Skip Navigation

The Senate Has Voted to Eliminate Public Media Funding

The Senate has passed the Recissions Act of 2025, which would completely defund public media. The amended bill now heads back to the House for consideration.

This vote threatens federal support for WITF — putting at risk the educational programs, trusted news and emergency communications our community relies on, both locally and from PBS and NPR.

Your voice is urgently needed. Call your Representative and donate today to help protect the future of local public media.

2 Republican Pa. legislators search for double voters across Ohio state line, again

After failing to find double voters from the 2020 elections and being turned away by law enforcement, the legislators are trying with 2024 data.

  • Jordan Wilkie/WITF
State Reps. Jamie Flick, of Lycoming County, and Eric Nelson, of Westmoreland County, are looking for double voters between Ohio and Pennsylvania and plan to share their findings with law enforcement. Experts say their methods will mostly find different people with the same name and birthdate.

State Reps. Jamie Flick, of Lycoming County, and Eric Nelson, of Westmoreland County, are looking for double voters between Ohio and Pennsylvania and plan to share their findings with law enforcement. Experts say their methods will mostly find different people with the same name and birthdate.

Correction: A previous version of this story missed a communication from Rep. Eric Nelson. He did respond to written questions ahead of publication for this story. This story was updated to reflect those statements, provided in addition to previous interviews.

Two Republican state legislators are using an inaccurate method for finding potential double-voters in the 2024 general election and say they plan to refer cases to law enforcement. They are working with the Voter Reference Foundation, a national group that has a history of baseless election fraud claims. 

State Reps. Jamie Flick, of Lycoming County, and Eric Nelson, of Westmoreland County, said they have a list of 700 people whose names and birthdays match 2024 voter files in both Ohio and Pennsylvania. Flick said he did extra research using the online white pages tool People Finder to identify 50 people who had addresses in each state that matched the voter records, then sent both lists to VoteRef.

The group has purchased the National Change of Address data from the U.S. Postal Service, Flick said. The representatives are hoping VoteRef can find voters who have addresses in both states that match voter registrations and voter histories. 

They have been waiting on VoteRef for over a month to complete an additional analysis. Once done, the lawmakers say they plan to share their evidence with federal, state and local prosecutors. Double-voting is a felony, meaning those accused of the crime face several years in prison and thousands of dollars in fines. 

Flick’s and Nelson’s effort joins a suite of Republican inquiries, audits and investigations into Pennsylvania election administration. Attorney General Dave Sunday is investigating whether any crimes were committed when hundreds of poorly filled out voter registration forms were submitted in several counties just ahead of the November election. Auditor General Tim DeFoor is checking to see if automatic voter registration through the Division of Motor Vehicles introduced any errors in the state’s voting files. 

Like Sunday, who was district attorney in York County when he ran for attorney general, and DeFoor, Flick and Nelson started their inquiry prior to the November 2024 election and have so far not detailed their findings. Last August, the lawmakers began comparing voter records from the 2020 presidential election in Ohio with corresponding records from 13 counties in Pennsylvania. 

In October, Nelson sent to the Westmoreland County district attorney a list of five names of people who allegedly voted in Ohio and Pennsylvania in the same election. 

Chief County Detective Ronald Zona, who was tasked with investigating, said Nelson’s list didn’t pan out. There just happened to be five Ohio voters with the same name and birthday as the five people on Nelson’s list. One voter shared her frustration with getting a call from Zona’s team with Pittsburgh’s TribLive, saying, “They’re just wasting my time over a bogus lie.”

But Nelson brushed aside the voter’s concern. “ If you didn’t double vote, then you don’t have anything to worry about,” he said. 

Flick and Nelson say they have been nurturing relationships with law enforcement with the hope that their findings launch criminal investigations. Since the lawmakers first started digging for double voters a year ago, several new Republican district attorneys won election, Sunday took over the state Attorney General’s office, and President Donald Trump won the presidency. Trump has since instructed the Department of Justice to search for voter fraud while also ordering investigations into former officials from his first term, including one who dismissed Trump’s 2020 election conspiracy claims at the time.

The odds of coincidence 

There were 5.9 million voters in Ohio and 6.9 million voters for president in Pennsylvania in November 2020. That means there are 41 trillion possible 2-person comparisons between voters in these states alone. 

Even if the chances of two voters across state lines having the same name and birthday are pretty small, it can still happen, especially when there are so many possibilities. That’s why comparing just the names and birthdays of voters is a bad metric for finding double voters, according to Marc Meredith, University of Pennsylvania political science professor. 

That helps explain why Flick’s and Nelson’s analysis found Westmoreland County voters whose names matched voters in Ohio, but it did not actually prove these people were guilty of voting in both states. Using the last four digits of a Social Security number would provide a much more accurate match, Meredith said, but that information is not part of the public record. 

Meredith’s research showed that the majority of possible double voter cases could likely be explained by clerical error. Voter history data is largely very good, Meredith said, but errors do happen, especially when poll workers are manually checking in voters or scanning in voter histories. 

Sometimes, a voter’s birthdate is entered incorrectly into the voter registration file, or vote history is accidentally credited to someone with a similar name, or just to the person one row down in the poll book, according to Jonathan Marks, an official with Pennsylvania’s Department of State. 

“ If you’re gonna accuse someone of double voting, you have to confirm that they actually voted,” he said. “This involves looking at records such as poll books, ballot envelopes and other things that would support the allegation.” 

That would need to happen in both Ohio and Pennsylvania to prove illegal double voting occurred, a time-consuming process, he said. 

Both Flick and Nelson voted against a Democratic-backed omnibus election reform bill that would likely clean up many clerical errors by introducing electronic poll books to track voter history at polling sites, which would also make it easier to prevent double voting. The bill had several other provisions and was passed on a party line vote in May. The GOP-controlled Pennsylvania Senate has not moved to consider the bill. 

Returning to the well of unreliable data 

Nelson and Flick also included inaccurate data in their analysis of 2020 voter data, according to the Pennsylvania Department of State. 

In October, the two lawmakers provided a list of potential double voters to the department. 

“ Roughly a third of them didn’t even have vote history in Pennsylvania,” said Marks, the department’s deputy secretary of elections. “So it was confusing why they were included on the list in the first place.” 

The Department of State, which manages the statewide voter registration database, has seen efforts like Flick’s and Nelson’s before. 

Back in 2013, Kansas’ then-Secretary of State, Kris Kobach, ran the Interstate Voter Registration Crosscheck Program. Kobach compared voter rolls across state lines matching first name, last name and date of birth in an attempt to find double voters. Kansas shared the data with Pennsylvania, which in turn forwarded the findings to about a dozen counties for further investigation.

“ We heard back from the counties and they — in more colorful words that I’m going to use — they basically said the data was junk,” Marks said.

A reliable, non-partisan organization already exists that can track duplicate voter registrations across state lines. It’s called the Electronic Registration Information Center, and until 2023 both Pennsylvania and Ohio were members. But in the last three years, nine Republican-led states including Ohio bowed out of ERIC over unsubstantiated claims of partisanship within the organization. 

VoteRef is part of a network that maintains without evidence there is extensive election fraud and that advocates for states to leave ERIC, while VoteRef simultaneously tries to build a product to detect problems, only with less data and data of poorer quality. 

Nelson and Flick both said they support ERIC. But both also mentioned concerns about ERIC’s data management and transparency, echoing claims made by VoteRef and Republican officials, though neither provided details or evidence. Nelson said it was up to ERIC to address those concerns in the hopes more states join the organization. 

“I support closing the election integrity gap which exists across state lines, whether it is achieved though improvements in ERIC or alternative means,” Nelson wrote in an email to WITF. “I am willing to work with many different organizations to improve our elections and learn more about the details of our process.”

ERIC’s executive director, Shane Hamlin, did not directly respond to those critiques. He said the organization is bipartisan and operates in a nonpartisan manner.

ERIC still has 25 member states and D.C. and is able to compare voter registration, motor vehicle, Social Security and National Change of Address data. That gives participating states more accurate and proactive reports to improve voter rolls and prevent election crimes.

A history of election denialism 

Nelson was one of the Republicans in the Pennsylvania House who signed a letter after the 2020 presidential election asking Congress to reject Pennsylvania’s electoral votes. It was part of President Donald Trump’s effort to overturn his election loss to Joe Biden. 

Nelson has been attending national voter integrity conferences since 2020, he said, which is where he was introduced to VoteRef. Like Nelson, VoteRef has challenged the legitimacy of the 2020 election. The group is funded by billionaire Republican donor Richard Uihlein and is led by Gina Swoboda. She is also the chair of the Arizona Republican Party and was an organizer for Trump’s 2020 campaign in that state. 

The group is in current litigation with Pennsylvania’s Department of State because it wants to publicly post the state’s voter registration files, like it has for most other states. The department says that would be a violation of its data usage policies and could lead to voter intimidation.

Instead of using VoteRef’s data, Nelson used Flick’s experience as a software engineer to parse through Ohio and Pennsylvania’s voter history files. The representatives asked VoteRef to use its purchased access to the National Change of Address data to match addresses in multiple states to a single person. 

To explain why they involved VoteRef, Nelson said neither he nor Flick are investigators, nor are they election integrity experts. 

“We really reached out to a nationally recognized organization to be able to say, ‘Hey, what are your thoughts about this,’” Nelson said. “It gives us another layer of vetting before, you know, we take that next step.” 

The goal is to be as accurate as possible with their list of potential double voters when referring names to federal, state and local law enforcement, Nelson said. 

 ”It’s important to recognize that the individuals that we’ve identified as potential double voters are across the spectrum in all parties,” Nelson said. “In one state they may be registered in one party and in another, they are a different party.  So it’s not a partisan thing, it’s an integrity thing.” 

Both representatives recognize that the potential pool of double voters is not large enough to sway statewide elections, or even the vast majority of local elections. Nelson said even doing the investigation is worth it because it amplifies the message that voting twice is bad. 

Nelson is the main driver of the effort, according to Flick. He says he is just the data guy, given his background in software programming. Flick founded, and recently sold, Susquehanna Software, Inc. 

Even with his ability to handle Pennsylvania’s unwieldy voter registration database, which is available to download for $20 and comes in four separate files for each of the state’s 67 counties, Flick said he needed VoteRef’s resources. 

VoteRef did not respond to questions for this story. Its website says it “is dedicated to ensuring transparent, accurate and fair elections” and does so by publishing voter data for 36 states. But state officials have dismissed the group’s findings, noting VoteRef’s inaccurate claims that there were more ballots counted in the 2020 election in Pennsylvania, Georgia, Michigan, Nevada, New Jersey, North Carolina, Ohio, Virginia and Michigan than there were voters who turned out for the election. 

Pennsylvania’s Department of State published a lengthy fact check of those claims on its website. It described how public voter registration files cannot be used to reverse engineer voter history on a prior Election Day, since registration files are constantly updated. Election officials in the other states also debunked VoteRef’s claims, showing the organization misunderstood election data and erred in its analyses

Nelson did not directly answer a written question about VoteRef’s history of inaccurate claims and analysis. He said he, along with Flick, will cross reference VoteRef’s findings with their own.

“Our goal is accuracy.  We are working hard to ensure a consistent, repeatable and reliable process and if we can learn from VoteRef’s analysis, we will seek to do so,” Nelson wrote. “We will then take the next steps to pass on those identified potential double voters to the appropriate authorities.”

Flick’s flawed methodology

After WITF’s first interview with Flick in early October 2024, he said he would share the code and methodology behind how he compared Pennsylvania and Ohio’s 2020 general election voter history files.

Then, in an email later the same day, he wrote, “I’m informed since I wrote this code as a state rep and not as the owner of a software company, the code is proprietary. That being said, I will send you a document of the logic and process I used.”

Flick did not answer who gave him that guidance — there’s no rule against voluntarily releasing the information. He also never shared a document of logic and process. Instead, he verbally described his process. 

WITF shared this process with University of Pennsylvania political scientist Marc Meredith, who has studied rates of double voting. Without adding additional data, like partial Social Security numbers, methods like Flick’s will result in a large percentage of matches that are simply different people with the same name and birthday, Meredith said. 

Flick’s methodology 

Step 1: Download and combine Pennsylvania’s voter files into one single file.*

Step 2: Download Ohio’s voter registration file.*

Step 3: Filter for registered voters who cast ballots in the 2024 general election (or 2020 for his first inquiry).

Step 4: Use first, last name, DOB and fuzzy matching on middle name.

Fuzzy matching means including people with the same middle name, including people who have a full middle name in one file but the same first letter of a middle name in another or when the column is left empty, but not including anything where name or initial don’t match.

Step 5: Rank matches on probability. Flick did not describe his exact methodology. He did say he excluded common names such as Smith, but did not provide a full list of excluded, common names. He did not describe how middle name fuzzy matching was weighted in his probability measurement of likely double voters. 

Step 6: Provide findings to the Voter Reference Foundation to compare the list to the USPS change of address database (ongoing). 

*Depending on when the voter registration file is downloaded, some voters from the last election may not be listed. This can happen when voters move after an election, if they are convicted of a felony, die, or otherwise provide cause for elections offices to clean them from the voter rolls. 


Help support the information and news you’ve come to rely on in central Pennsylvania. with a donation to WITF.

Click here to make your donation

Support for WITF is provided by:

Become a WITF sponsor today »

Support for WITF is provided by:

Become a WITF sponsor today »

Up Next
Regional & State News

Observed by the 7-10 fly-like insects: Adams County restaurant inspections July 5, 2025