The Second county in New Jersey, the poster child for FIRSTNET, found to have significant errors in election records.
Former New Jersey AG Election Investigation Fails to Report Major Red Flags in Monmouth County, New Jersey, November ’22 General Election Review.
Guest post by By Mark Demo and Erin Clements
Previously we reported on Mercer County, New Jersey, which had a catastrophic tabulator failure during the 2022 Midterms, prompting local citizens to do an in-depth audit of the election records from that county. The local team found that none of the county, state, and federal records reporting the number of ballots cast matched. The in-person voting records would have originated with the electronic pollbooks. In New Jersey, all electronic pollbooks are connected to the internet using FirstNet.
FirstNet is a nationwide cellular network under the control of private partners of the federal government. The fact that FirstNet is used for emergency services personnel was widely known, but it was only recently reported that FirstNet was also being used to connect election equipment, including electronic pollbooks and tabulators, at the precinct polling places. The Election Assistance Commission, the National Association of Secretaries of State, and the Cybersecurity and Infrastructure Security Agency were all found to have been prompting election jurisdictions to use FirstNet to transfer election data from polling places and county offices. When coupled with the Albert Sensors, which are installed on most networks that serve the Secretary of State and county election offices, the infrastructure exists to give the Federal Government and their private partners full access to the entire nationwide election system.
Now the election records in Monmouth County, New Jersey have been shown by three different teams to also be woefully inaccurate and inconsistent. When election results for the 2022 Midterms were published, a local citizen stumbled upon a discrepancy in the public records for a local mayoral race. When he brought it to the attention of the Monmouth County Clerk, she began an investigation which was ultimately referred to the New Jersey Attorney General, Matthew Platkin.
Platkin appointed former Attorney General Peter Harvey and his firm, Patterson Belknap Webb and Tyler LLP, to investigate the source of voting issues. The firm was also instructed to make recommendations for potential reforms based on their findings. A copy of the report written for the state can be found here.
As stated in the report and documented in local press (see examples here, here and here), six tabulator thumb drives were uploaded twice, causing the ballots saved on those drives to be counted twice. In one local Board of Education race, the wrong candidate was declared the winner. That error was later corrected based on hand counted results after the wrong candidate had been sworn into office.
Per the former AG’s report to the State, several factors caused the miscount in votes:
- Monmouth County’s voting equipment vendor, ES&S, failed to implement a software patch designed to prevent the tabulation of ballots more than once.
- Monmouth County personnel loaded some flash drives with ballot images into election workstations twice.
- Although an unidentified individual using a Monmouth County’s election software account ran a report that showed ballots being loaded twice on the day after the election, no one stepped forward to identify the issue before the county certified the election.
The state investigation wrapped up by stating the usual…they discovered no evidence of any fraudulent or willfully wrongful conduct by any Monmouth County election official or personnel, ES&S employee, or any other person. Nor did they observe any evidence of double counting of ballots on ES&S’s elections system before or after the November 2022 Election in Monmouth County, or elsewhere.
The software patch to prevent duplicate ballot tabulation was deployed by ES&S in November 2021 which means that every prior election, including the 2020 Presidential election, did not have the capability to detect duplicate ballots. The state’s report is unclear how thoroughly previous elections were investigated.
Not So Fast
However, research on the same election data performed by a professional election audit firm paints a completely different picture of the November 2022 general election conducted in Monmouth County. Greg Sobocinski, a candidate in the Congressional race in New Jersey District 3, hired Ray Lutz of Citizens’ Oversight to conduct a review of the November 2022 election using their AuditEngine Ballot Auditing service. In a report to the Monmouth County Clerk, Lutz detailed serious issues observed with the ES&S records which were not identified in the former AG’s report. A copy of Lutz’ letter to Monmouth County can be found here.
In his analysis, Lutz agreed with the State that some ballot images had been loaded into the system twice. However, after removing the duplicate ballots, he found other serious issues that affected at least two more races that were never mentioned by New Jersey officials.
Lutz reported more than 800 instances where the votes marked on the ballot image did not equal what was recorded on the Cast Vote Record Summary (CVR). The CVR is the electronic record of the votes that were cast on each individual ballot.
For example, the race for the Long Branch Board of Education had 747 ballots where the votes recorded in the CVR did not match the ballots. An example of the section of the ballot for this race is shown below.
BALLOT IMAGE FOR LONG BRANCH SCHOOL DISTRICT BOARD OF EDUCATION
In this race, the voter was to make three choices. The ballot image clearly shows that the voter selected Armour, Brown, and Michaels. However, the CVR recorded only two votes for this race – one for Brown, which was correct, and one for Bennett, which did not exist on the ballot. The vote for Michaels was ignored by the tabulator. At least one other race had similar issues. There was no obvious rhyme or reason to when the ES&S tabulators chose to invent or drop votes on the ballots in these races.
Even stranger, the official Monmouth County election results closely matched Lutz’ analysis of the ballot images even though there were gross inaccuracies in the digital record created by ES&S.
When questioned, county election officials referenced a printing issue which triggered a hand count where one ballot type was printed without a header, shifting the ovals away from where the tabulator was expecting them to be, causing the tabulators to misread the ballots.
Upon further investigation, the explanation does not hold water for at least two reasons. First, the ES&S tabulators are supposed to go through a logic and accuracy (L&A) test prior to the election where test ballots containing every possible ballot type are fed through the tabulator and the results are checked before any real ballots are cast. If there was a printing issue with some ballot types, this should have been caught in the L&A test. Or perhaps the L&A test wasn’t done correctly?
Whatever happened with the L&A test, a missing header would not have caused the results in the CVR.
The tabulator is programmed with a digital map for each ballot style that tells it where to look for each oval so it can detect whether that oval has been filled in. Correctly located targets would look something like this:
BALLOT WITH CORRECTLY LOCATED DIGITAL TARGETS
If a header was missing from the ballot, all the ovals would shift up vertically from where the tabulator was expecting them to be. Given the layout of this ballot, this means that the shift would have caused the tabulator to look either in white space above the ovals, causing all the races to be interpreted as undervotes, as in the following example:
VERTICAL SHIFT THAT WOULD CAUSE ALL UNDERVOTES
Or the targets could have landed on the table gridlines, which the tabulator could interpret as overvotes:
VERTICAL SHIFT THAT WOULD CAUSE ALL OVERVOTES
There is no way to vertically shift the targets to come up with the results that were in the CVR – with one vote counted correctly, one vote invented, and one vote completely missing. Those errors could not have been caused by a “printing error,” but more likely a faulty configuration within the ballot definition files created by ES&S.
Grassroots Researchers find Deeper Issues with Election System Tied to FirstNet
Moreover, analysis of the November 2022 Monmouth County general election conducted by grassroots researchers in New Jersey yields additional questions regarding the lack of internal controls that exist between the Official Certified Results, the Cast Vote Record, the electronic Poll Book, and the election results reported by the County to the State Voter History file. For example:
- 3,801 more voters were checked into the Poll Book on Early Voting and Election Day (not counting provisional voters) than there were in-person voters reported by the County to the State Voter History file.
- 419 voters were reported by Monmouth County as voting in-person during Early Voting and on Election Day (not counting provisional voters) that were never checked into the Poll Book.
These records are all tied to the ES&S pollbook software, which are connected to the internet using AT&T’s FirstNet network.
The New Jersey State Investigation Creates More Questions than Answers
Though the New Jersey AG tried to close off any additional questions surrounding the November 2022 election through their internal investigation into Monmouth County, the unaddressed issues scream out for further investigation. If New Jersey election officials and vendors want to reduce citizens’ anxiety surrounding the accuracy of their elections, they must be honest with the public about the full scope of what went wrong. And they need to demonstrate a real commitment to getting to the bottom of every issue that was discovered. The public wants to see accountability and justice, not inadequate investigations that only serve to cover up a completely broken system.