Missing Data in Imputed Highest Grade Completed in the 2015-2018 NBER CPS Extracts
16 Pages Posted: 9 Sep 2019
In 2015, the Current Population Survey (CPS) eliminated three questions related to educational attainment. These questions are used by the NBER to calculate the variable, "Imputed Highest Grade Completed" (ihigrdc) in their Monthly Outgoing Rotation Group (MORG) extracts. Imputed Highest Grade Completed provides a convenient measure of years of education and is based on the credential oriented CPS variable that is coded from 31 to 46. Because the NBER continues to use coding that relies on these eliminated questions to calculate ihigrdc, this variable has a missing value for 27.5% of the observations in the 2015-2018 extracts. These missing values, in turn, lead to an average Imputed Highest Grade Completed of about 1.2 fewer years after 2014 than would result from using the whole CPS. We informed the NBER of this issue in January 2019; however, because this variable with missing values remains on their website for download as of this writing (July 15, 2019), we are posting this working paper to inform users of these data so that they may address the issue appropriately in their own work.
Keywords: National Bureau of Economic Research data, monthly outgoing rotation group, education, current population survey, missing data, coding errors, NBER
JEL Classification: C18, C8, C810, J00, J01, I2
Suggested Citation: Suggested Citation