Wage Imputation in the OES Survey: A Model‐Assisted Approach Incorporating Data from the Quarterly Census of Employment and Wages

Jane G. Osburn


Wage imputation in the BLS Occupational Employment Statistics (OES) Survey requires a process that matches non-respondent and respondent establishments on characteristics that best predict the wage levels of the non-respondent establishments in a given MSA /Industry /Establishment Size cell. The mean wage distribution of the establishments in the donor cell is then imputed to the non-respondent. Currently, the OES Survey procedures first define the donor cell for a given non-respondent establishment /occupation cell by the same time period /MSA / industry (four or five digit NAICS) /establishment-size as the non-respondent, and cells are then collapsed across industry and size groupings in the case of insufficient response. If insufficient response still exists after collapsing across industry and size groupings, the base level strata default from MSA to State, the donor cell is reformed by State / four/five digit NAICS Industry / Establishment Size, and the process repeats as for the MSA base-level strata.1 The baselevel strata eventually default to Nation in a similar manner.