To identify all workers by industry, earning 200% FPL, how do I identify all household members?

I’m seeking all workers by industry earning 200% FPL and below, for Georgia in the 2020 ASEC. I presume that I need to identify each household member in each household size scenario in the Census Poverty Threshold, and then join(merge) all of that data together for analysis. In a one or two-member household with no children, I think I can id the householder and spouse/partner. But in larger household sizes with unrelated/related sub-families, identifying the third, fourth, etc. household member doesn’t seem possible by using the PERNUM or NUMPREC variables. (Keep in mind that I have not downloaded my samples and variables and put them into R yet, I’m trying to map out my data gathering methodology first)

To more specifically describe my process, here’s my data gathering method to filter for a 200% FPL, 2-member non-married household, w/ no children, householder below age 65:

For Householder: FTYPE code 2 (non-family householder), AGE < 65, NUMPREC code 2 (household size of 2), NCHILD code 0 (no children), OFFTOTVAL <= 31,288

For Partner: FTYPE code 5 (secondary individual), NUMPREC code 2 (household size of 2), NCHILD code 0 (no children), OFFTOTVAL <= 31,288

But this type of filter won’t seem to id additional sub-family household members. What do I do?

Can you clarify a few things about your approach?

First, I’m not sure if you want to calculate:

  1. The number of individuals in each industry who personally make less than 200% of the poverty line, counting only that one person’s income, where the poverty line for each individual is defined by their family size.


  1. The number of individuals in each industry who are members of families collectively making less than 200% of the poverty line, counting all the family members’ incomes.

Second, can you clarify your unit of analysis for calculating poverty? The official census poverty calculations use a definition which counts the primary family and all related subfamilies as a single family unit. You can also treat related subfamilies separately if you wish. Differences between these methods are detailed on this page. Using households (instead of families) as the unit of analysis is less common, but you could also do it this way.

Third, how do you want to define family units? This is only relevant if you use families, not households, as your unit of analysis. There are different ways of defining family units (apart from the primary/subfamily distinction). There are two definitions currently available through IPUMS. The official Census Bureau definition is captured in the variable FTYPE and related variables. The other definition is based on IPUMS-derived family pointers, and is captured in FAMUNIT and related variables. A major feature of this definition is that it counts unmarried partners as part of the family, and you can read more about the differences at this page.