Senate 1 recount reveals problems with NC's election system | Eastern North Carolina Now

    Publisher's Note: This article originally appeared in the Beaufort Observer.

    Our system of government, from the local to the state to the national levels, is a republic. That means, of course, that the essence of the system is that We The People elect representatives to exercise the power of government which we give to them. Thus, the cornerstone of the entire system is that our elections will have integrity. That is not the case in North Carolina.

    The proof of that statement can be seen in the recent Senate District 1 election. Unofficial results reported on election night showed the winner to be Stan White. After all the counting was finished on election day, the county boards of election in the eight counties reviewed the results in what is called a "canvass" and counted the absentee, mail-in and provisional ballots. The winner was announced to be Bill Cook, with 32 more votes than Mr. White. The eight boards certified the results.

    Mr. White called for a recount. The recount was done on November 26 in each of the counties. While the winner did not change, the number of votes for each candidate did change. Cook apparently got 21 votes more than White, according to the State Board of Elections website.

    So the question becomes: Why were the results between the original count and the recount different?

    If one looks at that question honestly one must begin with an assumption: If the number of votes cast and counted was accurate why would the count not be the same, no matter how many times they were counted? Clearly neither we (the public) nor the elections officials have a good answer for that question.

    So do we really know who won? Do we know that the results would not be different if we counted again? The answer is clearly "no."

    Thus, we are left with a situation similar to going to the bank to make a withdrawal from our account. We give a check to the teller for the amount we seek to withdraw. When she gives us the money we count it. It differs from the amount of the check. So we tell the teller and she counts it again. She takes back some of the money she had originally given us and hands us the remainder. We count it again. Still the amount is different, not only from the amount on the check but also from the amount she had originally given us, less the amount she took out.

    So are we content to just take the amount we are now holding and walk out of the bank? Or do we get it right?

    In North Carolina, our elections system says "close enough is good enough, (for government work, apparently)." In effect, Mr. Cook walks out of the bank and into the Legislature.

    Consider some facts that we do know about Monday's recount:

    1. Apparently (we say that because the reports are unofficial) there were discrepancies between the original "certified" count following election day and the recount totals in several of the eight counties.

    2. Apparently (there's that word again) there were more discrepancies in Camden (one of the smallest counties) than in all the other counties combined.

    3. What is known, as of first light the next day, is that the State Board of Elections reports that White got 43,715 votes and Cook got 43,737 according to the recount. More or less. We just don't know for sure.

    But it gets worse.

    We contacted Camden County Director of Elections Elaine Best and asked her: "Why was there a difference in the tallies between election day and in the recount." She could not answer the question. "Put your questions in writing and send them to me and I'll get back to you..." was her response.

    She did get back in about two hours with the numbers. Here's a summary:

    1. There was one vote difference between the original and recount for both White and Cook at the Courthouse, Shiloh, and South Mills (White only) precincts.

    2. The major discrepancies were at the Board of Elections office, where 9 more votes were tallied on re-count in one stop and 9 less for absentees.

    3. Within those discrepancies at the BOE office, White picked up 8 in one stop and lost 3 in absentee while Cook picked up 1 in one stop and lost 6 in absentee.

    4. Provisionals were correct (same between both counts).

    So what happened in Camden? It's hard to say for sure.

    One might conclude from the numbers that 9 ballots got moved from the Absentee box to the One Stop box or vice versa. Nine more were counted in the recount in One Stop and nine less in recount in Absentee. But the changes in each candidate's tallies do not support that conclusion. Something else happened.

    We don't want to speculate about what actually happened. What we know is that at the "end of the day" White had 6 more votes on recount and Cook had 7 less. And most of those changes came at the Board of Elections office. And Mrs. Best has no explanation for how it happened.

    Now, if you've read this far you no doubt have a keen interest in such matters. So let's explain another anomaly about these elections.

    Machines, like people, make errors. But it is not usual for those errors to batch together at just a few locations. One can justifiably assume that if you run a certain number of ballots through the machine one time and then run the same ballots through again you'll get the same results. The system is based on a presumption that it is reliable. Reliability, by statistical definition is: If you do the same procedure over and over again you get the same results. If you don't then either something is wrong with the system or something is being done "outside of the system" to cause the results to differ.

    The way the system is designed, variations in machine reading should be corrected by a "hand-eye" count. That is, if the ballot is not properly readable the machine kicks it out. It should then be read by two people to determine the voter's intent and those "hand counts" added to the machine read totals. Viola. The variations are eliminated. Camden County says the hand counts are included in their totals so that means the machine errors should have been corrected.

    But in Camden County once you correct for reliability you are still left with an internal validity inconsistency. The problem is that you can't eliminate the possibility that some ballots were changed between the original count and the recount.

    We can't prove there were changes (or replacements) but the managers of the system cannot say with validity that the system did in fact prevent such changes. What the system can say is what changes took place and where. That is what we have.

    We don't know if there was anything done wrong in Camden County. What we do know is that the system itself did not work the way it needs to work to insure the integrity of the electoral system.

    And that is unacceptable, regardless of who won or lost the election. We all lost.

    The Observer sent a reporter to observe the recount in Pasquotank, because Pasquotank had the most unusal voting patterns in the November 6 election. We should have gone to Camden. But you may be interested in the recount process in Pasquotank that can be viewed in the video clips below.

    The first clip is the startup process:



    The next clip has random shots during the day of counting.



    The third clip is at the end of the day as the totals are put together:



    Editor's note: We allowed Ms. Best to review this article before posting. She made one correction, which is included herein. We also offered to post her explanation about how Camden's results could have been so much different than all the other counties, but she declined to do so. That offer still stands.

    We believe "close enough is good enough" is unacceptable in elections. They should get it right or explain why they didn't get it right. In fact, we believe the law should be changed and an Elections Board required to issue a public disclosure of all irregularities and a discrepancy in a recount should be included as an irregularity. If you run the same ballots through a machine more than once and you get different results then those ballots should be counted by hand and two observers (one of each party) agree on what the voters' intent was.

    And the State should tighten the "chain of custody" procedures used to protect ballots from the time they are cast until the final certification. Based on our observations we would contend the security of the ballots is far too lax in some places. Modern technology should be employed, including video recording and electronic security.

    Finally, we don't buy the argument that "well, the winner didn't change..." is sufficient. Election results should be correct. Period. Regardless of the margin of victory.

    Moreover, the role of observers, including the press, should be refined. Procedures should provide for demonstrable disclose of critical aspects of the process.