Sample Design And Estimation For The National Compensation Survey

Susan R. Black, Lawrence R. Ernst, and Jason Tehonica


The National Compensation Survey (NCS) is the new BLS survey for measuring employee wages by skill level. This paper discusses the sample design, estimation and variance estimation procedures for this survey. The survey uses a rotating panel design, with three stages of selection used in selecting each sample panel, namely: geographic area PSUs; establishments selected from industry strata; and occupations selected separately from each sample establishment. The weighting incorporates several stages of adjustment, including three stages of nonresponse adjustment. Among the issues discussed are a new method for weighting the occupational selections which correctly adjusts for employment changes over time; a method for redistributing the weights of nonrespondent establishments to respondent establishments that is different from the approach commonly used; and the allocation of the sample in each industry strata among noncertainty PSUs in a way that ensures that establishments in different PSUs with the same sampled employment have the same unconditional probability of selection.