Appendix C — Frequently Asked Questions
C.1 — The Basics
What is Actual Vote? Actual Vote is a system for independently verifying vote reporting. It captures the numbers printed on poll tapes at voting locations and compares them against the officially published election results. If the numbers match, the reporting was accurate. If they don’t, something went wrong in the chain from precinct to official results — and AV provides the evidence to investigate what. See Section 2.1 for the full definition.
What is a poll tape? A poll tape is the paper printout a voting machine produces after polls close. It lists every contest on the ballot, the candidates, and the number of votes each received at that machine or precinct. It is the first physical record of the count and the starting point of the vote reporting chain. See Section 2.1 and the Understanding Poll Tapes reference document.
What is America Counts? America Counts (AC) is the non-profit, non-partisan organization that builds and operates the Actual Vote system. AC is a project of Democracy Counts, Inc., a 501(c)(3). See Section 2.5.
What’s the difference between vote counting and vote reporting? Vote counting is what happens inside the voting machine — converting marked ballots into numbers. Vote reporting is what happens after the machine prints the poll tape — transmitting, aggregating, and certifying those numbers into official results. AV audits reporting, not counting. The boundary is the poll tape. See Section 2.1.
Why does vote reporting need to be checked? Because nobody else systematically checks it. Existing safeguards — audits, recounts, canvasses — overwhelmingly focus on whether scanners read ballots correctly (the counting layer). Almost nothing checks whether the scanner’s output was faithfully transmitted to official results (the reporting layer). Documented cases show this gap matters: Monmouth County 2022, Prince William County 2020, Butler County 2008. See the Comparison to Other Election Integrity Approaches section and Section 13.5.
C.2 — Legality
Is it legal to record poll tapes? Yes. Voting results are public information. Photography in public places is generally legal. Recording a publicly posted poll tape combines both principles. See Section 6.1.
Do I need permission from election officials? For Method 1 (recording publicly posted tapes after polls close), no. You are recording public information from a public vantage point. For Method 2 (poll observer recording) and Method 3 (in-office viewing), you may need a formal appointment or advance coordination with the elections office. See Section 7.
What if a poll worker tells me to stop recording? Comply, document the encounter, and move on. You can record from a public sidewalk or other public vantage point instead. Don’t argue at the polling place — report the encounter to AC afterward. See Section 10.6.
What if a property owner tells me to leave? Comply. Property owners can set rules on their property. Move to a public area (sidewalk, street) and record from there if the tape is visible. See Section 6.1 and Section 10.6.
What if the police are called? Stay calm and cooperative. Explain what you’re doing: “I’m a volunteer with America Counts, a non-partisan organization. I’m recording publicly posted election results, which is legal.” Show ID if asked. In Broward County in 2018, the responding sergeant confirmed the team’s legal right to be present. See Section 10.6.
Can AV evidence be used in court? AV recordings are admissible in legal proceedings under standard rules of evidence — they are relevant, probative, and authenticated through metadata (GPS, timestamps) and chain of custody. The Legal Pathways paper describes six procedural contexts where AV evidence can reach courts and regulators. See Section 6.4 and Section 6.5.
C.3 — Participation
How do I get started? Download the Actual Vote app, create an account, and practice with the practice poll tape. On election night, go to your polling place after polls close, find the posted tape, record it with the app, and submit. The whole recording takes 2–5 minutes. See the Training and Volunteer Onboarding section and Section 10.1.
What equipment do I need? A smartphone with the Actual Vote app installed, a charged battery, and a flashlight. A portable charger is recommended. See Section 10.1 and the Training and Volunteer Onboarding section.
Will the app run on my phone? Almost certainly. Actual Vote runs on any iPhone or iPad on iOS 15 or later, and any Android phone or tablet on Android 8.0 (Oreo) or later — roughly any phone sold since 2017. The only hardware requirement beyond that is a working camera, plus a little free storage (each recording is about 20 MB). If your device is older than that, use the camera-fallback method below. See Section 5.1.3.
What if the app isn’t working?
Use your phone’s native camera to record the tape, then email the video to j.flatley@democracycounts.org with the polling place name and precinct identifier. See Section 10.7.
What if tapes aren’t posted at my polling place? Document the absence — a brief video or photo showing the location without tapes. This is useful information. Not every location will have tapes posted, even in states that require it. Move to your next location. See Section 10.11.
Can I participate if my state doesn’t post poll tapes? Yes. Method 3 (in-office viewing at the elections office after the election) and Method 4 (public records request) work in states that don’t post tapes publicly. Contact AC for guidance on your state. See Section 7.
How can I do more than just record?
You can join or form a video-recording team (Section 10.2), become a transcriber, become an analyst, or start an Audit Committee (Section 10.3). Contact Jason Flatley at j.flatley@democracycounts.org to discuss. See the Training and Volunteer Onboarding section.
Can my organization partner with AC? Yes. AC works with partner organizations including Scrutineers, Common Cause chapters, League of Women Voters chapters, and local civic groups. Contact AC to discuss scope, jurisdiction, and coordination. See Section 10.2 and Section 10.10.
C.4 — How It Works
What happens to my recording after I submit it? It enters AC’s three-layer pipeline: vetting (a human reviews it for quality and privacy), transcription (the numeric values are extracted into a structured format), and comparison (the values are matched against official results). See Section 4.1.
Does a computer or a human read the poll tapes? Both, by design. Most tapes are now read by Actual Vote Assist (AVA), a pipeline that uses AI vision to read every number off the tape, reconcile it across multiple video frames, and compare it to the official results automatically — which is what lets AC check a whole county’s worth of tapes in minutes rather than over an afternoon. But the machine never works unsupervised. A person first sets aside recordings too blurry or incomplete to read, and tells the system what contests and candidates the tape should contain so it knows what it’s looking at. After the automated pass, anything the system was unsure about is routed to a human to confirm or correct, and reviewers can spot-check even the reads it was confident about. The guiding rule is that the system flags uncertainty rather than inventing a number: a bad recording produces fewer confident reads, never fabricated ones. See Section 4.8.
How are poll tape values compared to official results? AC obtains the official results from the county or state elections office (published online, via public records request, or both), builds jurisdiction-specific comparison logic that matches each precinct’s tape values to the corresponding official figures, and runs the comparison programmatically. Any discrepancy triggers the seven-check investigation process. See Section 8.7–Section 8.9 and the Doing an Actual Vote Comparison Analysis section.
What is the seven-check investigation process? When a comparison shows a discrepancy, the analyst checks seven possible explanations in order: (1) transcription error, (2) comparison logic error, (3) missing poll tapes, (4) duplicate tapes, (5) aggregation logic error, (6) undocumented vote reporting logic, (7) operational context. Most discrepancies resolve at checks 1–3. See Section 8.9.
What happens when a real discrepancy is found? AC asks the relevant elections office — a cooperative inquiry that presents the evidence and requests an explanation. The weight of the ask is proportional to how unexplained the discrepancy is: a likely attributable discrepancy (one with a working benign hypothesis) gets a lighter ask; an apparent discrepancy (no benign hypothesis yet) gets a heavier ask. If the office’s response is satisfactory, the discrepancy resolves to a match. If the office acknowledges the error, it becomes a true discrepancy. If neither AC nor the office can resolve it, the discrepancy becomes escalatable — ready to be handed to a court-capable party — and may be escalated through appropriate channels, which may include regulatory complaints, media transparency, supporting candidate action, public records litigation, federal civil rights litigation, or criminal referral. See Section 8.10–6.11 and the Doing an Actual Vote Comparison Analysis section.
Has AV ever found a real error? AV’s most complete analyses — the NC 2024 Primary and the 2024 General Election — found zero true discrepancies: every apparent discrepancy (one with no benign hypothesis) was resolved through investigation to a match, and no elections office acknowledged an error nor did any court confirm one. This is a substantive finding: it is evidence that vote reporting in the analyzed jurisdictions was accurate. AV’s methodology would have detected the errors documented in the case studies (Shelby County 2015, Monmouth County 2022, Prince William County 2020, Butler County 2008) had it been in use at those times. See the Case Studies section.
What is the Actual Vote Video Archive? A permanent, publicly accessible repository of every vetted poll tape recording. It preserves the evidence independently of government custody — recordings cannot be destroyed by election officials, lost in storage, or shredded. See Section 4.6.
C.5 — What AV Can and Can’t Do
Can AV detect ballot stuffing or vote-flipping? No. Those are counting-layer problems — they happen before the poll tape is printed. AV checks the reporting layer (tape to official results). If the counting is wrong, both the tape and the official results will reflect the same wrong numbers, and AV’s comparison will show a match. See Section 13.4 and the counting-layer section of the Taxonomy of Vote Reporting Vulnerabilities.
Can AV prove an election was stolen? No. AV can detect discrepancies between poll tapes and official results and provide evidence about whether vote reporting was accurate. It cannot determine intent (error vs. fraud), and it cannot detect problems in the counting layer. See Section 13.4.
Can AV prove an election was conducted correctly? AV can provide evidence that vote reporting was accurate in the precincts it analyzed. High match rates are strong evidence of reporting accuracy. But AV cannot verify precincts it didn’t cover, and it cannot verify the counting layer. See Section 13.4.
Does AV replace audits or recounts? No. AV is complementary to existing safeguards. Audits and recounts primarily check counting accuracy; AV checks reporting accuracy. A system with both is stronger than a system with either alone. See Section 13.5 and the Comparison to Other Election Integrity Approaches section.
What is the “matched poll tape caveat”? If a voting machine is compromised so that it miscounts ballots and prints a tape reflecting the miscount, AV will compare the wrong tape to the wrong official results and find a match. AV detects reporting-layer problems, not counting-layer problems. This limitation is discussed openly throughout the manual. See Section 2.3 and Section 13.4.
C.6 — The Organization
Is America Counts partisan? No. AC is non-partisan by policy and by structural design. AC does not endorse candidates, does not affiliate with any party, does not accept donations from campaigns or parties, and does not coordinate findings with any political organization. The methodology produces the same result regardless of which party benefits. See Section 13.2.
Who funds America Counts? AC is funded by non-partisan donations. It does not accept money from candidates, campaigns, political parties, or partisan PACs. It does not accept funding that would bias its findings. See Section 13.2.
Who runs America Counts? Jason Flatley is the CTO and primary operational contact. AC is a project of Democracy Counts, Inc., a 501(c)(3) non-profit. See Section 2.5 and Section 10.8.
How do I contact AC?
Email j.flatley@democracycounts.org or info@americacounts.us. During election week, AC monitors email closely. See Section 10.8.
C.7 — For Elections Officials
What does AC want from my office? AC wants to confirm that your office’s vote reporting is accurate — and in the vast majority of cases, it will. A clean AV result is the strongest independent endorsement of your office’s professionalism. Specifically, AC may request: confirmation that tapes will be posted (in posting states), official results in machine-readable format, documentation of your reporting structure, and in-office access to view tapes (for Method 3). See Section 10.4 and Section 15.
What format should I publish results in? CSV is the gold standard. Structured Excel (.xlsx) is acceptable. JSON or XML work. Searchable-text PDFs are acceptable as supplements but should not be the only format. Results should be published at the precinct × voting method × contest × choice level. See Section 15.
What should I do if AC contacts me about a discrepancy? Respond. AC’s inquiry is professional, specific, and non-accusatory — it presents evidence and asks for an explanation. Most apparent discrepancies have mundane explanations (re-scans after ballot jams, reporting grain differences, precinct consolidations). A prompt, cooperative response resolves the question and produces a positive outcome for both the office and the public. See Section 8.10–6.11, Section 10.4, and the Doing an Actual Vote Comparison Analysis section.
Does AC publish findings that make my office look bad? AC publishes what it finds. In practice, the vast majority of findings are clean matches — evidence that your office’s reporting was accurate. When discrepancies are found, AC contacts your office first and gives you the opportunity to explain before publishing. AC’s published reports include the explanation if one is provided. See Section 8.10 and Section 13.1.
C.8 — Technical Questions
What voting systems does AV work with? AV works with any system that produces a physical poll tape — which is the vast majority of voting equipment in use in the United States, including systems from ES&S, Dominion, and Hart InterCivic. The tape format varies by vendor and jurisdiction, but AV’s methodology is format-agnostic: the transcriber reads the numbers regardless of layout. See Section 4.2 and the Understanding Poll Tapes reference document.
How does AV handle different reporting structures across jurisdictions? Each jurisdiction reports results differently — different precinct naming conventions, different breakdowns by voting method, different aggregation levels. AC builds jurisdiction-specific comparison logic for each analysis. This is the most technically demanding part of the pipeline and the reason jurisdiction-specific knowledge matters. See Section 8.7 and the Doing an Actual Vote Comparison Analysis section.
What is a “comparison analysis”? The full end-to-end process of scoping an election, obtaining poll tape recordings, transcribing them, obtaining official results, building comparison logic, running the comparison, investigating discrepancies, and publishing findings. See Section 8.1 and the Doing an Actual Vote Comparison Analysis section.
Does AV use AI? AC is moving toward AI-assisted transcription — using vision models for a first pass on poll tape readings, with human review. The comparison and investigation steps remain human-driven. See the Training and Volunteer Onboarding section.
Where can I read AC’s published analyses? The NC 2024 Primary Report and the 2024 General Election News Page are AC’s canonical published analyses. They are available on the America Counts website alongside this manual. See Section 12 and our analyses.
C.9 — Common Misconceptions
“AV is just for people who think elections are stolen.” No. AV is for anyone who thinks election results should be independently verified — the same principle behind financial audits, building inspections, and scientific peer review. Most AV analyses confirm accurate reporting. The goal is confidence, not suspicion. See Section 13.
“If audits already exist, why do we need AV?” Because existing audits primarily check the counting layer, not the reporting layer. Virginia’s risk-limiting audit confirmed counting accuracy with 99%+ confidence while a 4,000-vote reporting error went undetected. Monmouth County’s audit reported “100% accuracy” while the wrong person held office. AV fills the specific gap these audits leave. See the Comparison to Other Election Integrity Approaches section.
“AV only checks a few precincts — what good is that?” AV checks every precinct where a tape is recorded. The more precincts covered, the stronger the evidence. Even partial coverage is valuable: a sample of precincts showing accurate reporting increases confidence in the broader result, and a discrepancy at any single precinct is worth investigating regardless of coverage percentage. See Section 8.2 and Section 13.4.
“Recording poll tapes is illegal / suspicious / intimidating.” Recording publicly posted poll tapes is legal (Section 6.1). It is no different from photographing any other publicly displayed government document. AC volunteers are trained in de-escalation and non-confrontational behavior (Section 10.6). The work is civic participation, not surveillance.
“If AV hasn’t found any real errors, what’s the point?” Finding no errors is a finding. It is independent evidence that vote reporting was accurate in the analyzed precincts — evidence that did not exist before AV produced it. The value of a smoke detector is not diminished by the absence of fires. See Section 13.4 and our analyses.
“AV is just one organization’s opinion about election results.” AV’s methodology is publicly documented (Section 8), its evidence is publicly archived (Section 4.6), and its findings are replicable by anyone with access to the same data. It is not an opinion — it is a comparison between two sets of numbers, both of which are public. See Section 13.1.