Hello, I am trying to compare the NHIS 2018 to the CPS ASEC 2019 (which gathers data from year 2018). In particular, I need to compare the industry of those who report working. I know the classifications differ across datasets but I was trying to find the variable in each dataset most suited to this purpose. If I understand it correctly, IND90LY from the CPS ASEC 2019 and IND1995 from NHIS 2018 would be best in this regard. Am I correct? If not, please advice on which other variables to use (if you know them!).
Please also note that, for the comparison, I do not need detailed industry classification. I want to just do some tabulations at a very aggregate level (think, for example, of the industry used in the BLS monthly job report such as Manufacturing, Wholesale, Retail Trade etc.). So, I am not interested in precision at, say, the 3-digit level, as long as the larger groupings are not too far off.
And, yes, I am aware that the universe in the two samples is different but maybe that’s a topic for another thread, here I am only asking about the industry variable. Thanks!
The harmonized industry variables (IND90LY in IPUMS CPS and IND1995 in IPUMS NHIS) are useful for making comparisons across time periods that employ different industry coding schemes. However, it sounds like you are only interested in a single year of data. To that end, the IND variable (which has the same name in IPUMS CPS and IPUMS NHIS) is more appropriate as it will provide the most year-specific detail. From each of the IND pages I linked previously, you can click to see the specific lists of IND codes for the year of interest. The codes differ both in detail and in the scheme from which they are derived. The NHIS codes are informed by the 2012 North American Industrial Classification System (NAICS), while the CPS codes reported in the data are the 2012 Census industry codes (these are derived from the 2012 NAICS as well though). Given your characterization of the broad aggregations you are interested in, this probably won’t make a difference. The IPUMS CPS page does not classify the 2012 Census industry codes into the groupings you mentioned, but this IPUMS USA page does (note that the substantive codes should be the same, but the missing data codes or handling of persons in the armed forces may differ).