What is your prescription for fair elections?

[garybeck]

please check the option that best reflects your opinion.

Assumptions:

"paper trail" implies voter verified. I think we all agree a piece of paper that is not verified by the voter is completely useless.

with "random audits," it is assumed that if the audit turns up different than the machine count, even by one vote, that a larger sample must be hand counted, and so on, and at some point an entire state must be hand counted. it is also assumed that the audits are genuinely random, i.e., the precincts are chosen on TV, just like the lottery.

this survey refers only to the electronic voting issue. of course there are other issues in regards to open and fair elections (machine allocation, partisan election officials, i.e. Blackwell...). It is assumed that those issues have been addressed and the question is just about making electronic voting safe.

[tompayne1]

1. although personally i am for 40-50 percent random audit.

[BlueEyedSon]

2. independent exit polling, IRV, paid vacation day or Sat elections

complete paper trail and auditability

[gademocrat7]

3. 100% counting paper ballots

it works well in Europe. We should demand this type of tabulation here.
No more Diebold,Sequoia held elections.

[Febble]

4. And a radical fix

of vote spoilage, as well as prosecution of those involved in voter suppression tactics of all kinds.

Fortunately the voting methods should also make voting cheap, and therefore more accessible - and hopefully more reliable. But systematic disenfranchisement of ethnic minorities arguably stole both elections for Bush (they certainly stole Florida for Bush in 2000, and if Gore had become president then, he'd be president now, no?)

[NoBushSpokenHere]

5. Nothing short of .........

paper ballots and I don't care if it takes a month to do the counting! I also want extended voting days - maybe even a week to vote! I want a panel of ALL parties to be governing body of elections - NOT a biased SOS! I also want the ability for people to register to vote right up to the day polls open! I want provisional ballots to count REGARDLESS what county or precinct they vote in! I want corporations OUT of the ENTIRE election process!

btw.......recommeded

[kster]

6. Hmmm.........Let me think.........

Paper ballots hand counted by the people and recorded at the precinct level.

[kster]

7. K&R.... nt

[mod mom]

8. HCPB with STRICTLY ENFORCED guidelines!

[Cocoa]

9. early voting

extended polling hours

higher pay for election workers

same-day registration

election-day holiday

[Bill Bored]

13. Go ahead -- cast your "vote" early.

Give up your right to a secret ballot (the machines have to know who you are so you can't vote again on Election Day, but they won't tell you that).

Leave your vote in a database in Walmart or MickyD's or some other insecure venue so that anyone wanting to rig the election will know exactly how you and the others have voted before election day. That way, they can prepare the appropriate hacks in advance by altering ballot definition files, central tabulators and anything else they can get their hands on to reverse an undesirable result.

And you'll probably have to drive a lot further to the Early Voting Center OR wait on line longer on Election Day because the machine count will be reduced since so many people decided to vote early! Kind of like they did in Ohio in 2004 in Democratic precincts, remember?

So go ahead. Vote as early as you like. Just don't expect it to count for anything.

Oh, and I agree with the rest of your suggestions, BTW!

[adolfo]

10. Simple

1.)Voter Verified Paper Ballots
2.)Mandatory Auditing of results (as a failsafe)
3.)Strict accounting standards for ballots (including serial numbers)
4.)All software code (optical scanners) opened to the public along with audits by an independent party. Similar to Nevada's gaming commission.

NO E-VOTING MACHINES.

[Bill Bored]

11. Other.

Once again, see this paper by Kathy Dopp on how to audit randomly:

<http://uscountvotes.org/ucvAnalysis/US/paper-audits/Paper_Audits.pdf>

It has to be a sliding scale based on the closeness of the election, the number of machines in the jurisdiction, the number of votes per machine, and the results of the initial audit (which may be perfect but still not detect concentrated fraud sufficient to change the outcome).

[garybeck]

14. one point I disagree on.

I don't think the closeness of the election is relevant to the audit. How do you know how close the election is? If the results are in question, isn't that the point? People keep saying the election can only be stolen if it's close. Or there should only be a recount if it's close. I completely misses the real issue. If most of the votes are counted on questionable computers, then it could be stolen even if it's not close. And how do you know if it's close anyway?

Audits must be mandatory, regardless of how close it is. No one will ever know how close it is unless there are safeguards in the system that are utilized in all cases.

[kster]

15. Thats the media's part in the election

theft,The media moving to wards election day has to make the public believe it is close in order for theft to happen without being detected, "Close" doesn't have any thing to do with the voting machines, if they can rig the election it does not matter if it's close or not.

What does matter is that the public believes that it was Close,so that * could win without the riots in the streets. "Close is the media's part of in the election theft.

It would be "nuts"if the media going in to the election was saying Kerry 53% * 47% then election day * won with a majority. THE MEDIA IS IN ON IT. And "close" is their part in it.

[garybeck]

16. exactly my point

the "closeness" of the election is irrelevant to whether or not there needs to be random audits. that's why many months ago I came up with this acronym:

MMRA

Mandatory
Manual
Random
Audits


the "mandatory" part is just as important as the "manual" and the "random."

VVPB without MMRA is like a security camera without any film. we need both: VVPB and MMRA to ensure fair elections.

[kster]

17. I've been to your site

I know we agree, only difference to me we have to shelve e-voting we cant work along with these crooks, we have to fight for PBHC, I think people are getting it slow but sure.

[garybeck]

20. it's a good discussion to have

it's not like I'm against HCPB...

a few things should be noted:

Elections can be stolen with HCPB.

Some of our ballots are so complicated it would be nearly impossible to count them without computers.

I think that if there are voter verified paper ballots (or trails) and mandatory manual random audits, it would be nearly impossible to steal the election. even if there wasn't open source code and even if Diebold is still running the e-voting. I have yet to have someone explain to me how an election could be stolen if there were VVPATs and MMRAs.

I'm all ears.

[kster]

22. This would be so much better

http://www.ejfi.org/Voting/Voting-73.htm#way


"I think that if there are voter verified paper ballots (or trails) and mandatory manual random audits, it would be nearly impossible to steal the election. even if there wasn't open source code and even if Diebold is still running the e-voting. I have yet to have someone explain to me how an election could be stolen if there were VVPATs and MMRAs".

Probably couldn't get stolen that way, but why should we have to fight so hard to get these common sense ideas in to place they are the ones selling the equipment to us, for our elections if they are not there already FIRE THEM.

[garybeck]

23. "fire them"

I'm all for that!

but we have to work with the laws and system we have... if we can prove they did something wrong then perhaps we can fire them, or put them in jail.

short of that, we can put rules in the system that will make it impossible for them to steal elections. that's all I'm saying.

it's like your kid with the cookie jar. right now we're leaving the top off the cookie jar and not even checking how many cookies are left when we get home.

if we lock the cookie jar, and put a video camera on it, they can't take any cookies without getting caught.

then, it doesn't matter if it's Diebold, ES&S, or my grandmoter who is counting the votes. No one can steal the election, period.

[kster]

27. Correct

Unless my kid has Fox news in his hip pocket and Fox tell's everyone that saw the video that they are conspiracy theorist and now move along, or he can call a terrorist alert and grab all the cookies he wants while the camera is looking the other way.

All I'm saying is we have enough to warrant some sort of investigation of 2004 from terrorist alert to the exit poll numbers,to date we have NOTHING.

[Bill Bored]

41. Gary,

Only in the case where there is no auditing at all does the margin not matter.

Keep in mind that the object is to make sure that the outcome of the election is correct, i.e., that the wrong candidate didn't win. So you start with a reported outcome and you have to prove that it's correct.

The reason why the closeness (margin) matters is that once you have an initial random audit of X%, and it comes out OK, the only way the reported outcome of the election can be wrong at that point is if the election is close enough so that enough votes can be shifted without detection in the remaining unaudited machines or systems to reverse the outcome. If the election is close, you have to do a lot more auditing to prove that the outcome is correct than if the election is a landslide.

So see if you agree with this:

If there's NO auditing whatsoever then closeness doesn't matter. The outcome can be anything the machines say and only something like an exit poll might reveal a problem. But as soon as you do your first random audit, the closeness of the election comes into play and it's harder to steal if it's not close because you'd have to rig more votes on more machines which would be easier to detect.

[Blue Shark]

12. The Dems steal the next three elections...

We will hold the WH, the US House, the Senate...and we call it even and start over.

[Amaryllis]

39. Kinda hard to do when the companies are all owned by repubs.

[yowzayowzayowza]

18. Other:

1. VVPB, and
2. Every counting device must
a. pass a 5% manual audit &
b. publicly post results.

Forget the open/escrowed source & pseudo-random precinct selection nonsense... every counting device should prove itz validity in every election and the tabulation process must be independent of the counting process.

[kster]

19. Why forget

the "open/escrowed source" just wondering?

[yowzayowzayowza]

24. Source availability...

is important in the event of a detected failure, but meaningless in proving the reliability of a particular process cycle.

[kster]

25. You are like fox news

Its a simple question. Why forget the "open/escrowed source" just wondering?

[yowzayowzayowza]

30. ... and you ....

a Fox News aficionado.

Source code does not prove that all machines are in fact running that version.

Source code does not prove that all machines are calibrated correctly.

Source code does not prove that all machines have not been tampered with.

Etc, etc, etc.

Auditing verifies reliability; source code proves squat.

[kster]

32. "source code proves squat"

Then why be so secretive about it? If I was tring to sell voting machine's I would give it up, For some reason they won't. Why?

[garybeck]

21. how can you do that?

I don't think it's possible to do a manual audit on 5% of a machine. consider this -- a machine is one of several in a precinct. How do you do a manual count on 5% of the votes in one machine? what do you check the numbers against? which 5% of the ballots do you check? assuming a machine only records its total, there is nothing to check. You could check all the votes in the machine if the machine recorded it's total.

the reason the precincts are targeted for the manual audits is because the totals for the precicts are available. all the ballots in the randomly selected precicnt must be hand counted and checked against the machine count for the entire precinct. There are no numbers below that available in many cases.

[yowzayowzayowza]

28. Delaying audits until...

after the election is reported and allowing 95% of precincts to pass unaudited is unacceptable.

A reliable system must routinely audit all counting processes, to wit:

1. Opti-scan readable VVPBs are required whether produced by human markup, computer assisted completion or DRE,

2. Every precinct must manually count 5% of their VVPBs on election night, and

3. Opti-scan VVPB counts of three batches:
a. manually audited ballots,
b. unaudited ballots, &
c. all ballots.
Preferably these batches should be presented in random order and each batch should be followed by a printout. Obviously, the manual audit total should balance, a+b=c and c=automated tabulation totals.

VVPBs and Opti-scan printouts are evidence of a verified election; DRE & automated tabulation systems are only convenient adjuncts with no legal standing.

Still auditing only 5%, yet a much safer and more reliable election process.

[kster]

29. Thank you,If I may

what percentage of American voters know what the hell you are saying, and is this fair to them that they have to rely on other people to make sure that their vote got counted?

[yowzayowzayowza]

31. They aren't relying ...

on anyone else to verify their vote counts. Every precinct does their own audit. Not rocket science er even puzzling to Fox News aficionados

[kster]

33. They are relying on

everyone else. If they can't see or know how their vote is getting counted.

[garybeck]

34. that does not explain

how to manually audit 5% of each machine (which you suggested before) or even 5% of each precinct. I don't think it can be done when you consider that a machine does not keep a running total, and each ballot can not be matched to anything inside the machine. Once the ballot runs through the machine, or the DRE casts the vote, there is nothing matching the stack of VVPBs to the machine count.

To clarify, let's say you want to run your 5% audit on one of your machines. You have a stack of VVPBs, and there is the machine sitting there. What exactly do you do?


Also, I did not suggest that the results be delayed for the hand count. I think the random precincts can be selected on election night as soon as the polls close and they can be counted right away. If however, a discrepancy is found, there will be a delay, as there well should be.

[yowzayowzayowza]

35. Ok... Itz the end of election day...

at Precinct X. The last voter has cast their ballot. You have a stack of Opti-scan ready VVPBs before you. Every twentieth ballot is placed in the audit batch and counted manually. The audit batch is run thru the Opti-scan counter, batch A. The balance of the VVPBs are run thru the counter, batch B. Finally, all ballots are run thru the counter, batch C. Your manual audit should match batch A. Batch A plus batch B should equal batch C. Batch C should match any automated tabulation numbers including DREs.

The key is that the legal ballot is an opti-scan capable VVPB. A computerized system which produces ballots that can only be counted by humans is asinine. The legal count is of those VVPBs. DRE counts and "Eballots" have no legal standing.

The machine being audited is the opti-scan counter not the machines of the DRE or computer-assisted ballot completion.

(Hand count) can be counted right away.

It is not even remotely possible to complete a full manual audit in large precincts on election nite.

[OnTheOtherHand]

36. "every counting device must pass a 5% manual audit"

I see a need for some testing of every device, but I also wonder: if you set a 5% ceiling, aren't you almost inviting a hack that starts stealing votes if more than, oh, 100 votes are counted? (In Ohio, that should preclude detection in every precinct, under the 5% rule, and still allow votes to be stolen in 98% of precincts statewide.)

(EDIT: Sorry, the first graf here was completely redundant -- I somehow missed an entire branch of the thread.)

[yowzayowzayowza]

37. See post 28; step 3. n/t

[OnTheOtherHand]

38. I'm not sure that is bullet-proof, but let me think --

it would certainly foil a lot of schemes.

I have no hands-on experience with op-scan. How long does it take to run a stack of, say, 500 ballots, and is it possible to characterize a typical rate of variation between runs?

[yowzayowzayowza]

40. Counting extra batches...

might be an issue in centralized systems, but shouldn't be a burden in precinct level counts.

If the opti-scan system cannot reliably count the same batch repeatedly, itz time to get another machine... not like therez a concern for deteriorating chadz... the humidity claims in Ohio 2004 were ridiculous.

We also need much better access to view the counting process, perhaps even mandated web camz.

[OnTheOtherHand]

42. if we're talking precinct-based counts, then

I would like to believe that having observers actually present would be more effective than web cams! A web cam is pretty easy to defeat, and there is no guarantee that anyone will watch it anyway. (Observers can also be defeated, but I would think they would stand a better chance.)

Do you know of any sources on op-scan read reliability? I take your point that there is no fundamental problem here, because the ballots should be robust (and it is never prohibitively difficult to just try One More Time), but I still wonder how reliable these systems are in practice. (Manual recounts, assuming they are accurate, actually answer a slightly different question than the one I have in mind. I mean "reliable" in the sense of returning the same answer every time. As I understand it, op-scans can't necessarily do even that if voters don't follow instructions in completing ballots, but it's entirely plausible that a manual count can reliably interpret the will of the voter when an op-scan run can't, so it will give a different result for that reason.)

Oh, "humidity claims in Ohio 2004" -- do you mean Hackett 2005? I wasn't aware of humidity claims in 2004. Superficially, the final Hackett numbers seemed to come out pretty close to where they should have -- the bigger precincts in Clermont were more Republican, but that has been true in other elections too. But maybe it warrants a closer look.

[yowzayowzayowza]

43. Oh, no.

Web camz could be an additional set of "many eyes" not a replacement for on site observers.

Recounting discrepancies are an issue in both manual and opti-scan methods; from NH2004 recount:

In the eleven wards recounted, only very minor discrepancies were found between the optical scan machine counts of the ballots and the recount. The discrepancies are similar to those found when hand-counted ballots are recounted. http://www.votenader.org/media_press/index.php?cid=413

Perhaps the standard opti-scan methodology should include a second pass to identify disparities? Even tho it exists in manual counts as well, why ignore the problem?

I think the manual / opti-scan interpretation differences would be a good thing for people to see regularly. It would help poll workers understand spoilage issues more clearly.

Humidity problem... less a disparity issue than handling & duty.

[OnTheOtherHand]

44. OK on cams, still reserving the right to insist on some full manual counts

I'm pretty comfortable with the idea that clean op-scan counts should have small error rates similar to those from hand recounts. And it seems likely that any possible disparities could be resolved by doing a second pass if necessary.

I just want to be sure we don't introduce a back door through which it might be possible to introduce "very minor discrepancies," all in the same direction, that could determine a close election. That's a potential problem no matter how we do the auditing (not an intractable one, I'm just not sure I have thought it all the way through).

Beyond that, I initially like your approach now that I finally understand it. It makes substantial demands on poll workers, but not as substantial as full hand counts would be -- and if we're gonna have precinct-based op-scan, we ought to kick the tires on it. So, has there been any discussion of your proposal "out there"?

[yowzayowzayowza]

45. Should the outcome merit...

or a candidate/citizen fund it, I've no prob with full recounts.

...substantial demands on poll workers...

Most precincts will only have to manually count a few dozen ballots which seems a small price to pay to verify the counting process both locally and before tabulation.

This exercise was to show that bogeymen disappear when the light comes on and the closet door hinge is not our enemy. We further our cause by demanding reasonable & reliable audits not hysterical manual bogeyman patrols. Alas, we're prolly on ignore for most concerned anyway.

[OnTheOtherHand]

46. I feel downright subversive, having a civilized dialogue in the open!

About the demands on poll workers -- they have one manual count and three mechanical counts, and they have to make sure they all reconcile. Not rocket science, but work.

I'm not so sure about the closet door hinge, however.... ;)

Thanks for giving me some ideas to chew on.

[Ellipsis]

49. Three different machines counting right? n/t

[yowzayowzayowza]

50. Definitely not.

The object is to test a single machine's veracity by counting
- all the ballots twice
- broken down into three batches (ran in random order)
- including a quantity of manually counted ballots in two of the batches.

Plz see posts 35 & 28(step 3).

[Neil Lisst]

26. 100% hand authenticated votes for president

We need a standard presidential election ballot that is handled completely separately from the ballot for everything else.

Everyone gets one, and they all get counted immediately and separately, and the paper trail is there for every ballot.

Canada uses paper ballots, and in their last major election took 4 hours to count them. We demand more of our Banks and ATM cards than we do of the process to elect president. This has to stop now.


------
TERROR ALERT!
http://www.webcomicsnation.com/neillisst/

[rhite5]

47. We have to get rid of the electronic voting machines entirely.

• Hand counting at the PRECINCT level is essential.

• The overall number of ballots must be reconciled with the number of signatures in the poll book the voters signed when they arrived.

• It must be on the SAME DAY, beginning immediately after the polls close. If it takes all night, that is ok. (But it won't -- see Note #3 below)

• Qualified volunteers can be allowed to act as vote counters, because poll workers are likely to be exhausted.

• Allow observers representing any interested groups. (political party, candidates, media). A video of the proceedings would be a good idea. Challenges need to be addressed on the spot. Counts should be verified by a second count and any discrepancies resolved.

• The results must be publicized at the same time they are communicated to the County Elections Office and to the Media. The numbers can be posted on the outside door so the public can see them.

• The raw ballots should then be put under lock and key and RETAINED at the Precinct, not transported anywhere else, not given to a courier, until the entire election is certified.

Notes:

1. Precincts used to be smaller than they are now. With the introduction of e-vote counting, many precincts got combined and consolidated. Let's break up the big precincts into smaller units so that the likely number of ballots on a high turnout election would be no more than 1000 in any precinct. (even now many precincts only have a few hundred ballots to count in small towns and rural areas).

2. Never have two or more precincts voting in the same location!

3. If there are Statewide, County and Local INITIATIVES also to be voted at the same election, these should be on a separate ballot on a clearly different colored paper. This color should be consistent throughout the jurisdiction.

Bottom Line: We must remove the Corporations from the control of elections. Elections need to be simple enough to be easy to understand. We should not have to require high tech competence in our dedicated election workers. Anytime a person working on elections finds part of the process mysterious it can easily become an open door to fraud. So keep it simple, like it has been for 200 years. It is not rocket science. These people know their jobs and they know their communities.

Elections are supposed to be the time (often the only time) the PEOPLE are in charge and express their preferences and pass judgement on past performances. That is what democracy is all about. When you give that right to the Corporations, you give away democracy.

[PATRICK]

48. My Rx

is for the principle party that will be the victim of unopposed fraud, i.e. the Democrats, to centrally and totally get the word out to every responsible representative to get behind the total package of the touchscreen issue.

The battles are waged like the Revolutionary War militia actions while the main army is hibernating at valley Forge trying to survive. Some are successful, some more than others. But wouldn't it have been far far easier if ALL the Dems locally had been brought up to snuff and on board
on the object of their career survival- the protection of the vote?

You still have to compete with the used software salesmen of Diebold, the bullying glamor of HAVA money and phony mandates, the distraction of the paper trail debate instead of the integrity of the companies and their secret domain software, instead of the tabulation and communication process also vulnerable in hyperspace electrons.

Those stubbornly demanding paper ballots are right about it being safer, cheaper and a simpler alternative simply because the current engagement of the party organization is so weak, diverse, complex and compromised.

What any elected Dem needs to know. HAVA is a GOP dream bill. GOP companies shove untrustworthy machinery- already shown to suffer numerous violations and convenient incapacities that have led to only Dem losses. The basic principles of programming for morons can show anyone in five minutes what can happen to vaporized ballots or tabulations in systems closed to verification and scrutiny and at the same time open to "glitches" and tampering of every imaginable sort.

There should be no need to debate this with Democrats- or the remaining honest Repubs. There should be no need for education and lobbying and protests. There should not be- and what we have instead is a weak debate on paper trail add on as if THAT is the whole issue.

Instead the high profile Dem comments on the past questionable election contain either a thin coat of whitewash(Ohio review) or a blurred focus such as is reserved for aging movie icons or some suicidal pretense of bipartisanship or whatever.

ONE directive, and individual strongarming and vocal support to all Dem leaders involved is what was and is still needed. Or the grass roots will bear the whole unequal burden of trying to put the pieces together- which will not happen everywhere or in time. I think Howard Dean is finding out that the party chair or the DNC itself is not capable of automatically doing this, as necessary and logical as it is.