This article describes a case study on the potential of using smaller geographical units in an area probability design, and reports the challenges of collecting a nationally representative sample for this hard-to-reach population. The Census Integrated Communications Program Evaluation (CICPE) was designed to evaluate the promotional campaign’s effect on Decennial Census participation for six race/ethnicity groups of interest. A nationally representative Core sample was designed to collect interviews for Hispanics, non-Hispanic African-Americans, and non- Hispanic Whites. However, it was impractical to include the rarer Asian, American Indian and Alaska Native (AIAN), and Native Hawaiian and Other Pacific Islander (NHOPI) populations in the Core design. For the Asian sample, we designed a separate area probability sample.
Traditional area probability sampling designs use counties or metropolitan areas as first-stage units, but smaller geographical units can better target hard-to-reach populations. The CICPE Asian sample used cities as the first-stage units.