Using Internal Current Population Survey Data to Reevaluate Trends in Labor Earnings Gaps by Gender, Race and Education Level
36 Pages Posted: 29 Jul 2009
Date Written: July 1, 2008
Most empirical studies of trends in labor earnings gaps by gender, race or education level are based on data from the public use March Current Population Survey (CPS). Using the internal March CPS, we show that inconsistent topcoding in the public use data will understate these gaps and inaccurately capture their trends. We create a cell mean series beginning in 1975 that provides the mean of all values above the topcode for each income source in the public use March CPS and better approximate earnings gaps found in the internal March CPS than was previously possible using publically available data.
Suggested Citation: Suggested Citation