survey participation, nonresponse
The lower than expected mail response to the 1990 Decennial Census increased costs and has been cited by some observers as cause to re-think the entire census design for 2000. This paper discussed results from two surveys that help to account for the low mail response. One, the Outreach Evaluation Survey, was conducted by the Census Bureau primarily to evaluate the effect of programs designed to heighten awareness of the census and to explain its uses and purposes. The survey comprised two waves of interviewing: one in the winter of 1990 prior to the start of outreach activities, and the other in late April and early May. The other survey, the Survey of 1990 Census Participation, emerged from a Joint Statistical Agreement between the National Opinion Research Center and the Census Bureau, to measure a variety of characteristics that might be related to census mail response. Interviewing for this survey occurred primarily in June and July.
This paper reports findings of a preliminary analysis of the data from the two surveys, with particular emphasis on examining correlates of respondents’ self-reports of census mail response. The paper presents and discusses the observed association of each of several survey variables with census response, with attention to interactions with race and ethnicity. Comparisons to similar surveys 1980 provide a context to assess how changes over time may have brought about the decline in response.
Fay, Robert E., Nancy Bates, and Jeffrey Moore. (2010). Lower Mail Response in the 1990 Census: A Preliminary Interpretation. Statistical Research Division Research Report Series (Survey Methodology #2010-13). U.S. Census Bureau. Available online at <http://www.census.gov/srd/papers/pdf/rsm2010-13.pdf>.