Many questions remain unanswered about the odd and confusing way this year’s Columbia County election results have been counted and reported. A lengthy article in the Register-Star featuring numerous quotes from Republican Election Commissioner Jason Nastke provided some context. But Nastke’s answers didn’t address the fundamental issues which plagued both the numbers and their reporting.
Republican Commissioner Nastke has agreed to further discuss what happened, and this site will report on that conversation. Democratic Commissioner Martin only acknowledged that the Countywide results were published late on Monday, and ignored other questions.
It took the Board just a few hours shy of a full week to report the results of Countywide races and propositions.
With the District Attorney’s race being the most closely-watched contest of this cycle, it’s perplexing that the Paul Czajka-Gene Keeler results were among the last to be disclosed. A week later, the Board’s results finally confirm that Czajka appears to have easily beaten Keeler by roughly 10,500 to 7,900 votes, or 58%-42%. Those numbers don’t include an unknown number of absentees yet to be counted beginning on November 18th.
But even this late reporting throws one more monkey wrench into the gears for voters who just want to know who won. The Board published its numbers as a table with no totals of each candidate’s various party lines.
In order to determine the winner, the public apparently has to do their own math (in the case of Czajka-Keeler) three ballot lines for the Republican and two for the Democrat to come up with a grand total. Other counties’ boards prefer to present both the detailed party line results and a combined total showing the victor.
The confusion began almost from the very start.
The Board threw a first batch of numbers up against the wall raound 10 pm on November 5th, showing only the Election Day machine results for countywide and local results. Meanwhile, numerous news outlets featured results from various precincts as they came in, which were updated up until about midnight.
However, though the Board had “tape” printouts in hand that night showing results from each of the three early voting locations in Hudson, Copake and Valatie, these were not included in those initial reports. Early voting results are supposed to be treated just like Election Day results, but they weren’t.
Making things more confusing, the Board then retracted these limited numbers the next morning, Wednesday the 6th. In an email response to this site, Democratic Commissioner Virginia Martin described that first batch of results as “not correct,” but did not respond to questions about what way they were incorrect. The Wednesday morning numbers appeared to be generally lower than the Tuesday night ones, as if some results had been overstated.
Sources indicate that the Board may have discovered a problem with some machines not having been properly “zero’ed” out before the start of Election Day. Manuals for election inspectors around the State describe a required process before polls open for making sure each electronic machine is starting at zero for each line on the ballot, and not retaining votes from some previous cycle.
Similarly, after being barraged with inquiries from candidates and the press, the Board did begin to release Early Voting results, segmented and not totaled with the Election Day results. The Board then proceeded over a couple of days to publish and retract these numbers several times, before taking them down entirely. Finally, late on Friday night (the 8th), a fuller set of combined Early and regular votes was published, again without giving totals for each candidate.
Complicating things even more, on the first Saturday of Early Voting, which featured a huge crush of voters, it was discovered that hundreds of ballots were not able to be scanned by the machines, and were placed into a separate bin to be dealt with later. The rumored cause of this failure is faulty printing of the ballots on the wrong paper stock, causing the machines not to be able to see them.
Some of these ballots reportedly were “spoiled” and cannot be included in the final hand counts because of overvotes (for example, people choosing three Town Board candidates, when they are only allowed to vote for two) or other issues. Normally, when a ballot is scanned, if such problems occur the machine will reject it, and the voter gets a chance to fix the problem then try again. With the scanner not recognizing the faulty ballots, it appears the normal error-checking was bypassed, and these votes get lost.
The Board knew of the existence of some 700 such ballots, but only brought in a skeleton crew to begin counting them after polls closed. Reportedy two towns were counted that night, Ghent and Copake, but these results were not made known until some three days later.
Election officials are said to have told candidates whose town counts had been finished that it “wouldn’t be fair to other towns” to release those numbers, even though such results don’t affect them. Later, the Board conferred with the State and was told they could, in fact, release the partial results, and did so; but then retracted them again.
That left some 550 unscanned early votes to be hand-counted, but for unknown reasons the counting process appears to have taken even longer than for the initial two towns, dragging on throughout the week.
More updates will be posted as new information comes in.
