Saturday, November 1, 2008

The Bradley Effect: What It Is and Why You Don’t Have to Worry About It Anymore

Imagine that you are a white citizen, voting in an election where a black candidate and a white candidate are running against each other. You are very racist and believe that black people make terrible leaders, so you decide to cast your vote for the white candidate, regardless of his/her positions. Imagine that a pollster asks who you plan to vote for. Even though you are racist, you don’t live under a rock, so you know that in the post-Civil Rights Era, being openly racist is socially unacceptable. Would you tell the pollster which candidate you plan to vote for?

Now imagine you aren’t racist at all, but you’re casting your vote for the white candidate because you support his/her positions. Yet when the pollster shows up, you fear that if you say that you’re voting for the white candidate, you’ll appear racist. What would you tell the pollster?

You might, regardless of truth, report that you support the black candidate in an attempt to conform to socially acceptable ideas.

Scholars and pundits call this phenomenon the “Bradley effect.” Historically, black candidates for office have polled significantly higher numbers than actual votes. The Bradley effect has a variety of possible explanations, but is usually attributed to the situation described above. White voters have misreported their votes in elections involving black and white candidates. They have overstated their support for black candidates because of the social desirability of seeming racially open-minded. Therefore, black candidates that have had huge percentage point leads in polls have still lost elections or won by very small margins.

The Bradley effect was named after Tom Bradley, the Los Angeles mayor who ran for Governor of California in 1982. Polls showed that Bradley had as much as an eighteen point lead prior to the election, yet he wound up losing. Similarly, Doug Wilder, a 1989 Virginia gubernatorial candidate, had double-digit lead in the polls, and won by only .2 percent. The Bradley effect was consistent in elections involving black candidates from the 1980s through the mid-1990s.

There is renewed interest in the effect with the possible election of America’s first African American president coming up in less than a week. News shows and political pundits, including those at CNN, ABC News, and RealClearPolitics, continue to argue over whether or not the Bradley effect will be present, especially in light of Barack Obama’s surge in the polls in the past month.

CNN discusses the Bradley effect:



CBS’ Web Exclusive on the Bradley Effect:


Even Al-Jazeera [English] picks up the story:



Yet the occurrence of a Bradley effect seems unlikely since it has not been found since the late 1990’s. In fact, some even talk about a reverse Bradley effect, because Obama outperformed expectations in many primaries. There are many reasons in general to why a Bradley effect may be less relevant today Daniel Hopkins, for instance, speculated that the racialized issues of welfare and crime have since decreased in salience. However, racism has not disappeared along with the issues.

A more likely explanation is the influx of younger voters, who tend to be more racially tolerant than their elders. As time goes on, those who were born after the Civil Rights Era become bigger parts of the electorate.

Another explanation may be that the Bradley effect may be embedded in the reports of undecided voters. Some of these voters may have actually made up their minds but are reluctant to voice their preference. These undecided voters may not be reporting their support for the white candidates for the same reason that voters in the past voiced false support for the black ones.

Political commentators are speculating about Obama’s numbers because of effects they found in the Democratic primaries. Some say that Hillary Clinton’s win in New Hampshire was an indicator of the Bradley effect, but in actuality there were many other factors contributing to Obama loss in this state, including some sampling problems with public polls and the reaction to Hillary Clinton’s emotional display the weekend before the primary.

Since it has not been truly seen for a decade, it is unlikely that the Bradley effect will impact the election come Tuesday. The only big concern should be the undecided voters, who are mainly whites. These whites may in fact be McCain supporters appealing to social desirability. This remains to be seen. Obama himself seems unconcerned with the idea. We'll just have to wait for Tuesday night to see who is right.

For Further Reading:

Chris Cillizza. 2008. “Race, Polling and the ‘Bradley Effect.” Washngtonpost.com. June 10, 2008.

Michael W. Traugott and Vincent Price. 1992. “Exit Pollsin the 1989 Virginia Gubernatorial Race: Where Did They Go Wrong?” Public Opinion Quarterly. 56(2): 245-253.

Daniel Hopkins. 2008. “No More Wilder Effect, Never a Whitman Effect: When and Why Polls Mislead About Black and Female Candidates.” Paper presented at the 2008 Meeting of the Society for Political Methodology.

John Heilemann. 2008. “The Color-Coded Campaign.” New York Magazine. August 18, 2008. 18-21.

No comments: