Hand-Count of Duffy-Heaberg Special Election Ballots in Tredyffrin Finds 61 Uncounted Ballots . . . Changes the Outcome of the Race

A week ago the polls closed, the votes were counted and unofficially Democrat Molly Duffy had won the Special Election against Republican Mike Heaberg by 40 votes.  Chester County Department of Voter Services listed unofficially Duffy receiving 2,266 votes and Heaberg with 2,226 votes.

Immediately following the closing of the polls, there was discussion of voting inaccuracies and talk of machine malfunctions.  We learned first hand from Steve Shapiro, the Judge of Elections at the W-2 precinct, that there were an additional 5 ballots found in the voting machine that were not counted.  There were reports of similar malfunctioning machines at four or five other precincts in the township. As a result, Chester County Voter Services conducted a daylong hand-count yesterday of all 17 voting precincts in Tredyffrin.  Duffy and Heaberg attended the recounting by Voter Services, as did their attorneys and representatives from the Tredyffrin Township Democratic and Republican parties.

Late yesterday, after recounting all the ballots by hand, it was determined that 61 ballots were not originally counted, changing the results of the special election.  Of the 61 ballots found not counted, nine additional votes went to Duffy and 52 additional votes went to Heaberg.  The new unofficial vote total indicates Duffy receiving 2,275 and Heaberg receiving 2,278 . . . a difference of 3 votes, this time in favor of Heaberg. If this total is accurate, it may be the closest supervisor election in the township’s history.

It is my understanding that there was vote count issues found in 12 of the 17 precinctsHow is this possible?  That strikes me as a very high percentage of malfunctioning machines!  However, at this point, it is unclear to me if the problems were attributable only to machine malfunctions or if there were other types of errors.  Who could have predicted that a 40-vote difference in favor of one candidate could change with a recount to favor the other candidate by 3 votes?  What is the probability of that happening?

Again, there is caution that the new special election vote totals are unofficial until certified. Do we believe that the hand-count is accurate and that this final vote count will standOr, will it take 4-5 weeks as previously explained, for the certification process?  Assuming the new hand-count number is correct; will the Democrats challenge the election results?

For me, I’m still stuck on how 61 ballots went uncounted . . . and how many times in past elections has this same scenario played out but may have gone unchecked?  It really makes one wonder.

Bottom line, until there is official confirmation on the special election results, I guess we just need to stay tuned.

Facebooktwitterredditpinterestlinkedinmail

33 Comments

Add a Comment
  1. I think – am not certain — that certification takes place as a whole (meaning for the entire county, not race by race) and even with this recount the certification will still not be for a little while.

    [Reply]

    Pattye Benson Reply:

    I was just told that because the ballots have now been hand-counted, the results will stand and cannot be challenged. This would imply that certification of the special election is but a formality.

    [Reply]

  2. As I said before, tred dems overplayed their hand. They should not have invested everything in a spec election for an indep. What a mess. Tred is now back to rep rule, and they will have very difficult races in nov.

    [Reply]

    Easttown Res Reply:

    yes, would you? tred has not been under all R rule since 1976. there have been 2 dems on the bos. not that you deserve an explanation of my comment, but I meant back to all R rule as of election night (I had thought a D was on the board)

    and, easttown residents are affected by the tred bos, in many ways.

    9-1-1 in case you need it.

    [Reply]

  3. I am completely speculating, but the fact that 85% of the uncounted ballots broke for Heaberg suggests that, whatever was causing the machine malfunction disproportionately impacted Republican ballots. In other words, there may have been some issue with the Republican ballots (maybe a printing issue?) that caused the machines to periodically malfunction when Republican ballots were cast. When the final precinct-by-precinct results are released, I will test this hypothesis by comparing the unofficial number of Republican ballots cast to the official number of Republican ballots cast.

    [Reply]

    give it a rest Reply:

    That is the absolutely logical conclusion. Which is why it was in so many machines. Must have been some alignment/printing issue on a Republican ballot. Thanks for weighing in.

    [Reply]

    Steve Shapiro Reply:

    If, and only if, all 61 previously uncounted ballots turn out to be Republican ballots, then, yes, that is what I am suggesting. I cannot think of any other explanation for how 61 Republican (but no Democratic or Independent) ballots were not counted, if, in fact, that is what happened. If there is a single non-Republican ballot among the 61, you can ignore my theory.

    [Reply]

    give it a rest Reply:

    Less than 2,000 R ballots used that day. As I said earlier, a poll worker told me that while they are supposed to be able to scan with either side facing up, one she was aware of was scanned facing up (as opposed to down) and it didn’t register. They discouraged people from doing it face up going forward. ( I may have that completely reversed, but it seems face down is more private so more likely the way voters would naturally do it).

    So this is PURE speculation….I didn’t see a D ballot and certainly didn’t compare the two. So yes — what if? Only ballots for R, only scanned face up?

    These are distinctions without a difference. Clearly the system is flawed. We do indeed owe Marian Schneider and her cohorts for protecting us in 2005 from going all-electronic, based on the fact that her group believed we needed a paper trail. I understand she was at the recount/hand count. Kudos for HAVING a paper trail. “your” is your signature error. Don’t change it :)

    [Reply]

    Steve Shapiro Reply:

    We are taught that the scanners are designed to accept a ballot no matter how the voter chooses to insert it. So, of the four possible ways to insert the ballot (face up/top first; face up/bottom first; face down/top first; face down/bottom first) all of them are supposed to work. The County technician who came to W-2 during the afternoon of the election said that he surmised, given some of the reports he was receiving, that the problem was occurring when a voter inserted the ballot face down and/or bottom first, and encouraged us to tell voters to insert the ballots face up/top first (in my experience, a large majority of voters insert the ballots face up/top first by default). The technician did not say, one way or the other, that the reported malfunction was only happening with Republican ballots.

    So if John’s information — that all of the uncounted ballots were R ballots — is correct, I guess the theory would go like this: Something about the R ballots caused the machines to malfunction when a voter inserted them face down and/or bottom first. Since most voters insert them face up and top first, the malfunction only effected a small percentage of the ballots. And, yes, the D and R ballots were definitely different. Compare the following:

    http://chesco.org/election/lib/election/sample_ballots/635_2011_rep_primary_specimen.pdf

    http://chesco.org/election/lib/election/sample_ballot/635_2011_dem_primary_specimen.pdf

    The problem is, as John correctly foreshadowed, there is no good way test this hypothesis. There is no way to identify which 61 ballots were not counted the first time (they all just fall into a big pile inside the machine) so it would be impossible to try to rescan the ballots that did not scan on election day. I suppose Voter Services could scan all of the ballots in each of the four alignments to see if they can re-create the malfunction, but it will be impossible to say exactly what happened to any given ballot on election day.

    Is anyone, perchance, going to the Chester County Commission’s meeting this morning at 10:00? I would like someone to raise the issue and see what the Commissioners have to say about it.

    [Reply]

    PC Watcher Reply:

    I too talked to a neighbor that is a poll worker. She said that putting them face down, bottom first resulted in a rejection from what she was able to observe. They can’t handle the votes and can onlyl advise voters on how to insert them. They encouraged people to put the top in first and that seemed to work regardless, though she told me (I think someone else said this) that they encourage voters to put it in face down so that the vote is private and not viewable by any bystander.
    JP — I think the county should insert and scan each ballot again to see if they can replicate the error.

    [Reply]

  4. It may be that the outcome of this particular special election cannot be challenged, but until the question of how 61 uncounted votes could appear, there can be no faith in the integrity of the voting system. This kind of problem is alleged to have occurred at 90 precincts in Chester County. There needs to be a plausible explanation provided by Voter Services and corrective action taken before the November election.

    [Reply]

  5. It will be interesting to see what kind of explanation the County will have for the discrepancy. 61 votes is a lot to be missing, especially during an off year election. I can’t wait to compare the unofficial results vs the corrected hand count.

    Furthermore, it raises the question of how anyone can trust the system if it’s broken. There have been a handful of elections recently that have been decided with less than 100 votes and the speculation of what “could have been” might get some a little rightfully upset.

    If nothing else, it’s going to be fun to watch all of this play out between now and November

    [Reply]

  6. Pattye —

    Have you heard from either campaign’s supporters what is happening? After the election, Molly’s supporters were on this blog touting the victory, but now they have become basically silent. I understand this is all fresh – and probably surprising – information, but I was wondering the take from the main players.

    [Reply]

    Pattye Benson Reply:

    All seems quiet — I think most people are just wrapping their heads around this new information. I have had several phone calls & most people find the idea that 61 votes uncounted as very disturbing. It really does make you question the entire process – past and present elections. There have been several races in recent years that were close and you have to just wonder, don’t you.

    [Reply]

    Steve Shapiro Reply:

    I would be more concerned if we had not uncovered the problem on election night. Mechanical problems probably are inevitable, but they can be discovered and corrected as long as the election workers follow the procedures for reconciling the votes. Here, the discrepancy created by the malfunction was obvious — the numbers simply did not add up. I would be interested in seeing the “return sheets” that the precincts in Tredyffrin submitted to the County. Each effected precinct should have flagged the discrepancy for Voter Services. If they did not, THEN I would start getting concerned about the integrity of the system.

    [Reply]

    Pattye Benson Reply:

    I agree Steve. However, I did ask someone why was it we only heard about 4/5 polls reported problems when apparently there were 12 of 17 with problems. The response I received was that the others did not report problems. Not sure what that means in regards to the process – just don’t know.

    Steve Shapiro Reply:

    As if on cue, I just saw this quote in a story in the Main Line Times:

    “Tredyffrin Township Democratic Party chair Dariel Jamieson said on Monday night that the count uncovered that there had been “61 uncounted votes,” 52 of them for Heaberg, nine for Duffy, and that 12 of the Tredyffrin precincts had had voting machine problems, not two or five as had been reported, but that only two precincts had reported those problems.”

    I find that final point very concerning.

    Source: http://mainlinemedianews.com/articles/2011/05/24/main_line_suburban_life/news/doc4ddbb4c838205935882895.txt?viewmode=fullstory

    Pattye Benson Reply:

    If precincts had problems with the machines, I don’t understand why that would not be reported. As the Judge of Elections for W-2, you reported our machine malfunction. It’s just odd.

  7. Well the campaign workers (from all parties) manning the polling places record the number of ballots that are handed out, so if the machine tallies X ballots but there were X+Y ballots handed out, that would signal the machine is missing some . . . . except for ballots deliberately not cast, i suppose. But the poll workers know how many ballots are handed out and they are recorded . . .. not a lot of oppy for ballot box stuffing with controls like that. . . i hope . . .

    [Reply]

  8. I talked to a poll worker today and she was telling me that it seemed to be about which way the ballots face when they went in the scanner. She said they only had one go in that wasn’t scanned to her knowledge, but that they encouraged people to put them in one way which seemed to work. It was when they went in face up (or down — not sure which was the problem) that there was a read error. I have to conclude as well that there must have been a problem with the Repub ballots to have the 61 misvotes break so dominantly one way.
    Are both candidates satisfied with the result being accurate or will there be a 3rd count?

    [Reply]

  9. !!! How could this happen……….??? Do you think we will ever have all the answers?

    [Reply]

    give it a rest Reply:

    Here’s the likelihood: All Republican ballots — with 52 votes for Heaberg and 9 for Duffy. That’s the only thing that makes sense. So it was the Republican ballot that wasn’t being read correctly. These were special election votes that we are hearing about — so the undervote must have been about Republicans who voted but weren’t in the counts?

    [Reply]

  10. If it turns out that all 61 uncounted ballots were Republican ballots, I would surmise that a printing issue with the Republican ballots caused the scanner malfunctions. That said, I doubt all 61 ballots will all turn out to be Republican ballots given that they broke 15% for Duffy.

    [Reply]

    give it a rest Reply:

    Given the original county numbers, Heaberg got 85% of the Republican voters in the primary (not the special election –which would have cross-over voters?) 1,967 out of 2,311 Republican votes cast. That’s why I think the 61 ballots were likely all Republican. The 85% fits.

    [Reply]

  11. Pattye
    You pointed out early on that 63 voters had not voted in the “special election.” I know I read somewhere that 2 people had voted without using ballots, so maybe there is something relating to the ballots that were not counted in the special election? That would be 61?

    There were 4,555 ballots, 4492 special election votes. Coincidence?

    [Reply]

    Steve Shapiro Reply:

    Regarding the “2 people [who] had voted without using ballots” — I think you are referring to my report that 2 voters in W-2 voted using the electronic machine. It is highly likely that at least a few other voters voted on the electronic machines in the other 16 Tredyffrin precincts. So there should be more than 2 people in all of Tredyffrin who voted without paper ballots.

    [Reply]

  12. Give it a Rest-
    I actually also noticed the same number earlier today and am surprised that someone else saw that. I sure hope someone didn’t fill in all the ballots that didn’t have special election votes with votes after the fact.

    It would be interesting to contact some of the people who had misread ballots and show it to them and verify the ballot was not changed.

    [Reply]

    Steve Shapiro Reply:

    All cast ballots are anonymous — there is no way of linking a ballot to the voter who cast it.

    According to the press reports, the County found 61 uncounted ballots, all of which contained special election votes (52 for Heaberg, 9 for Duffy). That means that, when the final totals come out, there should still be 63 undervotes in the special election from the ballots that were counted in the original tabulation. It would not be possible for someone to just fill in the undervotes without detection.

    [Reply]

    Steve Shapiro Reply:

    Yes, but the ballot number is printed on the bottom portion of the ballot, below a perforated line. Before the voter places the ballot in the scanner, an election worker tears off the bottom portion of the ballot containing the ballot number and gives it to the voter as his or her receipt. If the worker forgets to remove the receipt, the scanner rejects the ballot (because it is too long) and spits it back out. So, when a voted ballot goes into the scanner, it no longer has a ballot number. It has to be that way in order for the vote to be anonymous. If ballot numbers were left on the voted ballots, we would be able to go back and figure out how everyone voted.

    [Reply]

  13. I didn’t mean to imply that — what I meant is that perhaps there were no non-votes? But that doesn’t seem likely given the information since. It sounds (based on Pattye’s W-2 example — where the machine registered 517, the county shows 519, and in fact there were 522 (not sure about these numbers??) cast/ The votes in the system 4,555 should now be short 61 total? The “total votes” will be useful info. But given the hand-count with witnesses, by precinct, I don’t think any conspiracy is likely.

    [Reply]

    Steve Shapiro Reply:

    The math wouldn’t work if there were no non-votes in the special election:

    4555 total ballots reported cast on election night
    4498 votes reported cast in the special election on election night (don’t forget to include the 6 write-ins)

    As of election night, then, there were a total of 57 reported non-votes in the special election.

    During the recount, they found 61 uncounted ballots, all of which contained votes in the special election. So, the final numbers should show:

    4616 total ballots cast
    4559 votes cast in the special election

    If there were no non-votes, the numbers would be:

    4616 total ballots cast
    4616 votes cast in the special election

    and the County would then have to explain how 118 (57+61) formerly-uncounted votes in the special election were split between the candidates. They only have explained how 61 formerly-uncounted votes were split between the candidates.

    [Reply]

  14. I understand that we are all speculating right now as to why the ballots broke so heavily Republican, and that we do deserve an answer as to what happened.

    However, I do have confidence in the result if only because BOTH candidates and BOTH parties were represented at the hand counting (including with attorneys.) With it this close, if either side thought these results were incorrect, they would be raising a ruckus right now.

    That something happened is clear. That the tally is now correct is also clear.

    [Reply]

  15. FTW- Well summarized. Seems like yesterday GIAR and some others had the information, but JP thought it was too random to be true. I’m just glad both sides are okay with the hand count. 3 votes seems too close to be real, but it is what is it. I also agree that we are lucky we had a paper trail. Good job!

    [Reply]

Leave a Reply

Your email address will not be published. Required fields are marked *

Community Matters © 2019 Frontier Theme