Replicating lnkfw8wt in Stata

CPS
1

I downloaded two extracts from January 2019 (MIS 1) and April 2020 (MIS 8) and attempted to replicate lnkfw8wt using the Stata replication file available here: IPUMS CPS.

In the .do file, I replaced “mish!=4 & mish!=8” with “mish==1” since the only respondents who are eligible to match across all rotations are those in MIS 1 in January. However, the final_raked values did not match with lnkfw8wt. What else do I need to edit in the Stata replication file to create a 16-month longitudinal weight instead of the 1-month longitudinal weight? Can you share more about how lnkfw8wt was generated?

2

It sounds like you are trying to link persons between their first and eighth month in sample only (e.g., you do not care if they link across all months of the CPS panel–only the first and eighth month). Note that LNKFW8WT is for linking persons across the full panel (e.g., who are observed in all eight months). Accordingly, I wouldn’t expect your weights to match. Please email your complete .do file and a description of your linkage criteria to ipums@umn.edu so we can provide more targeted feedback