Here is some interesting news from Spokane WA, where ballot counting has been seriously delayed because election officials are hand copying tens of thousands of ballots. It's an interesting lesson in how vote-by-mail (Spokane is an all-VBM county in WA) creates higher operational requirements for accountability, transparency, and election integrity. Some readers may not be familiar with the practice of hand-copying VBM ballots, and ask: what's going on? The situation is that for some reasons (read the news article for speculation on why), thousands of Spokane voters did not follow instructions on marking their ballot, for example, putting a check mark over a bubble rather than filling the bubble. If a paper ballot has even one of these mistakes anywhere, the ballot can't be machine counted -- the optical counting device kicks the ballot back out. And because this is vote-by-mail where the voter is not present during counting, there is no voter to ask to re-do the ballot. Instead, local election officials (LEOs) have to simply guess what the voter meant.

This is called "interpreting the voter's intent" in order to count every vote that the LEOs think that the voter cast on the ballot. After making such an interpretation of a ballot, an LEO marks a new blank ballot, copying all the voter's marks to tidy filled-in bubbles that the scanners will count. After all the uncountable ballots have been copied to a countable ballot-copy, the voting counting can finally proceed.

I've said many times that election technology should provide (and as our efforts at TTV bear fruit, will provide) support for such interpretation, and do so with as much logging and transparency as possible. I think that most people would agree that confidence in an election result depends in part on knowing how many votes were created by LEOs on behalf of a voter, rather than the mark of a voter that is so unambiguous that a machine can recognize it. Such automation might also reduce the need for laborious copying, preserving for all to see, an image of the original ballot together with the interpretation provided by LEOs during the counting process.

But the scale of Spokane operation really has me squirming. Tens of thousands! I mean, sure, I believe that the process is being done diligently, with intense scrutiny by people independent of the LEOs (members of the public, good government groups, political party people). But over days and days of efforts, under pressure to get the election results out, I fear that exhaustion and human error may take a toll. And unless the public (or at least auditors) have access to each ballot in all 3 forms (what the voter provided, what an LEO transcribed, what the scanner counted) it is going to be very hard determine whether this large-scale transcription process introduced errors. If this process were happening, for example, in New York with several very close contests, I could see people pushing for hand re-count. Let's hope that in WA the margins of victory are larger that the errors that could have been introduced by transcription.

And in the meantime, I wish the best to Spokane LEOs plowing through this mound of uncountable paper, and I continue squirm, wishing we had already finished the TTV central-count technology that could really help today.

-- EJS