Lack of clarity about rules/enforcement of campaign finance laws? Check.
Mixed messages about when to expect your mail ballot? Check.
Problems with secrecy sleeves? Check.
Super-late election results? Check.
But listen, the Secretary of State says that none of this stuff was illegal. You'll remember that the state sent observers to look after our April election after the first few mix-ups. But the state, in its supplemental report, says City Clerk Kathryn Young was right on target with this election, stating that she "met and/or exceeded legal requirements."
But don't just take it from me. Read it yourself:
Supplemental Report on April 5, 2011 Colorado Springs Municipal Election
Colorado Springs City Clerk Kathryn Young, mayoral candidates Brian Bahr, Tom Gallagher and Buddy Gilmore, and city council candidates Lisa Czelatdko and Angela Dougan invited Colorado Secretary of State Scott Gessler to observe the upcoming April 5, 2011 municipal election.
Secretary Gessler instructed his staff to thoroughly and independently review election procedures; this review required open access to the Clerk’s office and other locations where election activity occurs. Wayne Munster, Deputy Director of Elections, Christi Heppard, Special Projects Coordinator, and Jerome Lovato, Voting Systems Specialist, have represented the Secretary of State’s office. During the initial visits Clerk Young, her staff, and election workers were welcoming, knowledgeable and well prepared for the election. Clerk Young provided the Secretary’s observers full access to her location, staff and procedures.
This Supplemental Report addresses observations made by Munster and Heppard on March 28, 2011 and April 5, 2011. For reference purposes, the Interim Report can be found on the Secretary of State’s website.
Review Timeline for Supplemental Report:
March 28, 2011
Team members: Wayne Munster and Christi Heppard
Processes reviewed: Logic and Accuracy Test
Voting Assistance at Health Care Facilities
April 5, 2011
Team members: Wayne Munster and Christi Heppard
Processes reviewed: Ballot Processing
Curbside Drop-Off Locations
The Secretary’s staff reviewed processes relating to undeliverable ballots returned to the Clerk. The final number of undeliverable ballots was 5,378 out of approximately 150,000 ballots mailed, which is in the normal percentage for undeliverable mail. Undeliverable ballots are not opened at any time by City Clerk staff or election judges.
Once the undeliverable ballots are received from the post office, they are scanned so that a digital image is captured, then alphabetized. Undeliverable ballots are kept in a secure location, accessible only by sworn staff and election judges.
As mentioned in the Interim report, Munster and Heppard confirmed the security camera was properly functioning. The camera digitally records the ballot processing room and captures election judges processing and counting ballots. The camera and room lights are motion-activated so that if an unauthorized person were to access the ballot processing room after hours, the lights and camera would be activated and begin recording.
A television monitor showing the security camera feed is located at the Security desk, which is staffed during regular business hours. On Election Day Munster and Heppard noted that security staff were at the desk until well after 7:00 P.M. In addition to staff manning the security monitor, the feed is digitally recorded so that a permanent record is kept.
Logic and Accuracy Test
On March 28, 2011 Clerk Young held a public logic and accuracy test to ensure that the voting equipment was properly counting ballots. In order to complete the test, Clerk Young issued blank test ballots to each Logic and Accuracy Board member before the meeting. The Logic and Accuracy Board members are:
Wynetta Massey Deputy City Attorney
Jesse James Interim Chief Information Officer
Denny Nester Interim City Auditor
Before beginning the test, board members marked their test ballots and manually tallied their results.
During the public logic and accuracy test, several members of the media were present. Using the counting equipment, election judges counted ballots for each board member; a results report was printed for each member’s ballots. Board members then compared their hand-tallied results to the voting machine’s results. Both sets of results accurately matched, ensuring that the counting equipment was working correctly.
Munster and Heppard observed that Logic and Accuracy Board members did verify their hand-counted results because they did not match the machine count during the initial comparison. In all cases after verifying the hand-tallied counts it was confirmed that the voting machine had tabulated the votes correctly and the board member’s initial count was incorrect.
After the logic and accuracy test was completed, seals were placed over ports and memory card slots to ensure no unauthorized access prior to vote tabulation. After conclusion of the test, ballots used were sealed and stored for retention.
A vendor representative on site during the test informed Munster and Heppard that the optical scan equipment, the ES&S 850, was capable of capturing and storing digital images of ballots. Clerk Young was informed that the 850 had this capability and of the drawbacks of utilizing it. For instance, the vendor representative informed Clerk Young that by storing ballot images of both sides of each voted ballot the data transfer of 180,000 images to the jump drive for ERM software download would take as much as six hours to complete instead of the 20-30 minutes without the images. This would have meant that the process to transfer the ballot images would delay the posting of results by several hours. Further, the vendor representative on site noted that she was unaware of any Colorado jurisdiction, or any other U.S. jurisdiction for that matter, attempting to capture and store ballot images for a municipal election. Finally, Heppard investigated whether the municipal clerk was required by law to capture and maintain ballot images. She determined that there is no municipal requirement in the city’s home rule charter to keep such images. In light of the aforementioned, the Secretary’s observers were informed that Clerk Young decided not to capture ballot images.
When ballots are returned, either from the post office or drop-off locations, staff batches the unopened envelopes into trays of 450 ballots. Ballots are organized so that all names and addresses face the same direction, in preparation for signature capturing. Each tray is then assigned a unique number. Next, ballots were taken directly to the signature capture and verification area.
Once ballots are received they are processed by an AccuVote Envelope Scanner that first reads a bar code with the voter’s information then stamps the voter’s identification number and batch number on the outside of the envelope. During this process, the system checks to make sure that no more than one ballot was returned by each individual voter. If the system detects that more than one ballot was received, the information will be noted on a report. Munster and Heppard did not observe any instances where more than one ballot was returned by a voter.
During this process the machine also scans the voter’s signature from each envelope. The signature is stored in the system, by batch, for comparison against other signatures on file from the voter.
After signatures are captured, ballots are returned, still unopened, to their tray to begin the signature verification process. A report listing the number of ballots processed is automatically printed and is kept with the top sheet of the ballot processing audit log. The Secretary’s observers kept a sample of the report.
Signature verification is a two step process. First, signatures are electronically compared. If an electronic match cannot be made, election judges manually compare signatures. Clerk Young uses signatures from the SCORE statewide voter registration system to compare against signatures on ballot envelopes.
For electronic verification, election judges use a software system to electronically compare signatures captured from the ballot envelope to the signature on file in the voter’s record. The system uses parameters to set a threshold for a verified “match.” If an electronic match cannot be made, the signatures are manually reviewed by election judges. Out of each tray containing 450 ballots, on average 200 signatures must be manually reviewed.
Muster and Heppard observed election judges manually reviewing signatures. In most cases, the manual review resulted in a signature match. However, out of each tray of 450 ballots, approximately 15 signatures could not be matched by electronic or manual review.
Munster and Heppard observed several reasons for match failure. First, there may not be an electronic copy of the voter’s signature in the database to match the ballot signature against. In these instances, ballots are removed from the tray and placed in a separate container. Clerk Young’s staff then makes arrangements with the El Paso County Clerk and Recorder to review signatures directly from voter records.
Another common reason for signature match failure is known as “envelope swap.” This occurs when people in the same household unintentionally signs his or her family member’s return envelope. In these cases, once election judges confirm that no voter signed more than one envelope, the ballots are sent for counting. This process is consistent with state law and rule.
If during the signature verification process election judges find the voter did not sign the envelope, the ballot envelope is removed from the tray and a letter is sent to the voter instructing him or her to come to the Clerk’s office and sign his or her ballot. This process is consistent with state law and rule.
Once signatures have been scanned and verified, the ballot tray is sealed by election judges and the seal number is noted on the tray’s audit log. Ballot trays are then either stored sealed in the vault or taken directly to the ballot processing room. At this point, all ballot envelopes remain unopened.
In addition to the processes listed above, Clerk Young uses Logic and Accuracy Board members to resolve instances where a voter’s signature cannot be verified electronically, after manual review, or by reviewing records at the county clerk’s office.
Heppard observed final signature resolution on Election Day. Approximately 50 ballots were reviewed by the Board. Board members were thoughtful with decision making when considering whether or not a signature could be confirmed as a match. Thirty three signatures that could not be verified will be turned over to the City Attorney and ultimately the District Attorney for investigation.
Preparation for Counting
Clerk Young uses approximately 16 election judges in the processing room to open ballot envelopes and prepare them for counting. Election judges are divided into two-person teams. Approximately six teams of judges open the envelopes and prepare ballots for counting. In addition, two teams of duplication judges are available to replicate ballots, when necessary. Finally, two judges serve as “supervisors” that oversee the process. More specific information regarding election judge duties follows.
Processing judges work in a team of two to prepare each ballot tray for counting. There are multiple steps that must be completed and preparation for counting takes approximately 1 hour and 30 minutes per tray of approximately 450 ballots.
Step 1- The processing judges compare the seal number on the ballot tray to the number listed on the audit log to verify the tray has not been opened.
Step 2- Processing judges count the number of ballots in the tray to ensure that the tray contains the correct number of ballots.
Step 3- Processing judges use an envelope opening machine that very quickly opens the ballot envelopes. Up to this point, no ballot envelopes have been opened during any point in the process.
Step 4- Once all ballot envelopes have been opened, processing judges begin removing the ballots from the envelopes. Judges remove the voted ballot in its secrecy sleeve from the envelope and place the envelopes and ballots in separate stacks.
Step 5- The ballot is removed from the secrecy sleeve and unfolded. It should be noted that some voters do not choose to return their ballot in the secrecy sleeve.
Step 6- The ballot stub is removed from the voted ballot.
Step 7- Ballots are placed in a stack and each is inspected for obvious tears, stains, or other problems that require duplication.
Step 8- If any ballots require duplication, processing judges place them in a large envelope and deliver them to the duplication judges. Information about duplication can be found later in this report. The number of ballots sent to duplication is noted on the audit log.
Step 9- The number of ballots in the tray is counted and noted on the audit log.
Step 10- Ballots are placed back in the tray and securely sealed to await counting. The seal number is noted on the audit log and both election judges initial the document.
Step 11- Processing judges then count the number of secrecy sleeves for each batch. This number is also noted on the audit log.
Step 12- Processing judges count the number of ballot stubs removed from ballots. If any ballots were returned without a ballot stub, judges use a placeholder stub. This number is noted on the audit log.
Step 13- Empty ballot envelopes, ballot stubs, and secrecy sleeves are stored in a separate container and kept as official election records.
Step 14- Sealed ballot trays containing ballots prepared for counting are either sent directly to counting judges or are stored in the vault to be counted the next day.
Two election judges, with the assistance of a vendor service representative, used high-speed optical scan tabulation equipment to tabulate votes. Counting judges are election workers dedicated to tabulating ballots and do not perform any ballot preparation functions.
Step 1- Counting judges verified that the tray’s seal matches the number placed on the tray by the processing judges.
Step 2- The tray is opened and ballots are placed in a machine called a “ballot jogger” that lines up ballot corners for efficient processing.
Step 3- All ballots in the tray are counted by the high-speed optical scanner.
Step 4- The optical scanner sorts out any ballots that were overvoted, completely blank, or otherwise unreadable.
Step 5- Ballots with overvotes where the voter’s intent can be determined and ballots that have tears or stray marks are sent to an election judge team for duplication.
Step 6- Duplicated ballots, if any, are processed by the machine and included with the batch total.
Step 7- After each tray is counted, counting judges compare the number of ballots counted against the number of ballots in the tray, as noted by the processing judges. The number of ballots counted is noted on the audit log.
Step 8- After ballots have been counted, they are returned to their tray and securely sealed. The seal number is noted on the audit log and the log is initialed by the counting judges.
Two supervisor judges, who are experienced and well-trained on election procedures, were on site to answer question from processing judges and to oversee ballot preparation and counting.
Two teams of duplication judges were available in the ballot processing room. Ballots that were torn, damaged, overvoted where the voter’s intent could be determined, or otherwise unreadable were sent for duplication. Ballots may be sent for duplication during the ballot opening process when judges see an obvious tear or mark, or during the counting process if the counting machine indicates a ballot is unreadable.
Each duplicated ballot is entered onto a two-part carbonless log. One copy of the log is retained with the original ballot and the other is returned to the processing judges so that a paper trail is established. Munster and Heppard observed several ballots during the duplication process and spoke with the duplication judges. The procedures followed were consistent with law and ensure that ballots are accurately duplicated.
Clerk Young provides a designated area for watchers to observe election processes. Watchers are credentialed before they can enter the area. Munster and Heppard observed one watcher on site on Election Day.
Health Care Facility Ballot Delivery and Resident Assistance
Clerk’s staff worked with all 95 facility directors within the city before the election to arrange for ballot delivery and assistance with voting for those voters who request assistance. Only five facilities utilized the services offered by Clerk Young.
On March 28, 2011 Heppard accompanied two members of Clerk Young’s staff to a local health care facility, Mackenzie Place. Ballots were transported in a sealed ballot transfer case. Election judges met with several residents regarding voting. Only one resident requested assistance with voting.
Before providing assistance, the election judges completed appropriate forms and documentation noting that assistance was provided. Election judges provided assistance with the voting process but did not, in any way, attempt to influence the voter’s choices. After the voter sealed the ballot and signed the ballot envelope, the election judges placed the voted ballot in the sealed and secure transfer case for transport back to the Clerk’s office.
The process used conformed with best practices relating to voting assistance at health care facilities.
Curbside Drop-off Locations
In addition to the drop-off locations deployed around the city, Clerk Young used two curbside ballot drop off locations on election day. Both locations are at the City Clerk’s office, one on each of the main cross streets.
Curbside drop-off locations were staffed by at least three people, typically two student judges and one election judge. A sealed ballot box was available and signage indicated where voters should stop their cars to deliver ballots. Elections judges periodically picked up full ballot boxes and replaced them with an empty, sealed box, as needed. Munster and Heppard observed both curbside drop-off locations several times on election day.
At 7:00 P.M. when polls officially closed, Heppard observed the ballot boxes being transported into the Clerk’s office for processing. While there are no laws or rules that govern curbside drop-off locations, the procedures implemented appeared to be secure in addition to providing a convenient method for voters to deliver their ballots on election day.
As mentioned above, the City uses a high-speed optical scan ballot counting machine, called the ES&S 850, to tabulate results. The 850 has a dot matrix printer with a continuous paper feed that logs every activity, including when the power is turned on or off, when ballots are counted, and when information is backed up.
On election night, Clerk Young released preliminary, unofficial results at approximately 7:40 P.M. and updated those results throughout the evening. Final unofficial results were available on Wednesday, April 6, 2011.
Based on observations of the election processes in Colorado Springs made by Munster and Heppard on March 28, 2011 and April 5, 2011, the Secretary’s observers believe Clerk Young has met and/or exceeded legal requirements. Observers are satisfied that Clerk staff is properly following the law and documented procedures. Accordingly, the Secretary of State is confident in the election process and final results for the April 5, 2011 City of Colorado Springs Election.
The Secretary’s observers found no evidence of any improper activity or misconduct, either intentional or unintentional. The Secretary believes that the April 5 Municipal Election was conducted with integrity and in accordance with all applicable laws and rules.