Reasons Behind Inaccuracies in the Census
Up until 1990, the overall accuracy of the census improved with each decennial census. In 1990, however, the trend reversed; the national net undercount, as well as the differential undercount of minorities, increased from the previous census. With a new communications plan that included paid advertising, a partnership program, and increased funding, the 2000 census was a significant improvement over the 1990 census. However, this time the census resulted in a net national overcount of one half of a percent and a Black-White (non-Hispanic) differential undercount of 3.0 percent.
Several reasons account for the persistent and disproportionate undercount of people of color and low-income people. Lower-income areas experience lower response rates for mail and door to door collection methods. People with lower education levels, lower literacy abilities, and difficulty with the English language may have difficulty understanding the census. These communities may generally misunderstand the importance of census participation. Furthermore, many distrust or are suspicious of the government because they fear that census responses may be used by immigration or law enforcement officials to deport or incarcerate them or their family members or may disqualify them for social welfare programs.
Rural residents are also difficult to count. Many homes are very remote and, therefore, inaccessible. Also, individuals living in rural areas tend to use post office boxes and/or general delivery “rural route” addresses rather than individual home addresses, making it more difficult to deliver and collect census forms.
Determining the Undercount/Overcount
The Census Bureau began developing methods to better measure census accuracy when, in 1940, the census grossly underestimated the number of men that would enlist for the draft because of the miscount of young black men. Using information from birth and death records, past censuses, and information on immigration, the Bureau developed a system of measuring the undercount called Demographic Analysis. This method has been used as an independent check on census accuracy at the national level since 1940; however, it can only give national undercount numbers.
The 2000 census included a procedure, called the Accuracy and Coverage Evaluation survey, to check for missing housing units and people who were not reported on a census form. However, concerns about the reliability of the accuracy evaluation led the Census Bureau to revise its estimates of undercounts and overcounts several times and ultimately to conclude that it could not use the estimates to make any corrections to the original census numbers.
In 2003, the Census Bureau announced that the 2010 census would not include the possibility of a statistical adjustment based on a coverage measurement survey, citing concerns about the methodology’s reliability and the Bureau’s ability to produce adjusted population figures in time to meet legal deadlines for apportionment and redistricting. The 2010 census will include a quality check survey of 300,000 addresses (called the Census Coverage Measurement survey), to estimate undercounts and overcounts for the nation as a whole and for states. The Census Bureau plans to publish the estimates, which will measure coverage by race and Hispanic origin, gender, age, and renter/owner status at the national level only, in 2012. It will not use that analysis to “adjust” the original population figures, however.
