Reply #141: Usrename:

[Febble]

Neither I nor OTOH need ANY convincing that more votes were cast for Gore than for Bush in Florida in 2000. On every possible occasion, I link to Walter Mebane's The Wrong Man is President and to to Wand et al's The Butterfly Did It.

The point OTOH is making, however, is that even if all those votes had been counted, the state would still have been too close to call before all the votes were in.

This may seem like a geeky statistical point - but statistical inference is what we are talking about here. The exit poll is a sample. Any inference from a sample has a margin of error, as TIA has been explaining for as long as I've been reading his posts. And in 2000, the actual margin, even if the all those overvotes and butterfly ballot votes had been counted, was within the margin of error of the poll (I'm not talking about voter suppression here, as those voters wouldn't have been in the exit poll anyway.)

So yes - of course - the call for Gore was "correct" in the sense that Gore did, as we now know, receive more votes (whether or not they "would have" been counted had the SCOTUS not intervened). But it was not "correct" statistically. I sometimes do experiments that get the "right" answer but are still not "statistically significant". And if that "right" answer is not "statistically significant" I can't get it published, even if it replicates what other people have shown to be true. To put it another way, the Florida exit polls in 2000 did not have the "statistical power" to call the state, even had the vote count been fair. It should not have been called, ever, until every vote was counted.

The tragedy, I remember thinking at the time, is that because they muffed it, Gore lost the psychological advantage. If it had never been called for either, I somehow can't see even Bush insisting that he was the winner despite loss of the popular vote. So you'd probably have President Gore today.

So my point is: the networks were not "wrong" because they called the state for Bush after they called it for Gore. They were "wrong" because they should never have called it at all. Not because Gore didn't win, but because the race was far to close for the statistical power of the poll.