PREPARED STATEMENT OF
 
CHARLES LOUIS KINCANNON
DIRECTOR
 U.S. CENSUS BUREAU
 
Before the Subcommittee on Federalism and the Census
Committee on Government Reform
U.S. House of Representatives
 
6 December 2005
 

On behalf of the U.S. Census Bureau, I would like to thank the House Subcommittee on Federalism and the Census for inviting me to testify this morning. 

Today’s hearing focuses on the question of whether the decennial census should count, for purposes of apportionment, all inhabitants of the United States or more narrowly define its task to count only citizens. Our testimony this morning does not address the merits of the question, but focuses instead on the practical effects such a change might have on operational considerations and the accuracy of the census.

The Constitution and the Census Act of 1790

The census is one of our nation’s oldest activities. It is constitutionally required and is used as the basis to apportion the U.S. House of Representatives and to delineate congressional districts within the states. Article I, Section 2 directs that an “actual enumeration” was to occur every ten years. The first census law, the Census Act of 1790, was written just two years after the Constitution was ratified. The Census Act of 1790 instructed “the marshals of the several districts of the United States shall be, and they are hereby authorized and required to cause the number of the inhabitants within their respective districts to be taken.”

The basic instruction to count inhabitants living in the United States has not been changed by any subsequent census law. It is the foundation of the Census Bureau’s decennial census task. Counting every inhabitant living in the United States defines the scope of our operations and prescribes the need for accuracy in the count. The Census Bureau has developed rational, operationally feasible procedures in order to count every person and does not separately count the number of citizens, legal residents, visitors on temporary visas, or illegal immigrants, although the citizenship issue is addressed in the American Community Survey sample. To make such distinctions for purposes of enumeration would not only require changes to the decennial census questionnaire itself but different procedures and methodologies to attempt to obtain accurate information about residency or citizenship status from all respondents.

Operational Issues

The decennial census is a complex and daunting task. It requires unparalleled cooperation in order to reach every state, county, city, town, neighborhood, and street, in order to count every household in America. It is the largest peacetime mobilization undertaken by the federal government, involving years of planning and testing; hundreds of thousands of enumerators; billions in federal expenditures; and the cooperation of every household in America. Securing this cooperation from each community, household, and person is difficult, but crucial to the accuracy of the census.

With each modern census, the Census Bureau has documented decreasing levels of cooperation and response rates. The mail response rate determines the non-response follow-up workload and is an important factor in the cost and overall success of the census. In 1970, the overall mail response rate was nearly 80 percent. In 1980, it fell to about 75 percent, and by 1990 it had fallen to 65 percent. Census 2000 held at nearly 65 percent and seems to suggest that the Census Bureau, with the help of unprecedented congressional and community support, as well as paid advertising, was able to hold this trend in check. But we know the American public is becoming increasingly wary of issues such as protection of privacy, identity theft and unwarranted government intrusion. For the 2010 Census, we project further declines in public cooperation, even if we repeat all the efforts undertaken in 2000. We believe some of our efforts to reengineer the 2010 Census, especially the use of a short-form only census, can mitigate this trend. However, asking about citizenship could negatively impact this trend and could impair the ability of the Census Bureau to conduct an accurate census.

The Census Bureau relies not only on individual cooperation but also on public support to maintain the high level of accuracy expected from the decennial census. Public support is built and demonstrated through media coverage, public endorsement, and word of mouth. In 2000, there was an unprecedented effort to engage help from members of Congress, local officials, community organizations, schools, and the media to encourage public support. We relied on these partners to help educate the public, especially hard-to-count communities, about the constitutional requirement and the uses of census data in the distribution of public resources, and to assuage fears about participating and the possible use of data against respondents.

Asking all respondents for additional information relating to citizenship might raise additional fears about responding to the census. Even U.S. citizens may be wary of answering questions about citizenship. Countering these concerns and fears could require additional efforts to secure public cooperation. Moreover, because the census is conducted through self-enumeration, the Census Bureau could not verify whether the responses were accurate. The mere act of asking about the residency status of an individual may confuse or discourage respondents, even legal residents and citizens, and could affect the overall accuracy of the census.

Conclusion

Accuracy is important because the census is used to apportion congressional seats, fulfilling the obligation as outlined in Article I of the Constitution. Mr. Chairman, thank you for this opportunity, and I hope that this information is informative and will help the Congress as it considers this issue. I would be happy to answer your questions and concerns.