Census 2000 Main  |  Data Bank  |  Reference Room  |  Group Corner

Census 2000 Dress Rehearsal
 

John H. Thompson
Associate Director for Decennial Census
Bureau of the Census
U.S. Department of Commerce

The Census Bureau Decennial Directorate has reviewed the Dress Rehearsal Report Card - Evaluation of the Standards for Success - and concurs with its fundamental conclusions. The Decennial Directorate has taken steps to address the operational issues highlighted by the Dress Rehearsal and raised by the Report Card. On this basis, we are confident that Census 2000 will be a success. We offer the following additional observations about the areas considered critical for measuring the success of the Census 2000 Dress Rehearsal and how the results apply to Census 2000 operations.
 

Completeness of the Master Address File (Columbia, South Carolina and Sacramento, California)

The Report Card indicates that the Master Address File lacked completeness for areas with city-style addresses, particularly in the Columbia, South Carolina Dress Rehearsal site. The Dress Rehearsal Master Address File did not benefit from a 100-percent block canvass and the local review operations were complicated for the smaller governmental areas in Columbia, South Carolina. For Census 2000, we have reengineered the process for developing the Master Address File. We are currently implementing 100-percent block canvassing operations where census enumerators are reviewing all of the city-style address areas to ensure completeness of the address file. Additionally, we have greatly simplified the address list review process for local and tribal governments to help avoid the types of problems we experienced in the Dress Rehearsal. In summary, we believe these actions will result in a complete and accurate Master Address File for Census 2000.
 

Questionnaire Mailing Strategy (Columbia, South Carolina and Sacramento, California)

The Report Card shows that a second, or replacement, mailing of the questionnaire package to all households, regardless of whether they responded to the initial questionnaire mailing, did increase the overall mail response rate during the designated time window. However, it also confirms our decision to forego the second mailing. The Report Card indicates that approximately 40 percent of the households that mailed back a replacement questionnaire also mailed back an initial questionnaire. The Dress Rehearsal processing was extended by 3 weeks as a result of this volume of multiple returns. For Census 2000 a workload of this magnitude would significantly delay data processing operations and potentially introduce significant errors into the census data. Therefore, we do not believe a second mailing could be implemented without endangering the success and accuracy of Census 2000.

Paid Advertising (Columbia, South Carolina and Sacramento, California)

As indicated in the Report Card, all three Dress Rehearsal sites showed a dramatic increase in the public's awareness of the Census 2000 Dress Rehearsal after it had been conducted. It is difficult to attribute the increased awareness only to paid advertising when we employed a variety of marketing strategies such as media public relations, promotions and special events, and special methods to encourage response from direct mail. The levels of increase in awareness are encouraging and indicate the importance of all components of the partnership and promotion programs.
 

Followup for Nonresponse (Columbia, South Carolina; Menominee County, Wisconsin; and Sacramento, California)

Proxy Respondents

It has been very difficult to assess the precise level of proxy responses that occurred in the 1990 census. This is because the 1990 Census questionnaire did not include an indicator of who actually responded to the questionnaire. We have included such an indicator for the Census 2000 questionnaire so that we can measure this phenomena more accurately. Also for Census 2000, we are examining enhanced quality assurance procedures that we anticipate will reduce the level of proxy response.

Unclassified Cases

For the Dress Rehearsal, unclassified cases resulted from two sources - mail returns and enumerator returns. The largest component of the unclassified returns were from questionnaires that were mailed back. These unclassified cases resulted from problems with our data capture system. We have alerted our data capture system contractor, Lockheed Martin, of the problems. They have taken corrective actions that will resolve this component of unclassified cases.

Therefore, considering only the unclassified cases that resulted directly from the nonresponse followup (enumerator returns), we were much closer to meeting our standard for Census 2000. We believe that the new, enhanced quality assurance procedures will significantly reduce the level of unclassified cases.

Coverage Measurement (Columbia, South Carolina; Menominee County, Wisconsin; and Sacramento, California)

For the Sacramento and Menominee County Dress Rehearsal sites, it appears that the coverage measurement survey performed as expected within the time frame allowed.

For the Columbia site, it appears that the Dress Rehearsal fell below our projection of population size. We plan to carefully review the final results of the Dress Rehearsal Post Enumeration Survey to ascertain the causes.