Citizen Audit Report Lauds Improvements in Post-Election Audit Drawing

<Full Report (.pdf)> <Press Release>  <Backup Data>

Secretary of the State and her Office significantly improve integrity and transparency of the random drawing

Report findings included:

  • Significantly fewer errors in the random drawing list in November 2014 compared to November 2013, finding:
    • One (1) missing district in 2014 vs. sixteen (16) missing in 2013.
    • Seven (7) non-existent districts on the list in 2014 vs. two (2) in 2013.
    • Three (3) municipalities incorrect in 2014 vs. fourteen (14) incorrect in 2013.
  • Improvements largely due to the diligence of the Secretary of the State’s Office, in leading compliance in 2014, with a law intended to fix past problems.
    • A compliance failure rate of 2.4% in 2014 vs. 67% in 2013.
  • We also appreciate that the Secretary and Secretary’s Office responded to our request to reverse past precedent and hold a public and transparent drawing of races to be audited.

Citizen Audit spokesperson Luther Weeks noted, “We applaud improvements made by the Secretary of the State and her Office. The drawing was significantly more accurate than last year and, for the first time, the Secretary chose races for audit publicly and transparently

<Full Report (.pdf)> <Press Release>  <Backup Data>

Random Drawing: 68 Districts in 40 Municipalities Selected for Post-Election Audit

Two members of the Citizen Audit Board assisted the Secretary of the State by randomly drawing 68 districts in 40 municipalities across Connecticut for the post-election audit.

See the Press Release from the SOTS Office: <press release>

Observation Report – Flaws Remain after Seven Years

Citizen Audit Finds Little Improvement In Election Audits

Post-Election Audit Flaws Remain After Seven Years and Twelve Audits

The report concluded that the official audit results do not inspire confidence because of the continued:

  • Lack of consistency, reliability, and transparency in the conduct of the audit.
  • Discrepancies between machine counts and hand counts reported to the Secretary of the State by municipalities.
  • Lack of investigation of such discrepancies, and the lack of standards for triggering such investigations.
  • Weaknesses in the ballot chain-of-custody.
  • An unsatisfactory improvement in the random audit drawing integrity vs. the November 2012 audit, as reported in our recent Districts In Drawing Study.

Coalition spokesperson Luther Weeks noted, “When compared with audits in 2011 and 2012 we found little difference, positive or negative, on the issues previously identified and the level of concerns affecting confidence.

“Without adherence to procedures and effective follow-up, if there was ever a significant fraud or error, it might not be recognized and corrected.”

Some officials follow the audit procedures and do effective work, yet a trusted audit requires that all the districts and all the votes be counted in the audit as intended.”

<Full Report (.pdf)> <Press Release>  <Review detail data and municipal reports>

Districts in the Random Drawing Study

<Full Report (.pdf)> <Press Release>  <Backup Data>

Citizen Study Finds State Audit Flawed From the Start

Post-Election Audit Flawed from the Start by Inaccurate List of Election Districts

Based on concerns with the integrity of the random drawing in previous elections, the Coalition initiated a project to thoroughly check the integrity of the list of districts in the drawing for the November 2013 election.

Coalition spokesperson Luther Weeks noted, “The credibility of our elections depends on the integrity of the Post-Election Audits. The integrity of the audits in-turn depends on the integrity of the drawing.”

The report found,

  • Fewer discrepancies in the random drawing list than were found in November 2012 and efforts by the Secretary of the State’s Office to improve the accuracy of the list.
  • Seventeen (17) voting districts missing from the drawing list or recanvass list. The selection of two (2) fewer districts for audit than required by law.
  • The audit law calling for the random selection and auditing of 10% of voting districts is not being faithfully executed, leaving the integrity and credibility of the audit and our election system open to question and expose it to the potential of future manipulation.
  • The law intended to fix past problems is not being followed. — A failure rate of 67
  • Fully complying with the law is complicated by the loose definition of voting district in the law.

Tom Swan, Executive Direction, Connecticut Citizen Action Group, added, “These problems can be fixed. The report includes recommendations to the Legislature and Officials for effective and efficient ways to provide accurate lists, and remedies when errors occur.”

Cheri Quickmire, Executive Director, Connecticut Common Cause, said “We look forward to working with the Secretary of the State’s Office to assist in choosing effective remedies to result in an accurate random drawing.”

<Full Report (.pdf)> <Press Release>  <Backup Data>

Nov 2012 Post-Election Audit Report – Flawed From The Start

<Full Report (.pdf)> <Press Release>  <Review detail data and municipal reports>

Coalition Finds Continuing Problems with Election Audit and A New Flaw

Post-Election Audit Flawed from the Start by Highly Inaccurate List
of Election Districts

 The report concluded, the official audit results do not inspire confidence because of the:

  • Lack of integrity in the random district selection.
  • Lack of consistency, reliability, and transparency in the conduct of the audit.
  • Discrepancies between machine counts and hand counts reported to the Secretary of the State by municipalities and the lack of standards for determining need for further investigation of discrepancies.
  • Weaknesses in the ballot chain-of-custody.

Coalition spokesperson Luther Weeks noted, “We found significant, unexplained errors, for municipalities across the state, in the list of districts in the random drawing. This random audit was highly flawed from the start because the drawing was highly flawed.”

Cheryl Dunson, President, League of Women Voters of Connecticut, stated,, “Two years ago, the Legislature passed a law, at the Secretary of the State’s request, which was intended to fix inaccuracies in the drawing. For whatever reason, errors in the drawing have dramatically increased.

Weeks added, “Some officials follow the audit procedures and do effective work. This year one town investigated discrepancies and found errors to correct in their election procedures – that is one value of performing the audits as intended.”

Without adherence to procedures, accurate random drawings, a reliable chain-of-custody, and transparent public follow-up, when discrepancies are reported, if there was ever a significant fraud or error it would not be recognized and corrected.

<Full Report (.pdf)> <Press Release>  <Review detail data and municipal reports>

Additional Details:

<Observer Report Form>

Districts in Random Drawing Study:

<Comparison Aug vs. Nov>
<Districts In Drawing Aug 2012> <Districts In Drawing Nov 2012>

Aug 2012 Primary Audit Observation Report

<Full Report (.pdf)> <Press Release>  <Review detail data and municipal reports>

Coalition finds 31% of Official Audit Reports Lack Critical Data

Municipalities failed to report data critical to audit evaluation. Increasing numbers choose paper only elections, avoiding scanners and audits.

The report highlighted concerns with two increasing trends:

  • An increase in missing and incomplete official reports. There are 16 of 52 (31%) reports with errors making it impossible to determine if machines had functioned properly. What basis is there to trust audits, with this significant level of error in reporting?
  • Up up to 19 towns avoided optical scanners and audits by conducting paper only elections. Such voting is not audited, not transparent, and error prone based on past observations of hand counts.

We conclude, based on our observations and analysis of official audit reports submitted to the Secretary of the State, that the August post-election audits still do not inspire confidence because of the continued:

  • Lack of integrity in the random district selection and race selection processes.
  • Lack of consistency, reliability, and accuracy in the conduct of the audit.
  • Weaknesses in the ballot chain-of-custody.
  • Missing or incomplete reports, lacking critical information.

Cheryl Dunson, League of Women Voters of Connecticut’s President, stated, “When officials submit reports missing basic data such as the number of votes counted by hand or by optical scanner, it defeats the purpose of the audits – on what basis can we determine the accuracy of optical scanners without such information?”

Luther Weeks, Coalition Executive Director, stated, “In the past, many officials argued that their staffs cannot accurately count ballots by hand in the ideal conditions afforded by audits, yet now an increasing number choose to forgo optical scanning and report results based on unaudited hand counts performed in the more trying conditions of election night”.

The report found little change from past observations and conclusions and little progress by officials in improving post-election audit integrity. Because of ongoing shortcomings in the performance of post-election audits, the Coalition continues to urge our state election officials to offer more guidance to local personnel and to require more consistency in the conduct of the audit.

Weeks noted, “Once again as a public service we are providing online images of the official audit reports from local registrars, along with our digitized data so that anyone can audit our work.”

<Full Report (.pdf)> <Press Release>  <Review detail data and municipal reports>

Apr 2012 Primary Audit Observation Report

<Full Report (.pdf)> <Review detail data and municipal reports>

Based on the size and relative simplicity of this audit, we provide an abbreviated report. We summarize the statistical information and observations, while we make no new recommendations.
As we said in our Nov 2011 report, readers of past reports will note little change in our observations and conclusions; little progress by officials in improving post-election audit integrity; and little change in our recommendations which have been, for the most part, not addressed by current and past Legislatures and Secretary of the State’s Offices.
Compared to past audits there were significantly fewer ballot count and race count differences in this audit. The fewer number of ballots and the single race to be counted contributed to this improved result. We are skeptical that this audit is an indication of future improvement.

Items of note include:

  • Three towns selected for audit unsealed and combined their ballots from all districts, necessitating a town-wide count for the audit. This highlighted that ballots are not required by law to be sealed under a secure chain-of-custody during the audit period, only for the fourteen (14) day “lock-down” period originally applicable to lever voting machines.
  • Despite the simplicity of the counting, several towns were unable to completely or correctly complete the official audit report form.
  • One town’s report showed significant differences between the manual and machine counts for votes. The audit report form from this town was among those that were incomplete.

The audit usually begins on the 15th day after each election and primary, however, the Secretary of the State’s office held the random drawing on May 15th, 21 days after the primary. After the election, there was a question raised by the Secretary of the State’s Office, that the law might not require an audit of a presidential primary. Such primaries are covered by different sections of the Connecticut statutes, than sections that cover other elections and primaries. It was determined to conduct a presidential primary post-election audit, keeping with the precedent set in 2008 by Secretary Susan Bysiewicz. Those same variations in statute provide the for the legal use of paper ballots without optical scanners. This delay necessitated three towns selected for audit to count all their ballots in every district as the ballots had been unsealed and combined for storage.

Anticipating a very small primary, a number of municipalities used paper ballots without scanners as a means of saving the expense of programming and testing optical scanners. Only polling place optically scanned ballots are audited.

 

Nov 2011 Election Audit Observation Report – Data Available for Public Review

<Full Report (.pdf)>  <Press Release>
 <Executive Summary>  <Audit Integrity, Organization, and Chain of Custody>
 <Audit vs. Scanner Counts>  <Recommendations>
 <Review detail data and municipal reports>

Coalition finds continuing problems with audit integrity
Provides calculations and official data on the web for public review and verification

For the first time, in the interest of public information and transparency, we are making all official municipal audit reports and the data we complied available for everyone to review on the web. Citizens can see the reports from their own town, other towns, and perform their own audit of the Coalition’s data entry and calculations based on those official reports. <See detail data and municipal reports>

The November post-election audits still do not inspire confidence because of the continued:

  • Lack of integrity in the random district selection and race selection processes.
  • Lack of consistency, reliability, and transparency in the conduct of the audit.
  • Discrepancies between machine counts and hand-counts reported to the Secretary of the State by municipalities and the lack of standards for determining need for further investigation of discrepancies.
  • Weaknesses in the ballot chain-of-custody.

Most of our general observations and concerns remain the same as reported in previous reports. However, in this round of audit observations, we note that:

  • Reported discrepancies have decreased, approximating the Nov 2009 post-election audit discrepancies.
  • A significant increase in memory card problems have been reported.

Once again, in the November 2011 random selection, without investigation, we discovered, according to local registrars, two districts selected for audit were not included in the election. This despite the passing of a bill in 2011 by the General Assembly, at the request of the Secretary of the State, that was supposed to help remedy this situation.

 In two towns, for a total of ten selected districts, officials did not select and audit the full three races required by the law. While likely unintentional, if such situations are not investigated and remedied when they occurr, they  provide an opportunity for errors or deception.

 Among our greatest concerns are the discrepancies between machine counts and hand-counts reported to the Secretary of the State by municipalities. We can find no acceptable explanation for attributing these discrepancies either to humans or to the voting machines. In many cases, these discrepancies are not thoroughly and reasonably explained. In other cases the explanations make no sense or contradict the data in municipalities’ reports.

We believe that the lack of organization, planning, and ad-hoc counting procedures used by many municipalities were not sufficient to count accurately and efficiently. Therefore, we believe that the procedural guidelines provided by the Secretary of the State still need further clarification and improvement.

The Secretary of the State should expect officials to be able to organize audits to produce accurate hand counts, but not, in normal circumstances, to have the skills to analyze and explain differences between accurate hand counts and machine counts. Such differences should be investigated by technicians, with appropriate expertise, such as the University of Connecticut VoTeR Center

Audit and chain–of-custody procedures are not followed. This report does not question any individual’s integrity. However, a safe, credible system of security procedures should not permit a single individual any extended opportunity to handle records and ballots unobserved.

Nov 2010 Election Audit Observation Report

<Full Report, Press Release, Detail Data>
<Executive Summary>
<Audit Integrity, Organization, and Chain of Custody>

<Audit vs. Scanner Counts>
<Recommendations>

Coalition calls again for legislature to act.
Citizen observation and analysis show little, if any, improvement in
November post-election audits.

The Coalition noted significant differences between results reported by optical scanners and the hand count of ballots by election officials across Connecticut. Compared to previous audits, the Coalition noted little, if any, improvement in the attention to detail and in following procedures in the November 2010 audits.

The report concluded:

The November post-election audits still do not inspire confidence because of the continued:

  • failure in the integrity of the random district selection process,
  • lack of standards for determining need for further investigation of discrepancies,
  • weaknesses in the ballot chain of custody, and
  • lack of consistency, reliability, and transparency in the conduct of the audit.

The level of discrepancies reported has increased from the November 2009 post-election audits.

Among our greatest concerns are the discrepancies between machine counts and hand-counts reported to the Secretary of the State by municipalities.  In many cases, these discrepancies are not thoroughly and reasonably explained.  In some cases the explanations make no sense or contradict the data in municipalities’ reports. We believe that the lack of organization, planning, and ad-hoc counting procedures used by many municipalities were not sufficient to count accurately and efficiently.  We find no reason to attribute all errors to either humans or machines.

In our August 2010 Report, we noted a newly recognized problem: “the inaccurate list of districts used in the random selection process which is required by law to be based on all of the districts used in the election.  This directly impacts the integrity and credibility of the entire post-election audit.” Once again, in the November 2010 random selection, without extensive investigation, an advocate quickly discovered a district missing from the list of districts in the random drawing.

In this report we note another inadequacy in the audit law. One district initially selected was not audited because ballots were impounded based on a complaint to the State Election Enforcement Commission. … This represents a new formula for an individual to preclude auditing of a specific district to avoid the discovery of an error or fraud. Our existing recommendations to improve the law, had they been enacted by the Legislature, would have eliminated this opportunity.

This year the audits were particularly challenging for the Coalition, our observers, and election officials.  In this election the law mandated that the audits be completed two days prior to certification, requiring that all municipal counting must all be completed between November 17th and 22nd, where in previous large scale audits there was often a period of two to three weeks for the counting. In addition, the districts for audit were chosen just two days prior to the start of the counting period, on November 15, 2010, while in past elections the selection has occurred six or seven days prior to the audit period.

The short time for completion and the short time interval after the drawing, caused challenges for the Coalition, our observers, and election officials. Several audits occurred on November 17th and 18th which violated the Secretary of the State’s procedures which require three business days notice to the public and the Secretary of the State’s office.

<Full Report, Press Release, Detail Data>
<Executive Summary>
<Audit Integrity, Organization, and Chain of Custody>

<Audit vs. Scanner Counts>
<Recommendations>

Bridgeport Recount Report and Recommendations

Votes were miscounted and miscalculated adding votes to each candidate, but not changing winner in the race for governor

<Full Report, Press Release, Detail Data>

From the Press Release and Executive Summary:

Each candidate for the governor’s race gained votes in the recount when compared to the officially reported results, as follows: Foley (+174), Malloy (+761), and Marsh (+19).  These differences parallel candidate shares in the initially reported results. Counting of all ballots in the governor’s race resulted in differences in many counts, totaling 1,520 votes miscounted, of these 1,236 were initially under reported and 284 were initially over reported.

Luther Weeks of CTVotersCount states, “We found evidence suggesting just about every type of counting and calculation error one would imagine from officials who worked extended days and counted under pressure. We also found differences between numbers of ballots cast and numbers of voters checked-in at the polls.”

In one district, the ballot case contained Provisional Ballots cast at that district but sealed in ballot containers and not brought to the attention of registrars for evaluation or, if applicable, for counting in the election.  This represents a communication and training problem between the moderators, head moderator, and registrars. Provisional Ballots should not have been sealed with the other ballots and from the Moderator’s  Returns should have been recognized as having been voted from the Provisional Ballot Inventory Form.

Connecticut Citizen Action Group Executive Director, Tom Swan said, “In a closer election, similar errors could have changed an election result, precluded a necessary recanvass or caused an unnecessary one.”

In several districts the counts of ballots differed from the check-in list counts, and in several districts, our counts of the check-in lists differed from the counts on the moderators’ reports.  In many districts check-in list counts closely paralleled ballot counts, yet some counts were over by 21 ballots or less or under by 30 ballots or less. While some variation can be expected due to the difficulty of interpreting some check-in marks, we found several districts where the check-in lists were clearly not counted accurately. In one district, results were under by 117 ballots less than voters checked- in based on our counts.

Cheri Quickmire, Executive Director, Common Cause of Connecticut said “While the recount showed no evidence of fraud, there were numerous errors.  It is time for elections officials to be accountable to the citizens of the state. It is time to professionalize elections in Connecticut.”

From the Press Release and Recommendations:

Cheryl Dunson of the League of Women Voters of CT stated, “Voters deserve better.  Municipalities must provide sufficient numbers of ballots. Election officials need more training, organization, and more time to determine and report results. There is also a need for more enforcement of our election laws.”

We suggest possible components of a comprehensive solution that deserve consideration:

    • Mandatory formulas for minimum ballot orders

Adoption of emergency plans: It’s important to note that, simply printing more ballots only reduces the chance of the specific problem that occurred in Bridgeport. There are other causes that could result in a municipality having to scramble to photocopy ballots or perform hand counting such as a massive power failure or ballots lost in a fire, flood, or accident shortly before or during Election Day.

    • Enforceable laws, regulations, and proceduresfor the monitoring of turn-out to provide sufficient warning of ballot shortages
    • Enforceable laws, regulations, and procedures for the creation, security, and accounting for all ballots and for the creation and handling of emergency photocopied ballots
    • Additional state-provided training and improved manuals should be available for election officials
    • In addition to head moderators, the Secretary of the State should have the power to order discrepancy recanvasses
    • Allow more time for the initial reporting of results, investigation of minor discrepancies, and for the accomplishment of recanvasses and the certification of results
    • Enforceable laws, regulations, and procedures for the accounting for and public verification of the reporting of election results from counting teams, polling place officials, and head moderators, to the Secretary of the State’s Office
    • Require detailed reporting of votes for cross-endorsed candidates

<Full Report, Press Release, Detail Data>