Don't Believe The Hype---Polls and the Bradley Effect

Last week, banging out similar political ideas and issues (read: commiserating) with a colleague during a break, we hit upon an "A-ha!" moment of truth which, I fear, will play a major, albeit covert role in the '08 election, possibly blowing away all indications of an Obama lead in current polls, but which no one will talk about or admit to. Voters' dirty little secret; their prrrrecious!

My colleague, a forty-something African-American woman of southern heritage and southerly, sunny disposition, took me aside to a remote café table, sat me down with some mild force applied to the elbow, and honestly decreed: "Look. I'm gonna tell you how it really went down. This is what happened to all those white Hillary supporters who all of a sudden, in a hissy fit, decided to swing their votes McCain's way. They shut the doors, closed the windows, sat down at their dining room tables and quietly, among themselves and their families decided: 'I'm not voting for some n#*%er for president. If I can't have a woman, I'm gonna go for the default: the old, outdated WHITE guy Washington insider. Anything but a black man in the White House." Her words, not even paraphrased (I have a good memory).

Having been brought up and lived in a majority African-American county outside of Washington D.C. in the 80s and 90s, I'd been privy to a child's handful of similar "A-ha!" moments---frank, honest, tough talk by members of the black community regarding racial issues and underlying, covert bias by the white majority toward anyone of color. And so not only did I believe my colleague when she laid it down with brutal honesty, I agreed with her. Because, being white (an immigrant, but still white), I have been spectator to those private, racially-divisive, dinner table conversations; some initiated by my own, conservative family.

There is some reason to worry. Throughout recent political history we've seen the jarring loss of voter support for a black candidate on Election Day, despite positive indications from the polls predicting otherwise; because no voter would willingly admit his/her racial bias when asked his/her preference by the Frank Luntzes or Celinda Lakes of the business.

Since 1982, people have talked about this "Bradley Effect," (named for Tom Bradley, former mayor of Los Angeles) where even last-minute polls predict a wide margin of victory for the black candidate, yet he goes on to lose. Bradley, then mayor, lost his race for governor despite large leads in the polls---the assumption being that voters lied to pollsters about their support for an African-American.

But pollsters and political scientists say concern about a Bradley effect is misplaced. It obfuscates what they argue is the more important point: there are several ways that race complicates polling. For example, Barack Obama could theoretically score a landslide in the southern states, based on positive, but withheld polled opinion of his ethnicity. And then there's the reverse Bradley Effect in which support for a black candidate is understated particularly in regions where it's socially acceptable to exhibit distrust of blacks. Factor in also those voters who are not captured by the polls.

Scott Keeter, Pew Research Center's director of survey research, says pollsters historically have a harder time reaching voters with lower levels of education. Less-educated whites are the kind Mr. Obama has been courting in this election. Conversely, young people are more likely to answer surveys, and they tend to favor Mr. Obama. The race of the questioner, as well, affects the poll and what those surveyed are willing to say or not say to a white or black pollster. (As does, of course, the way in which one frames the question.) It's not a coincidence that both camps employ their own survey-taking foot soldiers. These are the kinds of basic intangibles we Bachelor's Degreed elitists learn in your basic Stat 100 class.

There is plenty of hand-wringing for Obama supporters and fretters yet to be done--remember, the exit polls in 2004 predicted President Kerry. A majority of polls indicate the Obama/Biden ticket is leading by 6-8%---numbers within the usual sampling error allowed in most polls.

However, many insiders argue that race does not play a huge role in the election this year, because the economy has emerged as such a gargantuan issue, and Mr. Obama is not primarily identified by his race...according to polls. So then, what are we to make of all these numbers?

Ask Flava Flav. He'll tell you.