3 — Understanding Poll Tapes

3.1 — What Is a Poll Tape?

Poll tapes — also called results tapes or election tapes — are long strips of paper printed by vote tabulating machines. They record a comprehensive set of information about the tabulation process and its results at a specific precinct for a specific voting method. Poll tapes are printed at the end of voting: after polls close on election day, or at the close of an early voting period.

Poll tapes are the starting point of the vote reporting chain. The vote totals printed on them are the ground-truth numbers that get aggregated to produce official county and statewide results. An Actual Vote analysis compares these ground-truth numbers to the officially reported results to check whether the aggregation was done correctly.

A poll tape resembles a cash register receipt — it’s a continuous strip of thermal paper that can be several feet long, printed by a thermal printer built into the tabulating machine. The paper itself is typically white with black thermal printing. Election workers are responsible for posting the poll tape at the polling location or securing it in an official envelope that becomes part of the official record, depending on state law.

Poll tapes exist because the vote totals stored digitally on a tabulating machine’s USB flash drive must also exist in a human-readable paper form. The tape is the paper-based verification of what the machine recorded electronically. When Actual Vote compares poll tape totals to official results, it is comparing this paper record against the digital record as processed by the county’s election reporting software. Discrepancies could theoretically occur at either stage — at the point the tape is printed, or later when the digital data is transferred and aggregated.

3.2 — The DS200 Precinct Scanner

What It Is

The ES&S DS200 Digital Precinct Scanner is one of the most widely deployed vote tabulating machines in the United States. It is the primary machine encountered in Actual Vote analysis work and the focus of this chapter. The DS200 is certified by the U.S. Election Assistance Commission (EAC) under the Voluntary Voting System Guidelines (VVSG) and has been deployed across multiple system versions, including EVS 5.0.0.0, EVS 6.0.4.0, and EVS 6.1.1.0.

The DS200 serves dual roles: it operates as a precinct-level tabulator on election day (scanning ballots one at a time as voters feed them in) and can also function as a central count tabulator for extended voting periods such as early voting or absentee ballot processing. It reads both one-sided and two-sided paper ballots using optical mark recognition (OMR) technology, evaluating both valid ballot marks and extraneous marks such as smudges or folds. It also accepts printed ExpressVote summary cards (see below).

Sources: ES&S DS200 Operator Guide, Chapter 1; EVS 6.1.1.0 Poll Worker Guide, pp. 6–7; EAC Certificate of Conformance for EVS 6.1.0.0.

Key Hardware Components

Understanding the physical machine helps contextualize what you’re reading on a poll tape.

• Touch screen: The operator interface for opening and closing polls, selecting report types, entering security codes, and monitoring ballot counts during voting. The screen flips up on a laptop-style hinge that also functions as a security lid.
• Thermal printer: Built into the front access compartment. Prints all reports — including poll tapes — on continuous thermal paper rolls. The front access compartment is opened via a blue lever and is locked with a barrel key during voting to prevent unauthorized access to the printer or paper.
• Paper ballot feed: The main input slot where voters insert hand-marked paper ballots. Ballots may be inserted in any orientation.
• ExpressVote card feed: A separate input slot for ballot marking device (BMD) summary cards.
• Front access compartment: Houses the thermal printer, paper roll, and primary USB flash drive port (in some configurations). Locked with a barrel key during voting.
• Rear access compartment: Houses the primary USB flash drive (election media) and backup USB port. Sealed with a tamper-evident wire seal before election day.
• Ballot box base: The lower portion of the unit. Holds approximately 2,500 ballots in the main bin, plus approximately 100 in the emergency/auxiliary bin. Secured with key locks.

Two different keys are required to operate the DS200: a standard silver key for the ballot box locks and a barrel key for the scanner locks. Both are needed to set up and close down the machine.

Sources: EVS 6.1.1.0 Poll Worker Guide, pp. 6–8; Fairfax County Opening Procedures guide.

ExpressVote Integration

In many jurisdictions, the DS200 works alongside ES&S ExpressVote ballot marking devices (BMDs). The ExpressVote is not a tabulator — it is an accessibility device that allows voters to make selections on a touchscreen and then prints a paper summary card with a barcode encoding those selections. The voter feeds this card into the DS200 for tabulation, just like a hand-marked paper ballot.

On the poll tape, ExpressVote cards and hand-marked ballots are indistinguishable — both contribute to the same vote totals. The Ballot Status Accounting Report (discussed in Section 3) separately tracks the total number of ExpressVote cards used, but the Voting Results Report does not break out BMD cards from paper ballots. In jurisdictions using both input methods, the poll tape reflects combined totals, and you cannot determine from the tape alone how many votes came from each method.

Some DS200 configurations also support multiple ExpressVote hardware versions (HW 1.0, HW 2.1, and the larger ExpressVote XL), all of which produce summary cards that the DS200 can read.

Sources: EVS 6.1.1.0 Poll Worker Guide, pp. 7, 10; ES&S Operator Guide, Chapter 5; Verified Voting: ES&S DS200.

Other Tabulators You May Encounter

While the DS200 is the focus of this manual, poll tapes are produced by all precinct-level tabulating machines from all major manufacturers. The field names, layout, and formatting differ, but the underlying data is analogous: precinct identification, contest-by-contest vote totals, machine metadata, and timestamps.

Other ES&S equipment:

• DS450: A central count scanner processing approximately 75 ballots per minute. Produces accumulation tapes combining results from multiple precincts processed in batch. Not a precinct-level device and not typically encountered in field recording work.
• DS850: A high-volume central count scanner processing approximately 300 ballots per minute. Same function as the DS450 at much higher throughput. Uses commercial laser and dot-matrix printers rather than an integrated thermal printer.

Other manufacturers:

• Dominion Voting Systems: The ImageCast Precinct (ICP) and ImageCast Evolution (ICE) are precinct-level optical scanners broadly comparable to the DS200. They produce results tapes with similar information but different formatting and field names.
• Hart InterCivic: The Verity Scan and older eScan are precinct-level scanners with their own tape formats.

All of these machines are subject to EAC certification testing against the VVSG. The core principle — that a printed results tape is the ground-truth paper record of votes tabulated at a specific location — applies universally regardless of manufacturer.

Sources: Verified Voting: ES&S DS200; EAC certified voting systems; Texas Secretary of State EVS voting-system examination report.

3.3 — What the DS200 Prints

The DS200’s integrated thermal printer produces a variety of reports throughout the election day lifecycle. Understanding the full catalog of report types is important both for identifying what you’re looking at on a tape and for understanding the context around the results tape that is the primary focus of Actual Vote analysis.

The Opening Sequence

When the DS200 is powered on and the Election Security Code is entered, the machine prints a series of reports automatically before voting begins. Poll workers collectively refer to this output as the “Opening Tape” — it is a continuous strip of paper containing multiple distinct reports printed sequentially. In Fairfax County, Virginia, for example, opening procedures require two officers to sign the Opening Tape, defined as the Configuration, Status, and Zero Totals Reports taken together.

The opening sequence typically prints:

1. Configuration Report (also called the Status Report or Hardware Configuration Report)
2. Ballot Status Accounting Report (opening version)
3. Zero Totals Report (one or more copies)

Each of these is described below.

Sources: Fairfax County Opening Procedures guide (Office of Elections, Rev. 05/21); EVS 6.1.1.0 Poll Worker Guide, p. 9.

The Closing Sequence

When a poll worker presses the Close Polls button at the end of voting, the DS200 automatically prints another set of reports. The specific reports printed at closing are determined by the election definition — the election programmer configures which reports the machine generates automatically. A typical closing sequence may include:

1. Ballot Status Accounting Report (closing version, with actual counts)
2. Voting Results Report (multiple copies — typically 3)
3. Write-In Images Report (if configured and write-in ballots were cast)
4. Event Log Report (if configured to print on close)

Each copy of the Voting Results Report must be inspected and signed by at least two election officers before being secured.

Sources: ES&S Operator Guide, Chapter 5; Fairfax County Closing Procedures guide; EVS 6.1.1.0 Poll Worker Guide, p. 11.

Configuration Report in Detail

The Configuration Report (also called the Hardware Configuration Report or Status Report) is the first report printed during the opening sequence. It is the most information-dense report the DS200 produces and serves as a snapshot of the machine’s state before voting begins. Its contents include:

• Voting Device Information: Current date/time, time zone, firmware version, touch screen status, power source (AC or battery), battery charge percentage, modem status (whether a modem is detected — some DS200s can transmit results via modem), compact flash card status, and scanner operating status.
• Storage: Available flash memory, available ballot storage capacity, and available backup storage.
• Firmware Details: Separate version numbers for the Paper Processing Engine, Ballot Transport Engine, Tabulation Engine, Scanning Engine, Printer Firmware, Scanner Firmware, and Operating System. Also includes the date firmware was last updated and the Unknown Count and Protected Count.
• Election Qualification Information: Election Name, Election ID, and Election Qualification Code (EQC) — the cryptographic credential confirming the election definition is authorized.
• Election Information: Jurisdiction name, election name, election date, polling place ID and name, election type, number of precincts served, ballot styles available, ballot dimensions, and target type (e.g., oval).
• Election Settings: Ability to re-open polls (yes/no), number of result tapes to print (typically 2–3), whether the Event Log prints on close, report detail (Public or Detailed), report level (Polling Place or Precinct), candidate sort order, display candidate order, whether to combine over and under votes into a single field, and ballot image storage setting (all, write-ins only, or none).
• Exception Ballot Settings: The machine’s configured response to each type of ballot exception — Blank Ballots, Overvoted Ballots, Undervoted Ballots, Unreadable Marks, and Crossover Ballots — each set to Accept, Reject, or Query (see “Ballot Exception Handling” below).
• Diverter Settings: Whether a ballot diverter is installed and configured to separate write-in ballots, overvoted ballots, or blank ballots into a dedicated bin.

The Configuration Report is not transcribable for vote comparison purposes, but it is valuable context for understanding how the machine was set up. If you have access to the opening tape bundle, reviewing the Configuration Report can explain why a results tape looks the way it does — for example, why overvote and undervote counts are present (Detailed format) or absent (Public format), or why ballot image records are or aren’t available for post-election review.

Sources: ES&S Operator Guide, Chapter 7 (Configuration Report sample fields); EVS 6.1.1.0 Poll Worker Guide, p. 11 (Voting Device Status screen).

Ballot Status Accounting Report: Opening vs. Closing

The Ballot Status Accounting Report is printed automatically at both the beginning and the end of voting, and the two versions serve different purposes.

The opening version is a baseline: it confirms zero ballot counts across all ballot style codes, lists the election configuration settings, and documents the machine’s starting state. It includes the timestamp, unit serial number, jurisdiction and election identifiers, Poll Opened date and time, and ballot counts by code (all zeros).

The closing version contains the same fields but now populated with actual counts: total ballots cast, ballots by style code (showing how many ballots of each style were scanned during voting), and the total number of ExpressVote cards used. This closing version serves a reconciliation function — the total ballots on this report should match the Total Ballots field in the Voting Results Report. Any discrepancy between these two reports is an anomaly worth investigating.

For multi-day voting periods (such as early voting), a Daily Ballot Status Accounting Report can be printed on demand from the Tools menu. This shows daily ballot counts broken down by ballot style, allowing poll workers to reconcile paper ballots at the end of each day without closing the polls.

Sources: ES&S Operator Guide, Chapters 5 and 7; EVS 6.1.1.0 Poll Worker Guide.

Ballot Exception Handling

The DS200 can be configured to handle ballot exceptions — ballots with potential issues — in three ways for each exception type:

• Accept (Always Accept): The ballot is accepted and tabulated without warning the voter. The exception is silently recorded in the machine’s data but no on-screen alert appears.
• Reject (Always Reject): The ballot is physically returned to the voter. It cannot be processed until the issue is corrected.
• Query: The DS200 displays a warning screen describing the issue and giving the voter a choice — typically “Cast Your Ballot with Mistakes” or “Correct Your Ballot” (return for corrections).

The exception types are:

• Blank Ballots: No voting marks detected on the entire ballot. The screen reads “You have made no selections on this ballot.”
• Overvoted Ballots: More marks than allowed in a contest (exceeded the “Vote For” number). The screen identifies which contest or contests are affected.
• Undervoted Ballots: Fewer marks than allowed in a contest. This includes leaving a contest entirely blank.
• Unreadable Marks: Marks that the scanner cannot confidently interpret — partial fills, stray marks, smudges, or folds that interfere with the scanning optics.
• Crossover Ballots: In primary elections with party-specific sections, marks appearing in more than one party’s contests.

The Configuration Report documents the setting for each exception type. How these settings interact with the results tape: in Detailed format tapes, overvote and undervote counts appear for each contest regardless of the exception handling setting. In Public format tapes, these counts are omitted entirely. This is one reason Detailed-format tapes are preferable for analysis — they show the full picture of how ballots were processed, including the exceptions.

Sources: ES&S Operator Guide, Chapter 5; EVS 6.1.1.0 Poll Worker Guide, p. 10.

Complete Report Type Catalog

The following table lists every report type the DS200 can produce, when it is printed, what it contains, and its relevance to Actual Vote analysis.

Write-In Images Report

The Write-In Images Report is a distinct report printed automatically at closing whenever write-in ballots were cast and the machine is configured to do so. It is part of the closing sequence and is printed alongside the Voting Results Report.

The report contains digitized images of all write-in ballot selections captured during the voting period. When a voter casts a ballot with a write-in entry, the DS200 captures an image of the write-in field and stores it on the USB flash drive with a randomized filename to protect voter anonymity. (The DS200 also physically stamps a pink circle on the trailing edge of the ballot, face-down, so that write-in ballots can be physically sorted from regular ballots during closing procedures.)

At the central count or accumulation station, election officials review the Write-In Images Report — either from the printed tape or from USB drive data on a computer screen — and compare handwritten entries against the certified List of Declared Write-In Candidates maintained by the Secretary of State’s office. Only entries matching a declared write-in candidate count as valid write-in votes. Ballot image storage is configurable in the election definition: save all ballot images, save only write-in ballot images, or save none.

Write-in handling is a common source of small discrepancies in Actual Vote analysis. The poll tape records the raw write-in count (all write-in entries), while official results may reflect adjudicated write-in totals (only write-ins matching declared candidates). Understanding this distinction is important when investigating such mismatches. See also Section 4 (Write-In Handling) for details on how write-in votes are processed and how they appear on the Voting Results Report.

Sources: ES&S EVS 6.1.1.0 system documentation (EAC certified system); Fairfax County DS200 Opening & Closing Procedures Guide.

Event Log Report

The Event Log is a report the DS200 can print at closing or generate on demand from the Tools menu. It exists in two versions:

• Event Log Summary: A high-level timeline of significant machine events. Useful for quickly identifying when polls opened and closed, whether any errors occurred, and whether any unusual events took place.
• Event Log (Full): A granular record of every machine event with timestamp: machine initialization, Election Security Code entry, poll opening, each ballot scanned, ballot exceptions (overvotes, undervotes, blank ballots, unreadable marks) and how each was resolved (accepted, rejected, returned to voter), report printing events, poll closing, and any error conditions including paper jams or system warnings.

The Event Log is not a primary transcription target for Actual Vote analysis, but it is valuable for investigating anomalies. If questions arise about timestamps, ballot counts, or what happened at a precinct during voting, the Event Log provides an audit trail that can reconstruct the sequence of events. Whether the Event Log prints automatically at closing is determined by the “Print Event Log on Close” setting in the election definition (visible in the Configuration Report).

Source: ES&S EVS 6.1.1.0 system documentation (EAC certified system).

Public vs. Detailed Report Format

A critical distinction that affects what data is available for transcription: the DS200 can print reports in one of two formats.

• Public format: Displays candidate names, candidate vote totals, and write-in totals for each contest. This is the format typically used for tapes that will be publicly posted.
• Detailed format: Displays everything in the Public format plus undervote totals and overvote totals for each contest. This is the more useful format for analysis work because the additional fields provide internal verification checks and richer data.

The report format is hard-coded into the election definition by the election programmer — poll workers cannot change it at the machine. The Configuration Report (printed at opening) shows which format the machine is set to use, in a field labeled “Report Detail.” If you are looking at a tape and see overvote/undervote counts for each contest, you have a Detailed-format tape. If those fields are absent, it is a Public-format tape.

For Actual Vote analysis, Detailed-format tapes are preferable because the overvote and undervote counts provide additional verification points. However, Public-format tapes are fully usable for the core comparison of candidate vote totals to official results.

Sources: ES&S Operator Guide, Chapter 7; EVS 6.1.1.0 Poll Worker Guide.

Polling Place vs. Precinct Report Level

The DS200 can also generate reports at two different aggregation levels:

• Polling Place level: Results for the entire polling location, combining all precincts served at that location.
• Precinct level: Results broken out by individual precinct.

Like the report format, the report level is configured in the election definition. The Configuration Report shows the setting in a field labeled “Report Level.”

For Actual Vote analysis, precinct-level reports are strongly preferred because official results are typically reported at the precinct level. A polling-place-level report from a location serving multiple precincts cannot be directly compared to precinct-level official results. If you encounter a polling-place-level tape, note it in your metadata — it may still be usable if the polling place serves only one precinct (which is common), but it requires verification.

Sources: ES&S Operator Guide, Chapter 7.

3.4 — Anatomy of a DS200 Results Tape

The Voting Results Report — commonly called the results tape or poll tape — is the report that matters most for Actual Vote analysis. This section describes every field on a DS200 results tape in detail.

A results tape is printed after polls close (or after an early voting period ends) and is a single continuous strip of thermal paper containing several distinct sections printed sequentially.

Section 1: Metadata Header

The results tape begins with identifying information that links the tape to a specific machine, location, and election. This metadata is essential for matching the tape to official results during analysis.

• Timestamp: The date and time the report was generated. For a closing results tape, this should reflect the official closing time or shortly after — typically 15 to 45 minutes after the posted poll closing time (e.g., 7:15–7:45 PM for a precinct that closes at 7:00 PM), reflecting time needed for final ballot processing and closing procedures. Timestamps significantly outside this window may indicate a clock error (see Section 10, AM/PM Errors). For opening reports, the timestamp should be around 6:00 AM or whenever polls open in that jurisdiction.
• DS200 Firmware Version: The software version running on the machine (e.g., 2.17.0.0 or 2.30.0.0). This field encompasses multiple firmware components; the tape typically shows the primary firmware version. The Configuration Report provides granular detail including separate versions for the Paper Processing Engine, Ballot Transport Engine, Tabulation Engine, Scanning Engine, Printer Firmware, Scanner Firmware, and Operating System. A firmware update does not affect the Protected Count. If the same Machine ID shows different firmware versions across elections, or if two machines at the same polling place show different versions within the same election, this should be noted as an anomaly.
• Hardware Status Information: Scanner board, motherboard, and modem status indicators. This section confirms all components are functioning. The modem status field is present even on machines without an active modem connection.
• Protected Count: A cumulative count of all ballots ever scanned by this specific machine across all elections and testing cycles it has processed. This number increases monotonically and never resets — not across elections, not after firmware updates, not after reinitialization. It is distinct from the Public Count (see below). The Protected Count is especially valuable for cross-election machine tracking: if a DS200 with Machine ID X shows Protected Count 14,582 in the 2022 General and 15,891 in the 2024 Primary, you know that machine scanned 1,309 ballots between those two snapshots across all intervening elections and tests.
• Public Count: Displayed on the DS200’s touchscreen during voting and printed in some report sections. Unlike the Protected Count, the Public Count resets to zero for each new election. It increments by one each time a ballot is cast. Poll workers verify the Public Count is at zero before entering Voting Mode. At closing, the Public Count equals the number of ballots cast on that machine during the current election.
• Machine ID (Unit Serial Number): A unique identifier in the format “0000DS” followed by a 10-digit code (e.g., 0000DS0318411525). This identifier remains constant across all elections for the same physical machine. If the same Machine ID appears on tapes from different elections, it is the same physical tabulating machine. The Voting Device Status screen on the DS200 touchscreen displays this as “Unit Serial #” without the “0000DS” prefix (e.g., “0316380143”) — both formats represent the same machine. Machine IDs are valuable for tracking equipment history, identifying recycled machines, and establishing quality baselines across elections.
• Location Information: County name, election name (e.g., “Wake County, NC 2022 General”), and other geographic identifiers as configured in the election definition.
• Precinct Code: The precinct identifier in a format determined by the jurisdiction. Precinct code formats vary widely: Wake County, North Carolina uses a two-part format (e.g., “0010 01-10”) where the first four digits count precincts alphabetically and the second part is the actual precinct code. Other jurisdictions may use simple numeric codes, alphanumeric identifiers, or entirely different conventions. Understanding your jurisdiction’s precinct coding system is essential for matching tape data to official results.
• Election Date: The date of the election being processed.

The metadata section also reflects the security chain that produced the tape. Before any reports can print, poll workers must enter the Election Security Code — a case-sensitive alphanumeric code programmed into the election definition. The machine also validates the Election Qualification Code (EQC), a cryptographic credential stored on the USB flash drive that confirms the election definition is authorized. The EQC appears on the Configuration Report but not on the results tape itself. Sources: ES&S Operator Guide, Chapters 3, 5, and 7; EVS 6.1.1.0 Poll Worker Guide, pp. 8–9, 11; Fairfax County Opening/Closing Procedures guides.

Section 2: Ballot Summary Counts

Following the metadata, the results tape provides aggregate ballot counts before the contest-by-contest breakdown.

• “Total Number Voting”: The total number of voters who used this machine — equivalently, the number of ballots scanned. This is the most important single field for quickly distinguishing a results tape from a zero tape: if this number is greater than zero, you are looking at a results tape.
• “Total Ballots”: A redundant ballot count used for verification. In a properly functioning machine, this should equal “Total Number Voting.” A discrepancy between these two fields is an anomaly worth flagging (see Section 10).
• “Ballots for Code”: The number of ballots cast using each ballot style code assigned to this machine. A ballot style code represents a specific combination of contests — voters in different districts within the same precinct may receive different ballot styles (for example, one set of voters is in School Board District 3 while another is in School Board District 4, so they get different ballots even though they vote at the same precinct). Understanding this field requires knowing how your jurisdiction assigns ballot codes, which varies widely.

Worked example — simple case (single ballot style): In a precinct where all voters receive the same ballot, the tape shows one “Ballots for Code” entry that equals Total Ballots. If 487 people voted, you’d see:

Total Number Voting: 487
Total Ballots: 487
Ballots for Code 0042: 487

Worked example — split precinct (multiple ballot styles): In a precinct that spans two school board districts, voters receive one of two ballot styles depending on their address. The tape shows two “Ballots for Code” entries that sum to Total Ballots:

Total Number Voting: 487
Total Ballots: 487
Ballots for Code 0042: 312
Ballots for Code 0043: 175

The 312 + 175 = 487 check is a useful verification. If the individual code counts don’t sum to Total Ballots, something is wrong.

Worked example — Wake County, NC (complex coding): Wake County uses 127 ballot codes organized into four blocks of codes, one for each voting method: ABSEN (absentee), OS (one-stop/early voting), PROVI (provisional), and TRANS (transfer/transitional). On an election day precinct tape, you’d typically see a single ballot code from the election day block. But at an early voting site that served multiple precincts, the tape may show dozens of “Ballots for Code” entries — one per ballot style offered at that site — reflecting the many precinct/district combinations that voters at that site could belong to. Each code maps to a specific combination of contests, and the codes themselves (e.g., “0010 01-10”) encode area and sub-area identifiers within the county’s precinct numbering system.

The key point: if you see many “Ballots for Code” entries on a tape, it doesn’t mean something is wrong. It means the machine served voters with multiple ballot styles, which is normal at multi-precinct polling places and early voting sites.

Section 3: Contest-by-Contest Vote Summary

The core of the results tape. For each contest on the ballot, the report lists:

• Contest Code: An internal identifier (e.g., “1378_US SENATE”). This is a machine-generated reference code assigned by the election definition software and is not typically used in public reporting. It can be useful for cross-referencing when comparing multiple tapes or matching to election definition documentation.
• Candidate Vote Totals: Each candidate is listed with their party prefix and vote count (e.g., “DEM-Cheri Beasley: 476”, “REP-Ted Budd: 286”). These candidate-level vote totals are the headline numbers that get reported to the county and then aggregated into official statewide results. They are what an Actual Vote analysis directly compares to official results.
• “Write In” Count: The number of write-in votes recorded for that contest. (Write-in handling is discussed further in Section 4.)
• “Over Votes” Count (Detailed format only): The number of ballots in which voters selected more candidates than allowed in that contest. Overvotes are typically not counted as valid votes for any candidate in that contest. How the DS200 handles overvoted ballots depends on the election definition’s exception settings: the machine can be configured to Always Accept overvoted ballots (tabulating the ballot with the overvote silently recorded), Always Reject them (returning the ballot to the voter), or Query the voter (displaying a screen asking whether to cast the ballot as-is or correct it).
• “Under Votes” Count (Detailed format only): The number of ballots in which voters selected fewer candidates than allowed in that contest, including ballots where the voter left the contest entirely blank. High undervote counts in a major race can indicate voter confusion, a ballot design issue, or simply a legitimate choice to skip the contest. In minor or nonpartisan races, high undervote counts are normal and expected.
• “Total Votes” for the Contest: A sum of all votes in that contest (candidate votes plus write-ins). The arithmetic relationship between individual components and the contest total is a valuable verification check during transcription: the sum of all candidate vote totals plus write-ins should equal Total Votes. If it doesn’t, either a transcription error occurred or the tape itself has an anomaly worth investigating.

Note that in Public format tapes, the Over Votes and Under Votes fields are omitted entirely for each contest. You will see only candidate names, their vote totals, write-in count, and contest total. In Detailed format tapes, all fields are present. The Configuration Report (from the opening sequence) indicates which format the machine was set to use. Some jurisdictions configure machines to combine overvote and undervote counts into a single field — the Configuration Report’s “Combine Over and Under Votes” setting controls this behavior. Sources: ES&S Operator Guide, Chapter 7; EVS 6.1.1.0 Poll Worker Guide, pp. 10–11.

Section 4: Write-In Handling

When a voter casts a ballot with a write-in entry, the DS200 performs several actions that affect both the printed tape and the physical ballot:

• The write-in vote is counted and included in the “Write In” total for that contest on the results tape.
• The DS200 physically stamps a pink circle on the ballot (on the trailing edge, face-down side) to mark it as containing a write-in entry. This allows election workers to physically sort write-in ballots from regular ballots during closing procedures.
• If ballot image storage is enabled, the DS200 captures a digital image of the write-in ballot and stores it on the USB flash drive with a randomized filename to protect voter anonymity.
• If configured, the DS200 prints a Write-In Images Report at closing, containing digitized images of all write-in ballot selections captured during the voting period.

At the central count or accumulation station, election officials review the Write-In Images Report (either from the printed tape or from USB drive data on a computer screen) and compare handwritten entries against the certified List of Declared Write-In Candidates maintained by the Secretary of State’s office. This adjudication process determines which write-in votes are counted as valid.

Write-in handling is a common source of small discrepancies in Actual Vote analysis. The poll tape records the raw write-in count (all write-in entries), while official results may reflect adjudicated write-in totals (only write-ins matching declared candidates). Understanding this distinction is important when investigating such mismatches.

Ballot image storage is configurable in the election definition with three options: save all ballot images, save only write-in ballot images, or save no images. The Configuration Report’s “Save Ballot Images” field indicates the active setting.

Sources: ES&S Operator Guide, Chapter 5; Fairfax County Closing Procedures guide; EVS 6.1.1.0 Poll Worker Guide, p. 10.

Section 5: Certification and Signatures

The certification section appears at the very end of the tape and is a critical administrative and chain-of-custody component of the election record.

• Certification Statement: A formal declaration, typically reading something to the effect of: “WE, THE UNDERSIGNED ELECTION OFFICIALS, DO HEREBY CERTIFY THAT THE ABOVE RESULTS ARE A TRUE AND ACCURATE ACCOUNT OF THE ELECTION…” This statement affirms that the recorded totals reflect actual votes cast on the machine during the election period.
• Blank Signature Lines: Space for poll worker names and signatures. The number of lines varies by jurisdiction and election (typically 2–5 workers). All signature lines should be completed by authorized election workers before the tape is secured. When multiple copies of the results tape are printed (typically 3), each copy must be individually inspected and signed.
• Date and Time: Some tapes include the date and time the tape was signed off, providing a record of when closing procedures were completed.

Signature compliance is tracked as part of the Actual Vote analysis. Proper signatures on all designated lines indicate the tape was produced as part of standard, authorized closing procedures. Missing or unsigned lines may indicate procedural compliance issues and should be flagged during transcription, even though missing signatures do not directly affect the vote totals being compared. The pattern of unsigned tapes matters as much as any individual instance — see Section 10 for a discussion of large-scale signature compliance failures.

Multiple copies and their disposition: The DS200 typically prints three copies of the Voting Results Report at closing (the exact number is configurable in the election definition via the “Number of Result Tapes to Print” setting). Each copy serves a different purpose: one is attached to the Statement of Results (SOR) for official chain of custody, one or more may be publicly posted at the polling place (depending on state law), and additional copies are retained for the Chief Judge or election office records. Sources: ES&S Operator Guide, Chapter 5; Fairfax County Opening/Closing Procedures guides.

3.5 — Zero Tapes vs. Results Tapes

A critical distinction that everyone working with poll tapes must understand.

• Results Tapes: Contain actual vote totals and are the basis of an Actual Vote analysis. Results tapes are transcribable and are what you will primarily work with.
• Zero Tapes: Confirm that the machine started at zero before voting began. Zero tapes verify proper initialization but contain no vote data for comparison with official results. Zero tapes are not transcribable for vote comparison purposes.

Both the zero tape and the results tape are printed by the same machine on the same type of thermal paper, using the same contest layout and the same formatting. They look similar at first glance. However, the critical difference is unmistakable: all numbers in the zero tape are zero, while the results tape contains the actual vote counts.

Quick identification: Check the “Total Number Voting” field. If it shows zero, you are looking at a zero tape. If it shows any number greater than zero, you are looking at a results tape.

The Opening Tape Bundle

In practice, the zero tape is not printed in isolation. When polls open, the DS200 prints a continuous “Opening Tape” containing three reports on the same strip of paper:

1. Configuration Report: Machine hardware status, firmware, election settings, and all configurable parameters.
2. Ballot Status Accounting Report: Confirms zero ballot counts across all ballot style codes.
3. Zero Totals Report: Every contest listed at zero votes.

Poll workers sign this entire bundle as a single artifact. Additional copies of the Zero Totals Report can be printed from the Report Options screen, but only before the machine enters Voting Mode — once voting begins, zero tapes can no longer be generated.

If any value in the Zero Totals Report is non-zero, the machine must be reinitialized and a new zero tape printed before voting can begin. A non-zero initial tape is a serious red flag that may indicate the machine was not properly reset after a previous election or testing cycle.

When recording poll tapes with the Actual Vote app, focus on the results tape. Zero tapes are welcome to record and may provide useful context about proper machine initialization, but they cannot be used in the comparison analysis because there are no votes to compare. The Actual Vote system will filter out zero tapes automatically if you attempt to submit them.

Sources: EVS 6.1.1.0 Poll Worker Guide, p. 9; Fairfax County Opening Procedures guide.

3.6 — Precinct Results Tapes vs. Central Count Tapes

Not all results tapes are created equal. The distinction between precinct-level and central-count output is foundational to the Actual Vote methodology.

Precinct Results Tapes

Precinct results tapes are produced by precinct-level scanners (such as the DS200) deployed at individual polling places. They capture votes from a specific location for a specific voting method. These are the ground-truth documents for Actual Vote analysis.

The official results hierarchy reinforces their importance: if there is a discrepancy between precinct-level tapes and centrally accumulated results, the precinct tape constitutes the official record. This is the legal and operational basis for Actual Vote’s comparison methodology — precinct tapes are the authoritative source against which official reported results can be checked.

Central Count Tapes

Central count tabulators — such as the ES&S DS450 (75 ballots/minute) and DS850 (300 ballots/minute) — process batches of ballots from multiple precincts at a central location, typically after election day. They produce accumulation tapes that combine results from many precincts. These tapes are generated at election central after USB drives from precinct scanners are collected, or after physical ballots are transported for batch processing.

Central count tapes are not precinct-level documents. They generally cannot be used for precinct-level Actual Vote analysis, though they may be useful for county-level verification in some circumstances.

Central count tabulators also differ physically from the DS200: the DS450 and DS850 use commercial laser and dot-matrix printers for report output rather than the DS200’s integrated thermal printer, and they include features like sortable output bins (for routing overvoted, blank, or write-in ballots to separate bins) and uninterruptible power supplies.

Why This Matters

Actual Vote’s methodology is specifically designed to detect discrepancies that occur during the aggregation step — when precinct-level numbers are combined into official county and statewide totals. This is the step that happens between the precinct tape and the centrally reported result. By comparing the two endpoints (precinct tape vs. official result), Actual Vote can identify whether the aggregation was done correctly.

Sources: Verified Voting: ES&S DS200; EAC certified voting systems; Texas Secretary of State EVS voting-system examination report; ES&S system architecture overview (WI_EVS5000_OVR00).

3.7 — Thermal Paper

Poll tapes are printed on thermal paper — the same technology used in cash register receipts, parking garage tickets, and shipping labels. Understanding the properties and limitations of thermal paper is directly relevant to Actual Vote’s mission, because it determines how long a poll tape remains legible and why early recording matters.

How Thermal Paper Works

Thermal paper has a chemically treated coating on one side that darkens when heat is applied. The DS200’s built-in thermal printer uses a printhead with tiny heating elements that selectively heat the paper to form text and graphics. No ink, toner, or ribbon is required — the image is formed by a chemical reaction in the paper itself.

The treated side of the paper can be identified by scratching it with a fingernail: a gray streak will appear. When loading a paper roll into the DS200’s printer, the treated side must face the right side of the printer opening. If the roll is loaded backwards, the printer operates normally but produces blank output — a known troubleshooting issue documented in ES&S materials as “reports are generating but no printing is visible on the paper.”

Longevity and Fading

Thermal paper images are inherently impermanent. Under standard storage conditions, thermal printing remains legible for approximately 5 to 7 years. Premium archival-grade thermal papers can last 10 or more years. Under poor conditions, fading can begin within weeks.

The primary accelerants of fading are:

• UV light: The single biggest threat. Tapes stored in rooms with windows or fluorescent lighting degrade faster than those in darkness.
• High temperature: Heat causes the chemically treated coating to darken uniformly, reducing contrast between printed and unprinted areas. Tapes stored in hot warehouses or vehicles during summer months are at particular risk.
• Moisture and humidity: Fluctuations in humidity accelerate chemical degradation. Optimal storage humidity is between 30% and 50%.
• Chemical contact: Contact with plastics (especially PVC), adhesives, solvents, hand sanitizer, and some inks can cause premature fading or staining.

Optimal long-term storage requires a cool, dark, dry, temperature-controlled environment — conditions that election offices may or may not maintain for their archived tapes.

Implications for Actual Vote

Every day between printing and recording, a poll tape degrades. Election-night recording captures tapes at peak legibility — they are freshly printed, the contrast is sharp, and no environmental degradation has occurred. Tapes obtained months later via appointment at a Board of Elections office or through a public records request may have significant fading, especially if they were stored in non-climate-controlled facilities.

This is a concrete, practical argument for recording poll tapes as early as possible. It also means that faint or hard-to-read digits on an older tape should not be assumed to be zero — they may be a 1, 6, or 8 that has partially faded. When working with older tapes, increasing screen brightness or using photo editing tools to enhance contrast can sometimes recover lost detail.

Sources: EVS 6.1.1.0 Poll Worker Guide, pp. 8, 12; thermal paper longevity research (Telepaper, Sure Paper, CNSunrise industry documentation).

3.8 — Physical Security and the Poll Tape

A poll tape does not exist in isolation — it is the final output of a chain of physical security procedures designed to protect the integrity of the voting process. Understanding this chain helps contextualize what you see on the tape and why certain fields and procedures matter.

Tamper-Evident Seals

The DS200 is secured with multiple layers of tamper-evident seals that are verified against a Machine Certification Form throughout the election day lifecycle:

Before polls open:

• The Ballot Compartment is sealed with a barcoded zip-tie seal. Poll workers verify the compartments are empty, then reseal with a new zip-tie seal.
• The DS200 lid is secured with a green seal whose number must match the Machine Certification Form. The seal is cut and placed inside the designated pouch.
• The DS200 screen is locked with a red barcoded sticker seal whose number must match the Machine Certification Form.
• The front access compartment (containing the printer) is locked with a barrel key after the opening tape is printed.

After polls close:

• The red barcoded sticker seal is removed and affixed to the Machine Certification Form.
• Poll workers record the Public Count and Protected Count from the DS200 screen onto the Machine Certification Form.
• Each copy of the Voting Results Report is inspected and signed by two officers.
• Memory sticks are removed from both the Battery Compartment (behind the screen) and the Printer Compartment (in front of the screen) and given to the Chief Judge.
• The black machine lid is re-secured with a new red zip-tie seal and the seal number is recorded on the Machine Certification Form.
• The Ballot Compartment is opened and voted ballots are transferred to sealed containers with multiple officers present.

The signature section on the results tape (Section 4, Section 5) is the final link in this chain. When Actual Vote analyzes a tape, the integrity of that tape depends on these procedures having been followed correctly.

Sources: Fairfax County Opening/Closing Procedures guides (Office of Elections, Rev. 05/21).

USB Flash Drive Security

All election results are stored digitally on USB flash drives in addition to being printed on the poll tape. The DS200 uses two USB flash drives:

• Primary (Election Media): Located in the rear access compartment under a locked panel with a provisional wire seal. This drive stores the election definition, all ballot images (if configured), audit logs, and election results. It is updated each time the DS200 scans a ballot. It must be FAT32 formatted, with a capacity of 1 to 8 GB. The DS200 requires ES&S-approved media (specifically Delkin brand USB drives) and will reject non-approved USB media — a vendor-specific security requirement that enforces supply chain control.
• Backup Memory Device: Located in an auxiliary USB port under the rear locking cover. When polls close, this drive receives a complete copy of all primary media files, including an encrypted machine-specific key.

At closing, poll workers remove both memory sticks and deliver them to the Chief Judge for transport to election headquarters. At election central, the USB drive data is read by the Election Reporting Manager (ERM) software, which consolidates results from all DS200 units across the jurisdiction.

The poll tape is the human-readable verification of what these USB drives contain in machine-readable form. When Actual Vote compares poll tape totals to official results, any discrepancy could theoretically originate at either the printing stage (an error on the tape) or the digital transfer stage (an error in the USB-to-ERM pipeline). The existence of both records — paper and digital — is what makes independent verification possible.

Sources: ES&S Operator Guide, Chapter 3; DS200 Security Bulletin; EVS 6.1.1.0 Poll Worker Guide, p. 11.

Logic and Accuracy Testing

Before every election, each DS200 undergoes Logic and Accuracy (L&A) testing to verify that the machine is correctly reading ballots and producing accurate results. The L&A test process is standardized by the EAC and involves:

• Creating a test deck of ballots marked with predetermined vote patterns, including single votes, overvotes, undervotes, blank ballots, and write-in entries.
• Running the test deck through the DS200.
• Printing a Zero Totals Report before the test and a Results Report after.
• Comparing the printed results against the expected totals from the known test deck.
• Verifying that overvote counts, undervote counts, write-in handling, and blank ballot handling all match expectations.
• Documenting the test results, which are signed and certified by election officials.

L&A testing generates its own set of poll tapes. These are not election tapes and will not be encountered during Actual Vote analysis, but they are part of the security and verification framework that produces the election tapes you do analyze.

Sources: EAC Logic and Accuracy Testing documentation (2016); ES&S Operator Guide, Chapter 4.

3.9 — Reading and Transcribing Poll Tapes

Tips for Clear Recordings

Creating a high-quality recording or photo of a poll tape is essential for successful transcription. A blurry or poorly lit tape is difficult or impossible to read accurately.

• Record from top to bottom in one slow, steady pass: A complete video of the entire tape, recorded in a single continuous shot from top to bottom, is ideal. This preserves the entire structure and flow of the tape.
• Pause briefly on key areas: As you record, pause briefly on critical sections to allow for focused review — the header section (Machine ID, precinct code, election name) for identification, each contest block (to clearly see all candidates and vote totals), and the signature section (to verify compliance).
• Avoid panoramic mode on phone cameras: Panoramic mode creates blurry, distorted images that are unusable for transcription. This has been a documented issue in Actual Vote field work — volunteers using panoramic mode have produced recordings that cannot be transcribed. Use normal video mode, which produces clear, standard-aspect-ratio footage.
• If using photos instead of video: Take multiple clear, overlapping photos of the entire tape. Ensure each photo captures a section sharply. Overlap your photos so the full tape can be reconstructed from the sequence.
• Good lighting is essential: Thermal paper benefits from strong, direct lighting. If recording on election night and the tape is taped to a wall or window, use your phone’s flashlight or flash to illuminate the tape. Avoid shadows that obscure text. Test your recording on a short section before committing to the full tape.
• Stable camera position: Use both hands to hold the camera or prop it on a stable surface. Hand tremor and camera shake produce blurry text. A small tripod or phone holder is valuable for long tapes.

Transcription Process

Transcribing a poll tape into structured data requires care, accuracy, and an organized system:

1. Set up a transcription spreadsheet with clearly labeled columns: County, Precinct (or Precinct Code), Voting Method (e.g., “Election Day”, “Early Voting”), Contest (contest name or code), Choice/Candidate (candidate name and party), Poll Tape Vote Total (the actual number from the tape), Under Votes (if transcribing by contest, and if available on the tape), Over Votes (if available), and Write-Ins (if applicable).

2. For each contest, transcribe every value: Every candidate’s vote total, write-in count, over-vote count (if present), under-vote count (if present), and total votes for the contest (as a verification check).

3. Record metadata in a separate sheet or header section: Machine ID, Protected Count, Public Count, timestamp of the results report, firmware version, poll workers’ names (if visible), and any notes about missing signatures or anomalies.

4. Double-check any value that seems unusual: A very high number of undervotes in a major race, a zero in a race where you’d expect significant votes, candidate vote totals that don’t sum to the contest total, or any value that looks like it could be misread.

5. If a value is blurry or ambiguous: Note it in your transcription as uncertain (e.g., “6/8?” to indicate the value is unclear). Use context to resolve it — the sum of individual candidates plus write-ins should equal the “Total Votes” field. If you’re unsure about a single candidate’s number, you can sometimes deduce it from the other known totals. Flag uncertain values for a second review of the video, which often reveals the correct digit with fresh eyes.

6. Verify arithmetic relationships: Total Ballots should equal Total Number Voting. For each contest, the sum of candidate votes plus write-ins should equal Total Votes (or close to it, depending on how overvotes are counted). Total Votes across all contests should be reasonable given the Total Number Voting.

Common Transcription Challenges

• Blurry recordings: Thermal paper, camera focus limitations, and distance from the tape can make individual digits hard to read. The confusion between 3/8, 5/6, and 1/7 is common on thermal printouts. Use context (the total vote sum for a contest) to disambiguate. If a value remains truly ambiguous after review, note it as uncertain and flag it for follow-up.
• Faded tapes: Tapes obtained weeks or months after an election may have significant fading (see Section 7). Increasing screen brightness or using photo editing tools to enhance contrast can sometimes recover lost detail. Avoid assuming a faint digit is zero — it may be a partially faded 1, 6, or 8.
• Long tapes: A full election day tape with many down-ballot races can be several feet long. Budget sufficient time for transcription — rushing leads to errors. Break the work into manageable sessions.
• Stitching together partial recordings: If a video pauses or restarts mid-tape, you may have discontinuous footage. Construct a complete picture by cross-referencing screenshots from different parts of the recording and matching contest boundaries and vote totals.
• Paper roll changes mid-report: If the thermal paper ran out while the results tape was printing, the DS200 pauses the print job while the poll worker replaces the roll. After replacement, the DS200 reprints the last 18 lines from the old roll onto the new roll. When reviewing a tape, you may see what appears to be duplicated content at a splice point. This is normal DS200 behavior, not an error or anomaly. To reconstruct the complete report, find the last entry on the old roll and use timestamps to locate the repeated entries on the new roll.
• Tape damage or ink smudging: Tapes may be creased, folded, or smudged from handling during closing procedures or storage. Do your best to read through minor damage. If a section of tape is genuinely illegible, note it in your transcription rather than guessing.

Sources: ES&S Operator Guide, Chapter 7 (paper roll change procedures); Actual Vote 2024 NC Primary report (panoramic mode issues); Actual Vote 2022 Wake County report (transcription methodology).

3.10 — Poll Tape Anomalies to Watch For

During transcription, you may notice irregularities worth flagging even if they do not directly affect the vote total comparison. These observations contribute to the overall election quality assessment and may identify systemic issues worth investigating.

AM/PM Clock Errors

Tabulator clocks may be set to the wrong 12-hour period. For example, a machine initialized at 6:00 PM instead of 6:00 AM, or showing a closing time of 7:00 AM instead of 7:00 PM. This is not outcome-relevant — the vote totals themselves are unaffected — but it indicates a procedural issue and should be reported to the county Board of Elections.

Real-world example: During the 2024 North Carolina Primary, Actual Vote analysis found that two Henderson County precincts (Laurel Park and Flat Rock) showed tapes with a 6:00 PM opening time and a 7:00 AM next-day closing time — a physical impossibility indicating the machine clocks had their AM/PM setting reversed at initialization. Henderson County Board of Elections confirmed the error and pledged to implement corrective procedures for future elections.

More generally, any timestamp significantly outside the expected window (6:00 AM opening, 7:00–8:00 PM closing) warrants investigation. A results tape timestamped at 3:00 AM almost certainly reflects a clock error, not overnight voting.

Source: Actual Vote 2024 NC Primary report, Henderson County AM/PM anomaly finding.

Firmware Version Discrepancies

If the same Machine ID (same physical machine) shows a different firmware version across elections, note it. Firmware updates can affect how votes are counted or reported, though in practice this is rare.

More significant: if two DS200s at the same polling place (e.g., an early voting site with multiple machines) show different firmware versions within the same election, this should be flagged and reported. All machines in a jurisdiction should be running identical firmware for a given election. Discrepancies suggest inconsistent pre-election preparation.

Missing or Incomplete Signatures

The certification section should have poll worker signatures on all designated lines. Missing signatures, illegible names, or unsigned lines do not directly affect vote totals, but they may indicate procedural compliance issues. Flag these during transcription and note them in your metadata. Missing signatures might warrant a follow-up query with the county if the discrepancy is systemic (appearing across multiple precincts rather than just one).

Real-world example: Following the 2020 election, approximately 130 unsigned poll tapes were identified in Fulton County, Georgia, covering some 315,000 early votes. Georgia law requires the poll manager and two witnesses to be present for the printing, checking, and signing of each tape. This scale of unsigned tapes indicated a systemic procedural failure rather than isolated oversights, and became a significant point of post-election scrutiny. When Actual Vote encounters unsigned tapes, the pattern matters as much as the individual instance.

Ballot Count Mismatches

Verify that “Total Number Voting” equals “Total Ballots,” and that “Ballots for Code” is consistent with both. If these values do not align, a discrepancy exists that could indicate a machine malfunction, a data entry error in the election definition, or a legitimate difference in how the counts are defined for multi-precinct polling places. Flag and investigate.

Unexpected Precinct Information

If the precinct code on the tape does not match what you expected based on the voting location or your source documents, double-check before transcribing. Precinct code mismatches can indicate mislabeled tapes, a machine set up for a different precinct than it was physically located in (a serious error), or a data entry error in the election definition. Confirm the precinct code against your source documentation before proceeding.

Unusual Vote Distributions

While individual unusual values are common and not necessarily anomalous, systematic patterns can be informative: an unusually high concentration of votes for one candidate across multiple precincts, precinct-level results that don’t align with historical voting patterns (though demographic and campaign changes can explain differences), or a pattern of consistent write-in votes for the same candidate (which might indicate an organized write-in campaign or a data entry issue). These patterns may warrant investigation, particularly if they appear consistently across multiple tapes.

Contests Missing from the Results Tape

If a contest appears on the zero tape but is missing from the results tape, or vice versa, flag it. This could indicate a ballot configuration issue, a machine malfunction, or a printing error. The most likely explanation is that the machine’s ballot configuration was updated between the zero and results tapes, but this is unusual and should be verified with the county.

Configuration Report Mismatches

If you have access to the opening Configuration Report (from the Opening Tape bundle), check whether its settings match what you’d expect for the jurisdiction. For example: if the Configuration Report shows “Save Ballot Images: None” but the jurisdiction’s policy requires ballot image retention, or if “Blank Ballots: Accept” is set when jurisdiction policy calls for querying voters about blank ballots, this may indicate a misconfigured election definition. Configuration mismatches don’t directly affect the results tape totals but are worth documenting as election quality observations.

Paper Roll Change Artifacts

As noted in Section 9, when thermal paper runs out mid-report, the DS200 reprints the last 18 lines on the new roll. This can look like an anomaly — duplicated content at a splice point — but is normal machine behavior. If you see this pattern, it’s an artifact of a paper change, not an error in the vote totals.

Precinct-to-Central Count Vote Loss

The most consequential type of anomaly from the Actual Vote perspective: votes correctly recorded on precinct-level tapes that are absent or different in centrally accumulated official results. In Memphis, Tennessee in 2015, a central processing system lost approximately 1,001 votes that appeared on precinct-level poll tapes — roughly 40% of votes in affected precincts disappeared during aggregation. This is the exact scenario Actual Vote is designed to detect.

Sources: Actual Vote 2024 NC Primary report; Actual Vote 2022 Wake County report; news reporting on Fulton County and Memphis incidents; ES&S Operator Guide, Chapter 7.

3.11 — Poll Tape Access and Legal Framework

Poll tapes are public records, but the ease of accessing them varies dramatically by state and county. This section provides a high-level overview; for state-specific guidance, see Section 7 (Access Methods) and the state-by-state posting appendix (B.3).

State Posting Requirements

Some states have statutory requirements for public posting of poll tapes at polling places on election night. Others leave it to county discretion, meaning posting practices can vary within a state. Among the states with posting requirements or established posting practices are Florida, Georgia, North Carolina, Ohio, and Pennsylvania. California requires poll tapes to be available for public inspection for 48 hours following an election.

The proposed Poll Tape Transparency Act of 2010 (H.R. 5228) would have standardized these requirements nationally, calling for oral announcement of results at poll closing, public posting in prominent locations, display during regular business hours, and publication on official websites. The bill was not enacted, and no comparable federal standardization has been adopted since.

Methods of Obtaining Poll Tapes

There are five primary methods for obtaining poll tapes, listed here in order of chain-of-custody strength:

1. Poll observer/watcher recording on election night: The strongest chain of custody — the tape is recorded as it is printed or immediately after posting, before any handling, transport, or storage. Requires legal poll watcher status in many jurisdictions.
2. Public posting at polling place: Recording a tape posted on a wall or window at the polling place on election night. Strong chain of custody — recorded within hours of printing.
3. Elections office appointment: Visiting the county Board of Elections office to view and record tapes. Any state. Moderate chain of custody — tapes have been handled by election staff and stored between printing and your visit.
4. Public records / FOIA request: Requesting copies through the state’s public records or Freedom of Information Act process. Any state. Moderate chain of custody and potentially weeks of processing delay, during which tapes may degrade.
5. Central tabulation observation: Observing the central count process at jurisdictions with centralized tabulation. Availability and procedures vary widely.

Why Early Access Matters

As discussed in Section 7, thermal paper degrades over time. Election-night recording captures poll tapes at peak legibility. Every day of delay introduces fading, especially for tapes stored in conditions that are not climate-controlled. Early access also means stronger chain of custody — fewer hands have touched the tape, and less time has passed for potential alterations or damage.

This is one of the practical reasons Actual Vote emphasizes election-night recording and pre-certification analysis windows where possible: the evidence is freshest and most reliable at the moment closest to printing.

Sources: America Counts, “How to Research Poll Tape Posting Policy”; H.R. 5228, Poll Tape Transparency Act of 2010 (111th Congress); state election codes for FL, GA, NC, OH, PA, CA; FOIA.gov.

3.12 — Conclusion

Poll tapes are foundational documents in election administration and the starting point for Actual Vote analysis. They are also, for many people, mysterious — long strips of thermal paper covered in unfamiliar codes, abbreviations, and numbers. This chapter has aimed to make them less mysterious by documenting everything we know about how they are produced, what they contain, how they are secured, and how they can be read.

The most important principles:

Poll tapes are ground-truth documents. The vote totals printed on a precinct results tape are the authoritative record — if there is a discrepancy between a precinct tape and centrally reported results, the precinct tape is the official source.

Poll tapes are impermanent. Thermal paper fades. Record early, record clearly, and record completely.

Poll tapes are the product of a security chain. The seals, signatures, codes, and procedures described in this chapter are what give poll tapes their evidentiary weight. Understanding the chain helps you assess the reliability of the tape you’re analyzing.

Your work directly supports the mission of ensuring that reported election results match the votes actually cast. Treat every tape with care, transcribe it accurately, and flag anything unusual. The more we collectively know about how these documents work, the better positioned we are to serve as effective, credible, independent auditors of our elections.