A Qualitative Approach to the Study of BLS Establishment Survey Nonresponse

Jean E. Fox, Sylvia K. Fisher, Clyde Tucker, Roberta L. Sangster, and Christine Rho


Recently, a trend toward increased nonresponse ingovernment surveys has resulted in concern about theeffect of nonresponse on data quality and statisticalestimates. Several recent studies have examined this issue in the context of describing nonresponse, and identifying ways to reduce nonresponse (Christianson, & Tortora, 1995; & Ware Martin, 1994). Concerns about declining response rates, both in household and establishment surveys, have been documented in the research literature (Smith, 1995; Atrostic, et al., 1999) and have been duly noted by government survey program managers. The Bureau of Labor Statistics (BLS) has always recognized the importance of nonresponse relative to data quality and the accuracy of statistical estimates for both its household and establishment surveys. In response to these concerns, BLS embarked upon an intensive study of nonresponse issues associated with four of its establishment surveys. The study was designed to learn more about the nature of establishment survey nonresponse; it focused specifically on nonresponse trends, causes of nonresponse, patterns in nonresponse, and possible solutions to nonresponse. The ultimate outcome of this research is the identification and implementation of improved data collection procedures that will address establishment survey nonresponse problems.