Last week in Quebec, they held a provincial election. Quebec Liberal Charest lost and claims to have left politics, and Parti Quebecois Pauline Marois won, and formed a minority government. But although the election may turn out to be important to the fortunes of the carré rouge movement, it’s not the subject of this post. Instead, I want to talk about the dry topic of voting systems. You see, in Canada, they use hand-marked paper ballots, hand counted in public. Among other things, that process means that we can actually be sure who won. And if the elections of 2000 and 2008 are any guide, and the race stays as close as the pollsters sat it is, we might, on Wednesday, November 7, not be sure who won.
Hand marked, hand counted paper ballots are “the gold standard* of democracy.” Brad Friedman of Brad Blog writes:
- “No ‘specialized technical knowledge’ can be required of citizens to vote or to monitor vote counts.”
- There is a “constitutional requirement of a publicly observed count.”
- “[T]he government substitution of its own check or what we’d probably call an ‘audit’ is no substitute at all for public observation.”
- “A paper trail simply does not suffice to meet the above standards.
- “As a result of these principles,…’all independent observers’ conclude that ‘electronic voting machines are totally banned in Germany’ because no conceivable computerized voting system can cast and count votes that meet the twin requirements of…being both ‘observable’ and also not requiring specialized technical knowledge.
After the verdict in the case — filed by a computer expert and his political scientist son — Lehto wondered how it could be that open, observable democracy is seemingly an inviolable right for “conquered Nazis,” but not, apparently, for citizens of the United States…
It was the fully public counting of hand-marked paper ballots that gave evidence that the unofficial, electronically-scanned election night results in Minnesota’s recent U.S. Senate race were wrong. A hand-count settled the results of Washington State’s Gubernatorial contest in 2004. And in the 2006 Republican Primary election in Pottawatomie County, Iowa, a hand-count found that seven races had been tallied incorrectly by the county’s optical-scan system. Unfortunately, that sort of publicly observable counting has become the exception rather than the rule in this country, and it happens only rarely, in elections where the candidates can afford the extraordinarily high legal costs of a contest, or when the results are so obviously twisted that officials are left with little choice but to count the ballots by hand.
“Hand-counting paper ballots is recognized as the gold standard in state laws across the country,” Ellen Theisen of the non-partisan election watchdog organization VotersUnite.org told me. “Why settle for anything less?”
Theisen’s thoughts echoed Lehto’s interpretation of the findings of the High Court in Germany. “By letting software count our votes,” she said, “we give software control over our government.”
Here are some Montreal bloggers discussing how their paper-based hand-counted system works after last week’s vote was taken. They seem to be voting geeks, but maybe that’s a consequence of a minority (Anglos) being very conscious of the effects that an election can have. I like geekery, so I’m going to quote the detail:
KATE I just spent 14 hours scrutineering in a school gym like a sauna: at one point my glasses just slid off my face onto the table.
STEPH I’m curious to know more about the voting & scrutineering process. Are ballots secret? It seems like after they vet your attendance based on ID and address, then give you a ballot with a serial#, but they take note of that # which is also on your ballot. What is it they tear off your ballot after you vote? They didn’t tear mine off before I put my ballot in the box myself and the scrutineer threw his arms in the air like I broke all the rules and they weren’t sure what to do. I hope they didn’t throw out the entire box of ballots because of me.
JOSH Steph, the thing they tear off your ballot is called a counterfoil. The people who run elections keep track of how many votes are cast by keeping these separate from the ballot box. This helps ensure that no one deposits multiple ballots into a ballot box. It serves as a check that the number of ballots actually deposited in the box is the same number that were handed out by elections officials.
I can’t find the relevant regulation in Quebec, but here is a section of the Elections Canada Act indicating that federal elections scrutineers, at least, are instructed not to throw out a ballot box for the reason you describe.
Now comes the really keen geekery where they explain how the keep the ballots anonymous (your vote secret) while still being able to count them and avoid fraud!
JOSH Oh, and as for the “secret ballot” part of the equation, here’s how it works:
An official hands you a ballot, with the counterfoil portion intact. There is a number on the counterfoil only. They hand the ballot back to you. You mark your vote in private and fold your ballot up the way they told you to. That obscures who it is that you actually voted for. Without unfolding your ballot (leaving your privacy intact), the official takes the ballot back from you and tears off the counterfoil. Now, your ballot is free of any identifying marks, because the only number is on the counterfoil. You (or they) deposit your ballot in the box. the counterfoil is deposited elsewhere. There is now no way to trace which person deposited which ballot.
KATE Steph, Josh is correct. We did not record any connection between the voter and the number on their ballot. We simply recorded on the list that the person had voted by checking them off.
The counterfoils are part of the system that guarantees voting fraud has not taken place. After the doors closed, we had to account for every ballot we received, adding up the used ballots, the unused ones, any ballots damaged and replaced (one, in the course of the whole day). The counterfoils are kept in a separate envelope so that if a recount were needed, it could be verified that the number of counterfoils matches the recorded number of ballots cast.
It’s really a very good system, because everything is solidly witnessed by two people who don’t know each other plus in most cases others as well [but see NOTE below], but great care is taken to make sure nobody knows how any one else voted. In the case of a damaged ballot (one voter managed to tear off the non-counterfoil end of the paper in our instance), the first instruction is for the person to go back behind the booth and mark EVERY candidate with the same mark before we write NUL all over it and give them a new one. So even in the case of this mistake, we didn’t know how the voter initially marked the ballot.
JEATHER Though Kate didn’t mention it, they also seal up all the spare ballots, so you can also count that my polling station was given, for example, 400 ballots, and we used about 220 of them, so we should have had about 180 left over.
KATE jeather, I did mention the unused ballots. Basically you get N ballots, and at the end you count everything to make sure the used, unused and damaged ballots all add back up to N. There’s an additional check in that the scrutineer initials every ballot handed out, so that nobody could smuggle in a faux ballot and use that instead (I don’t see what advantage that would create for anyone, but it’s enough of a concern that every ballot’s authenticated by the initials).
I was very firm with my voters that they needed to refold the ballot before bringing it back out of the booth, and never saw a single result in anyone’s hand. But people did joke about keeping the pictures, yes. This was the first election in which small black and white photos of the candidates were included on the counterfoil.
"I was very firm with my voters" I love that part.
It could be that more of a “public” is needed than “two people.” Perhaps everthing should be videotaped. Or perhaps a lot more people could be involved, so that counting ballots turned into a convivial, commmunity-building exercise.
So, summing up:
- In a democracy, election integrity should matter.
- Hand-marked paper ballots counted in public are the gold standard of democracy because they make the vote observable without special technical expertise
- Quebec uses a gold standard voting system successfully
We have an election coming up, and hence potential for wrongly counted votes, or even election fraud. It’s not only the Presidential election that matters, but all the Congressional races, the state races, and the local elections, including referenda, recalls, and charter changes.
Your jurisdiction is going to be counting all those elections in the same way, whether through touch-screen voting, optical scanners, mechanical devices, or even still on paper (don’t change, though maybe you should consider the Canadian “counterfoil” system). So, if your jurisdiction is thinking about buying electronic voting machine, you might consider using the information in this post as a starting point to dissuade them. And if your jurisdiction already uses e-voting, perhaps you can roll back that decision. If you or your committee don’t have the budget to travel to Quebec to observe a working system, here’s a good book on how paper-based voting is done.
NOTE * Sorry for the metaphor.
NOTE The referendum on Quebec sovereignty in 1995 was the exception that proves the rule:
After the vote, at each polling station, a scrutineer counted the ballots while a secretary recorded the result of the count. According to the referendum legislation, the scrutineer was appointed by the “Yes” [pro-sovereignty] committee, while the secretary was appointed by the “No” [anti-sovereignty] committee. When the counting was completed, approximately 86,000 ballots were rejected by scrutineers as “spoiled ballots”, meaning that they had not been marked properly by the voter.
Controversy arose over whether the scrutineers of the Chomedey, Marguerite-Bourgeois and Laurier-Dorion ridings had rejected numerous ballots without valid reasons, mostly by being overly strict on what marks voters could use to indicate their choices (for instance, rejecting ballots with check-marks or “X”s that were crooked, too large, made with a pen instead of a pencil, etc.). In these ridings the “No” vote was dominant, and the proportion of rejected ballots was 12%, 5.5% and 3.6%. In the riding of Chomedey, an average of 1 of every 9 ballots were rejected. Thomas Mulcair, member of the Quebec National Assembly for Chomedey, told reporters after the vote that there was “an orchestrated attempt to steal the vote” in his riding.
Yes, even when hand marked, hand counted paper ballots are counted in public, humans can attempt to game the system. In Quebec, however, that the voting system was robust enough to withstand the attack. (If Mulcair was correct that the Parti Quebecois was out to steal the referendum to win sovereignty, we should also assume they gave it their best shot. Secessionists don’t win by taking half measures.) Further, the record was clear, because the paper ballots were the vote, and not a (software-driven, therefore vulnerable) paper trail of the vote, so if the results of the election bad been contested, the evidence would have been solid.