I was told to use weight EARNWT for variable WHYLEFT. These two variables don't overlap in terms of availability.


I asked a question recently about variable WHYLEFT. Joe Grover mentioned that “most variables that are restricted to Rotation Groups 4 and 8 (including WHYLEFT) are actually part of the Earner Study and should be analyzed using EARNWT weight variable”.

The issue that I am running into is that WHYLEFT is available from 1976 to 1993, while EARNWT is available from 1990 to 2013.

What weights should be used if looking at variable WHYLEFT prior to year 1990?

Thank you,




There were earner study weights created for the March Supplement prior to 1990, but these are generally considered to be unreliable, thus they were omitted from IPUMS-CPS. The Census Bureau recommends using the March Supplement weights (WTSUPP) multiplied by 4 in place of the earner weight for these years.

I hope this helps.