ATLANTA– Opponents of an Atlanta police and fire training center cheered as they marched into City Hall in September carrying sixteen boxes of petitions to force a referendum on the issue. “116,000 signatures – can you hear us now?” they asked, convinced they had enough.
But an analysis from four news organizations shows that the outcome — if city officials ever count the petitions — could be decided by a narrow margin.
Organizers of the months-long petition to “Stop Cop City” still say they have 116,000 signatures, but a hand count by The Associated Press, Georgia Public Broadcasting, WABE and The Atlanta Journal-Constitution was only about 108,500.
The news organizations found that nearly half of a statistical sample of a thousand entries could not be matched to an eligible registered voter in the city of Atlanta. Some signers live outside the city, some apparently made up addresses, and others gave far too little information – such as the 'Lord Jesus' who signed with the address 'homeless'.
Even with these issues, the analysis shows that it is still statistically possible that organizers met their target of 58,231 signatures. But additional legal and procedural challenges could undo this effort by sharply reducing the number of eligible signatories.
The fight over the $90 million training center has become a national dispute, with opponents mocking a facility they say will worsen police militarization and harm the environment.
Kate Falanga, a bar manager who signed the petition, called the proposed project “terrible” and said the land – part of a huge urban forest – should be used for something “better for everyone”.
“There are much better ways to spend that than on militarizing a police force that already has an incredible presence and seems quite unnecessary,” Falanga said.
Supporters, including Atlanta Mayor Andre Dickens, say the city must replace aging facilities and better train officers to prevent improper use of force.
“I believe this center will be the classroom space that redefines the way we approach policing and maintains the readiness of our first responders to tackle the challenges they face,” Dickens wrote to U.S. Senator Raphael Warnock in September, after the Georgian Democrat questioned referendum procedures. .
Both Dickens and city officials who would validate signatures declined repeated interview requests.
The referendum seeks to terminate the city's lease with the private Atlanta Police Foundation to build and operate the 34-acre complex.
“The people who signed these papers are real, they exist,” said Britney Whaley, who helped organize the petition campaign. “They want to see it on the ballot. They don't go away. We have enough signatures to transform politics in Atlanta.”
But not everyone who signed will be eligible for the citizen-led petition process, and the referendum is in legal limbo. The reporting partners began analyzing the petition submissions because officials had not counted the more than 25,000 pages submitted to the Atlanta City Clerk's Office.
Petitioners had to gather 58,231 signatories. That equals 15% of Atlanta's active registered voters in 2021. Each eligible signer must now also be a registered city voter.
With approximately 108,500 entries, almost 53.7% must be valid for organizers to be successful.
Overall, the analysis shows that as many as 52.7% of entries could be eligible. That's below the 53.7% threshold, but because the analysis is a sample, it has a margin of error of plus or minus 3.1% at a 95% confidence level. This means that between 49.6% and 55.8% of the entries must be valid. A complete count could therefore provide enough valid data to exceed the threshold.
The reporting partners analyzed a sample of 1,000 entries, proportionally took a random sample of pages from each of the 16 boxes, and then chose a random entry on each page. By matching names and addresses against voter rolls, the partners matched 47.5% of names to eligible voters in Atlanta.
Another 5.2% of entries are uncertain – for example, common names that match multiple voters, but without a matching address. These may qualify based on dates of birth redacted from copies of public petitions, or through signature comparisons. They can also qualify if signers provide additional information, a process known as curing.
Conversely, dates of birth and signatures may disallow some apparently eligible entries. In particular, signature comparison is a much-debated part of the proposed verification process.
The news organizations did not attempt to verify signatures, meaning they could not detect fraudulent signatures created using a voter's information. On at least a handful of pages there are signatures in what could be the same handwriting. The sample found no duplicate signatures, but it was statistically unlikely to detect people who had signed more than once. Dickens said that's why the city should examine petitions if there is a referendum.
“We have a duty to review these petitions and ensure that it is Atlantans who speak for Atlanta,” the mayor wrote to Warnock.
Whaley said petition collectors wanted everyone who might be eligible to sign.
“You might be a little vague on some of the details, right?” she said. “But what I won't do is let you walk away without signing my papers if there's any chance you're registered in the city of Atlanta. in 2021.”
The demographics of the petition's eligible signers were approximately 50% black, one-third white, and a majority aged 42 or younger, which closely reflects the city's electorate.
But people who lived near the training center, just beyond the city's eastern border, were more likely to sign the petition. The five zip codes of the cities closest to the location contained less than a quarter of Atlanta's registered voters but produced 37.7% of eligible entries.
In contrast, only 4% of eligible entries came from zip codes in the city's northern Buckhead region, where the electorate is whiter and more politically conservative.
Many of the people who were not eligible in the analysis do not currently live in Atlanta. The media partners found that 16.8% of petition signers had registered elsewhere in Georgia, sometimes outside city limits. Others were not voters in Atlanta in 2021 or 2023 or did not appear to be registered to vote at all.
Another 30% of eligible entries may be disqualified for legal reasons or city counting methods.
Former City Clerk Foris Webb III, who would lead the signature verification, declined to say how the city will handle submissions that do not have an address that matches voter registration. That alone could disqualify 12% of potentially eligible entries.
State law says the collectors who witness signatures must be registered voters from Atlanta and organizers had until Aug. 21 to submit petitions. A federal judge extended that deadline until September, ruling that people who live outside the city could witness signatures. The 11th U.S. Circuit Court of Appeals halted enforcement of that order. That's why the city refused to count the petitions when organizers dropped off boxes on September 11.
Even if the petitions are ultimately counted, only submissions submitted before August 21 and collected by Atlanta residents can be considered valid. These two issues could disqualify 20% of potentially eligible entries, likely ruining the effort.
Appeals judges will hear arguments Thursday about the deadline and witness testimony. The city also claims the underlying petition is void because it violates state law and would illegally cancel a contract.
Whaley said city officials are pretending to be “scared” and calling their actions to block the referendum “highly problematic and undemocratic.”
“I think they're at a point now where they're doing everything they can to save this project,” she said.
___
RJ Rico of The Associated Press; Amanda Andrews of Georgia Public Broadcasting; WABE's Chamian Cruz, Emily Wu Pearson and Jasmine Robinson; and Riley Bunch, Pete Corson, Rahul Deshpande, Stephanie Lamm, Charles Minshew and Justin Price of The Atlanta Journal-Constitution contributed.