The Bradley Effect was first reported in 1982, when LA mayor Tom Bradley ran against George Deukmejian for governor of California. While a handful of polls showed that the race was tight, many polls had Bradley up by as many as 10 points over Deukmejian on the eve of the election, and exit polls showed that Bradley had won by a comfortable margin. However, Bradley lost the election by a small percentage.
The Bradley Effect has since been noticed in other elections. As the New York Times‘ Eugene Robinson noted in 2008,
The polls said David Dinkins would beat Rudy Giuliani by more than 10 points in the 1989 New York mayoral race; Dinkins ended up winning with 50 percent of the vote to Giuliani’s 48 percent. That same year, the polls gave Douglas Wilder an 11-point lead over Marshall Coleman in the Virginia governor’s race; Wilder squeaked into office by less than half a percentage point. In 1990, the polls said Harvey Gantt would handily defeat incumbent North Carolina senator Jesse Helms; Gantt lost, and it wasn’t even close.
The normal explanation for this phenomenon is that many white voters are essentially racists, but that they are reluctant to admit this to pollsters, so they will lie to the pollster, but then vote with their heart during the election. This analysis is unfair. A better explanation for the Bradley Effect is not that white people are racists, but that some white people are afraid to tell pollsters that they might vote against a black candidate, because they do not want to be seen as racist. That is, they feel extreme pressure to voice support for the black candidate in order to make some sort of symbolic political statement. But while that pressure changes the way they answer a pollster, it does not necessarily affect their vote.
Robinson goes on to note that:
[P]reelection polls were quite accurate in predicting the black candidate’s vote. What happened was that the polls greatly underestimated the vote for the white candidates. Unusually large numbers of self-described undecided voters ended up making the same decision.
Thus the polls were correct, to an extent. However, the polls did not accurately reflect the undecided sentiment, and undecided voters overwhelmingly voted for the white candidate.
The Bradley Effect was noticed in some, but not all of the Democratic primaries in 2008. According to Wikipedia,
[R]esearchers noted that … opinion polls taken just prior to an election tended to overestimate Obama in states with a black population below eight percent, to track him within the polls’ margins of error in states with a black population between ten and twenty percent, and to underestimate him in states with a black population exceeding twenty-five percent.
That is, in states where the white population was high, Obama underperformed polling estimates, showing that the Bradley Effect was in force–white undecided voters broke for Hillary Clinton. In states where the black population mirrored the national average, polling appeared to be accurate. In states where the black population was higher than the national average, Obama benefited from higher black participation in the primaries than was expected, resulting in a reverse Bradley Effect.
In the general election in 2008, Barack Obama out-performed opinion polls, and many experts believe that a reverse Bradley Effect was at work. Black turn-out greatly exceeded expectations. In addition, white undecided voters broke for Obama. One explanation for white voting behavior is that many white voters wanted to congratulate themselves for overcoming racism by voting for a black candidate. Another factor is what might be called the Ventura Effect. In 1998, Jesse Ventura unexpectedly defeated Norm Coleman and Hubert H. Humphrey III to win the governorship of Minnesota on the Reform Party ticket. In his case, Ventura also exceeded preelection polls, and won because undecided voters broke for him. In many ways, his candidacy–though politically dissimilar to Obama’s–echoed many of the same themes as Obama’s campaign: “Yes, we can” end business as usual. In 2008, voters had a choice between a career politician who was more than a little long in the tooth (McCain), and a fresh face with seemingly fresh ideas (Obama). Undecided voters chose the fresh face. They chose to end business as usual.
At the moment, opinion polls show Obama neck and neck with Romney. However, are these polls an accurate reflection of how people will vote in 2012? Of course, many voters are already locked in, and will not change their minds before election day. However, about 10% of the population appears to be undecided. Will they break for Obama, or for Romney? Additionally, there is the question of turnout. In 2008, for example, Obama won North Carolina by a mere 14,000 votes out of more than 4,000,000 votes cast. All things being equal, just a small drop in black turnout would have flipped North Carolina to the GOP in 2008.
There are signs that in the 2012 election, the Bradley Effect may well be in play. People may be reluctant to tell pollsters that they will not vote for a black man–especially someone they voted for in 2008–because they do not want to appear as racists. If this is true, then instead of being close, as current polls suggest, Obama is actually 5 to 7 points below his current poll percentage in actual voter sentiment.
The first factor that raises this possibility is the Obama campaign itself. The way Obama and his campaign team are behaving, they must believe that a reverse Bradley Effect occurred in 2008–that many white people voted for Obama not because they supported his policies, but in order to congratulate themselves on how broad minded they were. Over and over again, the Obama campaign is stressing that any dissatisfaction with Obama can only be rooted in racism. In such an atmosphere, many people would be afraid to admit that they had doubts about Obama’s presidency, even though they may not be willing to vote for him. Thus, Obama’s own political campaign gives evidence that polling may well be biased in his favor.
Another factor that raises this possibility is the internals of the polls. Poll after poll shows extreme disappointment with Obama and his policies, and distress with how the economy and country are going. However, Obama is personally viewed quite well by many people. Are we really to believe that people are going to vote solely on the basis of how they feel about a candidate’s personality, rather than his policies and the state of the nation? Are people really going to vote for a man they would like to shoot hoops with, if this means that they will continue to pay as much as $4 for a gallon of gas, or continue to stand in the unemployment line? Likely, undecided voters will break against Obama, and undecided and Democratic voter turnout will be depressed. Many people may like Obama, and many people may be proud that he is president and proud that they voted for him in 2008, but it is difficult to believe that a majority of people would knowingly vote against their own self-interest in a presidential election merely because of personality or race.
Finally, some surprising states may actually be in play in 2012. One of these is Oregon. In 2008, Obama carried Oregon by 16 points. However, both Karl Rove and Nate Silver believe that Oregon may be much more competitive this year, and that in a wave election it could flip Republican. If Romney can truly be competitive in a state like Oregon, then something is drastically wrong with current projections and opinion polls.
Our own feeling is that Romney is in much better shape than opinion polls currently show, but that he has a long way to go to seal the deal with the voting public. Likely, the states now considered battleground states will split about evenly between Obama and Romney, simply because so much money is being spent in those states on negative advertising that there will be few if any undecideds left by election day, and it will come down purely to who has the better organization and can get the better turnout. At the same time, a handful of other states not c0nsidered by most people to be battleground states will be close or even flip to the GOP in 2012. These states have a larger pool of undecided voters and will be largely untouched by negative advertising. The undecideds in these states will tend to vote against Obama or stay home.
The picture is thus muddled. Romney is probably ahead of Obama, but cannot afford to rest easy. If things stay as they are, he will win the popular vote, but it will be too close for comfort. On the other hand, if the Bradley Effect is in full force, this will be a wave election, 58% to 42% in favor of Romney. And Romney could even win Oregon.