"Peter Lewycky" <[EMAIL PROTECTED]> wrote in message
[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
> How many may have wanted to vote for Buchanan but got Gore instead? I
> don't have the all the stats at hand but Buchanan got something like
> 3000 votes. Is it the claim by the Democrats that all these votes
> should've gone to Gore? There were 19000 spoiled ballots due to double
> marking this year. In 1996, 14800 ballots rejected for the same reason.
> Where's the evidence that this year's ballot was more difficult? The
> butterfly ballot was used previously and is in use in other
> jurisdictions including Cooke county Illinois. The comments by Daley are
> hypocritical and unworthy. Buchanan's Reform party membership increased
> from 1996 to 2000 by 110% in Palm Beach county. Comparing 1996 to 2000,
> there's no evidence of widespread error or fraud.
Oh come on. As statisticians, we can't and don't (usually) make
claims with absolute surety. However, it was a nice bit of quick
data analysis to show how anomalous the vote was. For all I know,
Palm Beach County is the Pat Buchanan stronghold of the world.
But, in a greater sense, it is not up to us as statisticians to
assess who was in the "right," or to assign blame, or especially
to recommend remedies. It *is* perfectly within our sphere
of expertise to demonstrate that something was odd, to speculate
about mechanisms that might explain the oddity, and to estimate
its size. Personally, I think that Adams has jumped the gun on his
regression estimates of the size of the oddity. But that doesn't
vitiate the simple analysis that shows that something was,
indeed, very odd. I agree that there is no evidence of widespread
fraud. However, even taking into account growth in Reform
Party membership in the county I think it disingenuous to believe
there's no evidence of *error*.
--Robert Chung
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
