Developers in Hopkinsville had just gotten sewer lines in place for a new shopping and entertainment complex in 2007 when it was put on hold after the release of U.S. Census population estimates.
Those numbers, created by the federal agency and checked by the Kentucky Population Research Center at the University of Louisville, showed that Hopkinsville and Christian County had shrunk since the 2000 census.
They were wrong.
Still, the false information scared potential retailers and restaurateurs for the planned New Town Station complex, said Hopkinsville Mayor Dan Kemp.
"The developer felt strongly that the estimate hurt the momentum for getting that off the ground" because retailers don't want to expand into shrinking communities, Kemp said. The complex — which was to include a big-box store, such as Target, a movie theater, bowling alley and restaurants — never materialized.
The incident illustrates how census figures and estimates play crucial and sometimes costly roles in everyday life.
A prominent number-cruncher says the Kentucky Population Research Center, which vets census estimates and makes projections for counties' future populations, has a history of "sloppy" work that carries serious consequences for communities. Population numbers are used, among other things, to determine federal school funding and the size of new courthouses and to attract new businesses.
Ronald T. Crouch — who retired in May as the director of the agency's sister organization, the Kentucky State Data Center — said he repeatedly questioned figures released over the last seven years by state demographer Michael Price, head of the research center.
"I've told people not to use the numbers, which really makes me angry because our job is really to help these communities," Crouch said.
Price acknowledged that he has made some mistakes over his 28-year career, but contends that his checks and projection methods are better now.
He said he didn't understand why Crouch is raising the issues publicly now.
"And over this whole period of time, he's promoted and used the forecasts," Price said. "My take is that my work has always been of the highest quality ... I've always gotten that kind of feedback."
Staying at U of L
Crouch said he pointed out to Price problems with some data produced over the last seven years. And Crouch provided the Herald-Leader with documents detailing problems with the reports, as well as his own correspondence with U of L officials about those concerns.
"It sickens me that KPR has, over the years, produced such poor population projections and put so little work in making sure the census population estimates are correct," Crouch wrote to U of L Provost Shirley C. Willihnganz in a Nov. 4, 2003, memo.
U of L's Urban Studies Institute is home to both the Kentucky Population Research Center, which has the contract with the U.S. Census Bureau to work on Kentucky's figures, and the State Data Center, which was created in 1980 to work with the research center to analyze trends and disseminate the data to the public.
Crouch said U of L officials' lack of response to his concerns suggests that the centers would be better off somewhere else. Each state must have an entity to work with the census bureau. They are usually based in government offices, libraries or universities.
"U of L has had seven years to clean up their act. I think the centers should be transferred to a state agency," said Crouch, who said he retired because of mounting frustrations with the population agencies.
U of L officials said they took Crouch's concerns seriously and, after he raised the issues, created separate budgets for the data center and research center.
The two entities receive state funding through U of L's state appropriations, but the research center also receives pay for contract work such as the Kentucky Kids Count report.
The university wants to keep both centers and plans to hire a senior-level tenured faculty member to oversee them, something the two entities haven't had before, said Mark Hebert, U of L's spokesman.
"We believe that this is a critical function for the state of Kentucky, and it's very important to keep that work," at U of L, Willihnganz said in a statement.
Déjà vu in Hopkinsville
Vernon Smith, a researcher with U of L's Urban Studies Institute who helped found the state data center in 1980, said Price and the research center should be more diligent in checking census estimates.
"They should be in contact with all these people and know what's happening and know the counties ... so when they look at the census estimate numbers they can see whether they look reasonable or not," Smith said. "The bottom line is that for Kentucky, there's money attached to those numbers. So it's extremely important that we have those numbers correct."
Christian County should be on top of that list, Smith said. As home to Fort Campbell, one of the largest military bases in the country, residents often are in flux. It also borders Tennessee, an enticing place for people to claim they live because that state doesn't collect income tax.
That affects U.S. Census Bureau estimates because they are based largely on income tax information.
Christian County leaders found that out in 1996 when census estimates showed the county losing people since the 1990 census.
They challenged those numbers — which is rarely done — and won, prompting the bureau to add more than 7,700 people to the estimate. As a result, county officials urged the Kentucky Population Research Center to pay particular attention to future estimates and check such information as utility hook-ups and employment data, according to a 1997 article in Hopkinsville's New Era newspaper.
But a decade later, it happened again.
Price acknowledged that his group didn't check utility hook-ups. He said he looked at other numbers instead, such as area birth rates, which didn't indicate steady growth.
"Those patterns and other things we look at weren't real supportive of the growth," said Price, who didn't support the 2007 appeal. Ultimately, the census bureau agreed with community leaders that it underestimated population growth.
Marie V. Bousfield, a Chicago-based independent demographer, said the number of appeals from communities in a state vary widely. When Christian County and Hopkinsville challenged estimates, they were among 67 local governments in 19 states to appeal the numbers.
Price said the few challenges from Kentucky counties and cities suggest that his office does a good job.
Census officials revised the figures months later to show Christian County and Hopkinsville gaining population, making its schools eligible for more federal funding.
Chris Sutton, assistant director of the Pennyrile Area Development District that helped with the appeals, said that he'd been disappointed that the county has had to fight the same battles over again.
"We have a dialogue going with KPR, we hope, so that next time the estimates come out, we don't have to go through the same challenge process," Sutton said.
Money at stake
Accurate numbers also can prevent communities from spending millions of dollars in unnecessary construction.
Projections that Price released in July 2003 pegged Gallatin County to triple in population by 2030. One of the fastest-growing groups, according to those projections, was school-aged children.
That prompted the question of whether new schools were needed.
But, upon closer review, officials said other measurements didn't support projections of sharp growth.
"When we looked at the numbers Michael Price put out, they showed that the county was getting ready to explode," said Dorothy Perkins, Gallatin County Schools superintendant. "We disregarded those numbers."
Price said he revised projections by 2004 to show more moderate growth in Gallatin and other Northern Kentucky counties, such as Boone.
He uses various formulas to calculate those forecasts. Still, population growth is affected by a host of external factors that are tough to predict, such as the economy and migration pattern changes.
"It has improved — the accuracy of our estimates, especially for the smallest of places," Price said.
Issues at U of L
When Crouch raised his concerns with U of L officials in 2003, he cited what he described as a longstanding pattern of problems that "were an embarrassment to KSDC and this university."
Willihnganz, in an e-mail response, said the associate provost for personnel would delve into Crouch's concerns about the agencies' funding and "integrity of the data that are being produced."
A March 2004 audit said the university's audit division "is not qualified to evaluate the quality of the forecasts made by KPR."
Crouch said that was all the university did.
Price's supervisor at the Urban Studies Institute, Bruce Gale, expressed confidence in Price's numbers and methodology, said Hebert, the U of L spokesman.
In another example, a 1988 report from the Population Research Center showed the population of 20- to 24-year-olds dropping in Fayette County in the future, apparently disregarding University of Kentucky and Transylvania University students.
"Yeah, there is an error in here," Price said. "Errors are made. That's why we revise ... It's not a systemic problem."