The latest Gallup poll has Kerry up by a point, 49-48.
However, it’s one of those famous weekend polls, which nearly always show weaker positions for Republicans. The reasons for this are based in demographics: Republicans tend to be more rural, have families, and be regular churchgoers – which means that Republicans are less likely to answer the phones and talk to a pollster for 20 minutes. Indeed, the internals of Gallup’s poll and previous weekend poll results have shown the same trends when compared with their other mid-week polls. Hopefully Gallup will run another poll mid-week as a point of comparison.
What this shows is that there appears to be a lot of volitilty in this race, and today’s poll numbers may or may not have an influence on what the actual situation is. Polls are only a tool for political scientists. They’re a scientific fa&ccedi;ade over what is basically an unscientific system – at the end of the day political science is somewhat of an oxymoron – you can’t be scientific about a system that’s inherently chaotic. What you can do is take data and draw inferences.
I still maintain that Bush will win. It’s still going to be close, but I don’t think it will be quite as close as some think. Our ally Down Under just showed that a race between a proponent of the Iraq war and an anti-war candidate that was supposed to have been down to the wire won’t necessarily turn out that way – given the similarlities between Australia and America, it’s only reinforcing my belief that the fundamental tenor of the electorate has changed due to September 11.
As I’ve said countless times before, and will again, the votes are what counts – which is why it remains even more imperative that every Republican work to get George W. Bush in the White House. I’m guessing that Gallup’s poll is off – but every Republican should be campaigning like Bush is 10 points behind.
According to the internals 65% of the respondents saw the second debate. Unlikely to be representative.
Once again, your rationale for disputing the poll data is “Republicans spend 48 hours in church every weekend and don’t answer polls.” This really seems to be a logical stretch. Most people that I know who spend most of their weekends at home are the very salt-of-the-earth types you describe. Most people I know who are not at home over the weekend are party animals, who are more likely to be young and blue-collar, thus more representative of Democratic voters. The reality is that the weekend poll could very easily be lowballing Kerry’s margins. Also, I don’t recall you disputing the validity of Gallup’s post-Democratic Convention poll, taken over a weekend, that showed Bush leading by four points. Was that the one weekend of the year where quaint and pious Republicans decided to play hooky from church?
The correlation between weekend polls and lowered Republican numbers is well established – Ed Goeas did work on it in 2000, and it comes out in NES studies all the time.
No, as if you have a weekend poll that traditional undersamples Republicans that shows Bush leading by four points, it’s usually a sign that his lead is bigger than that.
I’ve done quite a bit of analysis on the polls and have concluded (as have others) that polls done on weekends are more favorable to Kerry. For example, averaging 39 polls in September revealed the following:
1) A poll that included a Saturday showed a Bush lead of 2.8%; a poll that did not include a Saturday showed a Bush lead of 5.6%.
2) A poll that included a Sunday showed a Bush lead of 3.1%; a poll that did not include a Sunday showed a Bush lead of 5.5%.
For Rasmussen polls published September 1 through October 11, polls that include a Saturday show a 1.9% lead for Bush; polls that do not include a Saturday show a 2.6% lead for Bush.