I think Paul's idea of eliminating punch cards is probably a good one. But, this is
really only a problem with large voting districts. The error rate is about 32 out of
1000. Usually, the error is an undervote, i.e. somebody voted, but it was not
counted. For small districts, it would be rather easy to examine the ~4-5% of
undervote ballots. Florida counties with punch card problems have two
characteristics that are preventing a quick resolution: 1) they are rather large and
2) they didn't examine the undervotes in the original count or the state-law mandated
re-count; it's only in the third count where they are considering them, which is what
is so disturbing.
Paul Thompson wrote:
> snip
>
> There is another message, probably more important, to obtain. These methods for
> voting are simply prone to error. Punching holes in cards has been abandoned in
> every segment of data acquisition save voting. It is too easy to make several
> forms of errors. As such, we must really ask: Is it time to eliminate punch card
> voting methods? I believe that the answer is patently obvious.
>
> What then should they be replaced by? The system should be cheap, flexible and
> verifiable. I believe that the best system is optical scanner methods. Optical
> scanners are stable. They are fast. They are based on real things - pieces of
> scanner paper. The ballots can be quickly examined to see that they do not have
> double counts.
--
WWLD? (What Would Lombardi Do?)
Was "Name That Tune" Rigged?
Rodney Sparapani, Duke Clinical Research Institute
For addressing and schedule information
see http://www.duke.edu/~spara002
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
