- Joined
- Oct 17, 2007
- Messages
- 11,862
- Reaction score
- 10,300
- Location
- New York
- Gender
- Male
- Political Leaning
- Centrist
Two thoughts on the polling:
1. Even as he had enjoyed a strong reputation prior to the election, my view is that Mr. Newhouse made elementary mistakes that a high-caliber pollster should not have made. His underlying assumption was that the composition of the electorate would not resemble 2008 and would, instead, more closely resemble the 2004 vote. How he made that assumption in the face of ongoing demographic change (growing Latino and non-White share of the electorate) is remarkable. Demographic change in the short- and medium-term is highly predictable. To not factor in demographic change is perhaps as basic an error as one can make in any surveys. There's no other way of putting it, the error was an elementary one no serious pollster should make in the face of demographic change. A simple regression analysis of the Census data and evolving turnout would have shown that the Newhouse assumption was not a realistic one.
2. One would reasonably have expected that internal polls would have been questioned when they consistently provided different results than the multitude of public polls. I suspect confirmation bias precluded such questions and deprived the campaign of possible time that might otherwise have been available to refine its message.
1. Even as he had enjoyed a strong reputation prior to the election, my view is that Mr. Newhouse made elementary mistakes that a high-caliber pollster should not have made. His underlying assumption was that the composition of the electorate would not resemble 2008 and would, instead, more closely resemble the 2004 vote. How he made that assumption in the face of ongoing demographic change (growing Latino and non-White share of the electorate) is remarkable. Demographic change in the short- and medium-term is highly predictable. To not factor in demographic change is perhaps as basic an error as one can make in any surveys. There's no other way of putting it, the error was an elementary one no serious pollster should make in the face of demographic change. A simple regression analysis of the Census data and evolving turnout would have shown that the Newhouse assumption was not a realistic one.
2. One would reasonably have expected that internal polls would have been questioned when they consistently provided different results than the multitude of public polls. I suspect confirmation bias precluded such questions and deprived the campaign of possible time that might otherwise have been available to refine its message.