Juan Williams of NPR said that Obama may need a 5% lead in the polls to overcome what has become known as the Bradley effect; white voters saying that they will vote for Obama but actually do not because of the racial issues.
Polling Methods May Tilt Obama Results, Exec Says
Wednesday, September 17, 2008 7:31 PM
By: Ron Kessler Article Font Size
Most polls may be overstating Barack Obamas support by 5 to 10 percentage points because those surveyed may not be telling the truth about voting for him, ad executive Gabe Joseph III tells Newsmax.
It is possible that, when live callers ask for a respondents name, the individual is more likely to say he will vote for Obama, says Joseph, president of ccAdvertising, which conducts polls for a range of candidates and members of Congress. When automated dialers do the polling, respondents are more likely to give an honest answer because they think of the poll as being anonymous.
I believe the traditional pollsters, when they ask your name, institute a bias, Joseph says. We never ask anybody who they are. We do not have to. The only personal identification we obtain comes at the end of the survey when we ask about gender and age.
The difference is apparent when automated pollster results are compared with traditional polls that ask the individuals name, he says. In many circles, particularly within the traditional Democrat base, it is politically incorrect to tell a pollster the respondent will not vote for Obama, he says.
Its very difficult to ask a question about race and get an accurate answer, Joseph says. People are concerned that their name is going to get out. So the only way that weve been able to track it, is to actually do surveys where the respondent is not identified before elections, before primaries. If you do one beforehand, you can look at the difference and you see whats really happening.
The clearest example of the disparity was in the Democratic primary in New Hampshire. The polls showed Obama running ahead of Hillary Clinton by up to 13 points. When the returns came in, Obama had lost by three points to his fellow Democrat.
The phenomenon has been dubbed the Bradley Effect, first identified in a 1982 California gubernatorial election. Los Angeles Mayor Tom Bradley was the supposed front-runner in an open race. Polls showed the African-American Democrat running well ahead of white Republican candidate George Deukmejian. Yet, when the returns came in, Bradley lost by more than 50,000 votes.
To be sure, when it comes to Obama, not every primary election validated the Bradley Effect. But now, in every state ccAdvertising is tracking, Joseph has discerned a 5-point to 10-point difference between the number of respondents who say they will vote for Obama when asked anonymously, compared with results when pollsters request their names.
In ccAdvertisings most recent poll on Sept. 7, the company surveyed 130,000 homes in 13 targeted states. Overall, 48.5 percent of the respondents said they would vote for McCain, compared with 40.2 percent for Obama, a difference of 8.3 percentage points. Of the respondents, 11.3 percent said they had no preference.
The results showed a decrease in support for Obama of 4.9 percent and an increase in support for McCain of 6.3 percent over a previous poll taken before the conventions.
McCain won 11 of the states surveyed with 149 electoral votes: Arizona, Colorado, Florida, Michigan, Minnesota, Missouri, North Carolina, New Hampshire, New Mexico, Ohio, and Pennsylvania.
Obama won two of the states surveyed with 86 electoral votes: California and New York.
Polling Methods May Tilt Obama Results, Exec Says
Wednesday, September 17, 2008 7:31 PM
By: Ron Kessler Article Font Size
Most polls may be overstating Barack Obamas support by 5 to 10 percentage points because those surveyed may not be telling the truth about voting for him, ad executive Gabe Joseph III tells Newsmax.
It is possible that, when live callers ask for a respondents name, the individual is more likely to say he will vote for Obama, says Joseph, president of ccAdvertising, which conducts polls for a range of candidates and members of Congress. When automated dialers do the polling, respondents are more likely to give an honest answer because they think of the poll as being anonymous.
I believe the traditional pollsters, when they ask your name, institute a bias, Joseph says. We never ask anybody who they are. We do not have to. The only personal identification we obtain comes at the end of the survey when we ask about gender and age.
The difference is apparent when automated pollster results are compared with traditional polls that ask the individuals name, he says. In many circles, particularly within the traditional Democrat base, it is politically incorrect to tell a pollster the respondent will not vote for Obama, he says.
Its very difficult to ask a question about race and get an accurate answer, Joseph says. People are concerned that their name is going to get out. So the only way that weve been able to track it, is to actually do surveys where the respondent is not identified before elections, before primaries. If you do one beforehand, you can look at the difference and you see whats really happening.
The clearest example of the disparity was in the Democratic primary in New Hampshire. The polls showed Obama running ahead of Hillary Clinton by up to 13 points. When the returns came in, Obama had lost by three points to his fellow Democrat.
The phenomenon has been dubbed the Bradley Effect, first identified in a 1982 California gubernatorial election. Los Angeles Mayor Tom Bradley was the supposed front-runner in an open race. Polls showed the African-American Democrat running well ahead of white Republican candidate George Deukmejian. Yet, when the returns came in, Bradley lost by more than 50,000 votes.
To be sure, when it comes to Obama, not every primary election validated the Bradley Effect. But now, in every state ccAdvertising is tracking, Joseph has discerned a 5-point to 10-point difference between the number of respondents who say they will vote for Obama when asked anonymously, compared with results when pollsters request their names.
In ccAdvertisings most recent poll on Sept. 7, the company surveyed 130,000 homes in 13 targeted states. Overall, 48.5 percent of the respondents said they would vote for McCain, compared with 40.2 percent for Obama, a difference of 8.3 percentage points. Of the respondents, 11.3 percent said they had no preference.
The results showed a decrease in support for Obama of 4.9 percent and an increase in support for McCain of 6.3 percent over a previous poll taken before the conventions.
McCain won 11 of the states surveyed with 149 electoral votes: Arizona, Colorado, Florida, Michigan, Minnesota, Missouri, North Carolina, New Hampshire, New Mexico, Ohio, and Pennsylvania.
Obama won two of the states surveyed with 86 electoral votes: California and New York.
Race has nothing to do with it. I will not vote for Obama because of his party. If he was white, I still would not vote for him. If McCain was black, I would still vote for him. I wish Condeleeza Rice would run for president , I vote for her in a second. Colin Powell, I'd jump up and down!
