Comparing Voting Computers To Electric Meters

We often hear voting computers compared to ATMs. We have debunked< the notion that Voting Computers can be trusted like ATMs. Today an article by the Courant's consumer watchdog, George Gombossy, Once Again Meter Madness, has me considering how Connecticut’s Voting Computers and Electric Meters are the same and different.

Update:  Courant Editorial calls for Independent Audit <read>

We often hear voting computers compared to ATMs. We have debunked the notion that Voting Computers can be trusted like ATMs. Today an article by the Courant’s consumer watchdog, George Gombossy, Once Again Meter Madness, has me considering how Connecticut’s Voting Computers and Electric Meters are the same and different.

I suggest reading Gombossy’s article 1st and then returning here for the comparison: <read>
Same: Voting computers and electric meters are complex pieces of equipment that the ordinary citizen and voting official do not understand.

Different: The case of an electric meter is often transparent with a counter you can read that records usage by the customer. Voting computers have internal meters controlled by software that nobody can see or read as votes are accumulated.

Same: Both are sealed with tamper evident seals.

Different: Meter seals are there to keep the customer from stealing electricity. Voting Computer seals are there to keep insiders from stealing democracy.

Same: Meters are read and audited by employees of the electric company that is charged (no pun intended) with charging customers accurately. Voting computers are read and audited by election officials charged with running elections with integrity.

Same: When their electric meters are audited, the customer may be restricted to standing at a distance which precludes the actual observation of the function of the meter. When voting computers are audited the public may be restricted to standing at a distance which precludes the actual observation of the marks on the ballots being counted and the results being tabulated.

Different: The Courant’s watchdog has spent several columns investing and bringing the important issue of electric meter accuracy to the attention of the public. The Courant’s editorial page has, in the face of contrary evidence, touted the accuracy of our voting computers.

Different: The Attorney General is calling for independent testing of a suspect electric meter that has passed two tests by the electric utility. The Secretary of the State, some registrars, and at least one State Representative are considering calling for the elimination of manual recounts of our voting computers, even as some of those recounts and audits show differences in the voting computer results and the manual hand count of the voters intent.

Same: Just because many electric meters are tested and work, it does not mean that all electric meters will work correctly all the time. Just because many voting computers are tested and work, it does not mean that all voting computers will work correctly all the time.

Different: When your electric meter does not work, you get an odd, suspect, transparent bill that can be a trigger to you or the electric company to look for an explanation. When your voting computer flips votes, unless it is audited carefully, nobody will ever know.

Different: All electric meters of the same model are the same unless there is a mechanical flaw in one meter. Voting computers are programmed separately for each election, each district, and each race – each is a unique opportunity for error or fraud involving many voting computers.

Stamford Times: Advantages of new voting machines uncertain

Fortunato said it is too early to tell whether the new machines will save their owners money. Ballots for the optical scan machines cost 50 cents each but older model lever-action machines required costly maintenance.

“I think maybe at the end of next year, after you have the primary in August and election in November, you’d be able to determine was this more costly,” said Fortunato. “I don’t think you’d be able to do that right now.”

This article sets the complete wrong context and criteria: Is the paper cost of optical scan cheaper or more costly than paper for lever machines or DRE (touch screen) voting machines?

The voting machines and the entire voting system is worth nothing unless it provides integrity and confidence that the intention of the voters will be carried out. Cost of election administration per ballot? approximately $3.00; cost of the last Senate campaign? approximately $40 per ballot; cost of one mailing from your State Senator or Representative? approx $1.00; Cost of sufficient audits to protect our vote? $0.25; Value of preserving our right to choose our leaders? Priceless. <read the article>

Audit 54% of the Vote?

Last Monday, with less than 90 percent of the vote counted and the opposition leading by just 50.7 percent to 49.3 percent, President Chavez congratulated his opponents on their victory. They had defeated his proposed constitutional reforms, including the abolition of term limits for the presidency.

… An extremely large random sample – about 54 percent – of the paper ballots are counted and compared with the electronic tally.

Venezuela Is Not Florida by Mark Weisbrot <read>

Secretary of State Questions Electronic Voting

Electronic voting systems …still aren’t good enough to be trusted with the state’s elections, … she admitted having doubts as to whether the electronic voting systems will ever meet the standards she believes are needed … “I don’t rule out the ingenuity of some computer science student now in the eighth grade,” but what’s available now isn’t as transparent or auditable as the paper ballot systems they replaced…“When the government finds a car is unsafe, it orders a recall,” she said. “Here we’re talking about systems used to cast and tally votes, the most basic tool of democracy.”

Read the entire story highlighting California Secretary of State Debra Bowen <read>.

Continue reading “Secretary of State Questions Electronic Voting”

TalkNationRadio: Memory Card Junk Data and LHS Replacement – Integrity of Testing Procedures in Doubt

UConn is charged with testing a random sample of memory cards sent to registrars to be used in the election. If the actual cards used are replacements sent by LHS after the random card is sent to UConn, then there is no guarantee that the cards tested by Dr. Shvartsman were the same as those used in the election.

Once again Dori Smith presents significant information. Memory card failures are not just a Florida phenomenon. <listen>

Alex Shvartsman of the University of Connecticut’s Voting Research Team joins us to discuss his findings of “junk data” on memory cards delivered to the polls for the November 6, 2007 State and Municipal Election. The team will release their report on the memory card failures shortly.

Perhaps even more disturbing is instances uncovered in Dori’s work that indicate widespread failures during pre-election testing. Although its a small sample it indicates that problems may be much higher than those in Florida:

Continue reading “TalkNationRadio: Memory Card Junk Data and LHS Replacement – Integrity of Testing Procedures in Doubt”

Bysiewicz To Consider Elimination Of Manual Recounts

Last year the Legislature passed PA 07-194 mandating audits of the optical scan machines. Last year, Secretary of the State, Susan Bysiewicz promised advocates that regulations would mandate that recounts be manual counts and did not need to be included in the law. Now with less than half the mandated audits complete, according to the New London Day she is reconsidering that promise. Just one example of why it is not safe to rely on regulations and procedures to accomplish what should be in the law: <read>

Bysiewicz said if those results continue, she would recommend that future recounts be done by feeding the ballots into a different optical scanner from the one used during the election. Election workers would have to count only ballots that could not be read by the machines.

It seems she may not recommend that change for audits, just for recounts. But don’t voters have even more interest in seeing that their votes were counted as intended in a recanvass?

The audits were specifically designed and specifically mandated to answer the question, how does the voter know the machine counted their vote the way they intended,” Bysiewicz said. “The only way to do that is to hand-count them.

Recall that races involved in recounts are currently exempted from audit. So, without action by the legislature, the most important races to verify would by law be exempted from a hand verification.

Also the current state law does not give audit observers the right to visibly see the ballots such that they can visibly verify that the marks on the ballot are counted correctly. Indeed, how is a voter to know. Maybe by ‘faith’ when the ballots are ‘independently’ counted by the local election staff conducting the election?

“It is important voters have faith that their vote will be recorded accurately,” Mrs. Bysiewicz said, “and that’s why the independent audits are so vital.” <source>

Remember the election might be overturned in a recount and the reason might be because the people failed to accurately program and test the optical scanner: <recently in Riverside>

FAQ: Framing The Issue: Did the Machine Perform Flawlessly?

(Note: I have benefited from reading and contemplating the concept of framing issues from the linguist George ‘Don’t Think of an Elephant’ Lakoff founder of Rockridge Institute and cultural anthropologist Jeffrey Feldman, founder of the frameshop. While contemplating press reports on the recent election in a moment of sudden inspiration, I realized everyone has been asking and answering the wrong question – incorrectly framing the issues and our concerns with electronic voting. Here is my meager attempt changing the frame and starting toward more accurate understanding.)

We are asking the wrong questions when we ask if the audits proved that the “Machines Performed Flawlessly” or if the “Machines Incorrectly Counted The Votes”. We are using a misleading frame.

Continue reading “FAQ: Framing The Issue: Did the Machine Perform Flawlessly?”

Voting Machine Mess Can’t Just Be Fixed by Congressional Bills

Excellent article by Norman Ornstein of the American Enterprise Institute.  It is not just progressives that understand the problems with electronic voting.

In addition to outlining some of the problems, Mr. Ornstein also articulates very well the business reasons why voting machines and ATM’s are different, not just in their function but in the market limitations and economics.  As a former software product manager and buyer, I can vouch for the accuracy of that aspect of his analysis.

He suggests that Apple or Google solve the problem while others often propose open source solutions.  I don’t expect Apple or Google to do anything so significant without a profit motive, any more than I expect Walmart or the Hartford Courant to do so either.   My longstanding belief is that some things are best done by private enterprise, others by government and that either can be screwed up (with us being the ones at the wrong end of the screw).  Obviously the current method is not working for the benefit of Democracy.

Read his article <read>. Some excerpts below: Continue reading “Voting Machine Mess Can’t Just Be Fixed by Congressional Bills”

Blank Memory Cards and No Problems In Recounts?

Update: The Secretary of the State’s Office was offered four hours to comment on this entry before publication, but said they needed more time to gather information. We will post their response promptly.

New York Times: Voting Machines Are Put To The Test – reviews the optical scan performance in the municipal elections, with information from Secretary Byseiwicz <read>

I found a couple of items quite interesting. The following is not a direct quote from Secretary Bysiewicz but from the context seems to be information she discussed with Times. There is no indication of any other source associated with the information:

In a report after that audit, UConn researchers found, after a hand count, that in at least one race, the machine at a polling place in East Hartford counted six more votes for one candidate than the person had actually received. The discrepancy did not change the outcome, but they warned that it could in the future.

No such problems have turned up in 39 recounts of the Nov. 6 election,

This seems contradictory to the recent news describing recounts that resulted in a change in 17 votes in New Caanan and another error in Riverside that overturned the results. I suppose it depends on what is meant by “No such problems”, these may be different in some ways from those in East Hartford.

Also for the first time we learn that UConn has discovered some issues with the cards tested in the pre-election test. It seems that in some cases either a set of blank cards were sent to registrars or sets inconsistent cards, some blank, were sent to registrars.

Mrs. Bysiewicz said that in addition to reviewing the audit results of this past election, the University of Connecticut audited more than 300 memory cards prior to the election and will audit another batch now that the election is completed. The memory card, which she described as the “brains” of the new machines, posed the biggest security concern for UConn researchers, who warned in October 2006 that if someone gained access to the card, the results of an election could be altered.

Mrs. Bysiewicz said the pre-election audit of the cards found that none of the cards had been tampered with, but that a handful were blank, meaning that they had not been programmed with the proper ballot information.

No report yet on how many blank memory cards were discovered in pre-election testing by the registrars. If the same percentage of 6 out of 300 holds, then if each registrar in 695 towns tested at least two cards then about we could expect that about 28 would have been discovered.

Recount Reaffirm’s Wisdom of Recounts in Close Races

That’s not the tittle of article, nor what it says. It says:

with 14 towns holding recounts. New Canaan is included in that and held a recount on Tuesday to reaffirm the 25-vote difference between Town Council candidates Beth Jones and John Emert. That recount showed Jones edging Emert by seven votes.

No. It reaffirmed the winner, not the difference. If Jones had won by 17 votes, perhaps the article would not be so rosy. Then we might have Emert winning by 1 vote. What has been reaffirmed is the wisdom of recounting close races!!!

Read the entire article which includes some thought provoking quotes by voter Jody Eisemann, Secretary of the State Bysiewicz, and George Cody, President of the Registrars Of Voters Association Connecticut. <read>