Contact Us

February 20, 2001
For Immediate Release

Contacts: John Chambers

Census Undercount Forum
Tuesday, Feb. 20, 2001 • 9:00am EST

Need RealPlayer 8.0? Download it here.


Findings Underscore Need for Modern Scientific Methods

Washington, D.C. (February 20, 2001) -- Today, the Presidential Members of the U.S. Census Monitoring Board convened a forum to release the findings of several highly respected independent researchers who were commissioned to study undercounts in previous censuses in order to decipher the potential impact of a Census 2000 undercount. The studies follow the Census Bureau's recent release of preliminary estimates that show between 2.7 to 4 million people were missed in the 2000 count. Following are some of the researchers' findings: (The studies and the forum can be viewed in full at

  • Due to the 1990 undercount, health policy planners were unprepared for 400,000 uninsured Americans, nearly half of whom were African Americans and Hispanics (192,000), according to a study conducted by Dr. Lee Cornelius, a University of Maryland Professor and a minority health insurance expert. Given that per capita health care expenses came to $2,400 in 1996, Cornelius found that the undercount translated into $960,000,000 in unforeseen health expenses in that year alone.
  • Use of unadjusted census data distorts measures of access to health care by inflating the rates of mortality, morbidity, injuries and accidents making minority communities less attractive for private investment, development and inflating provider-to-population ratios which results in doctor and hospital shortages, according to a study conducted by Dr. Darrell Gaskin, economic researcher at the Institute for Health Care Research and Policy at Georgetown University Medical Center. Gaskin also found that:
  • Preventable hospitalizations, a measure of access to health care, is overestimated at a rate six times higher for African Americans and eight times higher for Hispanics as compared to whites.
  • All counties are not affected (by access to health care) the same way. In particular, rural counties and counties in Texas, California, Georgia, Colorado, Mississippi, and New Mexico are the most adversely affected.
  • According to a study conducted by American University Professor and Constitutional Scholar Dr. Allan Lichtman, if the 1990 census had used corrected data, minority voter opportunities would have been expanded in 9 of the 10 most heavily undercounted states. In those nine states, at least 44 state legislative districts and one Congressional District would have significantly increased minority voter representation.
  • Census undercounts distort transportation policy, planning, funding allocation and governmental programs, according to Dr. Paul Ong, UCLA Professor and urban planning specialist. Ong found that the 1990 census missed more than 500,000 commuters who travel to work in 22 metropolitan areas.
  • Children in poverty are among the hardest hit by unadjusted census data, according to a study conducted by Professors Donald Hernandez and Nancy Denton, of the State University of New York at Albany, which found that at least 500,000 and as many as 2,000,000 children were missed in the 1990 count. Children in poverty rely on federal programs that rely on census data, such as Medicaid, Head Start, Foster Care, Adoption Assistance and Social Service Block Grants.
  • In a study that compared birth, death and school enrollment records to the results of the 1990 census, Professor Beth Osborne Daponte, of Carnegie Mellon University, found that more than 20% of infants were missed in the 1990 census (815,000). Daponte also determined that age misreporting was not a major factor of the infant undercount because there was not a significant number of one-year-olds counted twice.
  • In a case study that examined the funding effects of an undercount on city services, University of Southern California Professor Christopher Williamson found that the City of Long Beach lost at least $10 million in government services due to the 1990 census undercount. The 18,350 local residents missed in the count, resulted in a loss of about $1 million per year in annual federal funding ($56 per undercounted person).
  • Nearly 3,000 American Indians living in the Albuquerque metro area were missed in the 1990 census, according to a study conducted by Dr. Ted Jojola, of the University of New Mexico, which assessed the impact of undercounts and adjustments on urban and reservation Indian populations. Reservation Indian programs are more apt to be closely tied to census numbers because of their U.S. Federal requirements for formula funding and tribal program development. Jojola also found that social services, such as domestic violence shelters, were adversely affected. American Indians were the highest undercounted group in 1990 and preliminary estimates show that they will be the most undercounted group in 2000.
  • As a result of 1990 census undercount, the Portland Public School District, one of the best counted districts in the 1990 census and nationally recognized for its school and urban planning, was still off by as many as 1,250 students in their enrollment forecasts for the decade ending in 2010, according to a study conducted by Dr. Barry Edmonston, Director of the Population Research Center at Portland State University. That many students is enough to fill two new secondary schools.

"These studies present overwhelming evidence that the undercount, if not corrected, will have damaging effects at every level of our society," said Gilbert F. Casellas, Presidential Co-Chair of the Monitoring Board. "If millions of America's poor, minorities, children, and newest residents are disenfranchised over the next ten years, they will not only be left behind, they will never catch up."

Census Monitoring Board Member Everett Ehrlich added, "The scientific methods that are praised for measuring the errors in the census are every bit as good at correcting those errors as well. We must continue to allow the professionals at the Census Bureau to determine the accuracy of the census."

"Americans need to understand that their rights to representation and a fair share of federal funds are at stake," said Census Monitoring Board Member Lorraine A. Green. "And any suggestion that adjusted numbers are somehow more accurate for one purpose, but not another, is nothing more than a modern day separate-but-equal solution."

The bipartisan Census Monitoring Board was established in 1997 to monitor Census 2000 operations. Its findings are reported to Congress every six months.

To top.

U.S. Census Monitoring Board
Presidential Members
4700 Silver Hill Road
Suite 1250 – 3
Suitland, MD 20746
Phone: (301) 457-9900
Fax: (301) 457-9901