To help us better understand our voting procedure, Steve Shapiro, the Judge of Elections for the W-2 precinct, has graciously provided us detailed notes below. Steve’s explanation offers us the step-by-step guide to the voting procedure and to the close out at the end of an election day. As I read his notes, it is clear that Steve takes extra steps to protect the voting system process and to insure the quality and accuracy of election results. I know that all that read Community Matters, join me in thanking Steve for his thorough explanation . . . a great community service. Thanks Steve!
Judge of Elections
W-2 Precinct, Tredyffrin Township
Here is an overview of the voting process and the end-of-night procedures:
Before we open the polls, we set up the paper ballot scanner (the M-100) and the electronic machine (the iVotronic). We print out a tape from each showing that no votes have been cast on either machine.
During voting hours, we look up each voter in the Poll Books. If we find them, we have them sign the Poll Book. We also record their name in a different book, called the “Numbered List of Voters Book,” which contains a separate numbered line for each voter. The Numbered List of Voters Book allows us to easily determine, at the end of the day, how many voters signed-in and received a ballot (either paper or electronic). If a voter chooses to vote on a paper ballot, we record the number of the ballot next to his or her name in both the Poll Book and the Numbered List of Voters Book. If the voter chooses to vote electronically, we so indicate in both books. Here a sample of the Numbered List of Voters Book from our poll worker manual, click here.
After a paper ballot voter makes his or her selections, he or she takes the completed ballot to the M-100. An election worker removes the strip at the bottom of the ballot that contains the ballot number, and gives it to the voter as receipt. Now that the ballot number has been removed, the ballot becomes anonymous (it cannot be traced back to the voter) and the voter places the ballot into the machine.
If there are no problems with the ballot, the M-100 processes it and drops the ballot into a locked compartment for safe-keeping. If there are any errors on the ballot that prevent the scanner from reading it, the M-100 displays an error message and returns the ballot (at least, that’s what it’s supposed to do). If there are any over-votes on the ballot (for instance, a voter votes for 3 candidates in a race where he only may vote for 2), the M-100 will display an error message and ask the voter whether he wants to either accept the ballot as-is (thereby invalidating the vote in the race in which he over-voted) or take the ballot back and correct it. Here is the summary of the possible M-100 scanner errors from the poll worker manual – click here.
If a paper ballot voter makes a mistake and needs to correct his or her ballot, we issue a new ballot, change the ballot number for that voter in both the Poll Book and the Numbered List of Voters Book, and keep the spoiled ballot so we can account for it when we close the polls.
An electronic voter makes his selections on the iVotronic’s touchscreen and presses the “Vote” button when he or she is done. The iVotronic will not let a voter over-vote or otherwise mess up his ballot, so we never have to deal with spoiled electronic ballots.
At the end of the night, the first thing we do it process the absentee ballots. The County delivers absentee ballots to us in sealed envelopes with the voter’s signature on the outside of the envelope. We first check the Poll Books to make sure that the voter did not vote in person and that the signature on the envelope matches the signature in the Poll Book. If everything checks out, we open the outer envelope, and remove the sealed inner envelope that contains the ballot. We then mix up the sealed inner envelopes to anonymise them before opening them, removing the ballots and scanning them into the M-100.
Next, we close the voting machines. For the iVotronic, all that entails is a few presses on the touchscreen to lock down the voting function on the machine. We then print off a closing tape that shows the total ballots cast and the number of votes each candidate received. The tape shows only totals — it does not show how each voter voted. We also remove the data card on which the votes are electronically recorded.
The M-100 requires more effort. First, we print out a tape that shows the total ballots cast and the number of votes each candidate received. Then we remove the data card on which the votes are electronically recorded. Next, we open the machine and remove the ballots. The M-100 deposits any ballots with write-in votes into a different compartment. We review those ballots and hand-write the write-in votes onto a tally sheet. We then add the ballots with write-in votes to the rest of the ballots and count all of them by hand to see how many ballots we have (in a primary we have the added step of separating the ballots by party before we count them).
Once the hand count is completed, we fill in a “General Returns of Votes Cast” form. Here is a link to a sample General Return from the poll worker manual:
The form is basically a worksheet that allows us to ascertain whether we have accounted for all of the paper ballots. We add up the number of paper ballots cast, the number of spoiled ballots and the number of left-over ballots. That sum should equal the number the ballots the County delivered to us at the beginning of the day. We also add the number of paper ballots cast to the number of electronic ballots cast to come up with the total number of ballots cast. If, during this process, the numbers do not add up, you are supposed to flag that for Voter Services by writing a note in the “Remarks” box on the General Return.
At this point, it is my practice to check the total number of ballots cast on the General Return against the following to make sure they jive: (1) the number of voters listed in the Numbered List of Voters Book; and (2) the number of ballots cast as reported on the machine tapes. That is how we discovered the problem last week — the number of ballots cast on the General Return did not match the number of ballots cast on the tapes. I concede that neither the General Return nor the poll worker instructions direct us to make that comparison (maybe it should), but it seems like common sense to me. In addition, all poll workers take an oath before we open the polls, and I believe that checking all of the numbers for discrepancies is consistent with the part of the oath in which I swear to faithfully perform my duties to the best of my judgment and ability. A copy of the oath from the poll worker manual is here.
After we discovered the problem last Tuesday, we made a note in the remarks box on the General Return. I also emailed Voter Services and copied a representative of each party (that is not required, but it seemed like good practice).
As for the poll workers’ signatures, even if the numbers on the General Return do not add up and we leave a note for Voter Services, we still have to sign the General Return. So we are not certifying that the numbers are correct; rather we are saying that the reported numbers are what we counted. Likewise, although we have to sign the tapes from the machines, we are are not certifying that the numbers on the tape are accurate. We could not possibly know, for instance, whether the M-100 scanned the votes properly. Rather, by signing the tapes we are saying that these are the tapes we printed from the machines.
Finally, we place the General Return and the data cards into a small pouch, which we seal with a security tab. We place the paper ballots into a large pouch, which we also seal with a security tab. We then drive both pouches, along with all of the other election supplies, to Voter Services in West Chester. The unofficial results you see on election night are taken, I think, from the data cards (the tapes are just a print out of the information on the data cards).
3 CommentsAdd a Comment
Just as I did in my first post yesterday, Steve left out provisional ballots. I updated this in a later post. We had none of these in M6.
The M100 is supposed to separate ballots with write-ins from those without and place them into separate bins. This is not reliable, and we visually check each ballot. Write-ins are tallied by hand.
We did observe a difference between the Numbered List and the machine during the day, and the difference increased during the day. As Steve noted, there’s nothing poll staff can do about it, beyond reporting it on the General. Return sheet.
Steve, Thanks for a great written report on what is done when the polls close. This is what I have observed @ my poll in the past.
I will second your comment about matching the number of votes cast against the number from the list of voters book. It should definitely be included in the manual from voters services.
I believe these precinct election officials should be discussing the matter with voter services. It sounds like its less about the procedures and actions of the judges and more about those responsible at the county level for assuring that the machines and ballots are functioning as they should be. My biggest concern moving forward is how can the county assure ALL voters that future elections will not have a cloud of suspicion. It seems to me that some sort of failure occurred at the voter services level.