I am interested in corresponding industry codes from the 1940, 1950, 1960, 1970, and 1980 IPUMS-USA into SIC 1972. I see from a prior forum (Crosswalk between IND1950 and SIC and/or NAICS?) that the best method to do this is to create my own crosswalk using the IPUMS Codebooks.
I see that the 1980 Codebook includes the associated 1972 SIC codes. However, the 1950 and 1960 codebooks do not indicate which series of SIC codes are listed. Are they the 1972 SIC codes as well? Further, the 1940 and 1950 Codebooks do not appear to have any corresponding SIC codes listed. Do you know where I may be able to find the correlations between these years’ industry codes and SIC?
Finally, it seems that it would be more straightfoward to create a crosswalk from IND1950 or IND1990 to SIC 1972, rather than converting the distinct IND codes for each decade. Is there any way to do this?
The SIC codes are actually updated every decade, however they did not change very much between 1940 and 1960 (a brief history of SIC codes can be found in the “History” portion of this technical document). However, as you mention, it would be much easier to correspond IND1950 with SIC codes from 1972 and then use the resulting crosswalk for all of the decades you are interested in. The second half of the answer you linked to actually provides a pretty good framework for generating such a crosswalk. Since you are interested in the 1972 SIC codes, you would want to download the 1980 census file with both the IND and IND1950 variables. You can then select a single case from each unique IND1950 code. The result will be a crosswalk of single, unique IND1950 values and their corresponding IND codes (there may be multiple IND values associated with a single IND1950 code). You will then need to find the correspondence between these IND codes and the 1972 SIC codes using the codebook from 1980. All together this will yield a crosswalk between the IND1950 codes (which are consistent across decades) and the 1972 SIC codes.
Thank you very much for your helpful response! That is the procedure I went ahead and did, and it worked out well. It’s good to have the reassurance that it’s the correct way to connect the data.