Post-Election Audits | CTVotersCount.org

April Presidential Primary Audit – Does Not Make the Grade

Checks on State Voting Machines Do Not Make the Grade
Do Not Provide Confidence in Election System, Says Citizen Audit

From the Press Release:

Audits of the recent presidential primaries are so faulty that exact final vote tallies cannot be verified, says the non-partisan Connecticut Citizen Election Audit. Unless state and local election officials make changes, the same will be true for the November elections.

“State law requires audits to verify the accuracy of optical scanner voting machines as a check for errors and a deterrent to fraud. Local registrars gather officials to manually count paper ballots and compare their totals to the totals found by the scanners”, explains Luther Weeks, Executive Director of Connecticut Citizen Election Audit.

Issues reported by the group were:

Incomplete or missing official reports of vote counts from town registrars;
The lack of action on the part of the Secretary of the State’s Office to check that all required reports are submitted and all submitted reports are completed fully;
Of 169 municipalities required to submit lists of polling places before the election, the Secretary of the State’s Office recorded only 68, with 101 missing;
Poor security procedures to prohibit ballot tampering;
Not following procedures intended to ensure “double checking” and “blind counting” rather than having scanner counts as targets while counting manually;

“The public, candidates, and the Secretary of the State should expect local election officials to organize proper audits and produce accurate, complete audit reports. The public and candidates should expect the Secretary of the State’s Office to take the lead in ensuring the audits are complete. Yet, due to a lack of attention to detail and follow-through the audits do not prove or disprove the accuracy of the reported primary results,” Weeks said.

Checks on State Voting Machines Do Not Make the Grade
Do Not Provide Confidence in Election System, Says Citizen Audit

From the Press Release:

“State law requires audits to verify the accuracy of optical scanner voting machines as a check for errors and a deterrent to fraud. Local registrars gather officials to manually count paper ballots and compare their totals to the totals found by the scanners”, explains Luther Weeks, Executive Director of Connecticut Citizen Election Audit.

Issues reported by the group were:

Incomplete or missing official reports of vote counts from town registrars;
The lack of action on the part of the Secretary of the State’s Office to check that all required reports are submitted and all submitted reports are completed fully;
Of 169 municipalities required to submit lists of polling places before the election, the Secretary of the State’s Office recorded only 68, with 101 missing;
Poor security procedures to prohibit ballot tampering;
Not following procedures intended to ensure “double checking” and “blind counting” rather than having scanner counts as targets while counting manually;

“The public, candidates, and the Secretary of the State should expect local election officials to organize proper audits and produce accurate, complete audit reports. The public and candidates should expect the Secretary of the State’s Office to take the lead in ensuring the audits are complete. Yet, due to a lack of attention to detail and follow-through the audits do not prove or disprove the accuracy of the reported primary results,” Weeks said.

Of Prisons, Water, and Elections

A story about prisons claimed that officials look at a prison as a jug of water. Even with a small pinhole leak, the water will get out. They look for the slightest weakness in the prison, assuming prisoners (with lots of time on their hands, collective wisdom, and little to lose in trying) will find any weakness, no matter how small, difficult, and time consuming.

That is how we should look at voting systems

For justified trust and credibility it is critical that our elections be publicly verifiable.

I don’t have the original quite. A story about prisons claimed that officials look at a prison as a jug of water. Even with a small pinhole leak, the water will get out. They look for the slightest weakness in the prison, assuming prisoners (with lots of time on their hands, collective wisdom, and little to lose in trying) will find any weakness, no matter how small, difficult, and time consuming.

That is how we should look at voting systems: electronic, Internet, mechanical or manual. If there is a weakness in the system, someone motivated will find it and exploit it. When it comes to attaining publicly verifiable results, recounts, and audits, any opening for breaking ballot security or transparency, someone motivated will find it and likely exploit it

For more details, review our Common Sense Series post on Public Transparency and Verifiability <read>

Security Against Election Hacking

From Freedom to Tinker, Andrew Appel: Security against Election Hacking – Part 1: Software Independence <read>

We have heard a lot lately about the vulnerabilities of our elections to hacking. Both cyberhacking and unsophisticated insider attacks. Andrew Appel describes some common sense approaches to detect and deter error and fraud in our elections, covering three major vulnerabilities:

Incorrect or unavailable poolbooks.
Voting machines
Accumulation of results across polling places and jurisdictions

From Freedom to Tinker, Andrew Appel: Security against Election Hacking – Part 1: Software Independence <read>

Incorrect or unavailable poolbooks.
Voting machines
Accumulation of results across polling places and jurisdictions

Any of these computers could be hacked. What defenses do we have? Could we seal off the internet so the Russians can’t hack us? Clearly not; and anyway, maybe the hacker isn’t the Russians—what if it’s someone in your opponent’s political party? What if it’s a rogue election administrator?

The best defenses are ways to audit the election and count the votes outside of, independent of the hackable computers…

So the good news is: our election system has many checks and balances so we don’t have to trust the hackable computers to tell us who won. The biggest weaknesses are DRE paperless touchscreen voting machines used in a few states, which are completely unacceptable; and possible problems with electronic pollbooks.

In this article I’ve discussed paper trails: pollbooks, paper ballots, and per-precinct result printouts. Election officials must work hard to assure the security of the paper trail: chain of custody of ballot boxes once the polls close, for example. And they must use the paper trails to audit the election, to protect against hacked computers (and other kinds of fraud, bugs, and accidental mistakes). Many states have laws requiring (for example) random audits of paper ballots; more states need such laws, and in all states the spirit of the laws must be followed as well as the letter.

Read the full, brief article to understand the details of Appel’s recommendations.

In addition to paying attention to all these recommendations, Connecticut needs to attend to improving our existing post-election audit transparency, the security of ballots, and consider adding formal measures along these lines for check off lists and results reporting.

Warning: 15 states without paper records, half without audits

A Computer World article reminds us how much more there is to go to achieve verifiable, evidence based elections: A hackable election? 5 things to know about e-voting <read>

Voting results are “ripe for manipulation,” [Security Researcher Joe] Kiniry added.

Hacking an election would be more of a social and political challenge than a technical one, he said. “You’d have a medium-sized conspiracy in order to achieve such a goal.”

While most states have auditable voting systems, only about half the states conduct post-election audits, added Pamela Smith, president of Verified Voting.

Let us not forget that even states, like Connecticut, with post-election audits have a long way to go in making the audits sufficient to assure that election results are correct or confidence that incorrect results would be reversed.

A Computer World article reminds us how much more there is to go to achieve verifiable, evidence based elections: A hackable election? 5 things to know about e-voting <read>

Voting results are “ripe for manipulation,” [Security Researcher Joe] Kiniry added.

Hacking an election would be more of a social and political challenge than a technical one, he said. “You’d have a medium-sized conspiracy in order to achieve such a goal.”

While most states have auditable voting systems, only about half the states conduct post-election audits, added Pamela Smith, president of Verified Voting.

“That leaves a lot of gaps for confirming that election outcomes were correct,” she said. “In such a contentious election year, well, let’s just say it’s never a good thing to be unable to demonstrate to the public’s satisfaction that votes were counted correctly, whether in a small contest or large.”…

Twenty-three states used DREs without paper trails in the 2008 U.S. election, and 17 used them in 2012, compared to 15 states this year, according to information from the U.S. Election Assistance Commission and Verified Voting.

How not to audit, Chicago style vs. Connecticut style

This video has gone viral over the last week. It is a hearing of a public comment at the Chicago Board of Elections.

This would not normally, exactly, happen in Connecticut

This video has gone viral over the last week. It is a hearing of a public comment at the Chicago Board of Elections.

In summary:

The audit consisted of counting paper records.
The public was not permitted to observe the paper records, to confirm that they were counted correctly.
At the end of the count, the counters compared the result to the machine tape result.
It showed that the counts were way off.
The officials simply erased enough marks from one candidate and added enough to the other to make the audit count equal to the machine count.
And the Election Board apparently saw no reason for concern with the audit or with the election results.

This would not normally, exactly, happen in Connecticut

We do have procedures that require that observers be able to be close enough to see ballot marks etc.
In some audits when counts do not match, officials recount until everyone is confident in the result.
In some cases officials work to make the counts match in ways that are hard to follow, and provide little confidence in the reported results.
Some audits result in high differences between the scanner counts and audit counts reported on official forms to the Secretary of the State, most, with no confirmation, attribute the differences to “Human Error”.
Others reported without indicating “Human Error” result in the Secretary’s Office calling local officials to ask if they agree that the cause was “Human Error”.
We do not have an Elections Board that officially reviews and accepts the audit result.
We have a mandated audit report after each audit, with no specified deadline for its completion. The last such report released was for the November 2011 election.

EDITORIAL: General Assembly heading the wrong way on post-election audits

UPDATE: The bill passed the House unanimously, including several who responded to your emails with promises they would not vote to cut the audits.

The Connecticut Senate has passed S.B. 252. If the House passes and the Governor signs the bill it will be another national embarrassment for Connecticut, doing the wrong thing at precisely the worst time.

We have voter-verified paper ballots. To be valuable and provide confidence they must be used for strong, publicly verified, post-election audits. You can help. Tell your legislators that you want stronger audits, not weaker audits. Tell them to oppose S.B. 252. Then consider volunteering one day after each election and primary to observe with the Citizen Audit.

UPDATE: The bill passed the House unanimously, including several who responded to your emails with promises they would not vote to cut the audits.

S.B. 252 would cut our post-election audits from 10% of voting districts to 5%. S.B. 252 was originally intended to strengthen the audits, while providing savings for municipalities. The current version eliminates all the features in the original bill that would make the audits stronger. The current version saves, at most, just $15,000 more annually statewide over the original bill.

Those paying attention to the news in this primary season have heard many charges of potential election fraud, along with calls for post-election audits in Arizona, New York, and elsewhere. There is an embarrassing video of a faulty presidential primary post-election audit in Chicago, where the public was barred from observing the votes as they were being tallied by officials. Worse the counters had the original numbers in front of them. When their counts did not match by a wide margin, they added counts to one candidate and deleted counts from the other, so that the manual counts would exactly match the machine counts. Over the years there have been similar problems with the audits in Connecticut.

For the first couple of years in the Connecticut audits, some local officials barred the public from observing the ballots as they were counted, some made it very difficult for the public to determine the time and place of the audits. Secretary of the State Susan Bysiewicz fixed those problems in procedures, which officials have since followed, for the most part. Other problems have not been effectively addressed.

The Citizen Audit’s observations of Connecticut’s audits have shown that many are conducted well, yet many are not. Many are, at best, only marginally better than the one in the video from Chicago. Connecticut election officials often do not double check counts; they are often aware of the original totals as they count, and work to match machine counts rather than accurately count the paper ballot votes. Frequently officials use confusing, ad-hoc, non-transparent methods for combining totals from multiple ballot stacks and teams of counters. Local officials in their reports and the Secretary of the State’s Office in statewide reports attribute all differences in counts to “Human Error”. The last official statewide report released by the Secretary of the State was for the November 2011 election. Five years is too long to wait for reports on a critical aspect of democracy.

S.B. 252 is still entitled “An Act Concerning Post-Election Audit Integrity and Efficiency”. The original proposed bill was the result of a long negotiated compromise which would have strengthened our audits in return for a reduction in the audit of polling place optical scanners, a change long sought by the Registrars of Voters Association of Connecticut (ROVAC). Among other reforms to strengthen the audits, the bill would have subjected centrally counted absentee ballots to audit, subjected originally hand counted ballots to audit when there were large numbers of them, mandated investigations of significant discrepancies in counts, required stronger ballot security, and timely reporting of results by the State. Municipalities would have saved 40% of their current costs, reducing relatively low annual statewide audits costs from at most $150,000 annually to less than $90,000. The revised, one-sided, bill provides none of the benefits while providing just $15,000 more in annual savings across all municipalities in the State.

ACTION ALERT: CT General Assembly Should Not Be Weakening Election Audits

Amid charges of voting integrity lapses around the country, the Connecticut General Assembly is on its way to weakening, our already weak post-election audits. The Senate has already passed substitute S.B. 252. Please call your State Representative and ask them not to make that same mistake. Find your rep and their contact info at: https://www.cga.ct.gov/asp/menu/cgafindleg.asp Tell them:

As a concerned constituent, I urge you to oppose S.B. 252. The committee bill weakens our post-election audits. In a time of public concern with the primary process in several states, we should be strengthening, not weakening our post-election audits.

Call Today. The bill could be called for voting in the House at almost any time!

As a concerned constituent, I urge you to oppose S.B. 252. The committee bill weakens our post-election audits. In a time of public concern with the primary process in several states, we should be strengthening, not weakening our post-election audits.

Call Today. The bill could be called for voting in the House at almost any time!
Details:

For several years CTVotersCount, the Connecticut Citizen Election Audit, and others have been calling for stronger audits. The current audits do not subject all ballots to the potential for audit. The audits are poorly conduced and not reported upon accurately or in a timely manner by the state. (See Citizen Audit reports at: http://CTElectionAudit.org)

In 2015, we offered a compromise bill with several key reforms to strengthen the audits. The bill received uniform supporting testimony by the Registrars Of Voters Association of Connecticut (ROVAC), the Town Clerks Association, and the Connecticut Conference of Municipalities (CCM).

In 2016, the bill was improved to address some concerns of ROVAC. Yet the improved bill opposed by ROVAC and the Town Clerks. It was labeled as too costly by the Secretary of the State because it would require her office to report comprehensive results publicly. Interestingly, the Secretary is now touting an almost equivalent capability, recently implemented by her office, which has been under development for several years at $400,000+.

Instead of strengthening the audits, the Government Administration and Elections Committee has stripped all the strengthening provisions, leaving only the weakening reform of cutting the audit in half,

Testimony on Four Bills

This year we testified on four bills before the legislature. We supported two bills, and for a change opposed none. For the two we neither supported nor opposed,, we proposed changes to the same sections of the law addressed by the bills.

This year we testified on four bills before the legislature. We supported two bills, and for a change opposed none. For the two we neither supported nor opposed, we proposed changes to the same sections of the law addressed by the bills.

A bill we proposed, S.B. 252 would strengthen the current post-election audits, while adopting changes long sought by the Registrars of Voters Association Connecticut (ROVAC). It would also require that the electronic audits approved last year be publicly verifiable – without public verification we would simply be adding unverifiable “Black Box” auditing, to unverifiable “Black Box” voting (CT’s current manual hand counted audits are publicly verifiable . <Prepared Remarks> <Testimony>

We supported H.B. 5390 would require the Secretary of the State to archive directives and instructions on the web. We supported the bill which would do what we asked last year, when the Secretary’s properly identified directives and instructions were made enforceable. <Testimony>

We took no stand in favor or against S.B. 250 which would entail cutting off Election Day Registration (EDR) at 7:00pm rather than 8:00pm. We warned in our testimony that the Secretary of the State’s EDR procedures risked a serious election rights violation, due to the cut-off of incomplete registrations at 8:00pm, rather than letting anyone in line at 8:00pm have the opportunity to register and vote. From the questions asked during testimony of officials, it was clear that the Committee understood the issue we raised. <Testimony>

We took no stand on S.B. 251. It would allow officials to interrupt counting and paperwork completion on election night in municipal elections. It was similar to a change last year that would allow those same interruptions for State and Federal elections. We pointed out, as we did last year, that both parts of the law should provide an effective means for the public, candidates, and party officials to determine the date, time, and location of the continuation of counting. <Testimony>

Citizen Audit Cites Flaws in Official Election Audits

Again accuracy declined and write-in votes handled incorrectly
November 2015 Post-Election Audit Report

From the Press Release:

The Connecticut Citizen Election Audit has released its report on its observation of the November 2015 official post-election audits. The audits, required by state law, are intended to verify the accuracy of elections at the municipal level.

Citizen Audit spokesperson Luther Weeks stated, “After 9 years of official audits, voters should expect accuracy. Yet the audits have gone from poor to worse.”

The group’s observers found that official audit results do not inspire confidence because of continued:

Discrepancies between machine counts and hand counts of votes reported to the Secretary of the State by municipal registrars of voters.
Lack of investigation of such discrepancies, and the lack of standards for triggering investigations.
Lack of consistency, reliability, and transparency in the conduct of the audit.
Weaknesses in ballot chain-of-custody and security.

The group’s report noted:

28% of official audits cited “Human Error” in counting ballots and votes. Registrars of voters should be expected to take the necessary effort to count accurately.
Significant decreases in audit integrity, and accuracy.
In three towns audits detected districts where officials fed write-in ballots through scanners a second time on election night.
If the group’s recommendations from last year had been mandated and followed, all write-in ballots would have been counted accurately.

“Problems discovered counting write-ins two years in a row shows the value of the official audits. But the report also reveals the decline in official attention to the audits, demonstrating that independent citizen observation and reporting are essential to election integrity.” Weeks emphasized.

Again accuracy declined and write-in votes handled incorrectly
November 2015 Post-Election Audit Report

From the Press Release:

Citizen Audit spokesperson Luther Weeks stated, “After 9 years of official audits, voters should expect accuracy. Yet the audits have gone from poor to worse.”

The group’s observers found that official audit results do not inspire confidence because of continued:

Discrepancies between machine counts and hand counts of votes reported to the Secretary of the State by municipal registrars of voters.
Lack of investigation of such discrepancies, and the lack of standards for triggering investigations.
Lack of consistency, reliability, and transparency in the conduct of the audit.
Weaknesses in ballot chain-of-custody and security.

The group’s report noted:

28% of official audits cited “Human Error” in counting ballots and votes. Registrars of voters should be expected to take the necessary effort to count accurately.
Significant decreases in audit integrity, and accuracy.
In three towns audits detected districts where officials fed write-in ballots through scanners a second time on election night.
If the group’s recommendations from last year had been mandated and followed, all write-in ballots would have been counted accurately.

“Problems discovered counting write-ins two years in a row shows the value of the official audits. But the report also reveals the decline in official attention to the audits, demonstrating that independent citizen observation and reporting are essential to election integrity.” Weeks emphasized.

Brennan Center: Election Integrity: A Pro-Voter Agenda

Whenever we open a report with multiple recommendations we start from a skeptical point of view. We expect to agree with some proposals and disagree with others. A new report from the Brennan Center for Justice is the exception. We agree with every recommendation:
Election Integrity: A Pro-Voter Agenda

It starts with the right criteria, it has a great agenda, strong supporting arguments, and ends with an appropriate call to action

It starts with the right criteria it has a great agenda, strong supporting arguments, and ends with an appropriate call to action:

This history strongly suggests two overarching principles that should guide any further efforts to secure election integrity. Such efforts should have two key elements:

First, they should target abuses that actually threaten election security.

Second, they should curb fraud or impropriety without unduly discouraging or disenfranchising eligible voters.

Efforts that do not include these elements will just result in burdens to voters and little payoff.

One: Modernize Voter Registration to Improve Voter Rolls

Two: Ensure Security and Reliability of Our Voting Machines

Three: Do Not Implement Internet Voting Systems Until Security is Proven

Four: Adopt Only Common-Sense Voter Identification Proposals

Five: Increase Security of Mail-In Ballots

Six: Protect Against Insider Wrongdoing

We do not have to choose between election integrity and election access. Indeed, free and fair access is necessary for an election to have integrity. This report examined genuine risks to the security of elections, highlighting current vulnerabilities as well as those that will be faced in the future. Recommendations have been made about how to reduce each risk. We invite and urge policymakers to tackle these problems.

As examples, we particularly support its call for sufficient post-election audits and attention to detecting, preventing, and punishing insider fraud:

Require Post-Election Audits. Many machines now issue a paper record of a voter’s selection. But these records are of little security value without audits to ensure that vote tallies recorded by a particular machine match any paper records. Despite near universal expert agreement on the need for audits, some vendors have vigorously opposed these paper trails, contending that they increase costs and slow the voting process. Security experts also recommend that states pass laws for effective “risk-limiting audits.” These require examination of a large enough sample of ballots to provide statistically “strong evidence that the reported election outcome was correct — if it was.” Also, the audit process should not rely on any one individual who might be in a position to manipulate either the voting machine or the recount device. According to experts, these insider attacks are the most difficult to stop. Voting technology experts also say machines must be “software independent,” which is technically defined as when “an (undetected) change or error in its software cannot cause an undetectable change or error in an election outcome,” but practically speaking means that the election results can be captured independently of the machine’s own software. Auditors should be assigned randomly to further ensure the process is not being gamed. Finally, audits should be as transparent as possible. This not only is essential to garnering public confidence, but can show a defeated candidate that she lost the election in a contest that was free and fair…

It is not surprising that many instances of election fraud, both historically and in the present day, involve the actions of insiders. Recent abuses by insiders have included lawmakers lying about where they live, magistrate judges willfully registering ineligible persons, and legislators running fraudulent absentee ballot schemes. A pollworker in Ohio was famously found guilty of using her authority and training to conduct voter fraud and take certain steps to evade detection. Culprits have even included the chief election officer of Indiana. This is why election officials and workers should receive special attention because their insider status increases their opportunity to both abuse the system and avoid detection. Moreover, when organizational leaders are involved in wrongdoing, it can create a culture for fraud, encouraging others to commit misconduct.