Replies

I disagree with Rasmussen's analysis of possible outcomes, even leaving aside my problems with his methodology of using automated phones rather than live human beings to do his polling. He weights his polls by party affiliation based on the 2000 election, in which Republicans underperformed (turnout was 39%D, 35R, and 26I -- as if the millions of dollars that the RNC has spent the last four years on revamping and upgrading their GOTV model and four years of incumbency in the White House isn't going to have any positive effect on Republican turnout this year. Sorry, that dog won't hunt.

Here is how I see it.

1. If the turnout numbers mirror 2000, either because the Republicans once again screw the pooch in GOTV or because the Democrats improve their operation by as much or more than the Republicans do, then Bush is ahead nationally by about two points and the election will be close. This is what Rasmussen is showing, as well as other polls who use the 2000 turnout numbers, including Zogby and the ABC/Washington Post tracking poll. Even here, Bush may do slightly better than two points as long as his job approval rating stays at or above 50%.

2. If the Republicans poll even or within one point of the Democrats, Bush will win by about five to six points. The Battleground Poll assumes that turnout will be even, while Time's poll assumes that Republicans will be within one point of the Democrats -- and both show Bush leading by five points. Gallup also has Bush winning by five points in the poll they released earlier this week. I consider this outcome to be the most likely outcome on Tuesday, which is why I think Bush will win by about 52-46.

3. Last, if the Republicans improve their GOTV as expected, but the Democrats don't do as well -- a very distinct possibility this year -- it will be a blowout for Bush. The chances that Bush will blow out Kerry are not as good as the chances that he will win comfortably by 5-6 points, but I think they are much better than the chances that Republican turnout will be as bad or worse than in 2000.