>Well, if you wan to do the statistical analysis on "intent" what about the 180K+ voters, in predominately Republican areas of the Florida panhandle (where the vote ran 2 to 1 for Bush) who DIDN'T vote because the news had already declared Gore the winner before the polls closed (different time zone). That would be 60k votes more for Bush, and now its a 40K margin for Bush.

Voter dropout claims are often made, but almost always by partisans without statistical evidence. Bearing in mind I'm arguing as a statistician, not because I'm impassioned for either side, I disagree with you on this because:

1) The people who actually voted are quantifiable. People that never voted are not. As statisticians, we find quantifiable numbers are generally more meaningful, and are considered "hard numbers." Estimation can be made with hard numbers. And this is why the Gore team is still pursuing recounts, they have quantifiable data to work with.

2) There has really been no statistical evidence that early-calling causes voter dropout, in study after study. Voter dropout has been claimed by both parties numerous times, but there is virtually no evidence that voters tune-out (this is due primarily to other contests being on all ballots, in all precincts in the US).

The textbook example is considered to be West Coast Democrat voters supposedly tuning-out in 1980 when Reagan was called early, but no solid evidence has ever been shown to support the idea, AFAIK, despite what appeared to be an obvious case of voter dropout by Dem politicians.

Granted, early calls are still a very bad thing in terms of perception, and supposedly will stop after this. Even if it's only due to perception, it still conjures up the "fairness" issue that some of us have discussed.