Risk-limiting audit

A risk limiting post-election audit is one of several types of election audits. It is based on statistical principles, and is designed to limit the risk that election results are incorrect.^[1] Specifically, it enables election managers to limit the risk that computer error or fraud identified the wrong winners, without the need for a full recount.

There are two types of election audits: process audits, which determine whether appropriate procedures were followed, and results audits, which determine whether votes were counted accurately.^[2] Risk-limiting audits are one form of a results audit.

As of early 2017, about half the states require some form of results audit. Typically, these states prescribe audits that check only a small flat percentage, such as 1%, of voting machines. As a result, few jurisdictions conduct audits that are rigorous or timely enough for close election results.^[3][4]

In 2010, the American Statistical Association endorsed a specific method of election auditing, known as risk-limiting audits, to verify election outcomes.^[5]

By selecting samples of varying sizes dictated by statistical validity, risk-limiting audits eliminate the need to count all the ballots to obtain rapid confirmation of the outcome (that, is, who won?), while providing a comparable level of valid statistical confidence.

In brief, the process starts with election officials selecting a ‘risk limit,’ such as 5%, meaning they are willing to tolerate no more than a 5% risk that they identified the incorrect winner. Using formulae endorsed by the American Statistical Association, they identify a sample size for each race they intend to verify. The size of the sample depends primarily the margin of victory in the targeted race. Votes in the sample are then manually counted and compared to the computer-tabulated results. If the audit sample produces the same result as the computer-tabulated results, within the selected risk limit, the outcome is confirmed and the audit is complete. If the audit sample does not confirm the original winner within the risk limit, a larger sample is selected and counted. This process can continue until the sample confirms the original winner or a different winner is determined by counting votes from all the ballots.^[6]

In 2011, the federal Election Assistance Commission initiated grants for pilot projects to test and demonstrate the method in actual elections.^[7]

In 2014, the Presidential Commission on Election Administration recommended the method for use in all jurisdictions following all elections, to reduce the risk of having election outcomes determined by undetected computer error or fraud.^[8]

In early 2017, no state has yet implemented risk-limiting audits statewide. California^[9] and Colorado^[10] are leaders in the implementation of risk-limiting audit methods, and individual jurisdictions elsewhere may be using the method on the local election clerks' initiative.

Professor Phillip Stark of the University of California at Berkeley has posted tools for the conduct of risk-limiting audits on the university's website.^[11]

References