PROBLEMS AND LIMITATIONS OF CENSUS ESTIMATES

Despite the attempts by the Bureau to be sure the Census includes everyone, there are problems with the data that are especially pertinent to those of us in public health. The Post-Enumeration Survey conducted six months post-Census showed, in 1990, an undercount of populations at-risk, e.g., the homeless, minority males in poverty, and Native Americans. Other problems included:

• age-undercount of those individuals under one year of age;
• individuals whose race is classified as- "other," "non-Hispanic Black," and "Hispanic" as well as lower income individuals had the lowest Census form return rates;
• geographic location-urban areas were better delineated than rural areas; and
• population-since the population of the United States migrates freely between states and between cities, the stability of state and local estimates is affected.

Another problem of which we need to be aware is suppression of data. When a group analyzed in a table includes only a small number of persons, the Bureau may suppress the data through use of a threshold, which is generally 400 persons or fewer of a specific group. For example, in Illinois, 19% of the Asian or Pacific Islanders live in areas where they constitute fewer than 400 persons. While these individuals may have different social and economic characteristics, e.g., age, race, etc., than the other individuals in the area, their small numbers prevent these different characteristics from being enumerated. Conversely, a complementary threshold is used to limit the presentation of characteristics of the white population when the population of other races in the area is small.   Thresholds and complementary thresholds vary from 250 to 1,000 persons, depending on the table.

Census home

Next