It was reported last week that the St. Lawrence County Board of Elections had certified impossible numbers for the special election in New York’s 23^rd Congressional District.  Just hours before certification, I received from the Board, in a .pdf file, by e-mail, their results for each election district (precinct).  These contained, for six election districts (four in Canton, one in Massena, and one in Oswegatchie), more votes counted for the candidates than the reported number of ballots cast and, therefore, negative numbers for “blank ballots” or “undervotes.”  These negative numbers actually appear, in a computer printout, in the .pdf file.

The author has since learned that only the cumulative results for each contest, not the results for each election district, are actually certified to the State.  Thus it is not strictly correct that these negative numbers appeared in the certified results.  But they did appear in the district (precinct) totals from which the certified cumulative results were derived.

In response to the article, the Board of Elections, as always, has provided the data needed to analyze the situation.  Only the numbers for “total” ballots cast and for “blank” ballots have changed for the six districts in question; the vote totals for the candidates remain the same.  For comparative analysis, the Board has provided a breakdown of how many voters in each district voted in which manner (machine, absentee, affidavit, or special federal).

A comparison of these data confirms that there were enough “machine” voters (actual voters at the polls) to account for the vote counts for the individual candidates in these six districts.  This is what my audit of the poll books had shown, as reported in the original article.  Therefore, it is the “whole number” of voters listed on the computer printout in the .pdf file that has to be wrong, and the data newly provided by the Board of Elections confirms it:

• In Canton’s 2^nd district there were 143 votes cast on the machine, plus 8 absentees, for a total of 151; there is no way that the “whole number” could ever have been 134.
• In Canton’s 4^th district all 50 votes were cast on the machine; there is no way that the “whole number” could ever have been 46.
• In Canton’s 6^th district there were 241 votes cast on the machine, plus 17 absentee ballots and one special federal ballot, for a total of 259; there is no way that the “whole number” could ever have been 231.
• In Canton’s 7^th district there were 172 votes cast on the machine, plus 9 absentee ballots, for a total of 181; there is no way that the “whole number” could ever have been 148.
• In Massena’s 14^th district there were 367 votes cast on the machine, plus 12 absentee ballots, for a total of 379; there is no way that the “whole number” could ever have been 341.
• Only in Oswegatchie’s 2^nd district can the “whole number” be explained, as there were 254 votes cast on the machine and, in the elections other than the Congressional race, there were 20 absentee ballots, for a total of 274, which happens to be the “whole number” that turned up in the Congressional race (although it should have been 280, as there were 24 absentee ballots and two special federal ballots cast for Congress).

Thus, neither the “whole number” of ballots cast, nor the negative numbers for “blank votes” or undervotes, e-mailed to me by the Board of Elections just hours before certification, can possibly be explained in five of six cases.  No vote counting system should ever produce negative numbers.  If negative numbers for “blank” votes are allowed by the computer program, “phantom votes” can be entered into the count, and this is a grievous flaw in the system.

I have never suggested that the certified vote counts for the individual candidates are incorrect in any of these six districts (precincts).  Having not seen the ballots at the polling places, before they left the public view, I cannot say one way or the other.  All I am saying is that the electronic vote counting system utilized in much of the 23^rd Congressional District is not to be trusted because it is capable of producing negative numbers, and that no election results produced by such a system should be certified by the State.

For the record, negative numbers for “blank votes” appeared on the computer printout in the .pdf file not only for the Congressional election, but for five other contests also:

• In Canton's 7th district, for Proposal Number One
• In Canton's 7th district, for Proposal Number Two
• In Canton's 4th district, for Mayor of Rensselaer Falls
• In Canton's 4th district, for Town Clerk
• In Massena's 14th district, for Town Supervisor