* *** POSTMORTEM_2008-11-04 TXT - 7 Nov 2008 02:43:16 - JKNAUTH Postmortem on 11/4/08 Election at Precinct 01-43 ========== == ======= ======== == ======== ===== General Observations ------- ------------ Given the possible load problems expected for this election, the voting at 01-43 went more smoothly than I had expected. We had an early morning burst with possibly around 100 people in line waiting to enter at 6:30. I think this was somewhat smaller than in 2004; it certainly seemed less chaotic than in 2004. Perhaps the fact that it was raining reduced that opening queue. After that the flow was fairly steady thru the rest of the day, but light at the end of the day. Often thruout the day, particularly in the morning, we had most poll booths in use and sometimes overflowed onto the tables we had set up on the stage. The weather was pretty miserable all day -- cold and wet; that might have held down the turnout somewhat. However the biggest factor was certainly the heavy use of early voting in the preceding two weeks. That seemed to work very well. The BOE did a great job. I am sure all the election day pollworkers are very thankful for the reduced load on election day. For our precinct, as well as we know now, there were 1003 ballots submitted by absentee/early voting. In contrast only 819 ballots went into our M100 on election day. We also had 40 provisional ballots. I think there were probably around 2800 registered voters for 01-43 on 11/4, including a guess about how many might have registered after 10/1/08; this may well be an underestimate. Hardware and Supplies -------- --- -------- 1) Our most significant problems were in the hardware area. Initially the diverter in the ballot bin was not operating correctly, throwing most things into the right-side bin, or even trying to "stack" them in the non-existent middle bin, i.e., on top of the diverter. We did not know this was happening. That caused three jams in the morning. We switched to emergency bin use whenever a jam interrupted things. We called the BOE and a repairman was sent. After he made the repair, which he was able to do pretty quickly, there were no similar problems for the rest of the day. The apparent cause of the morning difficulties was that a pin had been sticking out incorrectly and was blocking the proper diverter movement. Possibly the problem cause was that there was a procedure problem during the M100 setup. The mechanical setup looked OK, with the M100 well seated back onto the diverter plug. However after the initial boot from OPEN/CLOSE POLLS, the key was accidentally turned to OFF instead of to VOTE. Probably this was because the key/lock was very hard to turn and the lettering was hard to see. The key was then turned to VOTE. The machine seemed to do a full reboot and then a diverter error message was displayed. We selected CONTINUE. A good zero tape was produced and the normal 0 voters showed on the display. We called the BOE to tell them what had happened and to get confirmation that things were OK. They said yes, so we didn't do anything else. Many initial ballots were subsequently read with no errors. However all this may have left the diverter in a bad state. Three jams later occurred and were cleared by removing each jammed ballot and then reentering it. Once the repairman reset the pin later in the morning, everything worked normally. We also fed in about twenty or thirty ballots from the emergency bin after the polls closed. All worked normally. The POLLS CLOSING procedure also worked normally. The modem connected on about the ninth try. We just let it keep trying until it connected, as in previous years. The three jams and diverter malfunction possibly resulted in a miscount of some ballots, as well as mixing the writeins and non-writeins. We put notes on the voted ballot box and reconciliation form to warn the BOE about the problem. Presumably the BOE will rerun the ballots for 01-43 thru a good machine to verify or correct the counts calculated by the scanner on 11/4/08. Hooray for using paper ballots, which allows such a recount. 2) There was an ink cartridge error during poweron of the AutoMark -- a should-not-occur since the ink cartridge was supposed to have been installed and tested prior to delivery. It turned out that the ink cartridge had come loose from its socket at some point. It was reinserted (with difficulty) and that fixed the problem. In any event, as usual, all day no voter ever used the AutoMark to actually print a ballot. No one has ever used the AutoMark in our precinct since the device was introduced in early 2006. 3) It's good that the BOE told us for this election to count ballots as soon as a ballot pack was opened. Three packages of the nine opened had only 99 ballots instead of 100. We also had one case in which two ballots were firmly stuck together and given to the voter as a single ballot. This is something we always have to watch out for; however this seemed to be a worse case than usual and was not detected by the usual procedures. The voter also thought there was only one ballot and marked the front of one and the back of the other as if this were only a single sheet. The M100 detected the double thickness and kicked it out. The two bad (stuck together) ballots were then marked as spoiled and a new ballot was then given to the voter. 4) The curbside radio was not particularly effective. We thought the bell used last spring had worked better. One problem was that the curbside voter might not hold down the button after pressing it; then whatever they said would not not be heard. Maybe having an inside bell ring when the the button was pushed would help in case the user did not hold down the button when speaking. Also, when we tested the device, the spoken words were hard to understand. In any event, for this election we had enough pollworkers so we had someone stationed out near the curbside almost all the time and really didn't stress the device's capabilities. The bell device had worked well in the preceding elections when no pollworker was at the curb. 5) The provisional labels we received at pickup were defective -- twenty duplicate sheets. Our coordinator brought us a replacement set on Monday and we had no problems with duplicates on election day. We did misplace one label, however, and had to write the number on the provisional envelope. 6) In general the Help Table notebook PC worked well. However sometimes it slowed down to a crawl, particularly when doing a series of backspaces. Our coordinator told us the same problem had been reported by some of the early voting centers. A reboot of the PC didn't correct the problem, at least not for long. 7) One of our poll booths collapsed. I didn't hear why. No one was injured. The booth was broken and we didn't try to reuse it. Procedures and Unusual Situations ---------- --- ------- ---------- 1) The yellow ATV forms were used a great deal and again worked very well for us. They were used only for unreported moves into our precinct. As it turned out, we didn't have to use any for data collection prior to database searches. The person doing pre-search preparation used paper from one of the small pads supplied by the BOE. We used the yellow ATVs only when we had to handle a definite unreported move. In fact we had several cases in which we received a yellow ATV carried in by the voter from another precinct. The PO in the other precinct realized they were filling out that form by mistake for an out-of-precinct transfer and just sent the voter to us with the partially completed yellow ATV instead of with the white ATV and label. We then handled it as an unreported move, just as if the voter had showed up on our doorstep to begin with. Of course we carefully checked the partially filled in yellow ATVs and completed them just as if we had done all the writing at our precinct. 2) We had many transfers into/out of precinct using the normal white ATV procedures. No problems were encountered. 3) The person who handled provisional ballots sometimes used the street book to look up precinct numbers if the PC was tied up. However those numbers were then double checked with the PC when it became available shortly afterwards. 4) The person who handled provisional ballots misunderstood the instructions in the manual and wrote the voter's residence address and associated precinct number on a few early ballots, not just the precinct number of the residence address. Later ballots were handled correctly. 5) I don't think I ever put a circle around the corrected ballot style on ATVs beside the crossed-out ballot style. I realized this only the next day. In any event, our precinct had only one ballot style to give out. I don't know if the circle had any use other than to highlight the change for the Ballot Table. 6) We had many cases in which a person said they registered via the DMV or Internet, yet they weren't in the database. For those who had registered within the last few weeks, we said the data entry at the BOE probably just had not caught up yet with the huge number of recent registration requests. However some people said they had registered many months ago. In all cases, we just had them vote a provisional ballot. 7) We found some people in the database, but not in our pollbooks although they should have been. In most cases this was probably related to timing of printing of the pollbooks vs. when the database copying was done. 8) We had a case in which a registree showed a voter card she had received within the last month, yet she was not in the database or the pollbooks. We called the BOE, but they could not find her either, not even in the history section. We issued a provisional ballot. This certainly sounds like exactly the type of thing provisional ballots were invented for: capture the data on election day and let the BOE research it later. 9) Of course we had many "out of county" voters. Most were people who had moved from some other county to Wake County and said they "had registered at the DMV", often long ago. They said they were asked at the DMV if they wanted to register and said yes. But the registration seems never to have been done. They each got a provisional ballot from us. I guess the good news is that this will at least be a second chance to get them registered, and also fix some known name misspellings and address changes. We also got people who had never lived in Wake County, but happened to be currently working here for a short time. They said they wanted to vote even though they knew they weren't registered. More provisional ballots for you to sort out. We didn't turn away anyone. 10) We also had a few people who stopped in to get registered, but didn't try to vote. They were given the appropriate registration request form. 11) We had several instances in which people were shown in the database to be registered in another precinct. However they now lived in our precinct and swore they had voted normally at 01-43 (i.e., they had been in our pollbooks) the last time they had voted, some time ago. I think this may have just been faulty memory. Anyway, we handled them as unreported moves to get them properly shifted into our precinct since they now lived here. 12) We had one case in which a voter found she had been marked in the pollbook as an absentee voter even though she said she had not voted. We confirmed in the PC database that she was flagged as an absentee voter (red background highlighting). However, we also saw in the database that there was another person (unrelated) with the same first and last (uncommon) names as our voter. That other person was not marked as having voted. We theorized that the other person had actually voted at an early voting site, but by mistake had been given the virtual pollbook sticker for the woman who lives in our precinct. Then the other person had not noticed the address mismatch on the printout she received at the early voting site. We called the BOE about the problem, but meanwhile had the voter use a provisional ballot. 13) After encountering the above problem, we wondered when absentee ballots were flagged in the PC database. The PC correctly showed that I had voted on 10/16/08. However when we checked several other pollworkers who had voted early, their votes were not yet shown. One said she had voted about a week before early voting ended, i.e., about a week after I did. I don't know if she voted at the same place I did. It sounds like the database cutoff for recording early voting was over a week before 11/4. Written by Jeff Knauth, Judge Reviewed by Elaine Williams, Chief Judge Frances Massey, Judge Sam Sneider, Ballot Table