The American Community Survey is a new survey that collects socioeconomic and demographic data describing population and housing in the United States. The current design of the American Community Survey employs three modes of data collection. Questionnaires are mailed to sample addresses. Nonrespondents are followed up using both Computer Assisted Telephone and Personal Interview methods. It is hoped that through offering an Internet response option some of the respondents who are currently enumerated by mail will respond via the Internet. This could reduce total survey costs. An Internet response option also has the potential to improve the quality of self-response data.

For this Internet Response Test, we selected a national sample of about 10,000 addresses. In the months of November and December of 2000 and January of 2001, these addresses received a mailing package similar to the standard mailing package used in the American Community Survey. These experimental packages included an invitation to respond via the Internet in lieu of mailing back the enclosed questionnaire. The primary objective of the Internet Response Test is to assess the impact of such an invitation on response rates. In addition, the test evaluates the comparative quality of Internet versus mail responses.