Does anyone have a method for handling missing data in CPS "EDUC" variable? 24 % missing.


#1

Too many missing cases in EDUC: 24 percent over 1968 to 2017. Are people inputing data?


#2

Note that cases with EDUC==001 “NIU or blank” all have an AGE<15 (in ASEC samples since 1980) and AGE<14 (in ASEC samples from 1962 through 1980). This corresponds to the definition of the universe defining who is eligible to respond to the EDUC question on the CPS questionnaire. Therefore, these “missing” cases are not due to item non-response, but rather due to the fact that some individuals are not eligible for the EDUC question. For analysis using the EDUC variable, you can simply drop all cases ==001 “NIU or blank” or limit the population as defined by the universe statement.