In article <[EMAIL PROTECTED]>,
Drake R. Bradley <[EMAIL PROTECTED]> wrote:
>With regard to the number of ballots rejected from the Florida (and
>other) precincts -- most of these were because either two holes or no
>holes were punched for the candidate. Selecting two holes might reflect
>confusion about which hole goes with which candidate (e.g., Gore vs
>Buchanan); selecting no holes might reflect the decision not to vote or
>an attempt to vote that wasn't successful (e.g., the chad didn't punch
>through, the filled-in circle wasn't scanned, etc.).
>Many of the currently rejected ballots would go away if a "no vote"
>option was available. That way, a voter who doesn't want to vote for
>president (or some other office on the ballot) could indicate this
>explicitly by selecting the "no vote" option (i.e., the voter can select
>Bush, Gore, or No Vote). To catch the two most common errors, each
>ballot would be scanned electronically before the voter leaves. A
>validity check would detect any double-punch or no-punch errors, and the
>voter given a chance to correct the problem either by filling out a new
>ballot or by modifying the present one. Once a "valid" ballot is
>obtained, it can be scanned for the actual votes cast. All of this needs
>to be done so as to preserve anonymity, or course, but that should be
>possible.
The present punched card ballots are not scanned electronically
at the polling place; there is no such electrical equipment there,
in general. Also, many of the errors do not admit of correction.
>With an explicit No Vote option and validity checking for double-punch
>and no-punch errors, an error that could still occur would be one where
>the punch reader (or optical scanner) fails to sense a selected response
>in one location (say for Gore) and at the same time erroneously reads
>one in a different location (say for Bush). Such an error would not be
>caught by the double-punch or no-punch error checking, and the voter's
>vote would be tallied for the wrong candidate. But how likely is this
>kind of error?
More than you think. I have voted in severl elections with
such a device, and one needs a fair amount of care. I check
my ballots after voting, and I am somewhat surprised at how
few do, or even know how to do so; I have never seen
instructions on this.
And in any event, this kind of error is not one which
>produces a rejected ballot, so it does not contribute to the large
>number of rejected ballots identified in Florida. (Such an error would
>only be caught on a manual recount, or perhaps on a second electronic
>scan with a different scanner.)
>It seems a shame that so many ballots are rejected. An inexpensive and
>workable system like the one suggested above could hopefully reduce
>these numbers substantially.
It is far from inexpensive; it is workable. The current
equipment costs a few dollars per device. The one suggested
would at least multiply the cost by an order of magnitude.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
