State and local elections in off-years provide pundits endless fodder for predictive analysis, even though flipping coins might be just as predictive. One prediction that seems to come true every time, however, is that computer-screen voting will remind us of the wisdom of hand-marked paper ballots. In Indiana, computer-screen voting machines switched voter choices — for the second straight election:
A vote switching problem that frustrated Greater Lafayette voters in November 2018 elections – and wound up part of a federal lawsuit filed in October – caused scattered problems again during Tuesday’s municipal Election Day.
Tippecanoe County Clerk Julie Roush traveled to West Lafayette Fire Station No. 3, 1100 Kalberer Road, shortly after lunch Tuesday to check the calibration on three machines, after receiving a call from a voter who said the voting machine he used kept marking an “X” for someone else when he touched the screen for the candidate he wanted.
Meanwhile, Robert Kurtz, a voter in West Lafayette, took a video of faulty vote recording on a touch screen at a vote center a Federated Church in West Lafayette. The video captured his attempt to vote for three of four candidates for three at-large seats on the West Lafayette City Council.
“When I touched a square next to a candidate’s name, the machine selected the square for the candidate above,” Kurtz said. “If I touched the square for the candidate at the top of the list, nothing happened.”
If the above video looks a little familiar, it’s because the Guardian also reported on touch-screen voting machines having the same problem. In 2012:
Yesterday, Bloomberg also reported that other issues arose in Pennsylvania, where the Guardian also noted problems seven years ago. Now the machine systems apparently are creating large numbers of phantom votes as well as frustrating voters with screen mistakes:
The first sign something was wrong with Northampton County, Pennsylvania’s state-of-the-art voting system came on Election Day when a voter called the local Democratic Party chairman to say a touchscreen in her precinct was acting “finicky.” As she scrolled down the ballot, the tick-marks next to candidates she’d selected kept disappearing.
Her experience Nov. 5 was no isolated glitch. Over the course of the day, the new election machinery, bought over the objections of cybersecurity experts, continued to malfunction. Built by Election Systems & Software, the ExpressVote XL was designed to marry touchscreen technology with a paper-trail for post-election audits. Instead, it created such chaos that poll workers had to crack open the machines, remove the ballot records and use scanners summoned from across state lines to conduct a recount that lasted until 5 a.m.
In one case, it turned out a candidate that the XL showed getting just 15 votes had won by about 1,000. Neither Northampton nor ES&S know what went wrong.
Actually, it’s clear what went wrong. Election officials keep making a choice to use expensive computer systems for the actual casting of votes where direct paper voting is both cheaper and more reliable. The decision to buy these expensive apparatuses is a choice to introduce more potential points of failure into what should be a straightforward process that requires no more technology than a functioning pen and printed paper.
As one computer-science expert told Bloomberg, it’s not just normal points of failure these machines risk, either:
Security experts say the cheapest, and to their minds, most reliable and hack-proof method to cast votes also happens to be the lowest tech: paper ballots marked by hand and fed through scanners (no chads) to tally the results. They have called for replacing computerized equipment—particularly paperless older models—with the decidedly Luddite alternative.
The devices have “raised far more security questions than paper ballots because you have a potentially hackable computer standing between the voter and the record,” said J. Alex Halderman, a computer science professor at the University of Michigan, adding that without sufficient research, these new machines could be “a waste of money.”
Part of the problem in this debate is Bloomberg’s characterization of hand-marked paper ballots as “Luddite.” There’s nothing wrong with choosing the most effective and safest technology for a task, nor is there anything anti-technological for concluding that computers might not be the best possible choice. We got into this problem because of the inexplicable and unsubstantiated assumption that computers were the only solution to the “hanging chad” crisis in Florida’s 2000 election recount … when those ballots were created specifically for ease of computer counting decades earlier, before scan technology became reliable. Those who think that computers are automatically the best solution and that anything less is “Luddite” should read this essay by computer programmer Ryan North at Medium last year to learn why that’s a very, very bad assumption in voting.
Besides, modern hand-marked paper ballot systems use computers too — but they use them for counting the votes off the ballots, while preserving the ballots as a permanent record in case of errors or need for recounts. Hand-marked paper ballots remove all of the failure points between the voter and the vote, except for the incompetence of the voter himself. Those optical-scan systems provide better security as well as instant confirmation of ballot acceptance while also allowing for fast vote counting at the precincts — all while retaining the actual hand-marked paper ballots for a reliable audit trail.
All of that makes hand-marked paper ballots the best technology for this particular task. The sooner that grandstanding election officials realize this and quit opting to aggrandize themselves by looking “modern,” the more secure and reliable our election systems will be.