A couple of weeks before this appeared someone (whose name I don’t remember) wrote a letter to the David Enterprise, our local paper, complaining about the length of time it took to count the votes on Election Day, and asking why the Yolo County Clerk/Recorder wasn’t using the more efficient, quicker electronic voting system to solve this problem? He called for an investigation into this.
I thought the call for an investigation was a bit extreme, and that his complaints, while understandable, deserved a response.
So here it is—my first, and probably only, op-ed piece ever!
The last election in Yolo County was different than the previous ones over the last 20 years. Instead of punch cards, we used ballots the size of a sheet of paper, and marked our votes with pens rather than punches. The secrecy envelopes were manila folders.
The results also took longer to tally and announce: instead of the election night results we used to get, initial counting of the votes took until nearly 4AM the next day.
A recent letter in the Davis Enterprise (“We can do better with elections”, June 18, 2006) suggests that all this was the fault of the Yolo County Clerk/Recorder.
My research as a professor of computer science at UC Davis, who teaches and studies computer security, takes me into unusual areas, including elections and electronic voting machines. So I know a fair amount about how Yolo County runs elections, and the constraints that the state and federal government impose.
Here’s what happened. Under state and federal law, a county must use a voting system that the federal government and the Secretary of State have certified. Shortly before the election, the federal government decertified the Data Vote system that Yolo County had been using.
Anticipating this, Yolo County Clerk/Recorder Freddie Oakley had been looking at several other systems. She chose the Hart Optical Scan system because it was a paper-based system; the systems that record votes electronically (called “DRE”, for “Direct Recording Electronic”) were in her opinion not yet reliable and accurate enough.
But the Hart Optical Scan requires the use of the large ballots we voted on. And to avoid the costs of having the ballot supplier supply the secrecy envelopes, the Clerk/Recorder went with the (considerably cheaper and reusable) manila folders.
The only other option would have been to go with a DRE system such as Diebold’s or Sequoia’s. However, both of those systems, and DRE systems in general, have lots of problems. They were designed without security in mind, and are only now being patched to provide some—rather like taking a dam or building with structural flaws and adding bracing to keep the structure from collapsing.
Worse, administrative lapses become much more dangerous. For example, setting up a DRE incorrectly can compromise every vote cast on that system. But it’s hard to set up a paper ballot incorrectly. Similarly, DREs can misrecord votes, and unless (a) the printed vote matches the actual vote and (b) the voter checks the paper to verify her vote, the misrecording won’t be detected.
And administrative lapses do occur; for example, in San Diego county, many poll workers took the DREs home the night before the election. This put all of those machines at risk, because once you have physical access to the machine, you can compromise it in a number of ways.
Given the uncertainty, the record of unreliability and errors with DREs, paper ballots seem like the most reliable way to record and count votes so far! Plus, election officials have been working with paper ballots for years, and are comfortable with them. The old saying, “if it ain’t broke, don’t fix it” seems particularly applicable here, to Yolo County.
While I too wish the counts had been done more quickly, I was glad to know that the Yolo County Clerk/Recorder made sure that every vote cast was counted, and counted accurately. Personally, I’d rather have the votes counted slowly and accurately than quickly and inaccurately. Under the circumstances, I think she did a good job.
Matt Bishop