It seems laughably obvious that this supposedly cutting-edge voting device will feel positively ancient in only a couple of years. It already looks like a cheap peice of crap to me, hardly something worthy of being integral to the American democratic process. And believe it or not, this photo was taken in 2004 — even though it looks a lot more like it’s from 1994 (think Windows 3.1).

In 2004, Behavior worked on a web site for the Smithsonian Museum of American History’s special exhibition Vote: The Machinery of Democracy. The exhibition focused on America’s “voting patchwork”, the broad range of voting technologies used state by state, county by county. It was an enlightening experience working on the project, and I encourage you to visit the site to learn about how we got to where we are now.

The current range of voting technologies in use today includes:

  • Paper Ballots
  • Gear-and-Lever Voting Machines
  • Punch Cards
  • Optical Scan Ballots and Readers
  • Direct-Recording Electronic Ballots

It’s widely assumed that the most modern technology available is obviously the best option — that is, that we should be using touch-screen direct-recording electronic voting machines. But maybe this isn’t the case — counterintuitively, perhaps an older technology is the best approach.

Here in New York State, we use the gear-and-lever voting machine, an error-prone 1940s technology that is still is wide use throughout the northeast, where close elections are rare. In 2000, many counties in other parts of the USA were still using punch cards, a technology that was cutting edge in the 1950s but which, no doubt due to the Florida debacle, has been almost completely retired by now. Much of the nation still uses “optical scan” technology — the old #2-pencil fill-in-the-bubbles we all know from standardized tests going back to the 1960’s. In fact, this technology has been in wide use for standardized testing for almost three generations of schoolchildren, and has been used for voting for almost 30 years.


Pencils Down

A colleague at Behavior, David Wu, told me the other day — without hesitation, and with definitive certainty — that he thought that the optical scan method was obviously the best choice, precisely because of the familiarity we all have from childhoods filled with standardized tests. The gears in my head started turning, and I was convinced instantly, especially when I thought of it in terms of the lifecycle of technologies.

I asked myself this: Of all the technologies currently in use (besides hand-written paper ballots, which are quaint but hard to count), which of them has had the longest history of practical use by real people? Or, to put it more simply, which technology is the least likely to quickly become obsolete, dated, and goofy?

For many Americans, going to the voting booth has always been like taking a trip one generation back in time. When the gear-and-lever machines were widely adopted in the 1960s, punch cards had already moved up to become the most high-tech means of electronic data storage and retreival in the real world of our offices and institutions. And then when punch cards were introduced in the voting world, most people in the real world had already moved on to optical-scan data collection, and even to direct-recording devices like ATMs and keyboard-and-monitor based computer terminals.

Optical-scan technology, however, defies this cycle. Even as newer technologies have come along — electronic voting machines with buttons, keyboard, mice, and touch screens — our schools continue to count on optical-scan technology to test our children, and the technology works just fine and without significant change in over 50 years of constant use. Everyone knows how to use it, it has a satisfying and trustworthy feel of direct access to the ballot via pencil-and-paper, a paper document/artifact is left behind, and the counting technology is pretty accurate and fast.

What’s more, no sophisticated software is required, no mysterious secret and proprietary algorithms or storage devices are used, no easily-hacked or insecure software, no buggy operating systems or unfamiliar UI paradigms (remember that millions of Americans have never used an ATM or a personal computer), and with all that, less distrust among voters that their vote will actually count.

Isn’t this obvious? Why on earth would we anybody consider using anything else?


5 responses to “The Best Voting Technology”

  1. Not only that, but you can hand-re-count the optical scan ballots.

    That’s the key to it all. Electronic tallying with human back-up.

    Americans have this odd fascination with “Teh Shiny”. Optical scanning and even paper ballots have worked well everywhere else in the world, even in highly dodgy countries (when the UN is there to inspect the procedure). Why be different for the sake of being different? The better question is “why be different when different is proven to be inferior”?

  2. To add, I find it *very* disturbing that many states are *mandating* a move to electronic (non-optical scan) balloting.

    Um, why?!?

    Sure, gear-based machines can break, paper ballots can be miscounted (hanging chads), but considering PC-based voting machines (and they *are* based on off-the-shelf PCs) rarely have a paper receipt, and therefore *cannot* be recounted, nor verified, why should they even be an option?

    Reliability will never be zero, but we do have an obligation to bring that percentage as low as we can. Touch-screen voting does not work toward that goal.

    See also: http://arstechnica.com/articles/culture/evoting.ars

  3. Unfortunately, and unbeknownst to most people, optical scanners can also be hacked just as easily as the ridiculous Diebold AccuVote TS voting machines. (See http://www.nyvv.org/paperballothacking.shtml as only one example) The optical scanner’s only edge up on the voting machines is the paper trail.

    A transparent, open-source, voting machine system that also leaves a paper trail is really the only way to have reliable electronic voting, and the Open Voting Consortium (http://www.openvotingconsortium.org/) seems to be the only group with an adequate design for such a system.

  4. Sean: You are right that the paper ballot optical-scan voting process is also vulnerable, but only to the extent that the counting machines can be hacked. In all fairness, however, this would by necessity be the case with all vote-counting techniques, except of course hand-counting.

    To be precise, in fact, in the case of direct-recording electronic ballots (i.e., touch-screens), the ballots themselves, the place where you actually record your individual vote, can be easily hacked … and both the ballots and the counting machines are “black boxes” with no accountability and no paper trail. That’s a far worse situation than the PBOS model where only the counting machine is vulnerable. So while I’m not necessarily disagreeing with you, I don’t think you are actually saying anything to disagree with my original contention that optical-scan is the best method precisely because of the paper trail. The “only edge” PBOS has — leaving a paper trail — is actually the critically significant advantage.

    I totally agree that the missing piece from the optimal paper ballot optical-scan voting system is an open-source counting machine, which is in fact exactly what the OVC recommends.

  5. kabababrubarta Avatar

    Nice site! kabababrubarta