I’m working on a project to examine the employment trend among workers with disability. My questions is very specific: for matched CPS ASEC (March CPS) samples from adjacent years, in order to compute statistics at the national level, which weight variable is the most appropriate to use?
It seems that my choice of weight variables for the matched ASEC data is either ASECWT or LNKFW1YWT (longitudinal weight for adjacent years). Am I correct or are there better weighting variables specifically designed for matched ASEC samples from adjacent years?
Thank you for any help you could provide!
You’re right in that either ASECWT or LNKFW1YWT are your choices for this match. Which one you use depends on what you are looking at, and in particular which year is the more natural “base period.” For example, if you are interesting in calculating flows out of the labor force, LNKFW1YWT is the correct one to use, since it makes the linked data representative of the first year. There is no such weight looking backward (which would be ideal for studying flows into the labor force), but you can use ASECWT without too much trouble. Just be cautious for calculating totals , since ASECWT doesn’t adjust for non-linkage. This page gives more info on the longitudinal weights in IPUMS CPS, including Stata code used to generate the weights. You can modify this code to make backward looking weights.