Michigan absentee ballot mistake

[ok222]

check this one out..................http://www.livejournal.com/community/ann_arbor_area/23038.html

[Obamarama]

1. Oh My God.

As fate would have it, I am on my way to my local Dem. party headquarters in about half an hour to sign up for some phone polling shifts. I've printed this off and am going to confirm this with them. Will report back later this afternoon.

[ok222]

2. Thank you

I was going to request a ballot to find out =)

[Name removed]

14. Deleted message

Message removed by moderator. Click here to review the message board rules.

[leftchick]

3. OMG! This is no "mistake"!!!

There are No mistakes with the crew in the WH! Who is the Democratic Representive in your district? You need to bring this to their attention at once!!

[annabanana]

4. Let's get this info to as many Michigan papers as possible.

I sure hope this gets picked up on PRONTO. (It's not too late already, is it?)

[leftchick]

5. excellent idea!

fax it to newspapers all over Michigan!

[ok222]

6. Let's all do it =)

[ScreamingMeemie]

7. ****Look here**** Problem was small and has been cleared up:

Roland posted this yesterday and also contacted someone. Here's the threads:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x947290

and

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x942216

[ok222]

8. Cleared up hmmmmmm

the person who sposted it got her ballott yesterday still there.......

[ScreamingMeemie]

9. This was yesterday that this was posted. I called a friend who is

voting here in my district and her ballot was okay. That's not to say that eyes won't be kept on it. I was just making the comment that a thread had been posted already. :hi:

[ok222]

10. oh ok I can't search so I didn't know

[ScreamingMeemie]

11. No problem. That's what DUers are for!!

Have a great weekend! :hi:

[TahitiNut]

12. COOL IT! (Here's a simple tutorial about optical scanning.)

This is an OPTICAL SCAN ballot. Optical scanning equipment works by registering the geometric position of the mark, in this case a "fill in the arrow" mark.

In this case, it's extremely likely that marks to fill in the arrow to the right of Kerry/Edwards would, in fact, be tallied for Kerry/Edwards. Indeed, Bush/Cheney voters would have to pretend there was a fractured arrow to the right of the Bush/Cheney entry and make a mark there as though they were completing the arrow.


Let's try to understand how the 'geometry' of the ballot is scanned in an optical scanning system. If you look closely, you'll always see marginal registration marks on any optically-scanned document. These timing/registration marks are the only other marks on such a document that the scanning equipment "sees". All the other printed information, including the fractured arrows or boxes to be filled, is printed with an ink that's 'invisible' to the scanning equipment. Imagine it were printed in red. Imagine you were looking at the document through a red filter (e.g. red-colored glasses). You wouldn't see anything printed in red, but you'd see the black marks.

This "ballot" should, in fact, be more upsetting to Bush/Cheney voters than Kerry/Edwards voters - if self-interest were the only basis for being upset.


If anything, imho, this demonstrates a reason we should be far more favorable towards optical scanning technologies than others. It's obvious to damned near anyone that this ballot is FUBAR. It merely requires eyesight - which is a 'sense' that's far more common than political sense these days.




On edit: An 'editorial comment' - There are two things that're frustrating to me about this mini-tempest. First, it's the general ignorance of both the public and those who should know better (elections officials and the media) regarding the distinct technologies relevant to such systems. The general failure to comprehend the basic technical/functional characteristics of optical scan, punch card, touchscreen, and other technologies causes more heat than light, more smoke than fire, and more rant than reason. We have some very real political issues we must deal with, and ignorance is our enemy. Second, that ignorance is compounded by both "faux experts" and a failure of both the public and the media to calmly identify and solicit the advice of objective and expereinced people who actually know about these technologies. Even a superficial review of Slashdot (where some real pros hang out) would demonstrate there's no failure of true experts to understand the issues of electronic voting systems. Instead, we tend to only listen to those who say what we want to hear. That's a lousy way to get reliable information.

[RedEagle]

13. Hmmmm....

Your analysis of the ability to recognize inks is valid for infrared scanners, an older technology. That is why it's a very, very good idea to make sure you mark a ballot with a medium that has carbon in it. Even some (and only some) inks have carbon in them, if you don't like pencil.

But visable light is newer technology and it will recognize a far larger range of colors and "inks."

While I can appreciate your argument that this should be more alarming for the Bush/Cheney gang, keep in mind that if the program tells the tabulator to scan the marks relative to position, (#1 = Bush/Cheney, #2 = Kerry/Edwards) and the position most likely to count is not the printed names of the candidates,(which the program may ignore) but the order in which they are ranked in the column to be marked, then yes, this is a huge problem, kind of like the butterfly ballot.

I can think of one other way, although not a very good one, to change the voter's intent on some ballots. If this is used on infrared scanners, and the ballot is pre-marked with ink by the Bush/Cheney choice, but the voter can't see it, (and I've talked to printers about this and yes, there is invisible ink with a carbon content) and a voter marks Kerry/Edwards with a medium without carbon, then the ballot will tally for Bush/Cheney and the system will not reject it as an overvote, because the system will only register the one vote.

In this case, contrary to what I've outlined above, the program would have to "see" that blank area next to Bush/Cheney, whether the arrows are there or not, as a space to look for a vote. If the space (and this holds whether it's marked by the voter or not) has carbon based marking on it, it will tally for Bush/Cheney. And if a Kerry/Edwards voter does not mark with a carbon content medium, the vote goes to Bush/Cheney and goes through the system without detection of the error. (In an infrared system)

The problem is based on what the programming tells the system to look for and where to look for it, not just whether something is marked or not.

I would advise people to make sure the medium they use to mark optical scan contains carbon. Some inks do have carbon in them. (Harmonyguy, help here!) The older optical scan technology, infrared, is most susceptable to this. And no, don't depend on the election workers to have the right marking utensils.

Some of this is just a general heads up. People who have not received a corrected ballot should take this ballot to their auditor and request a corrected one.

Thanks for posting on this topic because it's very important people pay close attention to these ballots. Seems to me we are seeing an awful lot of these kinds of problems this time around.

Some ballot printing is contracted out to local printers and some is through the vendors, fyi.

And if you vote absentee, please take another step to insure your vote is counted, if you can, and hand deliver it to the auditor. Do not mail it in if you can avoid that. That helps take one of the potential areas of fraud out of the system, ballots that get "lost" in the mail. (And there is a normal percentage of that happening without any fraud involved at all)

[TahitiNut]

15. I find it virtually impossible to imagine that the ballot parsing logic ..

... in the scanning system is heuristic. (By "heuristic" I mean, of course, that it can do any pattern recognition on the printed material of the ballot and dynamically establish an interpretive context.) In some election jurisdictions, I believe, the order of the candidates on the ballot varies from ballot to ballot, probably associated with some readable/scannable ballot format identifier. Clearly, from precinct to precinct and county to county the local candidates and local referenda vary enormously. It's impossible for me to imagine that given the necessity of pre-loading ballot-specific characteristics that any logic/investment would be wasted on heuristics.

At most, I think, the system would read the ballot's format code/identifier in order to verify the associated table-based interpretation.

At the scanning stage, the system would merely translate the scan into a record containing a series of coordinate-based fields indicating the geometric grid positions in which marks were sensed. The system would then associate such grid fields with a pre-defined and pre-loaded tabular definition for the particular ballot being scanned and, if real-time validation is available at the scan point, perform the straightforward overvote/undervote assessment (audit) to allow the voter to resubmit. (That validation function isn't always implemented, afaik.)


Obviously, I'd have to examine the specific system to ascertain more detailed system characteristics, but both the visible ballot design (with registration marks) and general knowledge of optical scanning technology would seem to bear me out.


The points you make about the potential for fraud and shenanigans in the systems and materials are well-taken. Nonetheless, optically-scanned paper ballots afford us (the voters) with the most important safeguard: a manually countable and auditable document. If we were to destroy all of the machines, we'd still be able to count the ballots with reasonable accuracy after election day. The same can't be said of either touchscreen or even "lever" voting machines (which have been used for 100 years). As I've always asserted, it's essential that a sample audit count of paper ballots be performed and validated before any 'electronic' tabulation is certified. We can never merely 'assume' that the machines/systems operate reliably.

[RedEagle]

16. Yes, optical scan provides evidence of voter intent

I should have been clear that this is very important.

IF we get audits of the paper ballots, then this can be a very good system.

Auditing is going to be key. Any system can be scammed- or just make mistakes.

Transparency is key. And it must be so to ANYONE- not just a few computer scientists. It simply can't be democratic if only a few understand or can audit the results.

Touch screens are only high-tech and very expensive lever machines. Same flaws in a different medium.

On edit:

We must insure that the original ballots are used in recounts and audits also.

Optical scan systems can print a report of the ballot which has the same flaws as the after-the-fact(election) print outs of touch screens- it's a programs interpretation of the vote, not the real ballot.