Karen L. Goldenberg, and Diane K. Willimack (2003) "Measurement Differences in Key Economic Indicators." Proceedings of the Section on Government Statistics, 2003, American Statistical Association.
Economic indicators such as employment, earnings, and work hours may be commonly understood terms, but as collected in U.S. government statistical surveys they take on highly specific meanings. Operational definitions (and sometimes underlying concepts) vary from survey to survey and across statistical agencies, and present different cognitive and recordkeeping demands for respondents. What happens to the "same" measures when they are collected and reported in different surveys? This paper examines the economic indicator concepts of total employment, production worker employment, work hours, and payroll as they are operationalized in self-administered establishment surveys conducted by the Census Bureau and the Bureau of Labor Statistics. We contrast questionnaires and collection procedures for the key variables across both surveys, and then compare published estimates of the same variables to assess whether they are statistically equivalent. To the extent that there are differences, we consider whether those differences can be attributed to the structure and the content of the survey questionnaires, and we discuss the implications of our research findings.
Last Modified Date: July 15, 2004