Use of Overlap Maximization in the Redesign of the National Compensation Survey

Steven P. Paben, Lawrence R. Ernst, and Yoel Izsak


Among the key products produced by the National Compensation Survey (NCS), which is conducted by the Bureau of Labor Statistics, are locality wage surveys. Consequently, a sample redesign of the NCS is necessary each decade following the release of new metropolitan area definitions. NCS uses a rotating panel design, with three stages of selection, namely: geographic areas; establishments, and occupations. We apply an overlap maximization procedure derived by Causey, Cox, and Ernst (1985) in the selection of noncertainty metropolitan areas in the NCS redesign. This procedure uses transportation theory to obtain an optimal overlap. Despite the optimality of this procedure, it has rarely, if ever, been used in production because it often produces transportation problems that are impractically large. For our NCS application, however, the transportation problems will be of manageable size. We present the expected overlap results for this procedure and compare them to the results using other overlap procedures, such as Perkins (1970) and Ohlsson (1996). We explain how we handle in the overlap procedure areas whose definition has changed in the new design.