|
... is being fraudulently characterized as dispositive rather than diagnostic.
Let's try to be clear. Ballot tabulation systems, when properly designed, will tally a vote for a specific candidate when the 'certainty' is very high. In other words, careful and responsible design will try to get a 5-nines (i.e. 99.999%) or greater degree of assurance that when a ballot is tallied for a specific candidate that a intense and unbiased manual forensic examination of that ballot will agree with the system's tally. In order to achieve such a low error rate for a tallied ballot, the system will "reject" (fail to tally) ballots where anomalies and variances are detected by that system.
The probability that what a system calls an "undervote" (just a system diagnostic message) is, in fact, an undervote (no vote in a particular race) is far, far lower. The system is not designed, nor should it be, to be error-free in such a diagnostic, since it's presumed (in system design) that a manual examination of the ballot will be dispositive.
When any such automated system "rejects" (fails to tally) a ballot, it does so with a 'diagnostic message' (more relevant to the system than the ballot) that goes a bit further than just 'rejected.'
This is where the rhetorical characterization of the automated system's behavior departs from the reality.
Depending on the ballot medium (punch card, mark sense, etc.), an "undervote" will, under manual examination, have a 2% to 10% likelihood of being a valid vote. The tighter the system tolerances in tallying a vote for a specific candidate, the higher the probability that a rejected ballot will, in fact, be a valid vote when carefully and objectively examined.
Example: Suppose a voter using an optical scan ballot circles the name of the candidate rather than fills in the box adjacent to that candidate's name. The system will reject (fail to tally) that ballot and will characterize the rejection as an "undervote". It's NOT actually an undervote. The voter's choice is clear and unambiguous! It's a vote.
Confusion arises, however, since the rejected ballots are only examined and tallied manually (under most state election procedures) when the count of such votes will potentially/possibly change the material outcome of the election. When the outcome would not be changed by a manual count of 'rejected' ballots, no such manual count takes place and the results of the election will be (over-)reported with the 'undervote' as indicated by the tabulation system.
|