It's not about HCPB, it's about CHECKS and BALANCES

[eridani]

So, you don't think that HCPBs need auditing? You know perfectly well the story of the fake terrorist alert in Ohio in 2004, during which the Repubs tabulated all the ballots behind closed doors. If they had been HCPB, would that have been OK with you? Yes or no, please.

ANY and ALL tabulation methods need required random auditing, including HCPB. People have been known to cheat and make mistakes with all known ways of measuring anything.

Now, for my county (second largest in the nation, and with 1000+ ballot styles), opscan is much more accurate than HCPB, which is the gold standard ONLY WHEN YOU ARE COUNTING A SINGLE RACE The best mandatory auditing method would therefore be auditing an array of single races with HCPB.

Now, for smaller venues (not sure what the cutoffs in size and complexity would actually be here), HCPB could work well. Mandatory auditing by HCPB would be acceptable, but I think it would be better to have a different method on the basic principle that checks and balances are good, just like paleontologists date fossil trees both by counting rings and 14C analysis. Someone last year gave a decent protocol involving weighing the ballots with scales designed to be accurate enough to count money by weight. I actually tried 120# card stock with a two-place lab scale, and found the variability to be within adequate limits. Or you might weigh first, and hand count later.

Checks and balances! Checks and balances!

[Eric J in MN]

1. They should put each ballot on a projection screen...

...so that a roomful of people can count along with the election judges.

[eridani]

2. Lots of potential mistakes in store there

At least in my county. At my polling place there are two different state legislative districts and two different congressional districts, and two different school boards. Same water department, though. Where do you find a room full of people, anyway? We have serious problems filling all our Dem observer slots in a heavily Dem county. Repubs are in worse shape. If we could find a room full of people, most would have been working all day and are tired. If their tally sheets don't agree as a result, then what?

[Eric J in MN]

7. There would be a separate room for counting each Congressional district.

I'm sure each party would find the necessary observers if votes were being hand-counted.

I would want to observe then.

When the votes are being counted by machine anyway, I don't have a desire to observe when there is so little to see.

[eridani]

10. It took WA state THREE FRIGGING WEEKS to hand count the 2004 gubernatorial race

And you expect King County to count dozens more races with 1000+ different ballot formats by hand accurately, let alone in a reasonable amount of time? What's your problem with Minnesota's current auditing system?

Minnesota Performs First Post-Election Review
State-mandated audit a success, officials and advocates say

By Sean Greene, www.electionline.org
December 14, 2006


Minnesota's first-ever post-election review- a manual count of votes from randomly-selected precincts in the state - drew raves from two sides that do not always see eye-to-eye, election officials and advocacy groups.

"I believe that Minnesota has done a most remarkable job at making every vote count and count correctly," said Janet Straub, a Minnesota resident and observer of the post-election review.

Incoming Secretary of State Mark Ritchie (DFL) said he was also impressed.

"I am excited to hear the very positive results from our first reviews. We can all feel a great deal of confidence in our election results - and only hope that other states can catch up to our system before the 2008 elections," he said in a press release from Citizens for Election Integrity Minnesota.

The law mandating this new procedure was spearheaded by State Representative Bill Hilty, DFL-Finlayson.

"We have a really good system in place in Minnesota. We have a 100 percent paper-ballot system. But even with these devices the only way to be confident in their security and reliability is to check them out," Hilty stated.

[Eric J in MN]

15. It takes Canada a few hours. NT

NT

[eridani]

16. So what? There is no relevant US equivalent real world example

Canada has a parliamentary system where they don't even vote directly for legislators. They have one, or at most two, choices on each ballot. Whatthefuck is King County supposed to do with 12 different state legislative districts, some of which are split between boardering counties, and at least one of which is also split between three different Congressional districts? Not to mention lots of different school boards, water districts, county council districts (my state LD is in FOUR different county council districts), fire districts, sewer districts and ballot initiatives.

And you never did state your objection to Minnesota's current auditing law.

[Eric J in MN]

23. Minnesota has a decent system.

But I'd prefer to know that each ballot was counted by people, at least for the major offices.

If some of the voting machines are defective and reverse the count for governor, we won't know in Minnesota unless those machines happen to be in an audited precinct.

[OnTheOtherHand]

24. actually, you would have a pretty good shot

As I understand it, almost 200 precincts statewide were recounted. A random sample of 200 precincts is likely to catch miscounts large enough to reverse an election unless the election is very close. It probably has about a 99% chance of catching a 1% shift in a gubernatorial race (but not in a smaller race). About a 90% chance of catching an 0.5% shift -- under that, a full recount is automatic in MN, isn't it?

From what I hear, Minnesota's audits are likely to get even better, which is good.

[Eric J in MN]

25. How many precincts are there in Minnesota? NT

NT

[OnTheOtherHand]

26. about 4100, but that doesn't matter much

The statistical power of the test mostly depends on how big the sample is, not the proportion. Connecticut's 20% audit sounds ginormous by comparison, but they have (I believe) fewer than 800 precincts, so their audit is actually smaller than yours, as I understand the two.

[Eric J in MN]

27. Are you assuming that if the governor's race is programmed wrong..

..in one machine, then it will be programmed wrong in every machine?

[OnTheOtherHand]

28. nope, I agree that wouldn't be a safe assumption

If I were assuming that, then you would only have to audit one machine(!) -- or maybe one machine per county.

But I do assume that in order to swing 1% of the vote statewide, a bunch of machines would have to be messed up. Say at least 2% statewide. Then each precinct you audit may have a 98% chance of not being messed up, but by the time you audit 200 precincts, the odds are very much in your favor. There are ways to 'beat' that logic, but it isn't as easy as a lot of people seem to think.

[eridani]

29. There is no such thing as error-free tabulation

If the audit sample is designed correctly, it should catch indications of any significant problem, which then can trigger a full hand recount.

[A Simple Game]

3. I don't see the problem, use separate ballots.

Do what ever you want on your local election ballots.

But I damn well want you to have a paper ballot and hand count it for the person you expect me to call president for four years! They could probably be written large enough for everyone to read, and be able to mark with a magic marker.

Who says the ballots have to be counted that night? Promise the results within a week, if it takes less time, you look good. Lock them up in a safe place and go back the next day and count them. Again do what ever you want with your local ballots, but don't screw up the ones that can affect me, and I promise to do the same for you.

[eridani]

6. I support the Kucinich bill requiring HCPB for presidential elections

Opscan plus mandatory auditing for everything else, except maybe governor and senators. The longer you have to store uncounted ballots, the more chance for error or hanky panky.

[Bill Bored]

17. I think Kucinich has it backwards.

The easiet races to hack are the smaller ones. The outcome of almost any statewide race can be confirmed with an audit if it's done properly.

However, I will say that since there are only 3 federal offices, if we hand count one, might as well do them all. I just think it would be better to start with the House because those are the easiest races to hack, unless they are gerrymandered. But in 2006, gerrymandering didn't stop the Dems from taking control. A lot of close races for the House that year!

[eridani]

18. By sheerest coincidence--

--the races that are easiest to hack are also the ones easiest to audit. A very helpful coincidence indeed.

[Bill Bored]

19. Well, easiest to hand count. Not necessarily to audit.

Actually, to audit them, you have to use a fairly high percentage. So why not go all the way to 100% on those and save the audits for the bigger ones where a small percentage may suffice?

[eridani]

20. Ok, good point. n/t

[kster]

35. Sounds fair to me

:toast:

[Land Shark]

4. this is bordering on disinformation, while being largely tru

1. anything counted in secret, even if by George and Martha washington isn't legit. That fact taht "anything" includes HCPBs really proves nothing.

2. similarly the fact that no system is perfect proves nothing, but if we know some sotires about HCPB that's actually a GOOD sign, not just evidence against it, because evidence-creation is a major security feature, one of the most important. Electronic is weak to nonexistent on evidence-creation.

3. The gold standard is single race HCPB because then they do it RIGHT. There's lots of ways to do HCPB w3rong on purpose or by accident, why would you blame bad execution on the HCPB system?

4. "for smaller venues HCPB could work well". Whatever the ratio of people are that are needed to do HCPB per 1000 ballots in a small venue, IT IS THE SAME RATIO PER THOUSAND BALLOTS in a big city. No real excuses for the big cities, they even have more money to pay help if they need to....

[Febble]

5. This

Electronic is weak to nonexistent on evidence-creation.

is certainly true if there no paper ballot is created, and even if a paper ballot is created, it isn't much use unless a good random audit protocol is mandatory.

Which is why at the minimum, a good bill needs to mandate a paper record of every vote that has the status of a ballot in the event of a recount or audit, as well as an audit protocol that ensures that, as eridani says, "check and balances" are in place to check the vote and balance the books.

I see absolutely no call for calling the OP "disinformation" or anything like.

[Land Shark]

8. how about misleading information instead? This poster knows I think she

does great work oftentimes, so the work is not taken in perhaps the harshness you attribute to it. No offense but I would not think those frames or formulations to be friendly to our movement generaly, just IN MHO

[eridani]

9. Point by point--

2. similarly the fact that no system is perfect proves nothing, but if we know some stories about HCPB that's actually a GOOD sign, not just evidence against it, because evidence-creation is a major security feature, one of the most important. Electronic is weak to nonexistent on evidence-creation.

Absolutely no direct recording of votes electronically, ever, under any circumstances. Using touch screens as ballot marking devices is another matter entirely. Though I think that the VotePAD system works better for more people, there are some who will find that a touchscreen works better for them. (I've joined the WSDCC disabilities caucus in order to educate them on election transparency. In return, I'm learning quite a bit more about accessibility and other issues of importance to the disabled community.)

You've expressed concern about people not differentiating between the use of touchscreens as ballot markers vs. vote tabulators. Didn't you greatly expand on a paper of mine explaining why DREs are not the same as ATMs, which people know and trust and tend to conflate with DRE voting? Is there some reason why we can't do the same about the distinction bewteen using a touchscreen as a creator of a paper ballot and using it as an unaudited vote tabulator?

True, we've been sloppy about conflating the terms "touchscreen" and "DRE". But why should teaching people to be more careful about language here be any more of problem than educating people about the distinction between "voter fraud" and "election fraud," or between a paper BALLOT and a paper "trail" or "record"?

3. The gold standard is single race HCPB because then they do it RIGHT. There's lots of ways to do HCPB w3rong on purpose or by accident, why would you blame bad execution on the HCPB system?

No, it's the gold standard for single races because its within the normal capacity of ordinary human beings. I'm blaming people here, people who get tired and cranky and are far poorer at doing complex repetitive tasks than computers are. Counting more than one race concurrently puts the tabulation problem in that ballpark very rapidly. It took three weeks just to do the WA gubernatorial recount, and you are proposing that King county should count all the many other races on ballots formatted 1000+ different ways in addition to that in the same way?

I'm a chemist, and I know a great deal about methods for measuring things and their reliability. Not only is my computer faster at doing peak integration and calculating final results, it's also better, with r values of 0.9999 instead of 0.99 (a factor of 100).

Does that mean that I actually trust the computer and the software? Not just no, but HELL, NO--ISO certification notwithstanding! I do at least one hand calculation for every sample set I turn in. Six months ago I turned up a random error of final results being off by an order of magnitude, and I then had to wipe my hard drive and reinstall everything. Mostly it's reliable, and the bugs left in the software are bugs that don't interfere with what I have to do, but, as Alastor (Mad Eye) Moody always used to say "Constant vigilance!"

Optical scanning is comparatively simple technology, and it's been banged up against reality hard enough and often enough to be pretty reliable. So much so that states rely on it for lotteries, not only because of speed but because of accuracy. Too bad we live in a culture that values gambling more than it values voting.

4. "for smaller venues HCPB could work well". Whatever the ratio of people are that are needed to do HCPB per 1000 ballots in a small venue, IT IS THE SAME RATIO PER THOUSAND BALLOTS in a big city. No real excuses for the big cities, they even have more money to pay help if they need to....

Except that cities have to find and organize and train the people. Since when is administrative complexity any less a threat to election integrity than software complexity? (Scale-up is every bit as much of a problem in administration as it is in chemistry, where one does not dare assume that because 1 gram of A and 0.5 grams of B and 2 grams of C gives you a useful reaction product, that 1 ton of A and 0.5 tons of B and 2 tons of C will react at all, let alone give you the same product.)

My LD has plenty of trouble right now finding enough Democratic election observers, and it's even more of a problem for the much smaller number of Republicans. Small counties do NOT have 1000+ different ballot formats. Andy Stephenson, being from King county also, and talking and working with many election workers, got this--and advocated precinct optical scanning with mandatory audits.

In addition to which, Repubs have been very successful in creating a political culture which denigrates the very existence of the concept of "public good," and has successfully stomped all over the living standards of average working people to the point that sheer survival, let alone paying attention to the public sphere, is very time consuming. In the broadest sense, election transparency work is linked to the re-creation of a more participatory public sphere, and we are nowhere near the point of getting enough people so involved to manage an all HCPB system.

[Land Shark]

11. lots of people recognize, from many perspectives, our need to rebuild community....

in many different walks of life, not just election activism. But it can be done, and it's an important part of successful HCPB like half of New Hampshire has (great video of that out there by the way)

THough you don't expressly say this, just because there is a challenge or a difficulty does not AT ALL mean we shouldn't do it that way. If that were the standard, we would lose the right to trial by jury because we would never insist on a "draft" or jury selection methods....

as far as your computer often being better and faster, that may well be true (even though you find errors by hand)with chemistry. But the first thing we need to do in elections is apply the values of DEMOCRACY. Because we tend NOT to do that, we are getting ourselves in trouble. Democracy is not always concerned with efficiency per se, we have three somewhat redundant branches of government checking and thus "hassling" each other. Couldn't we streamline that? Yes, Truman said if you want efficiency you get a dictatorship. That will happen by applying efficiency as a value and ignoring the values of democracy and human rights.

A lot of my ideas were better expressed here, in terms of coming up with rights first, as the foundation, and then going from there... http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=389x459022

I'd like to talk to you and learn more about your disabilities work. I'll PM you on that

[troubleinwinter]

13. Would I be understanding you correctly to infer that you support conscripting poll workers?

Like a "draft".

[eridani]

14. To me, democracy means also that accuracy is more important than efficiency

--that for every vote to be really counted, we should use the most accurate method. In the case of complicated ballots with a lot of side issues, that method happens to be optical scanning. (Separating the presidential race off for HCPB as Kucinich proposed seems like a good idea to me. Far more people are aware of and invested in the presidential race, which will help with the recruiting and training issue.) It's comparatively simple, and it has had a lot of real-world testing already. There is also weighing (too much hassle for King County, but potentially useful a lot of smaller venues), which is right now trustworthy enough to be used by bankers for counting cash.

What is real irony here is that you are the guy who so eloquently framed election transparency in terms of checks and balances. Now when I try to extend that beyond the people involved to the way that different tabulation methods, just by the simple virtue of being different from each other, add another set of checks and balances, you blow it off.

Re New Hampshire from the 2000 census--population 1,235,786, divided into 15 counties. King County, WA--population 1,737,034. Now, the half of NH that does HCPB would not happen to be the more thinly populated rural half, now, would it? Why do you suppose that is? Does NH have 1000+ different ballot formats?

[troubleinwinter]

12. Eridani, these are very good posts.

I am gratified to see some remarks about the challenges regarding the logistics of HCPB. Simply yelling for HCPB and slogans does not answer the real-life challenges of the fact that polling places on average have a 20% shortage of workers already. Poorer/inner city/minority areas have an especially difficult time attracting workers.

Thus far, I have seen no solution presented regarding the need for perhaps two to three times the current number of workers to perform hand counting. None. That doesn't mean that HCPB is 'bad'. But I do not want to see chaos and fiasco in our elections. Mandating HC without any planning whatsoever invites disaster.

Organizing, qualifying and training workers is important to take into consideration. Recruiting workers is already considered a "crisis" across the country.

May 10, 2006 -

County Short On Poll Workers

The Registrar-Recorder's Office has already recruited about 20,000 poll workers but still needs to recruit another 1,000 to 5,000 in the next month.

In a report to the supervisors, Registrar-Recorder Conny McCormack wrote that no other issue has dominated the long-term planning of election administrators nationwide in the past few years more than polling staff.

``The number of experienced, trained volunteers has diminished alarmingly with the aging of the traditional poll worker corps,''

The recruitment crisis is compounded by a recent change in state law limiting the size of voting precincts. As a result, the number of precincts increased by 10 percent this year -- from 4,635 in November 2004 to 5,065 for June 6.

http://earc.berkeley.edu/news/2006/May/CountyShort.php

Taking into account the current shortage of poll workers across the country and that the workers are "working a 14-plus-hour day", how will we recruit, train and organize this counting force that would double or triple the number of needed workers?

I agree that paper ballots, Op-scan with rigorous random auditing ("checks and balances"), hand counts in areas that choose it, along with the Kucinich bill is a practical option.

[Einsteinia]

21. Centralized or Decentralized Vote Counting

is actually a more important question than hcpb vs. 10% audit

I would rather have a 10% audit at the precinct, with citizen-controlled purview, on election night, than have a hcpb hand count at the county election headquarters.

We need to keep control of the one-to-one correspondence of the ballot cast to what is ultimately tabulated--by whichever means.

And as far as that goes, I think two or methodologies with separate and distinct oversight.

Read more: http://www.califelectprotect.net/Titanium.pdf

[eridani]

30. I favor decentralized counting as well--

--provided that the adminstrative end of it is well mananged. As I keep saying, administrative complexity is every bit as much a threat to election transparency as software or hardware.

[Einsteinia]

32. Then I think you'd like the Titanium Standard

which is in my prior link. Check it out. It's all about the fact that an audit formula is meaningless if it isn't in secure context, and lots more about non-corrosible checks and balances to secure the core of our democracy.

[Einsteinia]

39. Ahem, about hand -AND- computer counts

that are debated up the thread.

If you read the Titanium Standard (earlier in the thread), you'd see that both are required.

[Stevepol]

22. A lot of very good ideas! K&R!

[kster]

31. Bill Faulkner’s "checks and balances"

North Florida. Retired Navy aviator and veteran, Bill Faulkner, MBA, may have done the impossible. He devised a plan to return believable elections to Florida by turning optical scan forms into the ballot of record, to be counted by citizens in public areas where all can view the process taking place. This radical departure from the maze of today’s computerized voting harkens back to over 100 years of U.S elections history. But first, a little background.

We all want to vote and know that our votes are counted properly and that the true winner of any election won fair and square. Since 2000, it’s become virtually impossible to know what happens to our votes if we’re voting on computerized voting machines (touch screens) or on paper ballots totaled by the other computerized voting machine, optical scan readers. Lately we’re hearing terms like paper trails and verified ballots. These are just slightly more sensible than undervotes and overvotes. Casting aside bureaucratic jargon and the explanations of hired gun experts, we know this much.


When our votes enter a touch screen machine, we have no idea what happens to those votes.

When we mark a special paper ballot read by a computerized optical scan reader, we have no idea how the readers operate or if they’re operated properly.

If we ask to watch vote counting, we’ll almost always be told NO or, if allowed, we’re placed in a distant corner like a six year old having a tantrum.

Finally even if we’re given a pass to watch computerized vote counting, we end up observing a box covering a computer most of us don’t understand, provided by vendors (any of them) that seem to blame any problems with the election process on …. you guessed it … YOU, the citizen, the tax payer.. Comforting isn’t it.



www.scoop.co.nz/stories/HL0703/S00150.htm

[eridani]

33. This is a good plan overall.

However, where I live, counting so many different ballot forms accurately by hand is a real problem. Also, Faulker is counting on a great deal of citizen participation, and we still have a lot of work to do getting enough people to participate as elections observers.

I think that by far the most important strategy is to get mandatory auditing of paper ballots that are the official ballot of record in place everywhere. I hate the idea that any aspect of our election system should be owned by private entities as much as anybody, but this is going to be a long fight. Audits first, and then keep pushing for major overhaul of the election process which kicks the notion of private ownership of any of its aspects to the curb.

Another analogy--laboratory scales need to be accurate, or none of your results are any good. Given a choice between a schematic of the scale design publicly posted on the wall and a set of standard weights that you can use to check its performance on a daily basis--which is more important to have?

[kster]

34. Absolutely , I agree Audits and Hand Count,s, we need to get both

how hard would it be to include both, if a Citizen group request the ballots immediatly following the election they should not have to jump through hoops to get them, It should be automatic upon request.

Where you live there may be people who do not mind Hand Counting all the Ballots, those people shouldn't be denied Hand Counting all the Ballots, should they?

[eridani]

38. Where I live, there are NO people willing to use two weeks of vacation time--

--on counting ballots. We have enough trouble getting poll workers and elections observers for just a single day.

[kster]

36. I'm just curious what you thought,. should the people be denied access

to all the paper ballots` immediately following the election? I think its a perfect plan, your group wants to do audits, and my group wants to Hand Count all the Paper Ballots, talk about checks and balances it would be a two-fer, I would never deny your group being able to audit the ballots and I'm sure you wouldn't deny my group from Hand Counting all the Ballots, Would you?

[eridani]

37. It would be pointless to deny you the opportunity to flap your wings and fly,no?

With the ballot complexity that we have, it isn't possible to do accurate hand counts. We don't have the people, and if it took WA state three weeks to recount a statewide gubernatorial race, how many months before we can get done with ballots for state legislative districts that are split between three congressional districts and four county council districts?

Now, in smaller venues with simpler ballots, handcounting is much more feasible, and I have no problem with it. But the process still needs a separate audit.