From: Ed Kennedy <ekennedyx_at_yahoo_dot_com>

Date: Fri Nov 26 2004

Hello All:

I spend November 2, 2004, working as a poll worker in a precinct not too far from where I live. The polling place was at an elementary school in the music room behind the stage. During the March elections this was shared with another precinct and apparently didn't work out so well. This time we had the entire place to ourselves in a place that was approximately the size of a classroom. It had two external doors and two doors that opened onto a combined lunch room/general purpose room. I was part of a crew of 5 and we were quite fortunate to have a Vietnamese, Tagalog (Filipino) and Spanish speakers among the crew as the ballots were also available in those languages. Two of the crew including the inspector had worked before and were either current or retired San Diego County employees. One crew member was a high school student. High school students in San Diego County are required to put in a certain amount of volunteer/community service to graduate.

We all pretty much showed up about 0600 for a 0700 poll opening. I and at least two other people there got zero training and we were barely functional by 0700. Unfortunately the inspector blew off a refresher course from that weekend and hadn't studied his manual. All the pictures I see of polling places show nice metal and plastic polling stations with built in lights. Ours were built out of folded corrugated paper. They arrived flat, packed in large sleeves with the instructions printed on the sleeves. A couple of us flailed around and managed to get the 8 stand up polling stations built in short order. In the meantime we were also negotiating with the custodian to fully open the partially opened partition between the two halves of the room. Also we were moving furniture around, laying out paper work and so on all with a minimum of guidance. There was also a handicap station made out of cardboard. Votes were on paper ballots (hooray!) which were then feed through a Diebold optical scanner. What impressed me was that the scanner was held up at operating height by a cardboard structure as well. More on the scanner and station later.

People started lining up outside to vote around 0640. By the time we were barely ready at 0700 we opened the doors and about 20 people came in all at once. It wasn't until we were through the first 25 ballots that we realized that we were supposed to give the voter their ballot with one of the two stubs on it. The inspector was running around putting up signs, two people were working the index, the student was handing out ballots, pens, privacy folders and sending people to the polling stations. I handled running the ballots through the scanner. This went more or less continuously without a break until 1000. After a while, I realized people didn't 'get' the privacy folders at all so I instructed the person handing out the ballots to put them inside the privacy folders. This helped but wasn't a complete solution. Plenty of people handed me their ballots outside of the privacy folders. At first I would shove their ballot into the privacy folder immediately. After a while, I realized I could barely see them with my bifocals on and there was no way I could remember them and the people associated with a particular ballot. So I just let people hand them to me as they pleased. We didn't actually get the sample ballots and ballot marking instructions (in 4 languages) up on all the polling stations until around noon. In the meantime our inspector was struggling with his manual on how to deal with various situations and having a hard time keeping up.

The Diebold optical scanner seemed like a pretty good piece of equipment to my untutored eye on that day. I noted that on the front right there was a memory card slot. A memory card was present and was covered by a metal bar that was held down by what looked like two internal hex head screws. One screw had a transverse ~1/16" hole through it which was sealed with what was characterized as a spring seal. Unfortunately for San Diego, I found it to be broken upon taking the scanner out of its padded fabric case. A roving County troubleshooter ruled that as the tape read zero and the machine was otherwise sealed, that this broken seal didn't matter. On the top of the scanner were two features. On the left was a hinged door secured by an unimpressive 5 pin lock. The inspector had the key for these. Under the door was a small printer and a couple of buttons. The inspector pressed one button and the printer which basically had adding machine tape in it, ran off the contents of the memory card. I didn't get a chance to inspect it but I was told that all 15 races and 25 propositions had printed and showed zero for the totals The inspector and the assistant inspector signed that print out and put it away. On the right hand side of the top was the feed and read slot. Somewhere behind this was a micro-switch or something that detected the presence of a ballot when it was not quite 1/2" in. When the ballot was detected, internal rollers would grab it and pull it through the scanner rather quickly. There was an exit in the back of the scanner. A rather clever trap door arrangement had been built in to the cardboard holder which received the ballot and directed it down ward into a waiting cardboard box. The card board box was sized to fully occupy the cross section of the stand and two paper seals attached it to the inside of the stand so it could not be removed without tearing the seal. The scanner itself was secured with two paper seals to the top of the stand. To get at the 'trap door' and handle jams, one had to break the seals on the scanner to do so. I did that on one occasion. We had a log for broken seals which I maintained. I only found out at the end of the day that there were numbered, foil, numbered, tamper evident seals. On the front of the scanner was a two line LCD display. It gave a running total of the number of ballots successfully ran through the machine. The scanner was 'smart' enough to reject over votes by reversing direction and spitting the ballot back out. It would give a cryptic message on the two line display which included the word, "Over vote." I would quickly scan the ballot and find that it was usually a case where the voter had changed their mind and put an X through an already filled out bubble. In a few cases people had circled their choices or had inadvertently punched through the ballot. I would mark the ballot spoiled on both sides, hand the spoiled ballot back to the voter with a new ballot and get the voter to vote again. We had a special jam clearing page that had a line of dashed across the middle which I had to use a few times. It would be pulled in and pushed back out and worked fairly well. The machine ran on 120v but was supposed to have a 4 hour battery. It was the only AC electrically powered device at the polling place. We also had a Nex-Tel cell phone. Unfortunately this phone didn't handle the text messages it was supposed to and never was replaced despite a request.

In actual operation, I would line up the ballot and get it engaged into the feed slot. Then I would try and get the voter to push it the last little distances to where it engaged the reader. I regularly asked people, "Are you ready to cast your vote?" Then I'd get it ready and tell people, "This is your ballot so a want you to vote. Just give it a little push right here (on the bottom left of the ballot)." At first it seemed a hard concept for people to grasp but after a while most everyone was able to do it. People were impressed by the machine but a few people wanted to know what happened to the DRE's. I generally told them that they hadn't worked out. Most people seemed pleased with that especially when I explained that while their ballot were read right there and written to a memory card, the actual paper ballot was saved as a back up paper audit trail. A few people wanted to know if the machine knew if their vote was their particular vote. I assured them that this was a secret ballot. One person wanted to know if there was any assurance if the machine actually correctly read their ballot. I said no but referred him to Doug Jones' article on these machines http://www.cs.uiowa.edu/~jones/voting/optical/. The privacy folders did not work as the instructions indicated on them. First of all, they were brand new and too stiff to work as apparently designed. Then, once they had been broken in, the scanner kept trying to grab the bottom of the folder and drag it in with the ballot. There was a triangular ridge folded and glued on the top transversely at the leading edge of the ballot holder that kept it from being dragged in as well.

Part 2

As a group we had been quite concerned about poll watchers and exit polling because of all the press about Ohio. The County registrar had printed a flyer about how to deal with them and I was personally briefed on how to deal with them. I guess this has something to do with being big and not having too much grey hair. After all our preparations, a couple of college kids who were doing a scholarly study for someone came by talked with the inspector and one watched us do our thing for 15 or 20 minutes while the other one went outside and did a little exit polling. I don't actually know what the questions asked were or what was being observed. One voter came in and indignantly said that there was a person out there who was ASKING QUESTIONS! We assured him that A. He didn't have to answer any questions and B. We were aware of this person and they were within their rights. After their brief visit, the inspector told the rest of us that the observers had expressed the opinion that we were one of the better run polling places that they had observed. No else ever showed up. So much for press hype.

The County of San Diego provided fairly good support in my estimation. No, I haven't worked polls before, but in terms of other service situations they seemed pretty responsive. First of all, we had a roving support person. He visited our site 5 or 6 times as part of a circuit of polling places he was apparently responsible for. He had some good suggestions and really tried to help us with out problems as they arrived. He carried a cell phone and was linked in with the County Registrar's office. He was the person who discovered that the text capabilities of our polling place cell phone was broken beyond repair. He did ask us for another one but it never showed up. In addition to the rover there was a troubleshooter. When I showed the broken seal to the rover, he called in the troubleshooter who arrived about a half hour later and made the call on the broken seal on the optical scanner. I think he showed up one other time but as of the time I'm writing this (11/26/04) I can't clearly recall. The final layer of support was the provided cell phone. The County Registrar had set up a special phone bank to respond to calls from the inspectors. The inspector told us that his calls were picked up quickly and answers were quickly given.

I'd like to show up the actual ballot but the best I can do for now is this http://www.sdcounty.ca.gov/voters/Eng/vote_opticalscan/election_day.html and this http://www.sdcounty.ca.gov/voters/Eng/vote_opticalscan/absentee.html. As I've related in a previous posting http://gnosis.python-hosting.com/voting-project/October.2004/0216.html. The font size was a little smaller than I liked and the 'oval' was actually a lozenge shaped area. As Doug Jones observed in http://www.cs.uiowa.edu/~jones/voting/miamitest.pdf the instructions were a little unclear. In particular, there was no indication of what the oval looked like before being filled in. I am afraid that it was probably a little unclear that there were two sides to the ballot. To mitigate this, every ten of fifteen minutes, especially when we were busy, I announced that:

a.. The ballot has two sides this time.

b.. Please fill in the ovals completely. No X's, circles, lines, crosses or punch marks.

c.. Please make only one vote per candidate or proposition.

d.. Ballots are available in 4 different languages.

No, I was not directed to do this, but I wanted everyone to vote as well as they could. For the record, the ballot had two columns (I think) and each candidate or issue was surrounded by a box. Graphically, it was a little busy. With 15 candidates and 25 propositions, the ballot was completely filled up. Poll worker attitudes are pretty important in this. I know that I worked pretty hard with the provisional ballot voters to make sure they got all their information down. I told them that I wanted to make as sure as possible that there vote would be counted. Folks seemed to appreciate this.

Traffic (voter) was pretty intense until around 1000. It tapered off for the next two hours. Around 1200 we had pretty well caught up with everyone and we broke for lunch one at a time. Traffic stayed fairly low until 1400. It tapered back up by 1500 it was a busy as ever. By 1700, the place was packed and perhaps 20 or even 30 people were standing outside in the dark waiting to get in. Most people were able to vote in around 10 to 20 minutes. However, some people came totally unprepared to vote and took anywhere from 30 to 45 minutes to vote. A few people came in and only voted on the presidential election. Around about 1900 or 1930 traffic dropped way off and we were able to close the doors at the polling place by 2000, the normal closing time. If they had been a line we were required to stay open until everyone in that line (by 2000) had voted http://www.sdcounty.ca.gov/voters/Eng/vote_opticalscan/voter_bill.html.

The room had two doors. Voters came in the south door and faced a table nearly in line with this door. The voter rolls were there and this table typically had two poll workers at it. At right angles to the end of this table, there were the unmarked ballots, the absentee/provisional voting box and three boxes of additional ballots. At the opposite end of this table, we the actual optical scan ballot counter. Six to ten feet away from this second table and originally parallel with it were 8, standup, cardboard voting booths all in a row. When the rover came in the first time, he suggested that we reorganize the voting booths in groups of 4, two front and two behind. This seemed to work quite a bit better. On the opposite side of where I stood and behind both tables there was a handicap polling station. As this was a classroom, lighting for this layout was good. Voters usually left by the north door. About half way through the day, I realized that we could work in a couple of more sit down polling places at an existing table that was in the room. I put one of the transport sleeves on this table vertically and taped it down to the table and the wall. This brought our polling stations up to 11.

Disabled voters had kind of a mixed experience. During the morning, we didn't have a lot of disabled/elderly voters so the handicap polling places was usually available for them. In the late afternoon and evening, we couldn't really reserve this station for its intended use. Any and all voters used this spot. We had one person in a wheel chair come to vote. We accommodated her by handing the ballot to her on one of the removable tops of our transport boxes and she was able to mark the ballot that way. While she had no formal privacy, she and her husband were sort of off by themselves so it was the best we could do. In addition to that, we had one gentleman come into the room in his electric scooter. However, he preferred to stand at the regular polling places rather than sit down. Finally, we did have one blind voter aided by a relative come in and vote. I didn't know she was blind and cautioned her relative to not interfere with her voting. When he revealed that she was blind, I felt kind of silly. This is one of the things that touch screen voting (with a ballot printer) would have taken care of.

Provisional ballots were handled by means of a special grey envelope. The envelope was about the size of a large greeting card envelope and was printed on both sides. A great deal of information had to be filled out by the provisional voter and they had to sign the envelope. After they voted with a regular ballot I asked them to fold the ballot in half twice and put it in the envelope. For a few people I had to do this. There was a stub attached to the flap of the envelope that I would tear off. Then I generally wet the remaining flap myself and sealed the envelope for them. I asked the provisional voter if they had filled out both sides of the envelope and perhaps a third of them hadn't. After that was completed, I asked them to observe me putting it into the absentee voting box and told them to call the registrar of voters in a few days to see if their vote was counted. Folks didn't seem to have any questions. It is my impression that those people who got provisional ballots were generally young or what has been traditionally called racial minorities. Demographically in California there are no racial majorities.

We were able to close the doors of the polling place at 2000. With very little direction we stumbled through the canvassing process. The inspector ran a special card through the op-scan machine which indicated the end of polling. He unlocked the door to the printer on the scanner and proceeded to print two copies of the totals. The assistant inspector and the inspector signed one of the print outs. During that time the student poll worker and I were knocking down the cardboard voting stations and otherwise cleaning up. The assistant inspector and one other poll worker counter the various signatures in the roll book. Because I was the only one with a calculator and added up the page totals. Also, we opened the absentee/provisional ballot box, sorted the envelopes and counted them. We had 84 signatures for provisional ballots but had 89 actual provisional ballots. Actually we should have had 91 provisional envelopes but some people held onto their envelopes in such a way that I couldn't see them when they presented their ballots to me. I ended up running two provisional ballots through the optical scan machine. The voters would then say something to the effect of, "Oh yes, what do I do with grey envelope?" We tossed those envelopes into the spoiled ballots bag. By the way, even with 601 voters we had only about 6 spoiled ballots. The number of signatures of regular voters was less than the actual votes cast (by more than my misfeed ballots). As a former cashier, I immediately started to try and reconcile the signature numbers and physical ballot counts and was told not to bother. This was supposedly par for the course. The inspector seemed to thing this was just all a bunch of damm fool paperwork. In addition to that we also counted the ballots in the box below the op-scan machine (those agreed with the count on the display) and the number of remaining ballots. I am forcing myself here to reveal that I was disturbed by the fact that the boxes of blank ballots arrived at the polling place with their seals broken. I mentioned that to the inspector who seemed to think it was a trivial issue. Around 2100 three of the five poll workers had left. I helped the inspector to load his car with the voting supplies which he was taking to a nearby church that was a central gathering place. I had expressed some doubts as to the quality of canvassing that we had done and he said that if I wished, I could accompany him to this place and follow up. By this time, I was exhausted and in pain from a flare up of bursitis, so I passed on that opportunity. We both left about 2145. When I got into my car, I realized that we had left the lights on and hadn't told the night janitors that we were done. I took care of that and left the polling place at right around 2200. I jotted down a few notes before I went to bed and we had 113 more ballots than we had signatures. I do solemnly swear that I assembled and sealed the ballot boxes and they were empty at the beginning of voting. We had stubs from the stop of each ballot handed out and we needed to count those. We did include them in the ballot boxes. The ballot boxes themselves were suppose to have 1205 ballots but we never did check on that. I hope to follow up on these discrepancies now that the County Registrar's office has finished counting the write in ballots.

Our training consisted of a couple of hand outs that were mailed to us that seemed to deal solely with handling the voter rolls. There was effectively no training. If I hadn't studied the issues before and if we had not had three experienced poll workers, it would have been a disaster. I would have liked to see the sample ballots and voting instructions printed right onto the cardboard polling stations. It would have been nice if we had rotated positions during the day. We were told to report to the polling place around 0630. If we all hadn't shown up around 0600, we would not have been able to open at 0700. As it was, we were not really ready. I think everything need to be set up the evening before except for the roll books, the op-scan machine and the ballots. It was very chaotic.

Around about 1400 or 1500, we started to hear this terrible sound from the room next to us. It was sort of moaning and shrieking sound. At first I though someone was torturing cats! The inspector went to check on it and he returned to say that is was a woodwind and brass instrument class.