Democratic Primary Study
As anyone reading P&N regularly knows, polls showing Obama with a lead are regularly dismissed as a result of the so-called "Bradley Effect", which says that people will SAY they are voting for a black candidate when talking to pollsters to sound more PC, but will actually vote for the white candidate. Obama's poll leads have long been called misleading, both in his primary against Hillary Clinton and the current race against McCain, as a result of this effect.
However, the Bradley Effect theory is largely based on Tom Bradley running for governor of California in 1982, where he lost despite a steady lead in the polls. Not only has it been 26 years since that election, and not only is this a national election compared to a statewide one, but one case hardly makes for a good theory.
The study I linked to is a very good argument against the Bradley Effect, almost going so far as to demonstrate a "reverse-Bradley effect" in this election, at least in the Democratic primary. Almost everyone said during the primary against Clinton that Obama would do worse than the polls showed, yet in only 3 of the 32 primaries studied did Obama do worse than the polls predicted. In 12 of them (mostly in the Southeast), he actually did significantly better.
At the very least this suggests that Obama is not affected by the Bradley effect, and it goes a long way toward disproving the effect as the absolute law of political science some people seem to believe it is. But I also find the theory as to why there seems to be a "reverse-Bradley effect" with Obama in some areas of the country.
The explanation for the Bradley Effect assumes that the culturally correct response that voters think pollsters want to hear is that they support the black candidate. Alone in the voting booth, they are free to vote for whoever they want, but talking to another person they want to give the "right" response. But despite all the complaints about how everyone is too PC these days, what if the "right" response in some places isn't the PC one?
California was one of the 3 states where Obama did better in the polls than in the primary, which fits with Bradley's results there and also fits with the stereotypically PC nature of the state. The states where Obama's primary results were BETTER, on the other hand, were mostly located in the southeast, a place famous for being pretty un-PC, especially in regards to race. The theory put forward in this study is that folks in places like Georgia, where the error between poll and primary was 18 percent, think the socially correct answer is that they are voting for the white candidate.
Of course this could all be a bunch of crap, just like applying the Bradley Theory to the current race, but it's an interesting idea...and it raises some interesting questions about how racially PC we really are in this country.
As anyone reading P&N regularly knows, polls showing Obama with a lead are regularly dismissed as a result of the so-called "Bradley Effect", which says that people will SAY they are voting for a black candidate when talking to pollsters to sound more PC, but will actually vote for the white candidate. Obama's poll leads have long been called misleading, both in his primary against Hillary Clinton and the current race against McCain, as a result of this effect.
However, the Bradley Effect theory is largely based on Tom Bradley running for governor of California in 1982, where he lost despite a steady lead in the polls. Not only has it been 26 years since that election, and not only is this a national election compared to a statewide one, but one case hardly makes for a good theory.
The study I linked to is a very good argument against the Bradley Effect, almost going so far as to demonstrate a "reverse-Bradley effect" in this election, at least in the Democratic primary. Almost everyone said during the primary against Clinton that Obama would do worse than the polls showed, yet in only 3 of the 32 primaries studied did Obama do worse than the polls predicted. In 12 of them (mostly in the Southeast), he actually did significantly better.
At the very least this suggests that Obama is not affected by the Bradley effect, and it goes a long way toward disproving the effect as the absolute law of political science some people seem to believe it is. But I also find the theory as to why there seems to be a "reverse-Bradley effect" with Obama in some areas of the country.
The explanation for the Bradley Effect assumes that the culturally correct response that voters think pollsters want to hear is that they support the black candidate. Alone in the voting booth, they are free to vote for whoever they want, but talking to another person they want to give the "right" response. But despite all the complaints about how everyone is too PC these days, what if the "right" response in some places isn't the PC one?
California was one of the 3 states where Obama did better in the polls than in the primary, which fits with Bradley's results there and also fits with the stereotypically PC nature of the state. The states where Obama's primary results were BETTER, on the other hand, were mostly located in the southeast, a place famous for being pretty un-PC, especially in regards to race. The theory put forward in this study is that folks in places like Georgia, where the error between poll and primary was 18 percent, think the socially correct answer is that they are voting for the white candidate.
Of course this could all be a bunch of crap, just like applying the Bradley Theory to the current race, but it's an interesting idea...and it raises some interesting questions about how racially PC we really are in this country.
Facebook
Twitter