I’m looking at labor market flows (between full-time, involuntary part-time, voluntary part-time, unemployed, and not in the labor force) over a four month period using either lnkfwmis14wt or lnkfwmis58wt (depending on whether the the individual’s mis cohort is 1 or 5 at the start of the 4 month period) to get an idea about how many people the sample represents.
My question is how many people should make up the sum of these 2 cohorts, Should it equal the working-age population or should it equal a fraction of the working-age population?
I understand you are looking at a single 4-month period. Suppose you are looking at Jan-Apr, 2021. Your two cohorts consist of individuals in either MIS 1 or 5 in Jan 2021. That makes up 1/4 of the total CPS sample in that month. Note that LNKFWMIS14WT and LNKFWMIS58WT will only be nonzero in that month for people in those cohorts. The longitudinal weights of the linked individuals in the MIS 1 cohort will sum to the total of the weights in the full MIS 1 cohort (containing both linked and unlinked individuals). That’s about 1/8 of the total population (and again for MIS 5). So summing the longitudinal weights for both cohorts in January will give you about 1/4 of the working age population.
Thank you Matthew. That’s what I suspected after looking at the numbers. Thanks for the conformation.