Census

Every ten years since 1790 a census has been taken of the U.S. population. The process turns up much valuable information about American society, but its constitutional purpose is to determine how many members each state will have in the House of Representatives. The size of the House has been fixed at 435 since Arizona and [Page p70]New Mexico joined the Union in 1912 (except for a temporary addition of two seats in 1959 when Alaska and Hawaii became states).

U.S. Bureau of the Census

After each census Congress uses another process, reapportionment, to redivide the 435 House seats among the states according to population. As the population shifts, some states gain representatives at the expense of others. Then the affected state legislatures use the census data for redistricting —redrawing congressional districts as well as state legislative districts to make their populations as nearly equal as possible.

Equality among districts within a state is required by the Supreme Court's one-person, one-vote decision handed down in Gray v. Sanders (1963). Until then, some rural-dominated legislatures had drawn the lines to make farm districts much smaller in population than urban districts, in effect giving city dwellers less representation in the legislature. In 1964 the Court applied the same one-person, one-vote standard to Congress in Wesberry v. Sanders.

New computer technology, as well as legislation requiring more equitable representation of minority groups, greatly affected the redistricting process. Some congressional districts became so weirdly contorted that they gave new meaning to the term gerrymander, the shaping of districts to benefit a particular politician, party, or minority group.

Efforts to get an accurate census count of African Americans and other minorities, and then link their neighborhoods through racial redistricting, have proven to be particularly difficult and controversial. One congressional district in North Carolina, the Twelfth, included nearly every black neighborhood in the 175 miles between Durham and Charlotte and at times was no more than a strip along Interstate 85.

Conventional census methods usually undercount minorities and the poor. Many in those groups do not receive or respond to the forms mailed out by the U.S. Bureau of the Census. The 2000 census, for example, missed an estimated 3.3 million people.

For the 2000 census, the Census Bureau proposed to use statistical sampling techniques—like those used in polling —to solve the undercount problem. Ninety percent of households would be covered by questionnaire, telephone, or census visit. Enumerators would visit one in ten of the remaining households, and the additional data would be extrapolated from those surveys. Such sampling, the bureau said, could reduce the undercount from 2.1 percent to one-tenth of 1 percent.

The Democratic administration of President Bill Clinton supported the plan, but the Republican-controlled Congress tried to block it. Republicans believed that sampling would benefit Democrats because minorities, the homeless, and young people away from home tend to vote for Democratic candidates. They contended that the Constitution requires [Page p71]an actual head count. Some members of Congress sued to stop the sampling and largely succeeded. The Court ruled 5–4 in January 1999 that sampling estimates could not be used for reapportionment. The Court, however, did not foreclose using sampling numbers for redistricting within states.

...