E-Voting Undermines Public Confidence In Elections

Jeremiah Cornelius writes "Techdirt columnist, Timothy Lee, hit the metaphoric nail on the head, claiming that e-Voting undermines the public perception of election fairness - even when there is no evidence of wrongdoing. 'In a well-designed voting system, voters shouldn't have to take anyone's actions on faith. The entire process should be simple and transparent, so that anyone can observe it and verify that it was carried out correctly. The complexity and opacity of e-voting machines makes effective public scrutiny impossible, and so it's a bad idea even in the absence of specific evidence of wrongdoing.' Add to this the possibility technical faults, conflicts of interest and evidence of tampering, how long before the US vote is viewed as an electronic pantomime?"

Comments

i've been saying this for weeks

[circletimessquare]

the biggest threat to western democracy is not neocons, islamofascism, chinese technocrats, etc.

it's electronic voting

http://politics.slashdot.org/comments.pl?sid=413698&no_d2=1&cid=21986758

http://yro.slashdot.org/comments.pl?sid=409654&cid=21950000

democracy has plenty of problems, but one of democracy's greatest strengths is that by making the citizens it rules a part of the process, it inspires confidence in the government, it instills legitimacy

if you make the voting process opaque, you destroy confidence, you destroy legitimacy, you weaken people's faith in their democratically elected government, out of bad perception that their part in the process has been messed with, hidden

electronic voting must be universally rejected in all ways and all levels of government, asap

And the answer is so simple..

[spiffmastercow]

Open source the code, provide each registered voter with an encryption key sent through the mail (the real mail, so you have to send it to a real physical location), and maintain a database where each user can view his/her vote. Or, in the short term, just give everyone a receipt for their votes at the booth.

Solution already exists -- machine rejection

[Chris Burke]

There's already a simple solution to this problem. You stick the ballot into the Scantron machine. The Scantron machine tries to read the ballot. If the ballot is invalid, then it spits the ballot back out, and the voter either corrects it or has that ballot replaced.

This at least gets rid of the over-votes, and could get rid of the under-votes if the Scantron has a "possibly read a filled bubble but not sure" threshold (I really don't know).

One interesting tidbit from the 2000 Florida election that often gets ignored in favor of controversies over the felon lists and the nature of the butterfly ballots themselves, is that the machines they used were in fact capable of this! It was an optional setting, the machine could either take the bad ballot as-is but just not count it, or it could reject the ballot back into the voter's hand for correction. As you may have guessed, voting machines in the precincts with high rates of bad ballots had this option disabled, and ones with low rates of bad ballots had it enabled. But we weren't told that, and were instead just left to assume that the people in the high-error precincts were simply dumber than everyone else, and we just accepted it! But in reality, errors are common, but most get a chance to correct it when the machine spits it out.

That said, I do agree that the best thing to do is have the computer print out the ballot so as to minimize the possibility of error. It's really the best of both worlds: The accessability of a voting machine that lets you edit your choices, read more about propositions, and enforces rules like no over-votes, but you still get a human-readable paper ballot that serves as the vote of record and can be recounted by anyone with working eyes. And if you make the printed ballot machine readable -- I prefer an OCR-friendly font so it's the *same* markings that are both human and machine readable -- you can still use a machine counter to get your instant-gratification.

It's not that hard to design a working voting system that minimizes voter error, maximizes accessibility, and most importantly maximizes openness and transparency. Just... nobody that I can tell has actually come forward and put the pieces together in a real system intended to be sold.

Re:Solution already exists -- machine rejection

[CodeBuster]

nobody that I can tell has actually come forward and put the pieces together in a real system intended to be sold. There are probably reasons for this and a few come to mind fairly readily. First, the potential market for these machines is somewhat small and relatively fixed (i.e. it is a niche, a larger one perhaps, but still a niche). Second, the voting hardware has to maintain a high degree of accuracy and quality. For example, it will be made of quality materials and provable micro-programming with good documentation and easy configuration. All of these things would make the products expen$ive, at least to buy, and possibly to maintain as well. Finally, most governments or voting precincts are notoriously cheap and put in bids way below what it would actually cost to produce, provide, and maintain the equipment. All of these factors and perhaps others as well might explain why no private company has stepped up to the plate to provide this type of a system.

It might be well argued then, that given the importance of voting to our democracy and the relative frequency with which it occurs, that these product(s) and service(s) should be financed and standardized across all states and localities by the federal government. Generally I take the Libertarian view about the government providing goods and services (i.e. let the private sector do it) but in this case the combination of circumstances and the "public good" quality of the service makes it a good candidate for government funding (one of a very few exceptions to the private sector rule that I make).

Re:That's the plan

[ShieldW0lf]

Also, if they create a public perception that electronic voting is always uniformly opaque, complex and prone to fixing, they prevent direct democracy from ever taking hold and preserve the party system. Which is, of course, something all current and aspiring party politicians can unite behind.

The answer is:

1) To make peoples votes public information so everyone can see them immediately, know where everyone stands, and know if the votes were tampered with.

2) To allow them to directly vote on each issue if they want.

3) To allow them to vote for any person they want, and have that person cast their vote if they're not interested in directly voting on the issues.

4) To allow them to change the person they vote for at any time if they break faith.

You do that, and every person is in control of their own political power. That's the promise of electronic voting done right. I wouldn't think any person who covets power for its own sake is ever going to support something like this though.

In a way, confidence is what is most important

[Ungrounded Lightning]

Nice to see somebody noticing and describing one of the important pieces of the puzzle.

The purpose of elections in a republic is NOT because there's something "right" or "nice" about selecting the government officials and rules that are preferred by a majority of the voting population. (In fact, sometimes that's actually a bad idea. "Democracy" is often three wolves and a sheep voting on what's for dinner.)

The purpose of elections is to increase the stability of the country and pacify it internally. They do this by attempting to figure out which way the war would come out, if it were actually fought over the issue.

To do this, elections must convince the losing side that they can't reverse the result by resorting to force.

That means they don't have to be perfect - but they have to be convincingly good enough.

  - A wide race will be convincing. If the exact numbers are off it doesn't really matter.

  - A really close race may come out wrong. But if it's close it also means a war won't reverse the result: Too many additional people will get annoyed and oppose those who try the violent option. So the losers might exhaust the peaceable remedies: Recounts, courts, etc. Then they gripe about it non-stop until the next election. And EVERYBODY tries to fix the system to be more accurate and avoid this hassle next time. Repeat until the elections are believable and/or the margin is broad enough that there's no serious dispute.

But the easiest way for an election to be believably fair and honest is for it to be VISIBLY fair and honest. Count the votes behind locked doors or inside a software-driven black box and you substitute trust for visible honesty.

Once the people stop trusting the elections their stabilizing effect is gone. Then losers may think they are strong enough to reverse the result and (when the winners start doing things that hurt their interests) morally justified in making the attempt. Then you are just asking for civil "unrest", comities of vigilance, death squads, coups, and civil or revolutionary war.

So it's far more important that the election procedures be VISIBLY honest and their approximate accuracy known than that they be dead-on accurate.

Which is what we're seeing now. Computerized black-box voting killed the audit trail and enabled the possibility that a small number of people could introduce large and undetectable changes to the result. Then came a close election with important issues at stake. Regardless of whether the black boxes gave an accurate count or were corrupted, there was no way to SHOW they were right - or close enough not to matter. So the losers were unconvinced.

Repeat after four years, and again after eight, adding in a foreign war, massive government spending, "security" intrusions on civil rights, and attempts by media conglomerates to swing the election exposed by comparison to uncontrolled Internet communication. Now you're starting to approach a scenario where large groups of losers start thinking "Maybe the elections were stolen. Maybe we've been conquered. Maybe there are enough of us to reverse this. Maybe violence will work. Maybe the system is corrupted to the point that violence is the only answer. Maybe violence is PROPER."

This is WHY it is more important that the elections be VISIBLY, CONVINCINGLY accurate than that they just be accurate.

Vote Counting Doesn't Belong at the Booth

[Teancum]

I've said this before, I'll say it again, and I don't think most people will get it except perhaps among those here on Slashdot and other geek forums:

Electronic vote counting simply doesn't belong at the voting booth.

Instead, what should happen at the voting booth should be ballot preparation, not vote counting. I have absolutely no problem with a fancy machine that has cool graphics and touch screen voting options in a thousand different languages. If you want to go through the effort of putting that together for a somewhat reasonable price, knock your socks off. I don't even mind this end of it being completely closed and propritary.

But the vote counting part needs to be separated from the ballot preparation. The only thing all of this fancy hardware is really doing is to assist a voter in understanding the rules of the election. In other words, not vote for multiple people if multiple votes invalidate your ballot, remind you of races that you didn't vote in, allow you to clearly note who you are voting for (no missing chads or fuzzy and inconclusive ballots), and provide a very clean ballot that election judges can use later for the vote counting process. Ballots cast should be on some sort of physical medium that has irreversible marks (so an election judge can't modify the results afterward) and human readable so the voter can have clear confidence in how their vote was cast. Any computer scientist worthy of that title ought to be able to figure out how to make something simultaneously human and machine readable... we aren't talking rocket science here and this is a decades old solved problem in the field.

One of my largest problems with the Diebold machines is that they have the vote counting take place in the voting booth itself. This opens up not only the traditional forms of voting fraud, but it also opens up new vectors of attacking the system and requires far too much in the way of securing the machines in order to ensure the integrity of the data. Besides, most of the security protocols aren't followed anyway, and those that are followed are a joke in many cases and have multiple methods of being circumvented. By its very nature at least in America, voting is done in private and away from the eyes of election judges. This is also done to ensure the integrity of the vote cast (by stopping coercion). But this act, by its nature, means that the machines can be compromised during the act of voting itself.

Let's look at what problems these machines are trying to solve:

• Mechanical Voting Machine - these are from the 1930's and a few from even earlier. You press a lever and the vote is counted through some mechanical process. These have been notorious for breaking down during the middle of an election, or have the results manipulated by voting judges through multiple means. All states got rid of these machines decades ago, and for a very good reasons... yet the current Diebold machines (and others like them) simply are identical to these sort of voting machines but remade in the era of modern electronics. Every problem associated with mechanical voting machines can be found on the Diebold machines... and more.
• Punch-out Ballots - These received notoriety during the 2000 U.S. Presidential election when inconclusive ballots were cast, and rampant election judge tampering also took place. A common tactic in some areas where these were used was to punch through previously cast ballots to try and add some extra votes. Several methods of doing this were involved, and the "hanging chad" problem was actually a sign that such voting fraud occurred. This also suffered from mechanical break downs where a ballot could be cast, but a voter wouldn't necessarily know that their votes were invalid or inconclusive.
• Pencil bubble fill ballots - If you have ever taken a college admission test of some sort, you are more than likely familiar with these sort of ballots. This is where you have a little bubble or circle that yo