Risk-limiting audits are an effective tool that can be implemented in the state of Arizona


Paul Rosenzweig
Former Resident Senior Fellow, Cybersecurity and Emerging Threats
Elizabeth Howard
Brennan Center
Turquoise Baker
Brennan Center

Key Points

Arizona should require postelection audits by eliminating the audit stop order triggered by a political party’s failure to participate.

Arizona should replace the currently required postelection audit with the more effective and typically more cost-efficient risk-limiting audit.

Arizona should begin by establishing an RLA pilot program that will allow local election officials to test the procedure and give officials and the public time to provide feedback before issuance of uniform statewide procedures and documentation requirements.


In the face of record-breaking voter turnout and a global pandemic, Arizona election officials and poll workers successfully administered a safe and secure November 2020 general election. In large part, that was because of the coordinated effort of state, local, and federal officials to ensure that Arizona was well prepared for the election effort.

Two of the most important election security measures in use in Arizona today, paper ballots and postelection audits, were first implemented more than a decade ago. Currently, Arizona is one of at least 24 states that conduct postelection audits, which require hand review of paper ballots, prior to certification of the election results. However, as we discuss in detail below, there are two substantive deficiencies in Arizona’s current audit law: (1) local election officials are prevented from completing an audit if one or more political parties refuse to participate and (2) the type of postelection audit required (a “traditional” postelection audit) limits the efficacy and the flexibility of the audit.

A risk-limiting audit (RLA) is a check on the election outcome. Through the use of proven statistical methodologies, an RLA provides voters with confidence in the accuracy of election results. It can be conducted publicly and is designed to detect, and correct, counting errors or malicious attacks that change the outcome of an election.

Featured Publications