Skip to Main Content


Blogs related to important developments and innovations in election administration around the country.

Defining and Piloting Risk-Limiting Audits

Aug 09, 2018

As states look to bolster election security, many jurisdictions are refining their post-election audit processes or piloting new initiatives, including risk-limiting audits (RLAs). For those not familiar, a risk-limiting audit (RLA) provides strong statistical evidence that the election outcome is right and has a high probability of correcting a wrong outcome. Because RLAs are centered on examining the original ballot cast, a paper ballot record (paper ballot or voter verifiable paper audit trail) must exist. 

I have spent the past nine years working on RLAs and have led or observed pilot RLAs in jurisdictions in California, Colorado, Indiana and Virginia. For eight of these nine years, I served as the RLA project manager for the State of Colorado – the first state to implement a statewide RLA. I conducted RLA trainings throughout Colorado and have presented on RLAs at national, regional, and local conferences and meetings. In May of this year, I assisted Ball State’s Voting System Technical Oversight Program (VSTOP) with conducting the first RLA pilot in Indiana at Marion County. Most recently, I had the opportunity to observe the first RLA pilot in Virginia at the City of Fairfax. 

During this time, I have received consistent feedback from jurisdictions that have implemented or piloted RLAs which the EAC has compiled into a white paper released today: Risk-Limiting Audits – Practical Application. Here are some highlights: 

  • Jurisdictions like them. Jurisdictions that have piloted RLAs have positive feedback about implementing them and generally want to keep using them.
  • RLAs are efficient. RLAs saved time compared to other post-election audit methods.
  • Math works. Jurisdictions like to know the expected workload prior to auditing. RLAs provide this in advance since the number of ballots to audit is based on the margin of the selected contest(s) and the risk limit.
  • RLAs provide transparency. For ballot comparison RLAs, election officials get to compare the original ballot cast to how the voting system interpreted ballot by way of its cast vote record.
  • RLA preparation helps with other election processes. A byproduct of RLAs is that jurisdictions get to improve on processes and procedures they already have in place, such as physical security and ballot storage.
  • RLAs are adaptable. RLAs can be used for any type of voting equipment, as long as a paper ballot record exists. 

Research is ongoing for integrating RLAs for state and federal-level elections, and for testing and potentially implementing other statistics-based post-election audits. Colorado will continue to use RLAs, while Rhode Island and Virginia will implement RLAs this November and Washington State will implement an RLA in 2019. The EAC will continue to provide support to jurisdictions that are considering implementing RLAs and continue to provide resources such as this report.