What were the findings of the 2024 election audit versus the recount?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
Audits and recounts after the 2024 U.S. general election were different processes that, in the jurisdictions that completed them, overwhelmingly confirmed the original results: state and local risk‑limiting or statistical audits reported outcomes “accurate” and error rates near zero (Pennsylvania, Georgia, Virginia, Ohio, Wisconsin) [1] [2] [3] [4] [5] [6]. Individual recounts were limited, triggered by statutes or petitions, and in the cases reported they affirmed certified results (Alaska recount of Ballot Measure 2; Minnesota and New Mexico recounts or pending recounts) or changed totals only by tiny margins when they did differ [7] [8] [9].
1. What audits do and what the 2024 audits found
Post‑election audits — routinely risk‑limiting audits (RLAs) or statistical/2% audits — hand‑count samples of ballots and compare those counts to machine tallies to test whether the reported outcome is correct; multiple states reported that their 2024 audits confirmed the originally reported winners and overall accuracy (Pennsylvania’s two audits confirmed results and found only six discrepancies, largest changing two votes) [1] [2]. Georgia completed a statewide RLA for the presidential contest and announced the audit “confirmed the outcome” [3]. Virginia’s RLA sampling (1,878 ballots statewide, risk limit 10%) met its threshold and confirmed machine results [4]. Ohio’s post‑election county audits reported a 99.99% accuracy rate [5]. Wisconsin’s large post‑election audit found no machine miscounts and only a vanishingly small number of human errors (error rate reported as effectively zero) [6].
2. What recounts do and what the 2024 recounts showed
Recounts are full or partial re‑tabulations of ballots in a contested race; they are triggered by statute (automatic when margins are tiny) or by candidate petition. In 2024, several recounts occurred and generally reaffirmed certified results: Alaska’s recount of Ballot Measure 2 confirmed the certified margin (743 votes) and left the outcome intact [7]. Some state canvassing boards certified statewide results while leaving a few legislative contests to recounts that could still affect chamber control, showing recounts are used where margins remain critical (Minnesota example) [8]. New Mexico finalized its 2024 results with a small number of contests subject to recount [9].
3. How audits and recounts differ in purpose and scope
Audits are sample‑based, designed to provide statistical evidence the overall reported outcomes are correct; RLAs can stop early if the sample provides strong evidence for the result (Brennan Center context and RLA descriptions) [10]. Recounts are case‑specific, re‑tallying ballots in one contest to resolve a very narrow margin or a legal challenge; recount laws and thresholds vary by state and can require automatic recounts in a number of states (Verified Voting and NCSBE summaries) [11] [12]. Verified Voting explicitly frames audits and recounts as different tools that both build confidence but work in different ways [11].
4. Magnitude of discrepancies reported in 2024 post‑election reviews
Where audits or recounts reported discrepancies, they were minute and usually attributed to human error or ambiguous voter marks — Pennsylvania noted only six discrepancies in its RLA with the largest being a two‑vote change, and Wisconsin’s audit found only five human errors across the audited sample and no machine errors [1] [6]. Ohio’s statewide county audits summarized a 99.99% accuracy rate [5]. These findings indicate that, in official reporting, the processes detected only trivial differences and vindicated the tabulation systems.
5. Competing narratives and what the sources do not say
Election officials and secretaries of state framed these audits as proof of secure, accurate elections (Michigan, Georgia, Ohio, Pennsylvania statements) [13] [3] [5] [1]. Some advocacy groups and private analysts have pushed for additional or “forensic” reviews alleging broader anomalies; however, available sources provided here do not describe any state‑authorized forensic audit that overturned certified outcomes in 2024 — nor do these official reports find machine tampering [6] [13]. The Election Truth Alliance and similar outfits publicly call for more non‑routine audits, but their calls are not the same as state audit findings and do not appear in the official state audit reports cited [14].
6. Limitations and what to watch next
These state reports and summaries are authoritative for their jurisdictions but vary in level of public data and methodology disclosure; some offices publish batch manifests and audit software inputs while others issue summary statements (Georgia published batch manifests and machine tallies for verification) [3]. Laws governing recounts and audits differ across states — 48 states have recount provisions and standards for automatic vs. requested recounts vary — so national generalizations should be cautious (Ballotpedia/NCSL summaries) [15] [16]. If you want a jurisdictional deep dive, consult that state’s RLA or recount report and the underlying ballot manifests the secretary of state or elections board posted [3] [2].
Sources cited above are the official state reports and nonprofit analyses provided in the search results [1] [2] [3] [4] [5] [6] [11] [12] [13] [14] [8] [9] [15] [16].