Contents

Ballot Level Comparison RlaEdit

Ballot Level Comparison RLA, short for Ballot-level Comparison Risk-Limiting Audit, is a method used to verify election results by statistically testing whether the reported outcome is correct. It operates by mapping each ballot to a corresponding electronic record (the cast vote record, or CVR) and then using a carefully designed sampling process to decide, with a known and controllable risk of error, whether a full hand count would be required. Proponents describe it as a practical, evidence-based safeguard that protects the integrity of the vote while avoiding the expense and delays of statewide full recounts. See Risk-Limiting Audit for the broader family of techniques this method belongs to, and Ballot-level auditing for related approaches.

From a policy and governance perspective, Ballot Level Comparison RLAs emphasize the compatibility of electronic tallying with paper ballots. The approach seeks to preserve voter privacy, ensure traceability from ballots to results, and provide an auditable trail that officials and observers can review. In many jurisdictions, it is presented as a middle path between a full hand recount and a purely software-based tally, combining statistical rigor with a transparent audit process. See Cast vote record for discussions of the data used in these audits, and Ballot images for the physical or digital representations that may accompany the process.

What the method is

  • Core idea: Take a sample of ballots and compare the paper ballot image or actual ballot to the corresponding CVR entry. If the discrepancies are sufficiently small, the audit stops with a high level of confidence that the reported outcome is correct. If discrepancies appear too large, the audit escalates or expands the sample, potentially triggering a full hand count.
  • Risk limit: The process is guided by a pre-specified risk limit (often denoted alpha), which is the maximum probability that a wrong outcome would not be corrected by the audit. Lower risk limits require larger samples or more extensive verification.
  • Ballot-level work: Unlike some polling-style audits, the ballot-level comparison method ties each sampled CVR to an individual ballot, enabling precise discrepancy measurements at the ballot level. See CVR for the data structure used in these comparisons.
  • Privacy and integrity: The approach is designed to protect voter privacy while maintaining an auditable link between the paper record and the electronic tally. See Election integrity for a broader discussion of these goals.

How it works

  • Preconditions: An election system must produce a verifiable CVR for each ballot and maintain a reliable chain of custody for paper ballots. The CVRs and ballots must be capable of being linked in a way that preserves security and privacy.
  • Sampling: A statistically principled sampling plan selects ballots for verification. The plan is designed so that the chance of missing a wrong outcome falls below the chosen risk limit.
  • Comparison: Each sampled ballot is checked against the corresponding CVR. The audit measures the size and direction of any discrepancies.
  • Escalation: If discrepancies exceed thresholds, the audit may expand the sample or escalate to a full hand count. If the evidence is strong enough, the audit can conclude with the outcome confirmed.
  • Documentation: The process is documented in audit reports that outline methods, data, sampling, and decisions, helping observers assess credibility. See Audit reporting for examples of how these processes are communicated.

Variants and context

  • Ballot-level comparison vs. ballot polling RLAs: The comparison variant directly ties ballots to CVRs, allowing discrepancy-based decisions. Ballot polling RLAs, by contrast, rely on statistical sampling of ballots to estimate whether the outcome is correct without a direct ballot-to-CVR linkage. See Risk-Limiting Audit for a comparative discussion.
  • Jurisdictional practice: Different jurisdictions adapt the framework to their election infrastructure, including how CVRs are produced, how ballots are stored, and how metadata is maintained. See Election administration for governance considerations and State election laws for how risk limits are codified in law.

History and development

  • Origins: The concept of risk-limiting audits grew out of research by election scholars aiming to provide a verifiable link between machine outputs and paper ballots without resorting to full recounts in every case. Foundational work links to the ideas developed by researchers such as Philip B. Stark and colleagues, who articulated the statistical guarantees thatRLAs can offer.
  • Adoption and pilots: Over the past decade, several states and counties have piloted or adopted ballot-level RLAs, with variations tailored to local ballot formats, tabulation systems, and legal requirements. See Election reforms and Voting system for broader context on how different regions approach verification and auditing.

Controversies and debates

  • Pros and claims from a security-minded perspective:
    • Proponents argue that Ballot Level Comparison RLAs deliver strong assurances of correctness while remaining cost-effective compared with statewide hand recounts. They emphasize the transparency of the process, the auditable trail, and the fact that the method uses actual ballot evidence rather than relying solely on machine tallies.
    • Critics often point to the need for robust CVR accuracy, reliable ballot linkage, and the potential for misinterpretation of statistical results by the public. They may press for broader or different safeguards, such as post-election risk controls, signature checks, or more expansive hand-count verification.
  • Resource and implementation concerns:
    • Some opponents argue that RLAs incrementally increase complexity and require specialized expertise, which can slowdowns in certain jurisdictions. Advocates respond that structured training, clear protocols, and modern election-management software mitigate these concerns.
    • Costs and logistics are a frequent topic: RLAs can reduce the likelihood of a full recount, but they still require data integrity, careful chain-of-custody, and post-election procedures that must be funded and managed well.
  • Political and cultural debates:
    • In debates about election integrity, RLAs are framed by some as a technically sound safeguard that increases public confidence in results, while others view them as part of a broader set of reforms contested on partisan lines. From a practical, policy-first viewpoint, supporters emphasize that RLAs minimize disruption and maximize the value of the paper record as a backstop to electronic tallies.
    • Writings and commentary that dismiss RLAs on broad ideological grounds often conflate method-specific limitations with mischaracterizations of the system as a whole. Proponents argue that well-designed RLAs are evidence-based, transparent, and subject to public scrutiny, and that criticisms should focus on measurable outcomes and verifiable data rather than rhetorical framing. When critics rely on sweeping generalizations, supporters contend that such criticisms miss the core practical gains of the approach and distract from the material question: does the method reliably protect the accuracy of election results?

See also