Risk Limiting Audits

More than half of U.S. states do not require manual audits.  And according to experts, only a few conduct manual audits sufficiently robust to detect hacking. Manual recount laws generally apply only if the margin of victory is less than 1%, and bad actors can avoid them by flipping enough votes to exceed the specified margin.

It is thus imperative that all jurisdictions conduct robust manual audits or full public manual re-counts for all races in 2020.  Thus far, the only type of manual audit recommended by most experts is called a Risk Limiting Audit (RLA). RLAs were invented by UC Berkeley Professor Philip B. Stark and have been endorsed by the League of Women Voters, Verified Voting, and Common Cause (among others).

Additional Resources is maintained by a community of RLA experts. Our goal is to demystify the RLA and connect you to the experts and tools you need to run RLAs in your state.

Risk-Limiting Audits: Frequently Asked Questions and Answers

A “risk-limiting” post-election audit checks if the election outcome is correct. It can catch outcome-changing tabulation errors and correct wrong results. A “risk-limiting” audit limits the risk that the wrong election result will be certified because of a tabulation error.

Why do we need post-election tabulation audits?

For a thriving democracy, we need a voting process that we can trust. Unfortunately,

computer and human errors can produce large miscounts. In past elections, the wrong

candidate has been declared the winner because of a software glitch or programming

error. Since the 2016 election, there have been additional concerns that election results

could be hacked and changed. Post-election tabulation audits check whether a vote

tabulation error led to the wrong result. (Other kinds of audits can detect other errors.)

Would you provide examples of incorrectly reported election outcomes?

In North Kingstown, RI during the November 2016 election, a simple programming error led to the wrong result. The ballot scanner was programmed and tested using the wrong style ballot.  As a result, a measure to create community septic system loan program was defeated – 8471 votes to 5.  The results were so lopsided that Rhode Island election officials conducted a recount after reprogramming the ballot scanner using the correct ballot style.  After the recount was conducted, they discovered that the ballot measure passed, with a vote of 9481 yes to 4562 no.

In a municipal election in Palm Beach County, Florida during March 2012, a

“synchronization” problem with election management software allotted votes to the

wrong candidate and the wrong election contest. The error was caught in a post-

election audit and the results were changed by officials, after a court-sanctioned public

hand count of the votes.*.

How do conventional post-election tabulation audits work?

Election officials sample a percentage of the ballots, typically by choosing machines or

precincts at random. They manually tally the votes on those ballots and compare the

tally to the reported results. Generally, audits stop there. Conventional audits – say, a

flat audit of 3% of precincts – do not actually check whether election results are correct.  They just spot-check the accuracy of some machines.  Read more…

Verifying a statewide election could be this easy and cheap

Wisconsin election officials counted 135,712 ballots in the random voting-machine spot-checks after the November 2018 election, but used a method that did not confirm the winner in any race.

A risk-limiting audit of the same election would likely have verified the correct winners in the statewide races with only 37,841 ballots.

Article courtesy of Wisconsin Election Integrity & Karen McKim

Fact sheet courtesy of Susannah Goodman, Common Cause