The Bradley Effect

With just days left until the election, polls have Barack Obama building a sizable lead over rival John McCain. But polls aren’t perfect crystal balls, and politicos are abuzz over the latest potential hurdle between Obama and the presidency: a phenomenon called the Bradley effect.

The theory holds that voters have a tendency to lie to pollsters when they plan to vote for a white candidate instead of a black one. It was named after former Los Angeles mayor Tom Bradley, who was favored on the eve of the 1982 California governor’s race but dropped a nail biter to his white opponent, George Deukmejian. Some experts chalked up the skewed polling to a reticence to appear prejudiced–a notion bolstered by subsequent campaigns in which black candidates saw their leads evaporate in voting booths.

Since Hillary Clinton defied projections by edging Obama in the New Hampshire primary, analysts have warned that Obama could be in for another surprise on Nov. 4. But others say there’s scant evidence to back this up. V. Lance Tarrance Jr., a Deukmejian pollster in ’82, calls the theory a “pernicious canard,” arguing that Bradley’s numbers never accounted for Deukmejian’s lead among absentee voters. A recent Harvard study of 133 gubernatorial and Senate races between 1989 and 2006 found that the Bradley effect had abated in the mid-1990s as racially charged issues like crime and welfare receded. Another survey showed that Obama did better than expected, not worse, in the primaries.

Bias is a slippery variable to quantify and a potent one to overcome. At the end of a long campaign, however–and in the midst of an economic crisis–there’s little hard proof it will be foremost on voters’ minds on Election Day.