Findings
Published findings from America Counts analyses of vote reporting accuracy.
Each report on this page is an Actual Vote analysis — the published output of a single Comparison Analysis: one election, one jurisdiction (or set of jurisdictions), and the documented comparison between what each precinct's poll tape recorded and what was officially reported. Every report is built on the same methodology, makes the same kinds of claims, and is checkable against the same public archive of underlying poll tape recordings.
For the full methodology behind these reports, see Section 8 of the Technical Manual. For the philosophy guiding how findings are written and characterized, see Section 12: Findings and Reports. Curious what poll tape access looks like where you live? See Actual Vote in Your State.
Every number here is a shared accomplishment. Behind these totals are the volunteers who recorded poll tapes at their voting locations, the people who transcribed and checked them, and the neighbors who simply asked their elections office a good question — together building an open, verifiable record of how the vote was reported. It is what election transparency looks like when it's done in the open, by many hands.
And this is only what's published so far. Many more analyses from past elections are already in the pipeline, waiting for their reports to be written. You can be part of the next set of numbers: use Actual Vote in your next election, and your work will be counted here too.
Thirteen elections' worth of poll tapes, published by the county itself: 974 tapes, 13,986 comparisons and 1,668,865 votes, all reconciled with zero apparent discrepancies. Also a capability test — the whole corpus was transcribed for $83.87 and under four hours of machine time, which is what independent verification costs when a county publishes its tapes.
Cobb County's June 2026 General Primary Runoff: every candidate-vote comparison matched the county's certified results — 260 of 260, 80,449 votes, zero apparent discrepancies, across both election-day and advance voting. The advance vote became comparable after the county answered an open-records request with its per-tabulator cast-vote records.
75,740 votes audited in Florida's second-largest county, using official results captured before state certification. Zero apparent discrepancies — every compared value in the fully-captured precincts matched the official result exactly. A substantial share of transcribed values had no officially published precinct counterpart, and three precincts were excluded as partial captures.
76,530 votes audited across Georgia's three largest counties, completed before state certification. Zero apparent discrepancies — the single mismatch resolved as a write-in canvass event, corroborated by dated official snapshots. 70% of transcribed values had no officially published precinct value to compare against.
A field report, not a comparison audit. The North Carolina State Board of Elections did not publish precinct-level results for this low-turnout Second Primary, so no transcribed value had an official counterpart to compare against — hence the zeros below. Volunteers across four counties still recorded the election-day tapes: 12 transcribable submissions and 72 vote totals, archived and documented here.
321,502 votes audited across seven counties. Zero apparent discrepancies — every transcribed value matched the official results exactly. The methodology's most comprehensive validation to date.
483,415 votes audited in North Carolina's largest county. Zero apparent discrepancies. 79 first-pass differences resolved to matches with a plausible explanation — all accounted for by canvass mechanisms the county itself identified within a day of being asked, 29 of them documented by the county's own canvass minutes. The methodology's first fully documented investigation funnel, from first-pass mismatch to county explanation.
Each report is anchored in the publicly archived poll tape recordings it analyzed. Readers can verify any finding by following the submission links in the report back to the underlying videos in the Actual Vote Video Archive.
For a discussion of what these findings mean and don't mean — the matched-poll-tape caveat, the distinction between "no discrepancies found" and "the election was perfectly accurate" — see Section 12 of the Technical Manual.
