In 1982, L.A.'s black mayor, Tom Bradley, lost the election for governor of California, even though exit polls predicted he'd win. There were lots of reasons why Bradley lost, but some pollsters contend voters lied when asked whether his race was a factor in their decision. Their contention has, over the years, morphed into what's called the "Bradley Effect." For the past year, political scientists have been working on a study to see if there really is a Bradley Effect. KPCC's Brian Watt says they recently outlined their work at UCLA.
Brian Watt: UCLA, Stanford, and the opinion research firm Polimetrix are conducting what they call the "Cooperative Campaign Analysis Project." Last December, just as the presidential primaries were heating up, they began online interviews with 20,000 voters across the country.
By Election Day, they'll have interviewed those voters over the Internet six times. UCLA's Lynn Vavreck says the questions are designed to get at information she knows the voters probably don't want to share.
Lynn Vavreck: We want people to tell us if they have racial prejudices, if they hold antipathy toward a group. So there is an art to getting people to reveal these things about themselves that they may censor in normal conversation.
Watt: That art can sound more like a science. Here's Stanford's Simon Jackman explaining it in a PowerPoint presentation to his colleagues at UCLA:
Simon Jackman: A respondent would be assigned to any one of these four conditions. There were three list setups, a five-response control list, and then two experiments...
Watt: You get the point – or maybe not. So here's how it worked in the round of interviews conducted last month. The survey asked voters to think specifically about Barack Obama and say how many of the following characteristics are important reasons to support him – or not support him: his public speaking ability, his Iraq policy, his economic plan, his health plan, and his party affiliation. But then there was a twist:
Jackman: Some people saw that list. Other people saw a list where "he's black" was added to the list. Or, his "political experience" was the sixth item.
Watt: Now, remember, the voter didn't have to specify reasons to vote for or against Obama – but did have to say how many there were. The researchers could then see if the number went up or not when Obama's "race" was added to the list. So what did they find? Simon Jackman:
Jackman: We estimate that of people who are not voting for Obama, for 11 percent of those people, race is a factor. Not the factor, but it's a factor. But when you ask them flat out, just flat out naively walk up to them and ask them, basically you get zero or one percent at most, right?
Watt: The survey also ran a similar experiment that got at how many voters viewed "age" as a reason to vote for or against John McCain. And the survey said: for 42 percent of those not supporting McCain, age was a factor. So what does Jackman say about the so-called Bradley effect and the idea that some people might lie to pollsters?
Jackman: If you can't be for Obama because of his race, you can be for McCain. And you can say that in a survey: "I'm for McCain." Right? And it's not like you feel as though you've got to lie. I mean, being for McCain, that's not a racist, automatically racist, suspect position.
Watt: Jackman believes the country's economic woes will trump "race" in most voters' minds. He takes issue with a recent poll from some of his own colleagues at Stanford that says Obama's race will cost him 6 percentage points on Election Day. He puts the number closer to 2 percent.