On November 2, 2004 I spent the day as a poll worker in a Pittsburgh voting district. Pittsburgh is part of Allegheny County, which votes on lever voting machines. Having studied voting for over a decade, I had a pretty good idea of what to expect. However, I was surprised at how many procedural deficiencies I observed.

I decided a while ago that I wanted to be a poll worker for this year's election. However, as a new resident of Pittsburgh I wasn't sure how to go about volunteering. I examined the Allegheny County elections web site for information, but was unable to find any. Evenutally, I called the Allegheny County Elections Division. The gentleman who answered the phone was unable to provide any information about becoming a poll worker, but promised to leave a message for a woman named "Sue" to let her know that I was interested. I didn't hear anything about becoming a poll worker until six days before the election when Sue called me to tell me she needed someone to fill in for a poll worker who had to go out of town. Sue hurridly gave me the information about what I would have to do, where I would need to go, what time I would need to report (6:30 am) how long I would need to stay (probably until about 9 pm), and how much I would be paid ($95). Then she told me that there was an optional poll worker training class downtown on Saturday, and that I would be paid $5 more (perhaps enough to cover the cost of parking) if I attended it.

So, I went to poll worker training, eager to learn more. Poll worker training turned out to be a 90 minute session in which no questions were permitted to be asked until the end. We watched a slide show narrated by an experienced poll worker, who then showed us how to setup a voting machine at the front of the room. I was sitting in the back of the crowed room and wasn't able to get anywhere near close enough to actually see anything the presenter was doing to the machine. But I did learn that setting up the machine involved attaching a crank handle to the machine and cranking it approximately eleven times, that this cranking and the subsequent stowing and locking of the crank handle was critical, and that the failure to follow the instructions in the correct order could make it impossible to setup the machine.

Tuesday morning I got up bright and early, packed a bag lunch, and headed to my assigned polling place. I didn't know where to park or where exactly I was going, so I arrived a few minutes late. The district to which I was assigned shared a polling place with three other districts in the basement of an elementry school (the only part of the school that is handicapped accessible). In previous years, the four districts were each assigned a seperate classroom, but as the school enrollment had increased the school had fewer spare rooms available. This year, the only space suitable for voting was the school's small cafeteria. So, four districts, complete with eight large lever machines, four tables, and 20 poll workers were packed into the cafeteria. The voting machines had to be placed close enough to wall outlets to be plugged in. The tables were then placed adjacent to the each district's voting machines. Ideally, voting machines would be placed in a "secure" area behind a table, so that only poll workers and voters who had already been checked in would have access to the machines. However, such a layout was impossible in this small room. Voting machines were placed in front of tables and voters ended up lining up right next to voting machines.

When I arrived, I found the election judge and three other members of my district's election board. The board included a retired couple who had worked many elections. They were in the process of setting up the machines when I arrived. I watched them crank one of the machines a few times, unlock the doors, and declare the machine ready to use. I remembered from the training that a "zero sheet" was supposed to come out of the back of the machine to show that all the counters were starting at zero. The sheet had emerged partially from the back of the machine, so I pulled it the rest of the way out. However, the sheet had nothing printed on it. Something wasn't right. I remembered the training presentation about the crank, and the fact that the machine had to be cranked approximately eleven times, and that there were still a number of other steps that had to be done before the machine would be ready to operate. So I told the election judge (who had never been an election judge before) that I didn't think the machines were setup correctly, and we found the instruction sheet and attempted to figure out which steps still remained to be completed. As it turned out the crank handle had been attached to the wrong part of the machine, and so the counters had not been engaged at all. We were able to get one of the machines setup successfully, and printed a proper zero sheet. But there was no way for us to put the blank zero sheet back into the other machine, and thus no way to verify that the other machine was now properly setup. I also remembered from the training session that 50 mechanics were on call throughout the county to handle any sort of mechanical problems with the machines. So I called in and asked that a mechanic be dispatched. By this time it was 7 am, time to open the polls. Our district was not quite ready. We had one working machine, but still hadn't finished hanging our signage or setting up our table. But the other three districts were ready, so they opened the doors and voters started coming in.

The next two hours were completely chaotic. Voters entered the cafeteria and tried to find their district. Some knew their district numbers and only needed to find the right table. Others had missed the signs in the hallway listing which street addresses were assinged to which district, and did not even know what district they were supposed to be looking for. Voters lined up wherever they could find space. Around 7:40 our mechanic arrived and got our second voting machine working. The ballot was fairly short and most voters were able to vote in under a minute (about the same amount of time it took us to check each one in), so having only one working voting machine for most of the first hour of voting turned out not to be too much of a problem.

The poll worker training had left me somewhat confused about the proper procedure for checking in voters, but the retired couple seemed to know how this process worked, so we all followed their instructions. One pollworker found each arriving voter's card in an alphabetized card box. We split the registration book pages into two halves, and one of us took each half. A judge from another district later told us that we were supposed to keep all the registration book pages together in a looseleaf notebook and not let the voters see them. Apparently the split book was an old procedure that had been changed when the format of the registration book changed. After a voter signed the card we looked up the voter in the book, checked his or her signature, and asked for an ID if they were a new voter. We then removed the stub from the card and handed it to the voter, and sent the rest of the card back to the box. The poll worker who pulled the card from the box then wrote the voter's name down on each of two logs. The voters took their stubs and wandered around trying to figure out what to do next until one of the poll workers intercepted them and took them to a voting booth. The poll worker ripped the stub in half, gave one half to the voter and deposited the other in an envelope taped to the side of the booth, and pressed a button on the side of the machine to close the curtains and enable the machine for voting.

By 11 am there were no more lines. Voters were still coming in at a fairly rapid pace, but we were able to catch up on filling out forms in between voters. At this point I discovered that nobody had recorded the initial readings from the protective counters (the counters that record the total number of votes ever cast on each machine). I seemed to be the only member of my board who was even aware that these numbers were supposed to have been recorded.

Later when the rate of voters entering the polling place had slowed even further, some of the poll workers in the other districts suggested that everyone start counting the absentee ballots so that we would not have to do it after the polls closed. There was some discussion about it, and everyone agreed it would be best not to do it while the poll watchers were there. Some poll workers saw no reason to delay counting the absentee ballots and were intent on doing it as soon as the poll watchers stepped out. However, there always seemed to be at least one poll watcher present so they did not manage to get started. This turned out to be fortunate when an absentee voter showed up at one of the precincts in the early afternoon and asked to void his absentee ballot so he could vote on the voting machine.

During some of the slower periods I read through the various sets of instructions that had been provided in our district "suitcase." We were not provided with a pollworker manual. Instead, there were several sets of instructions, each dealing with a different part of the process (setting up the voting machine, counting absentee ballots, etc.). Nowhere in the instructions was I able to find a document explaining clearly the responsibilities of each pollworker (we each had a different title, but none of us knew exactly what that meant). Our judge spent a lot of time reading through the documents that pertained to counting the absentee ballots and closing the polls. But he nonetheless missed a number of steps. Of course, even if there had been clear instructions, it is not clear that all of the poll workers would have bothered to read them.