Recent Development in the Analysis of Repeated Survey Data
Edwin L. Robison
U.S. Bureau of Labor Statistics
The Current Population Survey (CPS) is one of the oldest, largest, and most well recognized surveys in the United States. It produces monthly household information about national labor force statistics of the civilian non-institutionalized population. Monthly nationwide unemployment rate is a key economic indicator. The CPS adopts a rotation sample design, which consists of a sample of eight rotation groups. Households in a rotation group are interviewed for 4 months, dropped for 8 months, and then interviewed for an additional 4 months. Currently, the U.S. Census Bureau and Bureau of Labor Statistics (BLS) have been conducting a systematic and comprehensive evaluation of the survey methods used in the CPS. In addition, we collaborate with Academia to work on the wide variety of researches related to rotating sample designs. Ongoing research in this context motivates our proposed Invited Session. Professors Eric Slud, Jun Shao, and Partha Lahiri will present their research involving theoretical and simulation-based validation for recently developed methods of generalized-raking calibration methodology, a new composite estimator with less bias, and multi-purpose bootstrap replicate weights in order to conduct different statistical inferences. Dr. John Eltinge will discuss these researches related to CPS and provide some suggestions on the alternative methods for our future research plans.