This is really great. But the subsequent information I’ve received from the Columbia County, NY commissioners in reply to my query is even better!
For a start, here’s the key parts of the story from Debora Gilbert at The Columbia Paper near Albany, New York.
Note, in particular, how both the Republican and Democratic commissioners concur on what should not be a partisan issue. They are doing a great service to their voters. Read the story and then I’ll share the even better news with you below that…
But they question the accuracy of the results for the new machines and see no reason to stop checking them by hand.
“The most accurate and reliable method is a 100% visual audit,” Elections Commissioner Jason Nastke (R) said Tuesday. He referred to multiple scanner miscounts in Greenport in a past election. “The machines are not completely reliable,” he said.
“Hand counting allows for voter intent to be taken into consideration,” said Election Commissioner Virginia Martin (D). “If someone has circled rather than filled in the ovals, it counts when the ballots are hand counted, but with machine counting, the only allowed discrepancies involve machine error, not human error.”
Optical scanners were viewed as a welcome alternative when compared to the more expensive and less secure touch screen machines that were the only other option during the push to have counties purchase and install new voting machines in 2009.
In a commentary published in the Times Union newspaper in Albany last week, Buffalo attorney Peter Reese wrote that the new voting system has led to “tens of thousands of invalid overvotes,” citing specific cases, including one near Buffalo that he was involved with. He blames New York’s HAVA compliance law.
Many voting activists favored the optical scanning system because it uses paper ballots that should allow an easy method of verification. Many hoped all the ballots would be posted online to allow easy oversight by citizens. But the law only calls for a small, random sampling of 3% of the machines to monitor accuracy. In Columbia County that would mean hand counting the results of one machine, said Ms. Martin. Both commissioners agree that a sample of this size is inadequate.
…
“The beauty of our situation is that the two election commissioners have agreed in advance to hand count ballots,” said Ms. Martin. After a primary, when one party might have the lead, it would be tougher to agree on something like this, she said.
“Although we’re looking into a different kind of audit, it seems easier just to hand count and you have great voter confidence”. We have a very strict chain of custody procedure, with redundancy built into our system. Those who see our accounting procedures know we leave nothing to chance,” Ms. Martin said.
I couldn’t be more delighted than to read the above. Well, actually, I could be.
From the story, as written, it was unclear whether or not Columbia was hand-counting in a central location, after ballots were moved from the polling place, or whether they were actually hand-counting at each precinct after the close of polls, with all parties, video-cameras and members of the public overseeing the counting and results posted decentrally at the polling place before ballots are moved anywhere.
So I dropped a note to both Commissioners Nastke and Martin to see if they were counting centrally (after ballots had been moved, and thus, questions about the secure chain of custody could arise) and, if so, whether they had any plans to try hand-counting publicly at the polling place in future elections, as per what we consider around here to be “Democracy’s Gold Standard”.
I received a very speedy — and quite encouraging — reply from Virginia Martin. Here it is in full…
Yes, we count centrally. Yes, we have talked about doing, and would like to do, pollsite counts immediately after the close of polls. There are folks in our county who are very eager to have us do so. The fact, though, is that we just haven’t yet been able to pull it together to do those at-the-polls counts. We’ve talked about trying it first in one or maybe two poll sites that are located in fairly close proximity to the Board, which would make it easier for us to oversee the counts as may be necessary. Then in subsequent elections we would expand our pollsite counts as proved feasible. It is indeed a goal.
As you can imagine, our ballot-custody procedures are exceedingly stringent and leave nothing to chance. We’re very confident that what we count is precisely what we should be counting.
We’re running four and possibly five elections this year (two is the norm for us), and I’m not sure if that makes the job theoretically easier or more difficult.
Yes, the proceedings are open to party representatives, candidates and their representatives, and the media. We haven’t turned anyone away.
Nice to hear from you. Thank you for your crucially important work in election integrity!
Virginia Martin
Democratic Commissioner
Board of Elections, Columbia County
Good news indeed! As long time readers know, we have long been calling for hand-counted paper ballot pilot projects, very much like the one Martin describes as being considered in her note. We’ll look forward to that in the future, if the democracy gods continuing smiling on Columbia County, NY.
As noted in the article, while Columbia uses hand-marked paper ballots, computerized optical-scan voting systems are simply not to be trusted, given their frequent inaccuracies and malfunctions and the ease with which they can be manipulated (as we have been documenting here for years.)
One recent example of such an op-scan failure in NY took place in 2010 in the South Bronx. But, incredibly enough, it came to light only a few weeks ago after the New York Daily News finally did a public records request for copies of the ballots from one of the precincts were anomalous results had been noticed by NYU’s Brennan Center for Justice.
As we detailed in late February, on the eve of the GOP primaries in Arizona and Michigan (where they use similarly unreliable optical-scan tabulating systems), the newspaper discovered that hundreds of paper ballots had been mistallied in both the 2010 primary and general elections at the one precinct they looked at.
An analysis by the Daily News found that the error rate of the op-scan systems during the September primary that year was 70% — with 69 errors out of 103 ballots tallied. In November’s general election, 156 out of 289 paper ballots had been mistallied by the op-scanners, for a failure rate of 54%. All of that, discovered almost two years after the elections.
With that in mind, Columbia County’s Martin looks pretty smart when we look back at her resistance over the years to move to the computer systems the state was forcing counties to move to as replacements for the lever systems used across much of NY.
In 2009, she stunned state legislators at the NY State Assembly’s Standing Committee on Election Law when she testified as to why she refused to rely on the computer systems they were forcing on her county (and all of the others in the state): “If Columbia County starts using software to count votes, I will not certify an election unless an appropriately designed audit of the paper ballots is conducted. So far, the State Board [of Elections] has not mandated an audit that audit experts agree will expose inaccurate counts.”
In 2010, in her must-read editorial Martin said that “Our state handed us a $50-million lemon when it required that we use computers to count votes.” She went on to detail her reasons for insisting on hand-counts: “Since I, as election commissioner, have to certify to the accuracy of any election run under my watch, that steers me in the direction of a more elemental process — a hand count under the watchful gaze of individuals who are invested in its accuracy.”
So, once again, kudos to the Election Commission in Columbia County, NY for taking the important steps to actually serve the voters by helping to ensure real confidence in their elections with hand-counted paper ballots. They are setting an example for the nation and I can only hope other election officials are taking notice.
[Hat-tip to @sjdorst on the Twitters for flagging the original Columbia Paper report!]
If it were actually random selection, and there were large victory margins, 3% could easily be a statistically significant sample.
The number of sampeds should really be at least on the order of 1 divided by the margin of victory squared. So, for example if there’s a 20% margin of victory, they should be testing 25 machines, 100 for a margin of 10% and, 10,000 for a margin of 1%.
A sample of 3% is likely to a resonable test for large elections, but not for small ones.
It would, of course, be nice if verification policies were created by people with a bit more statics sophistication…
Sounds to me like Columbia County isn’t there yet on the Gold Standard, but not for lack of trying!!
Hmmm
Needs a name
How about Bronze Standard? Bronze? Why not Silver? Here’s why: Both adhere to not using machines, but I’d like to keep Silver for when they actually start counting in the precinct. With Gold reserved for 100% precinct based, publicly observed and recorded counts.
And let’s go one further. Where there is both the Gold Standard enshrined in law, and the people riot against any attempts to weaken it, then we’ll have attained the Diamond Standard.
Wow! A bi-partisan election commission that actually gets it! How refreshing.
Now, if only the entire state of NY could have avoided throwing away $50 million on unreliable, e-voting systems, a large percentage of which, was no doubt gobbled up by foreign corporations like Dominion.
By hand-counting paper ballots before the press and public and publicly posting the results at each precinct, our election officials can do more than simply insure a democratic accountability that can come only with knowledge that the “winner” of an election actually received the most votes, but could produce a significant financial savings.
Moreover, the monies that would be spent on hand counts at the precinct level provide temporary local employment and are likely to be recycled into the local economy.
Whoopeeee!!!!
Nate (#1), the problem is actually getting a bona fide random audit. Better to do a bona fide hand count of all votes in the first place.
Congratulations and Thanks to the Election Commissioners in Columbia CO., NY. I have been suspicious of the HAVA legislation from Day #1. Any law that carries the fingerprints of Bob Ney and Diebold is not to be trusted.
Brad,
I just want to mention that the Help America Vote Act in the United States Code passed by Republicans after Rebecca Mercurio got involved is unconstitutional with regard to the conduct of state elections. The US Supreme Court struck down the statute passed by Congress to lower the voting age in state elections to 18 around 1970. The court said that Congress has authority over federal election procedures only and not over state, county, and municipal elections. The US Congress had to approve an amendment to the US Constitution to lower the voting age in state elections. That particular amendment was ratified by the states. That means that the Help America Vote Act has no authority over the conduct of state elections by state authorities. All the procedures including altering the requirements for voter registration are void and unconstitutional. Noone ever challenged the stupid Help America Vote Act as unconstitutional as regards the conducting of state and state subdivision elections. Brandon
“All the procedures including altering the requirements for voter registration are void and unconstitutional.”
Amendments 15, 19 and 26 cover state elections, and certainly give the SCOTUS enough to hide behind while justifying HAVA. Strangely, amendment 24 (the poll tax amendment) does restrict itself to federal elections.
This is fantastic news and the commissioners deserve our gratitude. I have a few concerns, very minor compared to those I have with other jurisdictions.
First, it’s important that the counts are carried out under circumstances that allow true, complete observation by people other than the commissioners. Second, any time anyone says something like “we’re very confident of our stringent procedures,” that should serve as a red flag. That confidence is irrelevant — what matters is the public’s ability to satisfy itself, not the commissioners’ being satisfied.
I’m also concerned about your first comment, by NateTG, which suggests moving from a good solution — hand counting the ballots — to a not very good solution — polling them.
If polling is good enough for verification, it must be good enough for the first count, right?
Using the polling logic, why not just take a random sampling of the ballots to begin with, to decide who won (within a statistical likelihood, as understood by statisticians and some small set of other college graduates)?
Thanks Brad, for your continued coverage of one of the right wings’ concerted effort to overthrow our government by ballot. ALEC, and the the rest of the anti-government republican extremists are trying to deflect the spotlight off of the real crime of election fraud by using voter fraud to muddy the waters. There are some who see this, but not enough of us to any good. thanks
Here’s absolute proof that Virginia Martin does it right. Columbia was the ONLY NY county in which Kerry improved over Gore. In 2004, Election Fraud in NY (and elsewhere) was massive. But not in Columbia, where Kerry’s recorded vote count increased by 14% over Gore’s while Bush declined by 9%. Columbia was the only county out of 67 in which Bush lost votes from 2000.
http://www.richardcharnin.com/N...LeverFraud.htm
Let’s do the math:
In 2000, Gore won the recorded vote by 60-35% with 4% to Nader.
In 2004, Kerry won a bogus 58-40% recorded share. He had 64% in the NY exit poll.
Returning Nader voters went for Kerry by 5-1 over Bush. So all things being equal, Kerry’s had 63-36% at minimum. But he did better than that since there was a net 2% defection of Bush 2000 voters to Kerry.
This graph displays the NY Bush Urban Legend myth: that he gained over Gore in urban counties. No, he stole urban votes. NY had a nearly 12% margin vote discrepancy from the exit poll.
http://www.richardcharnin.com/T...1_image001.png
Richard Charnin @ 10:
I’d be more careful, if I was you, in using “absolutes”. Columbia County, NY was not doing hand-counts until 2010. In 2004 — the time you refer to “in which Kerry improved over Gore” — they were still using lever machines.
Yes, but Virginia Martin was the Election Supervisor. And once again, Brad, you disregard the numerical evidence staring you right in the face. Like you did with Oregon.
Your concept of absolute proof is too strict.
I prefer: BEYOND A REASONABLE DOUBT.
DO THE MATH.
Lever machines are not the issue as much as election supervisors who do their job.
Unfortunately in NY 2004, only Virginia did the job right. You should give her more credit.
The RECORDED vote data shows clearly that Columbia was the ONLY NY county in which Kerry did better than Gore. Out of 67 counties. That tells me a lot. It should tell any impartial observer the same thing.
What are the odds that Kerry would gain in just ONE county without that county doing something special in its vote counting? And having a special supervisor?
Columbia’s success is replicated in all of Oregon’s counties. By the numbers. By the math. By the logic. BEYOND A REASONABLE DOUBT.
DON’T FIGHT THE NUMERICAL EVIDENCE. EMBRACE IT.
I had the unpleasant experience of hearing Brit Hume (FOX) speak to a group of CEO’s regarding the then upcoming 2010 mid-term election.
One chilling tidbit of information he offered is tha we should..and I quote… “ignore the exit polls”.