Paper Ballots Integral to Connecticut Election Security – ANNOTATED

A recent article in the Journal Inquirer is at best misleading:  Paper Ballots Integral to Connecticut Election Security <read>

Connecticut has some good election integrity practices, yet there are gaps and vulnerabilities.

Full disclosure, I am a resident of Glastonbury and have been a poll-worker here since 2013 and prior to that from 2008 in Vernon, Connecticut. I take no pleasure in writing this post. Yet, even when people you know and appreciate provide, in your opinion, inaccurate or uninformed information to the public, it is not appropriate to ignore it. There is some good information in this article, yet it is not entirely accurate.

I absolutely agree that Voter Marked Paper Ballots like we have in Connecticut are the widely recognized basis of election security and integrity. Yet they are just a start.

A recent article in the Journal Inquirer is at best misleading:  Paper Ballots Integral to Connecticut Election Security <read>

Connecticut has some good election integrity practices, yet there are gaps and vulnerabilities.

Full disclosure, I am a resident of Glastonbury and have been a poll-worker here since 2013 and prior to that from 2008 in Vernon, Connecticut. I take no pleasure in writing this post. Yet, even when people you know and appreciate provide, in your opinion, inaccurate or uninformed information to the public, it is not appropriate to ignore it. There is some good information in this article, yet it is not entirely accurate.

Below is my annotation in blue:

Paper Ballots Integral to Connecticut Election Security
Election officials in the state are pointing to the benefits of physical ballots as the national conversation around election security continues to draw focus.

I absolutely agree that Voter Marked Paper Ballots like we have in Connecticut are the widely recognized basis of election security and integrity. Yet they are just a start.

by Alex Wood, Journal Inquirer / January 4, 2021  
 
(TNS) — With election officials around the country under very public attack, Mark Dobbins, the Democratic registrar of voters in Glastonbury, wants Connecticut residents to know more about the procedures election officials here use to make sure that all legal votes — and only legal votes — are counted.

It is a sad consequence in the atmosphere of this election that officials are under attack. Anyone who threatens an official and makes false or uninformed claims should be ashamed. And, in the most blatant cases where physical harm is threatened, prosecuted.

One is the old-fashioned paper trail, which Connecticut election officials use for many records, including ballots.

“We use a lot of paper, and you can’t hack paper,” Dobbins says.

Not exactly, even though paper ballots are a good start at election security and auditability, anyone with access to the storage of voted paper ballots can hack them, better still, with access to blank ballots as well, hacking is pretty straightforward. As we have pointed out over and over, Connecticut has very weak ballot security. In the majority of towns a single official may access ballot storage undetected for hours. polling place voted ballots are normally sealed with numbered seals which scientists have demonstrated offer little protection from skilled experts and amateurs alike.

Security experts also point out that over confidence in security by officials is a warning sign of lack of security. <See Do Connecticut’s Tamper-“Evident” Seals Protect Our Ballots?>

In addition, the tabulating machines that count ballots aren’t connected to the Internet and can’t be hacked into, he says.

It is of course important that scanners are not connected to the Internet, however, Hari Hursti long ago demonstrated how the machines used in Connecticut can easily be hacked  <The Hursti Hack> by anyone who has physical access to them. In Connecticut scanners are subject to that same weak security as voted ballots. Even computers not connected to the Internet can be hacked by Foreign actors <for example the STUXNET Attack>.

He adds that the tabulating machines are useless without memory cards. When the cards aren’t in use, he says, LHS Associates, an election services company based in Salem, New Hampshire, holds them securely.

First, the machines themselves can be hacked. They have software which could easily be changed by changing the chips, especially by a rogue service person form LHS Associates. LHS employs a number of people who program the memory cards. There is no audit or observation of LHS programming or security. At least none has been publicly disclosed. Unlike Connecticut, many states require that all programming of memory cards be accomplished by officials and not outsourced as Connecticut does. Here is a UConn papers describing additional vulnerabilities <here> <and here>.

Gabe Rosenberg, general counsel to Secretary of the State Denise W. Merrill, says the University of Connecticut’s Center for Voting Technology Research, or VoTeR Center, takes the memory cards before and after the election to make sure there are no problems.

 Actually, UConn takes a sample of memory cards. Its not a scientific sample. Registrars choose the cards to send to UConn after the election and don’t send all that are asked for. UConn is a little slow in providing reports.  The memory card audit report was for the 2014 elections. Here it is latest report, we suggest reading the summary <2014> <all reports>   It’s hard to trust an audit that is not random and publicly verifiable. There are some problems reported in those audits.

Officials audit 5% of the state’s voting precincts, Rosenberg says. The results exactly matched the machine counts this year, and the historical error rate is less than 1%, he says.

Connecticut does an audit of 5% of the voting districts, yet excluded from the audit are Election Day Registration ballots and centrally counted absentee ballots. In this COVID year of increased mail-in voting there are clearly gaps in the audit. Looking at the UConn reports, the last report of the audit completed was for 2016, which was completed about 19 months after the election, so at best we may have to wait a long while for official confirmation of that claim. It might be a really long wait since the previously completed report was for the 2011 election. UConn excludes large differences between machine counts and audit counts, with the aid of the Secretary of the State’s Office attributing all significant differences to human error in the counting process.

Another security measure widely used in Connecticut elections is to have “many eyes watching everything,” Dobbins says.

ELECTION SECURITY

Against hacking: Use of paper ballots and other paper records, vote tabulating machines not connected to the Internet, audits of vote counts

Monitoring: “Many eyes watching everything”

Physical secuirty: Locked rooms for blank ballots and other election supplies, completed absentee ballots stored in “cages” within vaults, police escorts

I am sure all of these practices are in place in some towns in Connecticut. Perhaps most Absentee Ballots are in vaults. It is different for polling place voted ballots: Most towns hold them in locked rooms or cabinets that have single key locks, with keys available to several officials providing individual access. Few towns have cages for polling place ballots Few use vaults for polling place ballots. Serving in several polling places in Glastonbury I have never had or seen a police escort. From my experience police escorts are not prevalent for the custody of voted ballots.

For example, vote tabulating machines are tested before the election, and members of the public “are welcome to come and look over our shoulder and watch us do it,” he says. Likewise, the public can watch as ballots are counted, he adds.

Yes testing is important to find ballot programming errors. Public observation of testing is also important. However, no level of testing can prevent errors and fraud. Every computer is subject error and fraud. For instance, your laptop and smartphone were tested before you received them, there is no assurance that they did not contain errors or supply-chain fraud. Voting machines are no different.

Dobbins says election equipment — from blank ballots to office supplies to the personal protective equipment that election workers needed this year — is kept in “blue bins” in a locked room around election time.

Once again, the rooms in many towns can be accessed by single officials for hours undetected.  Those blue bins which are from the same manufacturer all have the same keys.

He says he and Lisbeth Becker, Glastonbury’s Republican registrar, are the only people with keys to the room — and that no one is allowed to go in alone. Any time he goes in, Dobbins says, he must be accompanied by a Republican or an unaffiliated voter.

My understanding is that such access by two individuals is not monitored, only procedure prevents an individual from lone access. In fact, just a few years ago, prior to the current registrars, I reviewed a couple of pages of the access log and noted a single individual, a Deputy Registrar, had signed in alone several times. He and others could have easily done that even without signing in. (That same deputy was later arrested by the town for computer security violations).

On Election Day, Dobbins says, the number of ballots at each polling place is noted at the start of the day — and every ballot has to be accounted for at the end of the day.

Due to the vastly increased demand for absentee ballots amid the COVID-19 pandemic, the secretary of the state’s office had absentee ballot applications mailed to every registered voter this year.

Some have wondered whether a new resident of an address could send in an application in the old resident’s name, then vote by absentee ballot under that name while also voting under their own name.

But if the impersonated voter were to vote elsewhere, Rosenberg says, the fraudster would be caught. He adds that state law makes such fraud punishable by up to five years in prison and a $5,000 fine — and that several federal laws prohibit it as well — putting a high price on casting a single fraudulent vote.

Responsibilities for Connecticut elections are divided between registrars of voters and town clerks. The clerks’ responsibilities include sending out blank absentee ballots to voters who request them, receiving the completed ballots back, and storing them until it is time for the registrars to count them.

Manchester Town Clerk Joseph V. Camposeo says his office had to add part-time staff members to handle the increased workload, including data entry when ballots were sent out and when they were returned. He says the ballots were stored in a “cage” in his office’s vault for extra security.

My experience is that Manchester has security practices that are well above average in Connecticut.

Manchester had absentee ballot “drop boxes” behind Town Hall and at the police station this year. When he collected more than one or two ballots from the police station box, Camposeo says, he would have a police escort on the way back.

©2021 Journal Inquirer, Distributed by Tribune Content Agency, LLC.

In summary, many of these practices are good ones, if followed by every official in every town. Yet, they are insufficient and especially vulnerable to insider attack, including vendors and a variety of town employees, not just election officials. Elections are more vulnerable to outsiders when the insiders are over confident.

Security is difficult. Connecticut has a good basis with Voter Marked Paper Ballots, no Internet connections, and no Internet voting. Security practices are much more difficult given the nature of a state with 169 towns each with two registrars from opposing parties expected to be knowledgeable in all aspects of elections, including election security. Many are understaffed and underpaid. It’s a lot to expect that each official can understand, implement, and monitor security, while following the best practices of other states with larger election jurisdictions. Many Municipal Clerks are in the same boat. Yet surfacing issues can be the beginning of improvement. On the other hand, the distributed nature of Connecticut elections makes it difficult for localized errors and fraud to result in an inaccurate state-wide result, yet local elections remain more vulnerable.

We need stronger uniform, enforceable, and enforced security procedures across the state. Among other things Connecticut needs stronger tabulation audits, audits of ballot and scanner security.

 

How Far Have We Come Since 2016?

Have not posted much since the Election, there has been plenty of true and false information to read from all sources. Time now for a little perspective.

Recall 2016, when Jill Stein and others demanded recounts and audits in MI, WI, and PA. They were largely thwarted by officials. Little was possible in PA with no paper records of votes to count. Potential Russian hacking of epollbooks across a county in NC which was never credibly investigated. Government conclusions that there was no evidence that election systems were hacked, without checking for any. Those involved in the recounts/audits, such as the were, conclude that the saw not evidence of hacking. Yet the government acknowledged there were 18 states with attempts (often mis-characterized as 21 attempts) at accessing voter lists. To many, including yours truly, Georgia represented the most questionable state with Secretary of State Kemp on the ballot for Governor, vulnerable, critical election data left on a server before the election – destroyed by officials precluding forensic analysis, voter purges, and no paper records of votes.

We have come quite a way since then in the 2020 election…

Yet we have much farther to go:..

Have not posted much since the Election, there has been plenty of true and false information to read from all sources. Time now for a little perspective.

Recall 2016, when Jill Stein and others demanded recounts and audits in MI, WI, and PA. They were largely thwarted by officials. Little was possible in PA with no paper records of votes to count. Potential Russian hacking of epollbooks across a county in NC which was never credibly investigated. Government conclusions that there was no evidence that election systems were hacked, without checking for any. Those involved in the recounts/audits, such as the were, conclude that the saw not evidence of hacking. Yet the government acknowledged there were 18 states with attempts (often mis-characterized as 21 attempts) at accessing voter lists. To many, including yours truly, Georgia represented the most questionable state with Secretary of State Kemp on the ballot for Governor, vulnerable, critical election data left on a server before the election – destroyed by officials precluding forensic analysis, voter purges, and no paper records of votes.

We have come quite a way since then in the 2020 election:

  • Georgia and Philadelphia now have new Ballot Marking Devices providing paper records for recounts.
  • Georgia has a Risk Limiting (tabulation) Audit (RLA).
  • Georgia by RLA and Philadelphia by recount, counted their paper ballots by hand.
  • Georgia also by recount used machines to recount them all.
  • Georgia officials, mostly Republican, defended the claim that Biden won. Even Governor Kemp.
  • Other mostly Republican states defended Biden’s win.
  • The Government beefed up cyber defenses and monitoring.
  • Online media improved their monitoring and response to false information.

Yet we have much farther to go:

  • We need Voter-Marked Paper Ballots everywhere. Despite claims to the contrary, Ballot Marking Devices do not provide voter-verified ballots. Tests and observations clearly demonstrate that about 90% of voters make no effort to check their ballots, that many who try fail to check well, and most officials would understandably not believe them if they complain.
  • Many RLA laws, like Georgia’s, are inadequate and poorly written. They like many audit laws fail to come close to the Principles and Best Practices for Post-Election Audits. They don’t specify enough details/requirements for RLAs. They mis-state RLAs. They allow election officials to pick races for audit after the fact, all but encouraging picking races with large margins, rather than close races.
  • Georgia’s RLA was really a complete hand recount. Much more valuable than a RLA.
  • Georgia’s hand recount did not count. That is why the official recount was machine recount, an extra waste of time. An adequate RLA law could eliminate the need for machine recounts and result in a full hand recount only when necessary – like it was in Georgia this year.
  • Georgia’s hand recount was not well planned, and not transparent. We can applaud the recount and the hard work involved, yet next time there should be detailed procedures published in advanced, oversight to be sure those procedures are followed, and transparency – observers were unable to see and verify results sheets, compare results to original batch totals, and see that the correct numbers were entered into the results accumulation system, designed for an RLA and not a full recount.
  • The audits and recounts only covered vote Tabulation. Full audits should include transparent audits of the chain-of-custody, eligibility, and other aspects of election administration.
  • Many of these same issues apply to other close states this year and to many other states as well.

Once again, we applaud officials who did exemplary work in trying conditions, especially those defending results they would have had otherwise.

In a future post we will go over some of these issues in more details, comparing to the standards in the Principles and Best Practices.

 

 

November Election 2020 Post-Election Audit Drawing

Yesterday we observed the Post-Election Audit Drawing by the Secretary of the State Denise Merrill. 38 districts in 31 municipalities will be audited by no later than December 4, 2020.
The Secretary’s Office is offering electronic auditing to municipalities, subject to scheduling limitations. Official Press Release <read>

Yesterday we observed the Post-Election Audit Drawing by the Secretary of the State Denise Merrill. 38 districts in 31 municipalities will be audited by no later than December 4, 2020.
The Secretary’s Office is offering electronic auditing to municipalities, subject to scheduling limitations. Official Press Release <read>

Making Every Voter Equal: Uniformity, especially for ‘curing’ mail-in ballots.

For years we have been complaining that Election Day Registration in Connecticut may well be a civil rights violation. Now we find that we may well have a similar problem contained in the rushed procedures for handling the volume of mail-in voting.

For years we have been complaining that Election Day Registration (EDR) in Connecticut may well be a civil rights violation. Now we find that we may well have a similar problem contained in the rushed procedures for handling the volume of mail-in voting.

The basic requirement is that all voters and potential voters be treated equally to make elections fair to everyone.  We are talking making them equal in just a single state. Obviously they are quite different from state to state. E.g. in Connecticut felons can vote, while in some other states they cannot, and in FL they have great difficulties restoring their right to vote. E.g. Connecticut normally requires a excuse for mail-in voting, many states do not, while others add barriers of one or two witnesses to submit an absentee ballot. E.g. other barriers are fewer polling places or fewer voting machines in minority districts in some states.  We could go on.

When it comes to EDR, since the beginning it has caused inequality in even year elections in Connecticut. This is likely unintentionally, yet that makes it no less unfair and unequal, and no less a civil-rights violation waiting to be challenged. It may have finally been addressed this year as we have seen no reports of problems. In past even years we have seen long lines in Hartford, Bridgeport, and most glaringly in New Haven and in some university towns. The problem is two fold. First such towns have many, many more transient populations of voters that causes more demand for EDR. Second, especially in larger cities there less budget for elections and apparently registrars not up the the larger planning and logistic challenges for those cities. We are not saying that city registrars are less competent than many of their peers, but a bigger challenge demands in addition to more compensation (which they have) more planning and execution capability.

When in comes to mail-in voting, especially the rushed and ever changing process this time – there are lots of areas demanding uniformity. Here we do not know all the answers. As far as we know, all the clerks kept up with demands to turn-around absentee applications and absentee ballots in reasonable time. As far as we know, all towns processed and counted the large number of ballots in a reasonable time. However, when it comes to “curing” ballots from what we have heard there was no uniform process. Likely some towns made some attempt at cures and others did not, overall not very uniform and likely favoring some voters and disfavoring others.

What is “curing?” We define it as any process that allows voters to correct a situation where their mail-in ballot would otherwise be rejected. Perhaps signing an inner envelope. Perhaps attesting that it is indeed their signature. Perhaps being given an opportunity to supply a new ballot or to attest that they put their ballot outside the inner envelope or neglected to submit the inner envelope.

There is a lot to uniformity here.

  • The process has to be uniform across the State, articulated in detail, published for everyone,followed by officials uniformly, observable by the public, able to see the details are followed (or not followed), and enforced.
  • What can be cured, how it can be cured, how voters are notified, and the deadlines specified.
  • And in practice it has to be uniform and fair. If the small town has a hand full of cases, finds them five days before the election, knows how to contact each voter even though they don’t have current phone numbers or emails, and does so – then its not fair if the large city has hundreds, finds them later under huge processing demands, does not give voters about the same amount of time to cure, and cannot give equal attention to locating each voter – its unfair.
  • If some towns call out the names of every such voter, give access to observers from each party, so they can call such voters – and other towns don’t do that or its somehow harder for large town observers to do the work or are given less time – its unfair.
  • If some demographics have a harder time submitting acceptable ballot packages or a harder time actually curing then it may be unfair.

Cures can work. Fairness might include:

  • Standards for how quickly after receipt or after election day cures must be detected and noticed to voters.
  • Mandatory contact methods: Call and leave voice mails or not, send emails or not, and mandatory letters.
  • Enough days for a voter to receive a postal letter and respond.
  • Uniform and multiple methods for response. Some states allow cures by internet application, phone, email, postal mail etc. Some only by personal appearance which is tough for those with disabilities.

Overall it means that the deadline for counting all votes would need to extend at least a couple weeks after the deadline for receipt of a mail-in ballot.  It really should not allow a voter to submit a new ballot after election day. Procedures should be followed that do preclude officials knowing how a voter voted before choosing to offer a cure.

We are not sure Connecticut is ready to delay counting for a couple of weeks. If so, then there is a lot in the law that needs changing: When we call for and perform recanvasses, certification dates, and in some cases when municipal winners are declared and take office.

 

 

No, its not the time for more electronics in Connecticut’s voting

An Op-Ed in the CT Mirror: It’s time to modernize the way Connecticut votes.

The main trust is that we should do more electronic automation of the election process in Connecticut such as electronic transmission of results and electronic pollbooks, and alluding to less pens and paper in voting.

Perhaps we can forgive the author for accepting at face value the claims of vendors and their customers that have sunk unnecessary millions into questionable technology. Sometimes it works well and saves time and effort, sometimes it doesn’t!

  • Lets start with electronic submission of results. That idea has a couple of basic flaws…

Our bottom line: Never change from Voter Marked Paper Ballots unless there is some dramatic technological breakthrough. Avoid connectivity for voting machines. Cautiously consider electronic pollbooks, with mandatory paper backup systems. Keep using our current AccuVoteOS until they really need replacing – perhaps better more economical alternatives will become available, perhaps they will comply with the new Federal standards expected soon.

 

An Op-Ed in the CT Mirror: It’s time to modernize the way Connecticut votes <read>We disagree.

The main trust is that we should do more electronic automation of the election process in Connecticut such as electronic transmission of results and electronic pollbooks, and alluding to less pens and paper in voting.

Perhaps we can forgive the author for accepting at face value the claims of vendors and their customers that have sunk unnecessary millions into questionable technology. Sometimes it works well and saves time and effort, sometimes it doesn’t!

  • Lets start with electronic submission of results. That idea has a couple of basic flaws.
    • First its risky. No voting system should ever be connected to the Internet, have wireless connectivity, or be connected to phone lines. All that risks hacking of the voting machine itself. Experts have cautioned election officials against any such capabilities. The leading voting system vendor, ES&S has been caught lying to officials, a government agency, and the public trying to hide that they had that capability.  CT helped develop and uses a data collection system where tapes of results from machines are data entered into a system not connected to our voting machines – it may seem like a lot of work to Head Moderator’s like the author. Yet overall its not that big a deal e.g.  it my town it takes a few hours work by two officials, not that much in comparison to the 60 or so that work 17+ hours on election day.
    • Second, many CT votes are not counted electronically.  Votes on hand counted ballots and write-in votes are not counted by machines but the data must in any case be entered and reported. In other states some of this data is counted electronically by copying unscannable ballots onto other ballots for scanning – this is a labor intensive, slow, and error-prone process.
  • On to electronic poll books. Once again, risky, expensive and not all they are claimed to be.
    • The University of Connecticut tested them and found all those offered by vendors to be lacking in security. The analysis is confidential due to (unfortunate, undemocratic) agreements with the vendors that allowed the testing.
    • They are not as fast as and they are as error prone as manual lookup and voter checkoff on paper lists.
    • Many tout the advantages of not having to print all that paper, yet every expert warns that a paper backup is necessary to keep voting going in the face of power outages, Internet outages, and software or hardware failures.
    • Like everything connected to the Internet they are vulnerable to hacking. In 2016 there was a huge failure in an entire county in NC. There is no evidence of hacking, yet that is only because there was no credible investigation of a trail that my well have lead to Russia.
    • Once again, a little time on the part of election officials doing data entry saves millions in hardware and software acquisition and maintenance – and could provide jobs to Connecticut residents.
  • Several years ago the Secretary of the State got bonding of $6,000,000 to buy electronic pollbooks and a scheme for wireless transmission of results. She wisely turned it back to the State.
  • When it comes less paper and pens, we agree with the author that our current system is secure and accurate. The alternative, pictured with the Op-Ed seems to be and electronic Ballot Marking Devise (BMD). They are risky and expensive.
    • Life time costs for acquisition and programming are at least double, perhaps triple that of optically scanners and paper ballots. One of our current scanners handles the volume of ballots in all but a couple of polling places in the State. A couple handle most central count absentee ballot locations. We originally bought two per polling place, with consolidation there are a number of extras around, they can be purchased very reasonably used online. In fact, the Secretary of the State purchase a number of spares a couple of years ago.
    • BMDs cost lots more because you need many more per polling place. Each must be acquired initially, maintained, programmed, and tested for each election.
    • BMDs and their more risky predecessor technology, DREs, are the cause of lines and polling places. Not scanners in Connecticut and in most jurisdictions.
    • BMDs are subject to hardware and power failures. To continue in spite of power failures there needs to be a sufficient supply of paper ballots in every polling place (that would presumably need to be counted by hand).
    • Finally, BMDs are risky. Research shows that voters do not and cannot reliably check the paper “ballot” they produce, and have a hard time convincing officials the BMD made errors, not the voter.
  • We agree that our scanners are old.
    • Unlike the author, the CTEletionAudit.org surveys of registrars after every election have not indicated any rise in scanner failures.
    • Newer technology scanners available today are marginally better then the AccuVoteOS scanners we have now. The create ballot images and files containing Cast Vote Records, both of which support more comprehensive, less  expensive audits.
    • Yet, the new systems are each more expensive and slower. Many more polling places would need multiple machines to process the volumes of votes we have in Connecticut. They are just as vulnerable to hacking and thus should never be connected to the Internet, phone lines, or wireless for electronic communication of results. And still those hand counted and write-in votes need to be reported manually.
    • Many more will be required for more central count absentee locations. ES&S provides high speed scanners. Two count all the absentee ballots in Rhode Island. But Rhode Island is not Connecticut. They count all there absentees centrally and also program and warehouse all their scanners centrally in one place in Providence. We count and manage everything in each of 169 towns. Those high-speed scanners are too expensive to deploy for local AB counting.

Our bottom line: Never change from Voter Marked Paper Ballots unless there is some dramatic technological breakthrough. Avoid connectivity for voting machines. Cautiously consider electronic pollbooks, with mandatory paper backup systems. Keep using our current AccuVoteOS until they really need replacing – perhaps better more economical alternatives will become available, perhaps they will comply with the new Federal standards expected soon.

 

New Paper: Evidence Based Elections

A new paper by Andrew Appel and Philip Stark: EVIDENCE-BASED ELECTIONS:CREATE A MEANINGFUL PAPER TRAIL,THEN AUDIT  Provides a thorough description of how the public can be assured of election outcomes, in spite of hack-able voting equipment.

The bottom line: The only reliable method available is Voter-Marked Paper Ballots, with strong security for the ballots, followed by sufficient post-election audits. Other technologies, including Ballot Marking Devices and Internet voting are insufficient.

Anyone interested in trustworthy elections should read this paper – especially those who think that expensive Ballot Marking Devices should be trusted. And those who think it is impossible to use technology to count votes accurately.

A new paper by Andrew Appel and Philip Stark: EVIDENCE-BASED ELECTIONS:CREATE A MEANINGFUL PAPER TRAIL,THEN AUDIT  <read> Provides a thorough description of how the public can be assured of election outcomes, in spite of hack-able voting equipment.

The bottom line: The only reliable method available is Voter-Marked Paper Ballots, with strong security for the ballots, followed by sufficient post-election audits. Other technologies, including Ballot Marking Devices and Internet voting are insufficient.

Anyone interested in trustworthy elections should read this paper – especially those who think that expensive Ballot Marking Devices should be trusted.

The vulnerability of computers to hacking is well understood. Modern computer systems, including voting machines, have many layers of software, comprising millions of lines of computer code; there are thousands of bugs in that code. Some of those bugs are security vulnerabilities that permit attackers to modify or replace the software in the upper layers,so we can never be sure that the legitimate vote-counting software or the vote-marking user interface is actually the software running on election day. One might think, “our voting machines are never connected to the Internet, so hackers cannot get to them.” But all voting machines need to be programmed for each new election: They need a “ballot-definition file” with the contests and candidate names for each election, and lists of the contests different voters are eligible to vote in. This programming is typically done via removable media such as a USB thumb drive or a memory card. Vote-stealing malware can piggyback on removable media and infect voting machines—even machines with no network connection. There is a way to count votes by computer and still achieve trustworthy election outcomes. A trustworthy paper trail of voter selections can be used to check, or correct, the electoral outcomes of the contest in an election…

If a BMD is hacked and systematically steals 5% of the votes in one contest and only 7% of voters inspect their ballots carefully enough to notice, then the effective rate of vote-theft is5% ?93% ,or 4.65%;this is enough to change the outcome of a moderately close election. The same analysis applies to a DRE+VVPATsystem.One might think:“not everyone needs to carefully verify their ballots;” if only 7% of voters carefully inspect their ballots, they can serve as a kind of “random audit” of the BMDs. But this sentiment fails to hold up under careful analysis…

in our hypothetical scenario in which a hacked BMD steals 5% of the votes, and 7% of voters carefully inspect their ballots (and know what to do when they see a mistake), then7% ?5% ofvoters will alert a pollworker; that is, 1 in every 285 voters will claim their paper ballot was mismarked—if the voters do not assume it was their own error. The BMD would successfully steal “only” 4.65% of the votes.One might think:“but some voters caught the BMD cheating, red-handed.” But nothing can be done. It is a rare election official who would invalidate an entire election because 1 out of 285 voters complained.

Special Session Election Bill – Safe and Reasonable

Last week the General Assembly passed and the Governor signed a bill to help deal with a flood of mail-in ballots with a special kicker for Bridgeport.

Our summary, its good, it could have been better, or a lot lot worse.

Last week the General Assembly passed and the Governor signed a bill to help deal with a flood of mail-in ballots with a special kicker for Bridgeport <read>

Our summary, its good, it could have been better, or a lot lot worse.

The Good

  • It provides an Election Monitor for Bridgeport. Bridgeport needs one for every election an primary, not just this election. My experience is that officials in Bridgeport generally know what they are doing. But the result often looks suspicious with many absentee ballot hi-jinks. The last time they had a monitor it resulted in the cancelling and rerunning of a primary. The question is, who will be that monitor and will they be up to the task?
  • It will provide two to three days extra time for clerks and registrars to do the per-election and some of the post-election work so that ballots can be counted more quickly after the election.
  • More time and more timely are good things. It will make all the work more orderly, efficient and accurate, while better satisfying the unfortunate demand for quick results.
  • In our opinion, it is realistic, complies with the spirit of existing law, and it does not compromise security in any significant way.

It Could Have Been Better

  • Realistically it only gives the registrars two additional weekend days to work a temporary staff that will be working full-out the next three or four days during the week. It gives the Clerk extra time to prepare for the work of the Registrars’ staff.
  • It requires that withdrawing a mail-in vote to instead vote at the polls be done by Friday at 5::00pm. It could have been done otherwise.
  • We would allow Clerks to organize ballot packets by district, street, and number as they come in. Not starting at some particular date. We have no idea why not, in any electon.
  • We would allow opening the outer envelopes starting the Monday, eight days before the election, hopefully avoiding much of the weekend work.
  • We would not separate inner and outer envelopes,thus delaying the withdrawal deadline. Its hardly a burden to take the inner out of the outer on election day.
  • We might have considered allowing, closely supervised, the unsealing of the inner envelope starting on the weekend.
  • We would have ordered much tighter security on ballot packages from receipt in the mail room for ballot box. We have been proposing that to the General Assembly for years.
  • THIS ALL SHOULD HAVE BEEN DONE WEEKS AGO.
  • We need monitors in New Haven and Hartford too. Both tend toward Election Day Registration disasters. New Haven has problem counting ballots on time, and we are not confident their officials are up to meeting the demands of the rush of mail-in ballots.

It Could Have Been A Lot Lot Worse

  • There were calls for scanning ballots before election day – other states do that, but it requires detailed procedures and stepped up security to accomplish safely.
  • There were calls for signature curing. That is calling, emailing, or mailing voters if their signature or packed would be rejected. That is all but impossible and likely a civil rights violation, unless almost all of the laws and deadlines for certification were changed. Other states that do that have taken years to claim to have perfected it. Its why one one of those states, CA too 51 days to count the primary – if they do that in Nov, their entire electoral votes may be disqualified.
  • There were calls  for  curing of ballots rejected by scanners. NO STATE DOES THAT. IT WOULD ALSO VIOLATE THE SECRET BALLOT requirement of the Connecticut Constitution.

 

“Delay” is a dirty word

WhoWhatWhy podcasd interview with Professor Ned Foley  <listen>

Ned is the leading legal expert on our presidential election system and how our country reacts to close elections.

In the second half of the interview he makes the point that counting votes after election day and waiting for certified results is a part of the process. He makes the point that the media and everyone else should avoid using the word ‘delay’ to describe results that are not complete on election night – they never are.

WhoWhatWhy podcasd interview with Professor Ned Foley  <listen>

Ned is the leading legal expert on our presidential election system and how our country reacts to close elections.

In the second half of the interview he makes the point that counting votes after election day and waiting for certified results is a part of the process. He makes the point that the media and everyone else should avoid using the word ‘delay’ to describe results that are not complete on election night – they never are.

237 Late Absentee Ballots in Enfield. Don’t Panic – Investigate

Hartford Courant Report:  U.S. Postal Service investigating why 237 absentee ballots showed up weeks late for Aug. 11 primary in Enfield

The U.S. Postal Service and the State Elections Enforcement Commission are investigating why 237 absentee ballots for the Aug. 11 primary showed up more than two weeks late at Enfield Town Hall…

The ballots showed up in batches more than two weeks after the Aug. 11 primary and were postmarked at the Enfield post office before being delivered on the same day. An initial batch of 65 ballots were suddenly delivered to Enfield’s town hall two weeks after the election, and then 49 arrived two days later, officials said.

“They were postmarked by a stamp by a person at the post office — not by a machine that they run through 1/4 u201a” Rosenberg said. “This is obviously a stamp.”

There were no legislative primaries in Enfield this year but the late-arriving ballots may have been a factor in the GOP primary for the 2nd Congressional District that includes Enfield. In that contest, Justin Anderson defeated Thomas Gilmer by 78 votes, according to results on the secretary of the state’s website.

This obviously may have potential implications for the November election,  may be a crime by postal officials or others, maybe not. Don’t panic yet. Some high level possibilities:

Hartford Courant Report:  U.S. Postal Service investigating why 237 absentee ballots showed up weeks late for Aug. 11 primary in Enfield <read>

The U.S. Postal Service and the State Elections Enforcement Commission are investigating why 237 absentee ballots for the Aug. 11 primary showed up more than two weeks late at Enfield Town Hall…

The ballots showed up in batches more than two weeks after the Aug. 11 primary and were postmarked at the Enfield post office before being delivered on the same day. An initial batch of 65 ballots were suddenly delivered to Enfield’s town hall two weeks after the election, and then 49 arrived two days later, officials said.

“They were postmarked by a stamp by a person at the post office — not by a machine that they run through 1/4 u201a” Rosenberg said. “This is obviously a stamp.”

There were no legislative primaries in Enfield this year but the late-arriving ballots may have been a factor in the GOP primary for the 2nd Congressional District that includes Enfield. In that contest, Justin Anderson defeated Thomas Gilmer by 78 votes, according to results on the secretary of the state’s website.

This obviously may have potential implications for the November election,  may be a crime by postal officials or others, maybe not. Don’t panic yet. Some high level possibilities:

  • It could be an innocent error or a series of errors that need to be prevented.
  • It could be a deliberate attempt at election fraud by someone or some group.
  • It could be a deliberate attempt to discredit mail-in voting.
  • AND it could have changed the result of a primary for U.S. Congress.

There is much that might be determined by the two investigations. Some questions that can likely be answered:

  • Who hand stamped the ballots? Why were they not run through normal processes? Where were they found? How did they come to be processed/stored together? Where might that lead?
  • Open the envelopes, leaving all materials together. Count the ballots. If legally submitted in time, would they have changed the result?
  • Are they from a single voting district? Likely mailed from a single mail box, or collected by a single mail carrier?
  • How many are D? How many R? Do the party affiliations closely approximate the D and R submissions for Enfield or the district(s) involved? (If a normal distribution then is less likely to have changed the result.)
  • Submit both the applications and inner envelopes to professional handwriting analysis. Do signatures match? Do they match those on file? Do they show that most of the ballots or applications were signed by one or a few individuals?
  • When were the applications submitted? E. G. were they submitted on the same day?
  • Contact all the voters involved. Did they submit their applications and ballots? When? Where? Did they have any help? In general, does their recall of who they voted for correspond to their ballots?

All that should give a pretty good idea if there was skulduggery or error. If skulduggery may well find the culprit and motivation.

For All Voters:

  • If at all possible, deposit your applications and ballots in the drop-boxes rather than the mail.
  • If you don’t receive your ballot by late October (Oct 20th?) call your Town Clerk or vote in-person.
  • Voting in-person is safer for your vote, with your ballot 2% to about 5% more likely to be counted (without fraud).
  • Yet in the age of COVID, voting by mail is safer for your health, choose carefully.