Steve Patterson, the mayor of Athens, Ohio, had hoped last year's census would bring the city some good news. Athens, home of Ohio University, was just shy of 25,000 people, and breaking that threshold would make the city eligible for grants and other funding reserved for larger cities.
But then the COVID-19 pandemic swept across the country just as the once-a-decade headcount was about to begin. College towns were already hard to count, but now universities were sending away students just weeks before April 1, the day residencies were officially tallied. University administrators, census counters and local government leaders like Patterson would spend the next several months trying to pin down where college students had lived in Athens before the coronavirus scattered them.
But in Athens and many other places, local officials fear the counts came up short. New data released by the U.S. Census Bureau last month showed that, in the decade between 2010 and 2020, Athens gained just 17 people.
That would be only 23,849 people — far short of the 24,984 that the Census Bureau had estimated lived in Athens as recently as 2019.
If the new census numbers are mistaken, they could be costly for the city. The difference between the 2019 estimates and the 2020 final numbers could result in the loss of $16 million over the next decade in federal Community Development Block Grant money, Patterson said.
"That's a significant amount of money that could be going to city infrastructure, to school districts, to social services," said Patterson, who is the president-elect of the International Town & Gown Association.
Many other college towns find themselves in similar situations. While some big university towns boomed over the last decade, others have unexpectedly low counts or even reported declines in their populations. The disappointing census numbers are particularly worrying in cities where colleges and universities dominate the local economy. They're concerning for higher ed, too, because university operations are closely intertwined with their local governments.
"Every great university has a great town that surrounds it, and every great town has a great university in it," said Stephen Gavazzi, an Ohio State University professor who studies town-gown relationships. "Once you start to see the deteriorating effect on one side or the other, it drags the other down."
College and local officials have been bracing for low census counts ever since COVID-19 started spreading in the U.S.
"There has been great concern for as long as people knew that there was a pandemic that was going to be getting in the way of the census," Gavazzi said. "The actual census numbers just simply confirmed what everyone feared was going to happen."
Once the problem became clear, though, institutions of higher education and local officials tried to mitigate it as much as possible.
At Penn State University, the university attempted to promote filling out census forms through social media and email messages from student leaders. It reached out to parents to explain that their children should be registered at their campus addresses if the students normally lived there during the school year.
"There has been great concern for as long as people knew that there was a pandemic that was going to be getting in the way of the census."
Professor at Ohio State University
Like many universities, Penn State provided the Census Bureau with its full directory of students living on campus. It also forwarded residential information about students living off campus.
But the cascading crises of 2020 made it hard to get the message out to students, as they scattered far from campus.
Many students spent the rest of the semester and summer at their family residences, said Charima Young, Penn State's director of local government and community relations.
"As a result, interest waned during the pandemic as they were bombarded with messaging and guidance about the pandemic locally and nationally," Young said. "The focus became safety, stability and survival for many individuals and families."
The university pushed students to participate in the census in part because of the impact it would have on federal funding for local municipalities, she said. It could make more money available for housing, education, transportation and healthcare initiatives in the greater community.
But getting students counted could directly impact their lives, as well, because the numbers could affect federal support for services such as hospitals, emergency response teams, police, public transportation and Pell Grants, Young said. "These are benefits they use and which really resonated with them, particularly Pell Grant funding," she said.
Despite those efforts, the census found that the population for the borough of State College, where Penn State's flagship campus is located, dropped in the last decade by 1,533 people. Last year's count of 40,051 residents was also lower than the Census Bureau's projections from the year before.
That's left borough officials weighing their options for how to respond, including a potential challenge to the official count. That process cannot begin until December, and it could stretch on for years. The adjusted numbers could be used for population-based funding formulas, but they would not affect the population numbers used for political redistricting.
For now, the main concern among borough officials is getting ahold of as much data as they can to see where any discrepancies occurred. They plan to look at hard-to-count Census tracts and see what their vacancy rates in apartments were, whether there were any issues with postal delivery of census forms to those areas, and other relevant information, said Douglas Shontz, the assistant to the borough manager there.
The borough is also asking for help from federal officials to see if they can get more information about how the Census Bureau handled the different challenges that affected college towns. And it would like financial help if it pursues a recount, because the Census Bureau currently requires municipalities to pay for those challenges, Shontz said. Those costs could be as high as $200,000.
Officials in East Lansing, the home of Michigan State University, had hoped that their city would finally break the 50,000-person threshold. The city was at 48,579 people after the 2010 census and, like many college towns, had seen plenty of new student housing being built in the years since. The 50,000-person mark was significant, because it would make the city eligible for more kinds of federal grants.
Instead, East Lansing's population dropped to 47,741 last year.
Michigan State had tried to help with the census counts by sharing student data with the bureau. It not only uploaded information on students living on campus, it also submitted the records of 17,000 students living off campus, said Janet Lillie, Michigan State's assistant vice president for community relations.
Now, though, Michigan State and East Lansing officials will have to investigate whether the census's low numbers resulted from an undercount or just a shift in student living patterns. The neighboring city of Lansing, Michigan's capital, also saw a drop in population. But nearby suburban areas recorded growth.
Figuring that out could be complicated, though, Lillie said, because the Census Bureau won't say whether it used the data that Michigan State shared to come up with its estimates for off-campus housing.
Still a hazy picture
Beyond a few prominent examples, it's hard to tell for certain right now how common undercounts were in college towns during the 2020 Census.
"Overall, we found that the 2020 Census results from the redistricting data are comparable to the population benchmarks we've examined," Census Bureau researchers said when the agency released its initial 2020 counts.
The analyses did not include any specific tests for college towns. But the agency compared the results of its hand count to projections of what various populations were expected to be. For example, it looked at how county populations compared with projections, as well as demographic groups such as people over 18 years old, people of Hispanic origin and various racial groups.
The bureau also plans to take a closer look at the accuracy of its 2020 count, a step it takes after every decennial tally. Part of that process this time will look at how to improve the accuracy of counts in privately owned student housing for universities.
What's more, the Census Bureau has not yet released the full findings from the 2020 count. Its initial batch of data contained the information most helpful for state legislatures, city councils and other government bodies that need to draw new electoral district lines to reflect the changing populations.
But the Census Bureau still has to release more granular data that would make it possible to see how many 18- to 24-year-olds lived in a neighborhood. That's important, because neighborhoods that have traditionally been college housing don't necessarily stay that way.
In Ann Arbor, Michigan, for example, it appears that the tracts where students tend to live at least maintained their population, said Lisa Neidert, a sociology lecturer at the University of Michigan who teaches courses on federal data and statistics. But available information doesn't show whether all of the students were counted correctly, she said.
"They've added a lot of housing downtown, and it could be that 65-year-olds find it appealing as well. The kids are gone, and they're done with their big lawns. They want to walk or take a cab somewhere," Neidert said. "So they live in what would be an attractive student neighborhood."
Michael Cline, North Carolina's state demographer, is also waiting for more information.
"It is too early to come to any conclusions about the 2020 census counts — whether it be in college towns or anywhere else," he said in an email. "I don't yet see any glaring issues with the counts."
He said the census counts for dormitories, for example, were in line with what universities reported their populations to be. Cline also said the federal population numbers for North Carolina cities seemed similar to the state's projections.
"Even when the census does count students, it seems to be messing up their race data quite a bit."
Demographer for the University of Virginia Weldon Cooper Center for Public Service
In neighboring Virginia, though, some demographers are concerned about college town counts.
Hamilton Lombard, a demographer for the University of Virginia Weldon Cooper Center for Public Service, said he found several instances of population counts that didn't match with what's going on in college towns.
In the city of Radford in southwestern Virginia, for example, students who attend Radford University make up more than half the population. There has been a "fairly significant increase" in the number of students living in dorms over the last decade, Lombard said, but the census figures showed a population decline for the city.
Lombard is part of the Demographics Research Group, which releases its own population estimates. The biggest differences between what the group projected and what the Census Bureau released, Lombard said, were in localities with colleges or jails.
"Even when the census does count students, it seems to be messing up their race data quite a bit," Lombard said. At the University of Virginia, the Census Bureau's count showed three times as many Asian students as were at the university, he said.
Finding the reason for those discrepancies is time consuming, and it requires researchers to look at individual census blocks for issues. In Charlottesville, which is home to the University of Virginia, Lombard's group found one large apartment complex that had a population of about 750 students. But the census figures showed a population for that area that was half that total, Lombard said.
That sort of granular analysis will be harder with 2020 census figures than with previous headcounts, though, because the Census Bureau has intentionally added "noise" in its data to protect the residents' privacy. That means the bureau swaps data about respondents from one area to a nearby area so that the public won't be able to single out individual people. That won't have much of an impact on the results at the citywide level, but it could skew results for very small geographic areas.
Lombard said he thinks the problems he's encountered with college towns and other places with lots of group quarters are not unique to Virginia. But he doesn't expect an outcry, either.
"One reason why demographers are probably a little reluctant to start criticizing 2020 census data is because they don't want to undermine people's trust in the data," he said. "But I think they're going to start bumping into it and realizing there's just problems with the 2020 Census."