Auros (auros) wrote,

  • Mood:

The polls are closed...

First note that turnout was clearly far higher than the registrar planned for. At 7am when we opened, there were six people lined up outside, and from then til about 10:30am, there was never a time when there wasn't somebody waiting to sign in. Then there was an hour lull, then another rush between 11:30am and 1pm, then a lull until about 4pm when the evening rush started. When we closed the line to new entries, at 8pm, we still had 28 people waiting to vote. Throughput for the machines, when everything was running smoothly, was about 1 voter every 2 minutes (with five machines -- so each person was taking about ten minutes to vote). And a lot of the time, things weren't.

We discovered the night prior to the election that one of our five machines was non-functional -- the lights were on indiciating it was receiving power, but you'd turn the power switch and nothing would happen. So it was completely out of service. When we opened at 7am, our Card Activator didn't work. This meant we had to actually put a couple of machines into the dreaded "manual mode", in which a ballot can be started simply by pressing the big yellow button on the back of the machine. We only switched two machines over to manual, because that way we could stand guard to make sure that nobody was trying to vote twice. Around 8am, the field inspector brought a new activator, and a bit later (maybe 10am?) he brought a replacement for the dead machine (which came with a printer, bringing us to seven total; we started out with five on the voting machines and one spare). One of the printers went wonky fairly early in the day (I really don't remember when -- Xta says around 4pm, but I could swear it was much earlier than that -- like, towards the tail end of the morning rush; in any case, there are notes on it in the Technical Exceptions Log, so the registrar's office will know; yay for paper records), and it was swapped out by our head clerk for the spare we had started with; he never provided a satisfactory explanation as to what went wrong with that one, and it's possible it was fixable, but at the time Xta and I were too busy dealing with voters to take a look at it.

Around 4-5pm, we started losing printers fast -- one of them jammed irreparably, and, ultimately, three ran out of paper. At about 7:15, our field inspector brought in one more printer (our eighth). If you've been doing the math, you've figured out that we ended up with five printers dead; we were down to two working machines, then back up to three for the last less-than-an-hour before the polls closed, and the 35 minutes or so it took to process the people still in line at that point (many of them on paper -- the paper was faster than the touchscreen machines, especially since some people were coming in with the sample ballot already marked).

Because of the situation with the printers, we had to start asking that as many people as possible use the paper-ballot option. The way this worked was that people would pry open the staples at the center of a sample ballot, take out the paper ballot included in there (cheap, newsprint-flimsy paper), and mark it as their ballot. This would then be sealed in an envelope and dropped in the ballot box. My understanding is that these will not be run through the optical-scan machines. but rather get transcribed by a human being into some more permanent form (either a real cardstock paper ballot that can go through the scan machines, or into a set of touchscreen machines with printers that can serve as repository for all the paper ballots).

Because we were concerned about running out of printer paper entirely -- at which point, anyone who did not have their sample ballot would simply be unable to vote -- we had to encourage even people who had some non-standard marks on their ballots to use that option. If the ballot was still clean, they could just use the standard blank ink bar, but if they had some marks like question-marks, exclamation points, etc, we asked them to do their best to make sure their intent was clear -- write the name of their choice in the margin, if necessary. I still think the whole transcription thing is sketchy, and introduces the possibility of human error, or intentional mischief with partisans or finicky bureaucrats trying to disqualify votes even where the intent is clear (like the victory-margin votes for Donna Frye in San Diego that were disqualified because either the bubble next to the write-in space wasn't filled in, or her name was mis-spelled as "Fry"). I called in to the non-partisan Election Protection hotline to let them know about the situation, and I'm hoping the county party will be able to send observers as well.

Oh, also, we almost ran out of the envelopes for the paper ballots! We were down to one left, and were considering cannibalizing our spare absentee-ballot envelopes (they're supplied so that if somebody comes in with a filled-out absentee ballot but had lost their envelope, they can still turn it in; we figured we could write on them that they were in-precinct paper ballots, not absentees) when the field inspector brought the eigth printer, and also gave us another twenty-five paper-ballot envelopes.

We need to ditch the touchscreen machines. Without the printers to provide a voter-verified, recountable record, they're clearly unreliable. But even with the printers, you're faced with the problem that stuff breaks too easily -- the power can go out, the printers jam, etc. I want us to just go back to using the optical-scan ballots for everyone. The technology of "pen on paper" is a lot more reliable than these gizmos, and it's less expensive to just have a big stack of extra paper optical scan ballots, than to have a bunch of extra computers to cope with the fact that one out of every six will break. These voting machines are being built, and certified, by people who have no frickin' clue what it means to do proper QA -- if anyone tried to sell ATMs at this level of quality to a bank, they'd get laughed out of the room. And honestly, it doesn't take that much longer to count the optical scan ballots, and the error rate on optical scan is, according to the studies I saw in the wake of the 2000 election, very low -- on the order of 0.1% (compared to at least 1% for punchcards). We could even bring in something like the InkaVote Plus (a machine which records its actual ballots as ink-marks on paper, but lets visually-impaired voters listen to an audio system, make selections, and have the machine actually print the marks; I don't see how the voter could verify that the ballot was marked correctly without help from a sighted person, but the exact same problem exists on the touchscreen-with-printer, or even with a basic paper ballot being penned in by a not-so-trusted assistant). Heck, the Australians, with their 90+% turnout (they fine you if you don't show up!) manage to get by with writing checkmarks (or numbers, for the "preference vote" IRV races) in boxes, and hand-counting the ballots. It takes a little longer, but at least they don't have to turn people away from the polls!

Turnout at our precinct: I'm estimating that there are about a thousand active voters in College Terrace. (There are ~1100 names in the book, but there are a fair number of duplicate entries, and entries that I know to be people who've moved out but have not yet been moved to the yellow "inactive voter" pages.) We had 484 people show up to vote at the polls, plus 103 absentee ballots turned in at the precinct. Add in the absentee ballots mailed in (we have about 350 registered absentee voters in our roster, and I know I've heard from the party that statewide, more than 70% of absentee voters do vote -- that's why we always try to get people to register as Permanent Absentee, during registration drives), and I'd estimate overall turnout in our precinct was at least 60%, and probably closer to 75%. (484 + (350 * 0.7) ) / 1000 = 73%.

According to the SecState election retruns data, statewide turnout was ~45%, and Santa Clara County was at ~50%. MacPherson's office had been estimating 55% statewide turnout ahead of the election, so clearly turnout actually fell short of expectations. That makes it especially pathetic that we weren't provided enough materials to efficiently deal with the turnout we got, even if, clearly, College Terrace is way more engaged than the average for our county. (It's remotely possible that my fellow precinct-captain and I actually boosted turnout by some non-trivial amount; College Terrace had turnout 10-15 points higher than the state average in the primary. I was highly gratified to see at least a dozen people come through clutching the endorsement sheet I had left at their door. I wish I had access to turnout numbers by precinct for the past several elections... I know the RoV does have that info -- the parties buy it and use it in trying to choose target precincts. In fact, info is also available on what elections each individual voter turned out for -- your ballot is secret, but the fact that you voted is not.)

Anyways. Despite losing the Governor and Insurance Commissioner races (as expected), we swept the rest of the statewide offices (I'm especially pleased with Bowen and Garamendi, both because I like them, and because the Republican candidates were seriously bad news), we got Jerry McNerney in CA-11 (!!!), we held off the takings-clause proposition and increased our margin against the parental-notification proposition (relative to the special election version), the local party won all three Mountain View City Council seats that were up for grabs... Sadly, we just barely missed on Measure A, and got beaten fairly badly on 87 and 89; I'm hoping that in all three cases, we'll be able to get something through the legislature (county Board of Supervisors, and the CA state leg) in the next couple of years -- though I suppose Arnie would veto Clean Money. Sigh.

Also, as you might've heard, we had a very good night, nationwide, at the federal level, and in quite a few states -- we won a bunch of key gubernatorial races, and every one of our target Secretary of State races, which will be critical in fighting GOP cheating in the '08 presidential cycle.

Direct money given to candidates and party committees: about 10% of my income for the past year.

Walking eight local precincts (the three that make up College Terrace, plus five more in Mountain View, either with endorsement sheets and other literature or GotV reminders), plus three out in CA-11, plus phonebanking another half-dozen precincts: about 10% of my waking hours for the past six months (~12 hours a week, on average?).

Working the polls to make sure that my neighbors don't get disenfranchised by kludgy technology: one full day out of my life.

Making Nancy Pelosi the first female Speaker of the House, and Harry Reid (almost certainly *knock wood*) the Majority Leader of the Senate: Priceless.

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded