EPEC NEWS

Analysis by Rick Naigle, Senior Analyst, EPEC Voter News:

EPEC Team data analysts reviewed voting data from the June 18th Primary election and found serious discrepancies. There is still time to fix the problems before November.

Virginia’s 2020 election experience during the COVID-driven shift to mail-in ballots was marred by allegations of mail-in ballot fraud, suspicious activity with defective thumb drives — and statistical oddities in the results, to name a few examples.

Back then, Virginia’s count controversies were overshadowed by controversies over count errors in the six battleground states. It will be different this time.

Virginia’s election officials can anticipate closer scrutiny in 2024.

EPEC Team data analysts reviewed voting data from the June 18th Primary election (including seven weeks of early voting from the Republican and Democrat primaries). We found serious discrepancies.

Our view: The issues from the June 18th Primary (including voting from May 3 – June 15th) must be addressed prior to the 2024 General Election to promote transparency and trust in the results. Some of our initial findings follow.

1. Voters Who Appear to Have Cast More Than One Ballot

According to Virginia’s Dept. of Elections early voting data from the June 2024 primaries (combining Republican and Democrat results), 454 voters appear to have cast more than one ballot.

OUR METHODOLOGY:

Of 298,597 unique voter names that cast ballots in the 2024 Virginia primary, the state’s own voting data show that 196,592 were tabulated.

The successful ballots reflect the sum of ballots with the following statuses:

—FWAB (federal write-in absentee)

—Marked (registrar has received voter’s mail-in absentee ballot)

—Pre-Processed (ready for pre-tabulation review).

• Using a filter to compare voters with a FULL NAME match, we found 454 voters with multiple ballots counted.
• This indicates voters who were registered more than once with variations of the same name. (Do they have the same social security number? Dept. of Elections officials would be wise to run a cross-check.)

• Using a filter for FULL NAME match, we found two (2) voters who may have submitted as many as five (5) ballots during June Primary early voting.
• Another three (3) voters may have submitted four (4) ballots each.

  (See summary data in the table below:)

Using a filter by FULL NAME, the voting records show as many as 24 voters may have submitted three (3) ballots during June Primary early voting. See summary data in the table below:

RECOMMENDATIONS

• Dept. of Elections must identify and remove duplicate voter registrations from the Registered Voter List prior to the 2024 Presidential Election.
• Dept. of Elections Watchdog organizations need access to unofficial election day voting data (similar to the Daily Absentee List it already provides to qualified groups) prior to certification.
• This will ensure independent verification of election results during canvassing and before certification.

2. Low Success Rates For Mail-In Absentee Ballots

• Out of the unofficial voting data from the primary, 197,077 voters with unique IDs showed one or more countable ballots in the Daily Absentee List (DAL).
• Five (5) appeared to have voted twice out of a voter population with a 52% “absentee by mail” voter success rate.

Some factors to consider in the lower success rate for tabulating mail-in ballots:

• Registered Voter List errors (deceased, moved out of state, incorrect mailing address, etc.)
• Voter indifference (historically low turnout for both parties’ primaries)
• Insufficient turn-around time for mailed ballots
• Voter error when filling out / returning ballots
• Processing errors

3. Other Indicators of Voting Irregularities

• We found time stamps in the Daily Absentee List (DAL) and voting data showing In-Person voting occurring when polls were closed, according to time stamps.
• We identified ballots rejected as late during early voting periods.

4. Irregularities Observed in the 2024 June Primary Daily Absentee List

• EPEC Team analysts identified records showing In-Person absentee voting (“On Machine”) before May 3rd, which was the start of early voting in the primary.

  See the summary data in the table below:

• Why does this matter?

  The appearance of votes being cast before the official start of voting creates the appearance of fraud. All ballots should be received and processed as allowable by law during the voting period specified for the election.

• We recommend Electoral Boards and the Commissioner of Elections take a look.

  See summary data table from Charles City County below:

• Irregularities like this contribute to a lack of voter confidence in election results — and raises questions whether the technology is secure and being used properly.
• The Virginia Commissioner of Elections and the Charles City County Board of Elections should determine how these entries occurred in the DAL.
• All ballots should be received and processed as allowable by law during the voting period specified for the election. In-Person early voting can only be conducted when polling places are open, and only during the hours when polling is allowed.
• The DAL record also shows in-person voting on a Sunday. Were the polls open or was this a record from the day before?

• Polling places are typically not open on Sundays.
• The date reflected above, from the Dept. of Elections data, indicate Goochland County was the only Locality in Virginia with June Primary early voting occurring in-person on a Sunday.
• The Commissioner of Elections and Goochland County Board of Elections should verify if this is accurate data.
• We also observed countable ballots without a Ballot Receipt Date.

• Why does it matter?

  A ballot must be received before it can be tabulated. Ballots with processing errors are suspect.

• This can be easily addressed with a software update, which would force an election official to fill the Ballot Receipt Date field before the record is complete.

  (Required fields are common in data-heavy environments for quality control.)

• Every ballot status in the process should be time-stamped: On Machine (in person), Provisional, FWAB, Marked, Pre-Processed, or Unmarked. None of these improvements would infringe on the voter’s anonymous ballot marks.
• We observed a large number of ballots in the 2020 General Election DAL without Ballot Receipt Date entries. This can impact public confidence in the outcome of the election.
• See summary data below from Election Day In Person Voting (via the DAL):

• The Dept. of Elections does not provide access to unofficial Election Day voting (like it does with the DAL) until after the election has been certified.
• No one has explained why this is necessary or what informs the policy.
• Voter participation organizations and civic groups must be able to access Election Day data in the same format, at the same level of detail, as is provided via the Daily Absentee List.
• In the summary data view below, we see by-mail ballots received and processed prior to, and after, the early voting period, which began May 3, 2024:

• Why does it matter?

  We do not know with certainty that anything is wrong with these DAL entries, but there is an indication that some localities are prematurely pre-processing ballots. We would not expect the Ballot Status to be Pre-Processed prior to the start of the early voting period.

Because of incomplete data in the file, we do not know with certainty when the ballot was Pre-Processed and moved closer to the tabulation stage.

RECOMMENDATIONS:

• Marked would be a more appropriate ballot status for the example above.
• The database must capture data and time of each process step in the record, which aligns with industry best practices. This enables performance measurement of the electoral process.
• Enable Date/Time capture of ballot status changes — not just the date when the ballot was received.
• See summary data below showing ballots received after Election Day:

• Why does it matter?

   

  By Mail ballots which have been postmarked by 6 PM on Election Day can be processed if received in the registrars’ office by or before noon on the third day after the election (or first workday of the next week if the Friday is a holiday).

   

  The record below shows a by-mail ballot that was counted after the deadline. Without additional information (postmark Date / Time, and Ballot Receipt Date), we cannot ascertain whether ballots are being counted after the statutory deadline. See summary data below:

SUMMARY:

EPEC Team analysts have found multiple voter registration IDs assigned to the same voter. Some voters appear to have multiple registrations assigned them within a single Locality; others appear to show multiple registrations in multiple Localities.

• Our question is this: how is the Dept. of Elections running checks on its voter lists to guard against these types of errors?
• If a Locality does not submit updates in a timely manner to the Statewide DAL, a voter with multiple IDs in multiple Localities might end up casting more than one ballot.
• Virginia’s Gov. Glenn Youngkin recently issued an Executive Order to “strengthen Virginia’s voter rolls and ensure better data accuracy.”

  The governor’s EO is aimed at preparing for the 2025 rollout of its new Voter Election Registration Information System (VERIS). His order appears to be aimed at improving voter-data quality overall.

• Voting for the 2024 Virginia presidential election starts in September. There is still time to address these findings and improve voter data quality.
• To do so would align Virginia with its commitment to modernizing voter data management as a matter of routine. It also builds transparency and trust in the results.
• We welcome your feedback on the findings. Please reach out with any questions.

  ---

  Source: Democrat Primary and Republican Primary Daily Absentee List as of 0600, 24 June 2024. #

  ---

   

  Sign up to receive the weekly EPEC Team newsletter.

  Email *

  Sign Up