Here are some of the things that I think we can still improve on with the TS unit. First, the last screen that comes up when a voter has voted a blank ballot was a definite improvement than not having one at all. The last election that I-Mark ran had an undervote ratio of about 10%. This last election we were down to 2.7%. 33 blank ballots were cast out of 1208. This is still about 1.5% higher than their current lever machine undervote ratio. Sequoia with their Advantage was about .0035%. 5 undervotes in about 1400 voters. 1). I believe if we put the (Attention: you have currently voted a blank ballot) screen after they have touched the cast button, where they are given one more opportunity to go back and then vote the issue or candidate and then touch cast again would decrease the number of undervotes further. 2). Another problem was that the poll workers were accidentally touching the EXIT button in the poll book application, causing the screen to go back to the Icon's. They then had to go back through the procedures of opening the poll book. If we could put the exit button in a place that is not so tempting to touch, and maybe smaller so it doesn't draw so much attention would help. 3). After a voter touches the start button on the instruction screen, some voters accidentally touch a response area on the next page that was not yet displayed on the screen. This gave the voter the perception that there was something wrong with the unit because there was already a vote cast when the screen actually came up. 4). Because the current size of the TS unit, it was very difficult for the poll workers to work with the unit to create and read the reports out of the unit when put in existing punch card voting booths. The disc's were also very hard for them to remove with accidentally pulling out the electrical plug from the unit. Can't we have the electrical plug on the newer units wired inside the machine so we have less chance of someone accidentally pulling the plug. 5). Accumulating the reports from the poll book is something that needs to be addressed. The response from people I trained from the Lake County Elections Office said that they thought it would be very difficult for the normal poll worker to be able to transfer results from disc, and generate reports. The option button is very same as we have it now, and they feel there are too many steps that need to be done before you can generate reports from the machines. Even with extensive training, they feel that there would be problems with most poll workers in transferring and generating reports the way we currently do this.
Overall, the machines functioned very well except for one poll book, which Jeff took out of service and installed another. Leading up to the election was very hairy to say the least. We had our poll workers training on Tuesday, a week prior to the election. It was hard for Bob D. and myself to get very detailed with the poll workers since we had no idea ourselves how the actual screens would look like since the software was still not functioning properly. Thanks to Jeff for finding all the bugs and to Tab for listening to Jeff to solve the problems, I was able to load the new software changes and program to the units of Wednesday, and do a small scale test on each machine