Yesterday was the final day for remote early voting in Los Angeles County. In addition to my experiences yesterday at the Pasadena early voting site, I went with a graduate and undergraduate student yesterday to an early voting location in Monterey Park (where we have also observed early voting in the past). I had a few observations to note, in addition to those I wrote about during the past week:

• The ballot was amazingly long, and it does seem to take most voters somewhere between 8 to 10 minutes to cast their ballots. We’ll have some hard data on this sometime in the near future, as some of my Caltech undergraduate students are doing a research project this fall collecting quantitative data on how long it is taking people to vote. But in any case, my ballot was 18 screens — and it took me probably about ten minutes to cast my ballot, check for errors (I did find that I had mistakenly not cast a ballot in a race), and then compare my screen with the VVPAT on the Diebold device. One thing to note about my VVPAT: the paper was slightly misaligned, and as a result one of the races that appeared on the screen was not visible on my VVPAT. Only when I went to the next screen would that final race be visible on the VVPAT (but of course by then my vote on that race was no longer visible on the screen). I wonder how often this sort of problem occurs with VVPAT devices …
• I personally still find it difficult to navigate between the sample ballot that Los Angeles County issues to every registered voter, the summary screen on the Diebold device, and the VVPAT. There are a couple of sources of confusion. First, the sample ballot is laid out for use with InkaVote — it has the same appearance as the vote recorder does for InkaVote users. But that is not what the Diebold summary screen looks like — which leads me to wonder what it might take to reprogram the Diebold summary screen so that it looks more like the county sample ballot? That might make the sample ballot more usable for comparison with the summary screen. Second, the Diebold VVPAT device is located in an awkward position relative to the screen, to the right, and much below the screen. It requires that many voters, myself included, stoop and peer at the paper trail — inconvenient and difficult to use. Third, the way the summary information is provided on the screen, relative to the paper trail, is very different which makes it difficulty to easily and effectively compare the information. This certainly is an area where some good research and redesign could simply the technology and improve the voting experience.
• We saw a lot of “assisted voting”, especially in Monterey Park — both elderly voters, and language minority voters. Clearly, figuring out ways that the technology can be used more easily by such voters, so that they can cast a private and secret ballot, should be a high priority.
• Finally, early voting was very brisk yesterday, both in the morning in Pasadena, and in the late afternoon in Monterey Park. I won’t be surprised to see that McPherson’s prediction of 44% California’s votes are cast by early or absentee voters is correct.