie evidence of being a card-carrying Kluxer.
