Election Auditing

The foundation of a functioning democracy rests on the trust citizens place in the electoral process. Therefore, auditing should be considered a cornerstone in the complex system of election administration.  These election details are expected to help promote voter and stakeholder confidence in the election administration process and consequently the legitimacy in the outcomes of elections.

Post-election tabulation audits are the most widely known and most investigated in terms of theory and methods, but audits can occur before, during, and after Election Day, and can cover processes such as voter registration, logic-and-accuracy testing, and ballot management.  Procedures around forensic audits, which have been promoted by election integrity activists, have not been identified and, as practiced, cannot properly be termed audits.

Post-election tabulation audits can be further categorized into three types:  tabulation of a fixed percentage of machines or votes, Risk-Limiting Audits (RLAs), and machine-assisted audits.  RLAs are the post-election tabulation audit that has garnered the most attention from researchers and policymakers in recent years. Four major RLA methods have been developed:  ballot comparison, batch comparison, ballot-polling, and machine assisted. 

Research topics that have been explored in this area include:

• How are post-election audits implemented?
• How do they perform/what are the results?
• How do they impact public confidence?

Despite great attention devoted to election audits in recent years, research in this area is still in its infancy.  Areas still in need of research include the following:

• How best to organize and hand-count ballots in audits
• The relationship between type of audit and results, efficiency, and security
• The impact of audits on voters, including their awareness of audits, methods of reporting audit results to the public, and the public availability of ballot images and cast vote records.
• Qualitative case studies that document how states perform audits.
• Developing a scientific basis for sample sizes in traditional audits and for newer types of audits (e.g., voter registration audits).
• Accuracy of voter registration files.
• Better understanding of how ballot management, in-person voting, and mail-balloting procedures are audited for accuracy.

Research into audits requires the availability and maintenance of data.  This includes developing protocols and agreements that allow data sharing between election jurisdictions and researchers, and protocols that secure election materials and ensure voter privacy.

We know little about election auditing, except as it relates to post-election tabulation audits, and even that is in its infancy.  There are many opportunities to engage in research and for election officials and researchers to work together to create broader transparency around election auditing to help improve operations and trust in the system, and to advance the field of election science.