The U.S. Census Bureau evaluated how well Census 2000 counted the population by conducting a coverage measurement survey known as the Accuracy and Coverage Evaluation survey (A.C.E.). The Census Bureau considered adjusting the Census 2000 population total of 281,421,906 to correct for coverage error on three occasions, but each time decided not to adjust.

As part of the preparations for Census 2000, the Census Bureau planned to decide in March 2001 whether to adjust Census 2000 numbers legally required to be produced by April 1, 2001 for the purpose of redistricting of the U.S. Congress (Prewitt 2000). Initially the A.C.E. measured a 1.2 percent net undercount, but after much debate, the Census Bureau decided not to adjust Census 2000 for redistricting. The reason for not adjusting was that discrepancies between the A.C.E. estimate and another method of measuring census coverage, Demographic Analysis, could not be explained by the deadline for the decision. Demographic Analysis(DA) initially produced a base estimate of a -0.7 percent net undercount (an overcount). Demographic Analysis subsequently produced an alternative estimate 0. 3 percent net undercount with an alternative assumption that net undocumented migration was twice as much as included in the base estimate (Robinson 2001). For the 1980 and 1990 censuses, the estimates of the percent net undercount from the two programs had been within 0.2 percentage points. The discrepancies in the estimates of Census 2000 coverage error caused the Census Bureau to not be sure that an adjustment would improve the census numbers (U.S. Census Bureau 2001b).