[Dooby]

Just tossing some analysis of the Newsweek poll from various sources:

Richard Baehr at The American Thinker has completed a thought-provoking analysis of the Newsweek poll. Short version, their sampling is flawed Longer version:

The Newsweek poll indicates that by a 61%-19% margin, registered voters in the survey believe that Kerry won the debate. The survey also indicates that 74% of registered voters in the survey group watched the debate. Stop here. Approximately 62 million people watched the debate on television. There are approximately 135 to 140 million registered voters in the country. So fewer than half of all registered voters saw the debate, but 74% of Newsweek’s survey group did. Three networks conducted traditional post debate telephone surveys on who won the debate. Kerry won the debate according to all three surveys, but by margins of 9 to 16%, and ratios of 3 to 2 or 4 to 3. The Newsweek survey group thought Kerry won by about 4 to 1, and by 45%.

Does this sound like a representative sample to you? Did Newsweek merely interview the Washington press corps or the CBS News team?

The Rasmussen survey interviews 1,000 people a night. Their survey showed no change in the results from Friday night (after the debate) from the the two previous nights. Only 6% of all those interviewed by Rasmussen on Friday night indicated that the debate would have any effect on their candidate preference- 3% more likely to vote for Kerry, 2% more likely to vote for Bush and 1% now more undecided.

The Democracy Corps, the Democratic polling operation of Stanley Greenberg and James Carville, also has a new post-debate poll that shows Bush is still ahead, but now by 2% instead of 4% as he was two weeks ago. This probably means Bush is ahead by more than 2% of course, given the pollsters’ bias (this poll has consistently been more favorable to Kerry than almost all other national polls throughout the year). The internals on this poll also show higher job approval ratings for Bush after the debate than before. Similarly two networks found that while more debate viewers thought Kerry won the debate than Bush, the margin between Bush and Kerry in the election preference of these same viewers did not change after the debate from before. Bush remained ahead in both cases. The Los Angeles Times found a 1% shift to Kerry among those who watched the debate, and the paper, to its credit, clearly states that this group which watched the debate is not a representative national sample of the electorate.

You can bet that the Newsweek survey will be a big story the next few days. Kerry is ahead (the nation is safe). The race is certainly not over. Bush clearly did not get a lift from the foreign policy debate that his camp might have expected. The Kerry camp, which was in despair, however, has gotten a major morale boost. But I still think Bush is in better shape than Kerry with just over four weeks to go.

UPDATE: The blog Political Vice Squad says the Newsweek poll is skewed from its earlier one. They decreased Republican sampling by 5 percentage points and increased Democratic sampling by 6 percentage points. Furthermore, on the first of the three nights, the poll was limited to the "Pacific and Mountain time zones." In other words, registered voters from the following states completely were excluded: Texas, Florida, Ohio, Pennsylvania, Missouri, Louisiana, Arkansas, Kentucky, Indiana, and the entire old South.