September 5, 2004
ET TU, RASMUSSEN?:
52% Believe Bush Will Win in November (Rasmussen Reports, September 3, 2004)
Fifty-two percent (52%) of Americans now believe that President Bush and Vice-President Dick Cheney will be re-elected this November. A Rasmussen Reports survey found that just 38% expect the Democratic ticket of John Kerry and John Edwards to emerge victorious.
Those numbers reflect a big change from the last time we asked the question. Following the Democratic Convention, data released to Premium Members showed that 47% of voters expected Kerry to win while 43% took the opposite view. [...]
The national telephone survey of 1,000 Likely voters was conducted by Rasmussen Reports September 3, 2004. Margin of sampling error is +/- 3 percentage points with a 95% level of confidence.
A lot of folks on the Left seem to be clinging to the Rasmussen and Zogby polls (as well as to building ledges) which show the race to be closer than TIME and Newsweek have it. But the fact that the Rasmussen numbers have hardly changed in the daily tracking poll is pretty hard to square with this huge swing.
Posted by Orrin Judd at September 5, 2004 4:07 PM
As I said elsewhere in the comments today, the tracking poll numbers are likely to change once the results of the GOP convention are fully digested. We're still getting in the results for the polls that were taken during and just after the con. Once people have had a couple of days to absorb everything, then you'll really see things start to pop.
And, presumably, 48% believe not.
Faith is so useful
Strangely, a non-negligible percentage of voters treat elections like horse races: they want to pick the winner. The evidence for this is in post-election polls, which consistently find that a greater percentage of respondents say they voted for the winner than actually did. So, combined with the fact that the Dem vote is more anti-Bush than pro-Kerry, this is bad news for them. Additionally, it may mean that many on the left give up on Kerry and vote their conscience (Nader or Green).
Of course, it could also mean that Reps get complacent, but this year, I don't think so.
I follow Rasmussen regularly and have since 2000. Numbers that include a holiday weekend are often flaky. The main thing with Rasmussen is to watch the trend over a period of time, not the horse race snap shot. If you graph the last three weeks or so, Bush is up 5 points, and that is the first sustained move by either candidate of this magnitude since Super Tuesday.
Rasmussen's numbers don't make much sense. He currently shows Bush with a 53% approval rating but only 48% in the tracking poll. Really? Kinda hard to believe in such a big divergence between those numbers.
Rasmussen bombed so badly in the 2000 presidential race. He called a big win for Bush and really looked bad when the actual results came in.
I have to think he's always shading his tracking poll numbers toward a wishy-washy dead-heat position. That way he can't be caught too far out.
Zogby is a self-identified liberal Democrat who released some, shall we say, odd numbers in the 2002 midterm elections. He had Bush behind by seven in mid-August, at exactly the same time Gallup was showing Bush three or four up. His latest two-up for Bush is a big swing from his previous numbers.
Polls are fun, no? And as the ancient cliche warns, only one counts.
Both Time and Newsweek polled registered voters, and made no attempt to find out who really intended to vote. Rasmussen says Republicans were over-represented in both polls. Rasmussen's rolling averages may lag a few days, but I'd have more confidence if his poll was showing Bush up by 11 points.
But I like the "who's going to win?" numbers, very much. This has the potential of demoralizing Democrats everywhere... meaning Bush might not just win, he might win with coattails.