Translate

Monday, November 19, 2012

The Aftermath (Part V): How do we count the vote faster?

[From a comment I made at the politics blog of the Bellingham Herald ]

There could be lots of legislation  proposed that would make the ballot and the chain of custody, etc. more secure and more easily counted or encourage greater and more accurate ballot participation.  But at least in Whatcom County, I'm convinced that trying to speed up the process would probably just result in inaccurate counts and disenfranchised voters.  Inside a flawed HAVA driven process, I truly think our Auditor's office makes a very strong effort to count all the votes in a fair and honest fashion.  The problem for us is "small county" mentality and "small county" resources with "big county" participation.  We've got 125.5K registered out of a max of 208K population. Over 105K cast their ballots, so half the residents of this county voted. Exclude those under 18, felons, and others unable to vote and you have a result that means that most adults in this county who can possibly vote did so.

At these vote numbers, with this high level of public participation in Whatcom County, we end up (in this election) with 10% of the ballots needing "duplication". This essentially means the Sequoia 400c Optech scanner couldn't discern intent on at least one vote, therefore a 'bi-partisan' committee has to decide voter intent for that ballot. This is a problem at 105K ballots - 10,500 ballots or more probably needed to be hand counted (and judged) and duplicated which costs us dollars and time. Probably election costs will near $500K for this cycle.

An education program would help with this.  The auditor's office should be given print space in the Herald, Whatcom Watch, Cascadia Weekly, etc. and/or perhaps the County should fund video productions in which the Auditor herself personally walks the user through marking ALL of the ballot cleanly and accurately.  They could produce such public service announcements that describe the vote counting process and the pitfalls of not marking your ballot so that it is machine readable and run them on BTV10.  For example, how to avoid an undervote or overvote, how to keep your ballot clean, how to print out some type of electronic draft ballot so that you don't create a mistake on your voting, how to request a new ballot if you manage to 'spoil' your first try.

The cynical will say the user is at fault. That those who can't fill out their ballot correctly shouldn't be counted. But this type of logic (besides being inhumane), neglects the flaws in the process of a complicated and lengthy ballot and disadvantaged citizenry.  In software engineering, we are all very familiar with the criticality of user interactions. We create user interfaces to allow a  product to be easily understood and accessed and the software industry spend millions each year doing so. Still, people make thousands training others on software products each of which comes with extensive documentation, context sensitive help, etc.

There really are different types of 'users' with different needs : 'Power users', occasional users, disabled users: all with different profiles for literacy, hand eye co-ordination, eyesight, patience, functionality needs etc. The needs and profiles of election ballot users are not really much different. If you think 'duplicating' 10% of the ballot is a problem now, imagine Whatcom County at one million votes sometime in the near future.

If you want a process to encourage as many people to vote as possible, you have to design the architecture of that  process from a systems perspective to handle that level of participation with less defects.  For example, what was the reason so many ballots needed to be duplicated? What are the specific categories of overvotes and undervotes? Please categorize all spoiled ballots and address the engineering processes necessary to create a 'cleaner' (e.g. more defect free) vote.  At least in Whatcom county, such analysis and (perhaps) subsequent user training might speed up the process of counting the vote.

No comments: