Paper Vote Canada

The classic Canadian example I usually use is the last Referendum on Quebec independence where it was 50.58% "No" to 49.42% "Yes".

There is another great example going on right now in the Minnesota senate race.
According to Daily Kos, "Today's latest results show [Democratic challenger Al Franken] is now trailing Republican incumbent Norm Coleman by 204 votes."

Wikipedia currently shows the tally at

Popular vote Coleman:1,211,562 Franken:1,211,356 Barkley:437,389

If you want that in percentages that's Coleman 41.988%, Franken 41.981%

That means if your voting machines have even a .01% error rate, they've already thrown the election. And the high-tech threat to Minnesota's optical mark-sense scanners? Dust.

Undecided Minnesota Senate Race Used Machines that Flunked Accuracy Tests - Wired - November 5, 2008

In an earlier posting, Wired writes

The problems occurred during logic and accuracy tests in the run-up to this year's general election, Oakland County Clerk Ruth Johnson disclosed in a letter submitted October 24 (.pdf) to the federal Election Assistance Commission (EAC). The machines at issue are ES&S M-100 optical-scan machines, which read and tally election results from paper ballots.

Johnson worried that such problems -- linked tentatively to paper dust build-up in the machines -- could affect the integrity of the general election this week.

ES&S Voting Machines in Michigan Flunk Tests, Don't Tally Votes Consistently - Wired - November 3, 2008

Say what you will about human failure modes, but dust usually isn't one of them.

Given that
1. Elections are often surprisingly close
2. Integrity of the count is paramount (your vote must be correctly counted)
3. Machines have many failure modes
4. A paper count by humans can be open and easily verified and rechecked

Then the best option to ensure confidence in election results is: hand-counted paper ballots.

(I don't know whether the Minnesota recount will require hand-counts.)

Labels: election, electronic voting, optical scan, usa

500th post.



What's interesting (and sad) is that Oprah blames herself for her voting problems.

First of all, if the machine doesn't record your vote, that's because the machine is badly designed. Second of all, it means you shouldn't be using machines.

It doesn't seem to occur to Oprah that the fault could lie with the machine.

Labels: election, electronic voting, usa

shape-shifting electronic votes are more than fantasy, according to reports from states including West Virginia, Missouri, Nevada, Georgia and Colorado. Whether by accident or design, touch-screen voting machines have "flipped" votes from a caster's chosen candidate to one he opposes.

Unlike the old days when campaigners hung around street corners haranguing voters with handouts and pints of beer, the electronic era presents a sophisticated challenge to democracy.

Now, says Crispin Miller, author of Loser Take All: Election Fraud and Subversion of Democracy 2000-2008, changes can occur seamlessly, without a breath of suspicion. Electronic glitches are only one of a range of mishaps, mistakes and dirty tricks that may decide outcome on Nov. 4.

Complaints about the electronic machines have mounted, along with calls for a return to paper ballots, like Canada's.

"More traditional systems are better," says Jeremy Epstein, a technological security expert and member of two Virginia legislative commissions that studied voting machines. "Paper-based and hand-counted ballots are fast, accurate and cheap. Studies show that machines are insecure and vulnerable to attack."

Fraud fears grow as [US] voters throng polls - The Toronto Star - October 21, 2008

(The article title is not great, something like "voting machine errors and voting surpression plague election" might have been closer to the mark.)

Labels: election, electronic voting, usa

Election Data Services provides the US November 2008 voting equipment composition (I'm tempted to say "breakdown").

I should mention that they use some confusing terminology.
To me electronic voting covers optical scan, DRE and Internet voting.
They consider electronic voting to cover only DRE (usually touchscreen) machines.

An optical mark-sense reader is an electronic device just like a touchscreen machine. It uses optical sensors to read a dot on paper, rather than to record a fingerprint. It is subject to most of the kinds of attacks that a touchscreen suffers from: you can compromise the software/firmware, there may be errors in the software/firmware, the optical sensors may be mis-aligned or malfunctioning, the paper path may jam, the power can fail, etc.

As well, if you record the order in which voters submit their ballots for scanning, you can reverse this to determine exactly who voted for whom, by going down the stack of ballots - once again the secret ballot is compromised.

It is true that IF AN ERROR IS DETECTED or IF A RECOUNT IS MANDATED, you can then hand-count the ballots (albeit going slightly crosseyed staring at tiny circles for hours).

Of course if you were a clever hacker, you would just program the scanner to distort the election by a margin smaller than that which would trigger any investigation. A similarly small error would also not be detected.

NOTE: some kind of rendering bug puts this table far down on the page.

Type	% Registered Voters
Punch Cards	0.10
Lever Machines	6.72
Hand-Counted Paper Ballots	0.17
Optically-Scanned Paper Ballots	56.17
Electronic (DRE / Touchscreen) Systems	32.63
Mixed	4.22

from 2008 Voting Equipment Study (PDF)

According to votingmachines.procon.org the numbers previously were

2004: 1% paper, 35% optical scan, 29.5% DRE
2000: 1.5% paper, 29.5% optical scan, 12.5% DRE

Labels: election, electronic voting, usa

The elections staff had collected electronic copies of the votes on memory cards and taken them to the main office, where dozens of workers inside a secure, glass-encased room fed them into the “GEMS server,” a gleaming silver Dell desktop computer that tallies the votes.

Then at 10 p.m., the server suddenly froze up and stopped counting votes. Cuyahoga County technicians clustered around the computer, debating what to do. A young, business-suited employee from Diebold — the company that makes the voting machines used in Cuyahoga — peered into the screen and pecked at the keyboard. No one could figure out what was wrong. So, like anyone faced with a misbehaving computer, they simply turned it off and on again. Voilà: It started working — until an hour later, when it crashed a second time.

...

so many printers had jammed that 20 percent of the machines involved in the recounted races lacked paper copies of some of the votes. They weren’t lost, technically speaking; Platten could hit “print” and a machine would generate a replacement copy. But she had no way of proving that these replacements were, indeed, what the voters had voted. She could only hope the machines had worked correctly.

...

In the last three election cycles, touch-screen machines have become one of the most mysterious and divisive elements in modern electoral politics. Introduced after the 2000 hanging-chad debacle, the machines were originally intended to add clarity to election results. But in hundreds of instances, the result has been precisely the opposite: they fail unpredictably, and in extremely strange ways; voters report that their choices “flip” from one candidate to another before their eyes; machines crash or begin to count backward; votes simply vanish.

An extensive New York Times Magazine report from January 6, 2008: Can You Count on Voting Machines?

And these are just the obvious, visible ways in which machines can fail.
There are many other silent ways in which the machines could fail internally that you would never detect.

You can move to optical mark-sense, but these are still machines:
* the poll workers need to get trained on them
* the paper can jam
* the scanners can fail
* the entire machine can fail

and on and on and on.

In case you think those are unlikely scenarios, they are already happening in advance voting in the United States.

The Jacksonville Times-Union reported long lines in northeast Florida, with at least two counties reporting problems with voting machines. In Duval County, 7 of 15 optical scanning machines used to count ballots had to be replaced, the newspaper reported.

Early voting suggests 2008 may see record turnout, expert says - CNN - October 21, 2008

Labels: election, electronic voting, usa

"People make mistakes more than machines," said Jackson County Clerk Jeff Waybright.

Dear Jeff Waybright,

You are way wrong. You are confusing consistency with correctness. If a machine is programmed to do something (programmed, by a person) it will do that thing, consistently. If what it was programmed to do is WRONG, it will do it CONSISTENTLY WRONG.

Yours Truly,

Someone who actually knows about machines

Above quote from More W.Va. voters say machines are switching votes in the Charleston Gazette, October 18, 2008. The story reports that machines are not correctly displaying votes (presumably because of touch screen misalignment, or other malfunction).

Labels: election, electronic voting

It's election day in Canada.

Please vote.

Remember there are new identification rules, roughly you need either a driver's license (or health card with photo and address in Ontario) or two pieces of ID, one with name & photo and one with address.

See Voter Identification at the Polls for more information.

In general, see

http://www.elections.ca/

for any information you need about voting today.

If you're new to the process, this very simple guide will walk you through (with the exception of the new identification rules).

Labels: canada, election