Lawrence E. Ernst and Chester R. Ponikowski (1998) "Selecting the Employment Cost Index Survey Sample as a Subsample of the National Compensation Survey," Proceedings of the Section on Survey Research Methods, American Statistical Association.
Abstract: A key aspect of the current integration of the BLS compensation surveys will be to select future establishment sample for the Employment Cost Index (ECI) survey as a subsample of the larger National Compensation Survey (NCS). Initially the transformation of the ECI to a subsample of the NCS will take place by initiating five subsamples of the current NCS sample over three years. This paper details our plans for the subsampling, which involves several interesting issues. For example, the NCS sample is actually two independent samples with some establishments selected twice, but the subsampling method insures that no establishment will be selected more than once in ECI. Also, the subsampling probabilities are determined with the goal that all establishments in an industry stratum in the universe with the same frame employment have the same overall chance of being in the ECI sample. Finally, we discuss the additional sampling complexities arising from the fact that the current NCS sample consists only of establishments with frame employment of at least 50, while the ECI has no minimum employment restrictions, necessitating supplemental ECI samples of the smaller establishments.
Last Modified Date: July 19, 2008