Voters put ballots into UNlocked boxes, not directly into scanners

[garybeck]

you may have read all the hubbub about the IRV vote in Burlington this week, how a third party candidate won. until now it all seemed great.

however, one of the other candidates just called me. he told me that he was told he only got 119 votes, when he knows he got much more than that. He said he could almost name 119 people that voted for him. he ran for mayor once before and came in second.

he said that he had reason to believe that the people who ran the IRV thing didn't like him... I started rolling my eyes until he told me this, which seems eerie:

We use Diebold opscans throughout our state. Every time I've voted, I've put my ballot directly into the scanner.

This time, he told me, the voters put their ballots into UNlocked boxes, and continued on their merry way. So when the polling station (precinct) was closed down, everyone there had access to the ballots, and who knows when or where they were actually fed into the scanners.

anyone else think this is strange? Would using IRV be a reason to not put the ballots directly into the scanners?

[Wilms]

1. Does seem odd.

I'm guessing you do precinct-based scanning. Correct?

Maybe a PM to Demodonkey would be an idea.

[garybeck]

2. yes, every precinct has had at least one scanner

and the voters shove it in there on the way out.

this time, the scanners were not in the precincts. it was a local election and it was IRV, so maybe there's some reason, but it does seem odd, and I think they should have been locked nonetheless.

[Wilms]

3. So we're not even clear why the scanners were not present.

Hopefully, the guy will call you back after checking procedure with the BoE.

[garybeck]

5. i'm going to talk to the folks who administered the IRV

I believe it was these folks, not the BOE:

http://www.electionsolutions.com/

[ayeshahaqqiqa]

4. In Newton County in my township

they have rarely, if ever, locked the ballot boxes. One reason that I go to the courthouse and vote early-I know the boxes are locked there.

[FogerRox]

6. ARRRGGGG

[diva77]

7. sorry to go off on a tangent, but it seems shocking to place your ballot

into a scanner as soon as you vote . Doesn't that mean that your vote is being tabulated while the polls are still open (or capable of being tabulated without you being able to observe) ? is the scanner hooked up to anything else?

[garybeck]

8. welcome to the machine

there is no observation of vote counting anymore. it's impossible to observe the vote counting when it's done by a computer. even if there's a group of people standing there, watching the machine, how do they know what is going on inside? how do they know the votes are being counted right? you might as well not watch.

you are correct.

[crispini]

11. Yes, our scanners count as you vote,

but they are not hooked up to anything. They print out a poll tape at the end of the day and then the paper ballots are available for verification or a recount.

[Wilms]

12. What scanners do you use, Crispini?

And can you tell me anything about dealing with paper jams, be they intake, on reject, or out-take?

Thanks.

[crispini]

13. My county uses ES&S optical scan at the polling place.

The newest model. I don't know what, if anything, they use at the office to count high volumes of ballots. I've never had paper jams occur at the polling place-- it's pretty easy to put things straight in. If it's an overvote or there's another problem with the ballot, it'll beep at you, you hit RETURN and give it back to the voter and they can get another ballot. If the voter's walked away already, then you can hit ACCEPT and it'll go on in and get voted. I'm not sure how it handles the counting if there are overvotes -- I should hope it would count everything except the race that you overvoted in, but who knows. As far as retrieving the ballots from the ballot box, we open up the side door and haul 'em out and box 'em up and seal the boxes and tote 'em off to the office. During the general election this took about half an hour and two carrying boxes. During our primary it took about five minutes. HTH!

[Wilms]

14. Never a jam. Wow!

How many scans do you think you've witnessed?

I've been pondering a VoteTrustUSA article about voting machine testing in CA.

http://www.democraticunderground.com/discuss/duboard.ph...

The CA test used the test deck up to seventeen times! Maybe more in the case of ballots that jammed and were returned to the test deck!

It seemed to me (and to the CA report writers who cautioned about "ballot fatigue") that may have been the cause of the jams.

Any thoughts?

[crispini]

15. This article?

http://www.votetrustusa.org/index.php?option=com_conten...

It should also be noted that, according to the Sequoia test reports, ballot “fatigue” was probably a factor in some of the optical-scan failures, because the same ballots were scanned at least seven times each during the test (twice to verify the test decks, and five more times to perform the scanner tests). Employing only 10 test decks to test 50 scanners calls into question the accuracy of the optical-scan reliability measurement.

It makes me wonder what they are calling a paper jam. I mean, I have seen our older model optical scanner (that we were using until this year) reject ballots which which were correctly marked. If they were not overvoted we just had the voter put it in again and it usually went through the second time. I did not see that happen at all this year with the newer optical scanner. I wouldn't characterize this as a paper jam, I would call a paper jam a ballot really getting stuck in the machine and not being able to use it at all. I have never seen this but I have only worked 3 or 4 elections and probably only seen 3,000 or so ballots cast.

However, as our elections administrator points out, if the machine goes down, there is a locked "holding box" on the ballot box. If there is a problem for any reason you simply open up the flap for the holding box and have your voters put their ballots in there and then they can be tabulated when the machine has been fixed. This is why he likes optican scan and I hope (cross fingers, knock on wood) he doesn't intend to change.

Here is the product I'm referring to:
http://www.essvote.com/HTML/products/m100.html

[Wilms]

16. Yes. That's the relavant snip.

Note however, that the snip came from the section of the report dealing with the first of two OpScan models that were tested.

When it came time to test the second model, they simply used the same test deck, and did mention that as a result, the ballots had been run up to seventeen times. My guess is more than seventeen because of rescanning rejected ballots. Some test.

Do you think it would take more than 10 seconds (that's from the 2002 VVSG) to clear a jam when you have to open the machine?

Thanks, Crispini.

Your help is HUGH!!11!

[crispini]

17. They never trained us how to open it.

I don't think it DOES open. I mean, not like a printer or a xerox does, where you can get in there and unwind some paper from the rollers. (Or, if it does, they didn't teach us that part.) We had two sets of keys. Key #1 opened the side doors to retrive the ballots, also the "emergency door" flap. You can see the lock for the emergency door in the picture below.



Key #2 was used on the top of the machine to turn it on. I don't think there is any way to flip that top part up, or if it is, they didn't teach us about it.

If it jammed, and actually stuck so that there was a ballot halfway in and halfway out and it wasn't going anywhere, I'd hop on the phone and ask the equipment office what to do next. I suspect they would have us fall back to storing the ballots in the emergency door until they could get a replacement out to us. I think we'd probably want to revote the whole batch anyway because who knows what the counter did when it failed that way?

I suspect this is a very uncommon occurrence, because I have been through 10 hours of election training from this office in the last 2 cycles, and they have never even mentioned the possibility that this would happen to us. And they have been very good about training us to work with the equipment.

[Wilms]

18. I'll be interested to see how ES&S machines fare in upcoming CA testing.

Sequoia and Hart scanners "looked" really bad in the "volume testing" (fifty machines and, I think, 400 ballots per machine). BUT, they kept reusing the test deck ballots.

It seems the test was really compromised. And, I fear, the case for paper ballots over DRE.

Again, I appreciate all your help as this OpScan thing has been on my mind of late.

[crispini]

19. Yes, it's obviously not a a fair test

to keep re-using the ballots. Our ballots as they come out of the package are somewhat heavy and stiff. It stands to reason that the ballot would soften and become a bit more flexible as they go through the machine. Also, ballots can certainly wear or get bent as you take them out of the box and stack them.

And IMO, no matter how much you compare the two, optical scan will always win. Why? If you have a DRE that goes down, you have a dead DRE and you can't vote on it. If you have a dead optical scanner, voting can continue and the voters can put their ballots safely in a locked ballot box until voting continues. In my mind that will ALWAYS make the optical scanner more usable.

Have you seen this website?
http://designfordemocracy.aiga.org/

[Wilms]

25. Very cool web site. Thanks.

Did you know ballot design can impact OpScan accuracy?

Counting Mark-Sense Ballots
Relating Technology, the Law and Common Sense

by Douglas W. Jones

http://www.cs.uiowa.edu/~jones/voting/optical /

[Land Shark]

23. I really LOVE those L&A tests, parallel testing, etc.

Land Shark's evil twin spends at least an hour a day L"M"FAO because the election officials run all these test decks but the way the evil twin "helps" the elections are all totally unaffected by test decks and pass all such inane, gradeschool "tests".

EXAMPLE: on the touch screen, press a sequence of locations over and above certain letters (so it's easy to find those locations) on the main voter-boot screen, and that will have the effect of flipping the number of votes that is equal to 20 times the number of touch screen presses that occur within ten seconds after the fourth press overall is completed. e.g., Presses 1-3 specify to activate the subroutine buried in the source code, and the fourth specifies which race to affect, with 0-9 specifying the first ten races from the top of the ticket on down, and then typing in 30 in the fifth and sixth positions would flip 600 votes. Probably not a good idea to flip 600 since individual machines don't do that many in a day.

Thinking L&A tests are somehow a protection is as illogical as counting all 52 blackjack "21" cards and verifying that there are 4 aces, 4 deuces, etc., and then concluding that there's no way the game could be rigged by Marked cards. (e.g. marks on the backside of the card that indicate identity of the card)

If you think about it, tests decks are something that would be marginally useful to detect scanner error in op-scans. And that's about it.

[Wilms]

24. Yes. In these tests, the "test deck" was OpScan ballots.

The DRE tests were done with "scripts".

I have thought that the internal code could be set to not futz around until more ballots have been cast than some pre-determined number--presumably a bit more than might be expected during a test.

Short of gross error in tallying, I wouldn't want to assume these tests reveal anything about DRE tallys, or even OpScan.

I think the main purpose of the volume testing was to check reliability.

The OpScan results probably should be ignored as a result of reusing ballots up to 17 or more times.

[salin]

20. That was my understanding of how it worked

when I voted by optical scanner in No Cal in the nineties. The ballot was counted - and then stayed in the machine/being collected for easy verification for recount.

[Wilms]

9. Not much in this article, but here it is for your file.

http://www.burlingtonfreepress.com/apps/pbcs.dll/articl...

[crispini]

10. I would wonder if the optican scanners could not be programmed

to count IRV. That would make sense to me. However, I would think that the elections department should at least have provided a locked ballot box. We have a lockable ballot bag that we use for our provisional votes, for example.

[Bill Bored]

22. Whoops! I sent almost a dupe of your post Crispini

Great minds think alike....and so do ours I guess.

[Bill Bored]

21. Since when does Diebold support IRV?

Why should the ballots be scanned if there is no IRV support in the voting system?

Tell me how, once the favorites' totals are known, the second, third and forth choices will be tabulated, and I'll shut up. But off the top of my pointy head, I don't think Diebold supports this.

Of course, the ballot boxes should be locked though!