RE: MORE Questions from election officials

From: John Payson <jpayson_at_circad_dot_com>
Date: Tue May 04 2004 - 20:20:04 CDT

Case one is handled because we support recounts of the paper ballot,
and we reconcile the computer records against the paper ballots
against the precinct voter counts. Doing the reconciliation in the
precinct helps and allowing inspection of the process is also useful.

To my mind, "recounts" are of very marginal value. If there are any systemic
problems, recounts--whether by machine or by hand--will do little to shed light
on them (suppose a handcount yields a different result from the machine
count--which is right?) Having a record of how particular paper ballots were
recorded is IMHO a major key to detecting and diagnosing tabulation problems
and would make recounts much more valuable.

For example, suppose that a stack of 50,000 ballots yields 30,123 votes for
candidate X and 19,573 for candidate Y when it's run through one machine; when
run through another machine, it yields 30,126 for candidate X and 19,571 for
candidate Y. Which count is right?

If the ballots are assigned ID's, then it would be possible to find out which
count is right by identifying the ballots (at leat three, but possibly many
more) that were scanned differently on the two machines. This list of ballots
could then be fed into a machine that was programmed to feed through ballots
until it found an ID on the list and then either set the ballot aside or stop
(so a human operator could set it aside). Inspecting the particular ballots
where the machines disagreed with each other would thus allow one to determine
which count, if either, was accurate, and also whether further analysis was
required (if, for example, there were 500 ballots where the two machines
disagreed, every single one of them was marked for candidate X, and yet the two
machine's totals were within a few votes, it shouldn't take a brain surgeon to
realize that neither machine should be trusted).

Has anyone here looked at the benefits of providing tracking ID's to ballots
(with the caveats that they should not be visible to voters, and that tracking
ID's should not be able to match up people's votes in different races)?
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Mon May 31 23:17:13 2004

This archive was generated by hypermail 2.1.8 : Mon May 31 2004 - 23:18:15 CDT