I am attempting to create a new variable representing the individual-level (ftype=2) poverty threshold which varies by year, conditional on year, sex, age, and farm status. While I can find the actual threshold for each combination of these variables by going to census webpage, I can’t apply them to microdata because the variable “farm” does not exist between 1968 and 1976. Is there some way to construct a “farm” variable for those years so that I can apply the individual-level poverty threshold to the microdata CPS extract?
Unfortunately, it is not possible to replicate the FARM variable for the ASEC samples between 1968 and 1976. One approach would be to use the variable INCFARM for householders and use the “$1000+ in farm sales” criteria used beginning in 1989 (since acreage is not available). However, this does not allow you to restrict to rural households. Ultimately, it is the researcher’s decision if such a constructed FARM variable is sufficient.
I hope this helps.