|
Abstract:
|
Since 1970, the U.S. Census Bureau has implemented an experimental program as part of the decennial census to evaluate a variety of alternative methodologies and questionnaire design strategies. For Census 2010, a robust program is planned that includes an ambitious test of race and Hispanic origin question changes, an alternative address collection for improved within-household coverage, and a 2000 Census short form-style form. In addition, other treatments include a deadline message with and without a compressed mailing schedule, and two alternative privacy messages. Finally, a nonresponse follow-up contact strategy experiment will examine the feasibility of reducing the number of interviewer contact attempts in order to save costs. This paper presents the sample design for each of these experiments, and includes a discussion of the unique challenges in achieving the optimal design.
|
