Understanding Factors that Contributed to the Large Error of Closure in Census 2000

Gregory Robinson, U.S. Census Bureau
Kirsten West, U.S. Census Bureau

This paper puts in perspective the relative contribution of the two components of the large 6.8 million error of closure, which resulted when the 1990 census-based population estimate of 274.6 million was compared to the Census 2000 count of 281.4 million: (1) the amount attributable to change in coverage from 1990 to 2000, (2) the amount attributable to error in the population estimates for 2000. Many users believe that the error of closure was largely caused by the understatement of growth by the population estimates for the 1990's. Now that the coverage studies for 2000 have been completed, we can systematically assess the contribution of estimates error and coverage change. The findings show that 60 percent or more of the error of closure is attributable to the dramatic reduction in net undercount from 1990 to 2000; the remainder includes that accounted for by estimation error.