Incident: Ranked-Choice Voting Software Failure in New York City Election

Published Date: 2021-07-02

Postmortem Analysis
Timeline 1. The software failure incident happened in June 2021 [116350].
System 1. New York City Board of Elections' ranked-choice voting software 2. Universal Ranked-Choice Voting Tabulator software package 3. Election Management System 4. Absence of administrative manager during the election season 5. Lack of training for employees on the software and ranked-choice voting 6. Delay in testing the software due to an impasse with the State Board of Elections 7. Resistance from Republicans on the State Board of Elections 8. Decision-making process on releasing ranked-choice voting results
Responsible Organization 1. The New York City Board of Elections was responsible for causing the software failure incident [116350].
Impacted Organization 1. New York City Board of Elections [116350] 2. New York State Senate and Assembly [116350]
Software Causes 1. The failure incident was caused by a human error where a staffer did not remove the test ballot images from the Election Management System, leading to the inclusion of over 130,000 test ballots in the preliminary count [116350]. 2. The New York City Board of Elections ignored offers of technological assistance from the supplier of the software used for tabulating ranked-choice votes, which could have helped catch the test ballots that were inadvertently added to the tally [116350].
Non-software Causes 1. Lack of leadership and operational vacancies within the New York City Board of Elections, including the absence of the executive director and administrative manager [116350]. 2. Political structure and nepotism within the New York City Board of Elections, leading to criticisms of ineptitude and corruption [116350]. 3. Delays and challenges in decision-making processes, such as the impasse with the State Board of Elections and resistance from Republicans, causing delays in software testing and training [116350]. 4. Pressure from various stakeholders, including Councilman Brad Lander, to change the plan for releasing results, leading to a compromise that ultimately resulted in errors and outrage [116350].
Impacts 1. The software failure incident led to the release of preliminary ranked-choice vote totals that were later retracted due to the inclusion of more than 130,000 test ballots, causing confusion and mistrust among the public [116350]. 2. The incident created outrage across the city and renewed calls for changes at the New York City Board of Elections, highlighting long-held frustrations about the agency's persistent blunders and scandals [116350]. 3. The failure to release accurate results in a timely manner caused delays and uncertainty in the election process, impacting the integrity and credibility of the election [116350]. 4. The software failure incident exposed the lack of leadership at the board, with key positions vacant or filled by individuals without sufficient experience, contributing to the chaos and mismanagement during the election [116350]. 5. The incident highlighted the board's failure to accept offers of technological assistance from the software supplier, indicating a lack of preparedness and coordination in implementing the ranked-choice voting system [116350].
Preventions 1. Proper training for employees on the software used for ranked-choice voting could have prevented the incident [116350]. 2. Accepting offers of technological assistance from the supplier of the software to ensure accurate results could have helped prevent the failure [116350]. 3. Resolving the impasse with the State Board of Elections earlier to allow for sufficient time to test the software and train employees could have prevented the incident [116350].
Fixes 1. Implementing reforms to the New York City Board of Elections to address long-standing issues of nepotism, ineptitude, and corruption within the agency [116350]. 2. Holding hearings and passing reform legislation by the New York State Senate to address problems at the board promptly and properly [116350]. 3. Providing proper leadership and staffing within the board, including having an administrative manager in place during critical election periods [116350]. 4. Accepting offers of technological assistance from the supplier of the software, such as training for election workers and support in tabulating votes, to prevent similar incidents in the future [116350]. 5. Resolving delays in testing the software and training employees on ranked-choice voting to ensure a smooth election process [116350].
References 1. New York City Board of Elections 2. Ranked Choice Voting Resource Center 3. New York State Senate 4. New York State Assembly 5. Michael Ryan 6. Pamela Perkins 7. Dawn Sandow 8. Georgea Kontzamanis 9. Christopher W. Hughes 10. Valerie Vazquez-Diaz 11. Douglas Kellner 12. Councilman Brad Lander 13. Councilman Fernando Cabrera

Software Taxonomy of Faults

Category Option Rationale
Recurring one_organization, multiple_organization (a) The software failure incident related to the New York City Board of Elections is not the first time such issues have occurred within the organization. The board has a history of blunders and scandals, including mistakenly purging about 200,000 people from voter rolls, forcing voters to wait in long lines, and sending erroneous ballots to thousands of New Yorkers in the past [116350]. (b) The incident at the New York City Board of Elections is not an isolated case, as other jurisdictions have been more responsive to outreach from the supplier of the software used for ranked-choice voting. The Ranked Choice Voting Resource Center offered assistance to various jurisdictions using its software, indicating that similar incidents may have occurred or been avoided in other locations [116350].
Phase (Design/Operation) design, operation (a) The software failure incident in the New York City Board of Elections was primarily due to contributing factors introduced by system development and procedures to operate or maintain the system. The incident was related to the ranked-choice voting software, where the board had mistakenly included more than 130,000 test ballots in the preliminary count, leading to the results being retracted [116350]. Additionally, the board faced delays in testing the software to generate the ranked-choice results due to an impasse with the State Board of Elections, which took more than a year to resolve [116350]. (b) The software failure incident also had contributing factors introduced by the operation of the system. The board ignored offers of technological assistance from the supplier of the software, which could have helped catch the test ballots that were inadvertently added to the tally on Tuesday [116350]. Furthermore, delays in training employees on the software and ranked-choice voting itself were reported, with some employees not receiving training until after early voting had already begun [116350].
Boundary (Internal/External) within_system (a) The software failure incident related to the New York City Board of Elections releasing inaccurate ranked-choice vote totals was primarily within the system. The incident was attributed to human error within the agency, specifically a staffer not removing test ballot images from the Election Management System, leading to the inclusion of over 130,000 test ballots in the preliminary count [116350]. Additionally, the board faced delays in testing the software and training employees due to internal issues and an impasse with the State Board of Elections [116350]. The failure was not directly caused by the software itself but rather by the agency's staff and internal processes [116350].
Nature (Human/Non-human) non-human_actions, human_actions (a) The software failure incident in the New York City Board of Elections was primarily due to non-human actions. The incident occurred because the board mistakenly included more than 130,000 test ballots in the preliminary count, which led to the results being retracted as they were no longer trustworthy [116350]. Additionally, the delay in testing the software to generate ranked-choice results was caused by an impasse with the State Board of Elections, which took more than a year to resolve, and delays in training employees on the software and ranked-choice voting also occurred [116350]. (b) Human actions also played a role in the software failure incident. The board's remaining leaders declined help with the ranked-choice software and delayed training for employees, creating confusion among the staff [116350]. Furthermore, the board ignored offers of technological assistance from the supplier of the software, which could have helped prevent the inclusion of test ballots in the tally [116350].
Dimension (Hardware/Software) software (a) The software failure incident in the New York City Board of Elections was not directly attributed to hardware issues. The incident was primarily caused by human error, specifically a staffer's mistake in not removing test ballot images from the Election Management System, leading to the inclusion of over 130,000 test ballots in the preliminary count [116350]. (b) The software failure incident was primarily attributed to contributing factors originating in software. The incident was a result of human error where a staffer failed to remove test ballot images from the Election Management System, causing the software to include these test ballots in the preliminary count, leading to the release of inaccurate ranked-choice vote totals [116350]. Additionally, the Board ignored offers of technological assistance from the supplier of the software, which could have helped prevent the error. The delay in testing the software and training employees on the software also contributed to the failure [116350].
Objective (Malicious/Non-malicious) non-malicious (a) The software failure incident in the New York City Board of Elections related to the ranked-choice voting can be categorized as non-malicious. The incident was primarily attributed to human error within the agency rather than any malicious intent. The failure was caused by the inclusion of more than 130,000 test ballots in the preliminary count, which led to the results being retracted as they were no longer trustworthy [116350]. The article mentions that the issue was not caused by the software itself but by a human error where a staffer failed to remove the test ballot images from the Election Management System. The supplier of the software, Ranked Choice Voting Resource Center, offered assistance multiple times to help run the ranked-choice voting election and ensure the accuracy of the results, but the offers were declined by the board [116350].
Intent (Poor/Accidental Decisions) poor_decisions, accidental_decisions (a) The software failure incident related to the New York City Board of Elections and the ranked-choice voting system seems to be primarily due to poor decisions made by the agency. The agency faced turmoil with the loss of key personnel, delays in training, and a lack of leadership during an intense election period [116350]. Additionally, the board declined offers of technological assistance from the supplier of the software, which could have helped prevent the inclusion of test ballots in the preliminary count [116350]. The delay in testing the software and training employees also contributed to the failure [116350]. (b) The software failure incident also involved accidental decisions or mistakes made by the agency's staff. The inclusion of over 130,000 test ballots in the preliminary count was attributed to human error, where a staffer failed to remove the test ballot images from the Election Management System [116350]. This accidental decision led to the release of inaccurate ranked-choice vote totals, causing outrage and confusion [116350].
Capability (Incompetence/Accidental) development_incompetence, accidental (a) The software failure incident in the New York City Board of Elections was primarily due to development incompetence. The incident was caused by human error where a staffer failed to remove test ballot images from the Election Management System, leading to the inclusion of over 130,000 test ballots in the preliminary count [116350]. Additionally, the board declined offers of technological assistance from the supplier of the software, which could have helped prevent the error. The delay in testing the software and training employees also contributed to the failure, showcasing a lack of professional competence in managing the ranked-choice voting software [116350]. (b) The software failure incident can also be attributed to accidental factors. The inclusion of the test ballots in the preliminary count was a result of human error, where a staffer mistakenly left them in the system. This accidental mistake led to the release of inaccurate ranked-choice vote totals, causing confusion and outrage across the city [116350].
Duration temporary The software failure incident related to the New York City Board of Elections' ranked-choice voting system can be categorized as a temporary failure. The incident was caused by contributing factors introduced by certain circumstances, such as human error in not removing test ballot images from the Election Management System, delays in testing the software due to an impasse with the State Board of Elections, and challenges in training employees on the software and ranked-choice voting [116350]. These factors led to the erroneous release of preliminary ranked-choice vote totals, which was later retracted, indicating a temporary failure rather than a permanent one.
Behaviour crash, omission, other (a) crash: The software failure incident in the New York City Board of Elections involved a crash where the system lost state and did not perform its intended functions. The preliminary ranked-choice vote totals were released, only to be retracted hours later, acknowledging that they were no longer trustworthy due to mistakenly including more than 130,000 test ballots in the count [116350]. (b) omission: The software failure incident also involved an omission where the system omitted to perform its intended functions at an instance. The board ignored offers of technological assistance from the supplier of the software that it would use to tabulate the ranked-choice votes, which could have helped catch the test ballots inadvertently added to the tally [116350]. (c) timing: The timing of the software failure incident was also a factor. Delays plagued the plan to train employees on the software and ranked-choice voting itself, with some employees not receiving training until after early voting had already begun. Additionally, the board struggled to decide how and when to release the results of the ranked-choice voting, facing pressure to release results earlier than initially planned [116350]. (d) value: The software failure incident did not involve a failure due to the system performing its intended functions incorrectly. (e) byzantine: The software failure incident did not involve a failure due to the system behaving erroneously with inconsistent responses and interactions. (f) other: The software failure incident also involved a leadership vacuum during an intense election with a new method of voting, causing tumult at the board for months. The board's structure, which empowers Democratic and Republican Party bosses to choose staffers, has been criticized for nepotism, ineptitude, and corruption, contributing to the recurring issues faced by the agency [116350].

IoT System Layer

Layer Option Rationale
Perception None None
Communication None None
Application None None

Other Details

Category Option Rationale
Consequence property, delay, theoretical_consequence (d) property: People's material goods, money, or data was impacted due to the software failure The software failure incident in New York City's ranked-choice voting system led to significant property-related consequences. The New York City Board of Elections mistakenly included more than 130,000 test ballots in the preliminary count, which resulted in the release of inaccurate ranked-choice vote totals [116350]. This error impacted the integrity of the election results and caused confusion and outrage among the public, as the initial results had to be retracted and recalculated, affecting the trust in the election process and potentially undermining the legitimacy of the outcome. Additionally, the delay and errors in releasing the results could have financial implications for the candidates and parties involved in the election process.
Domain government The software failure incident reported in the news article [116350] is related to the government industry. Specifically, it involves the New York City Board of Elections and the ranked-choice voting system used in the citywide election. The incident led to the release of inaccurate preliminary ranked-choice vote totals due to the inclusion of more than 130,000 test ballots in the count, causing confusion and outrage among the public and officials. The failure highlighted issues within the government agency overseeing the election process, including leadership vacancies, delays in training employees on the software, and a lack of responsiveness to offers of technological assistance from the software supplier. The incident also brought attention to the longstanding criticisms of nepotism, ineptitude, and corruption within the New York City Board of Elections, as well as the challenges in implementing reforms in the election system. Therefore, the failed system in this incident is related to the government industry, specifically in the context of election administration and voting processes.

Sources

Back to List