Question about CPS weight



I recently used multiple-wave CPS data to estimate poverty status in the US. I consider to apply WTSUPP variable in my analysis because it is a person-level data set and, as IPUMS suggests, most of person-level data should be used this weight. However, I am not very sure about it. I have three general questions:

Do the poverty variables come from the CPS ASEC data? Or these variables come from CPS monthly survey (e.g., CPS March)?

Do the poverty variables, e.g., OFFTOTVAL and OFFCUTOFF, be constructed by earning-related variables such as EARNWEEK or HOURWAGE?

If I would like to make crosstables and get correct estimates for both demographic variables and poverty variables, and also if these variables, by chance, should apply two dfifferent weight, e.g., WTFINL and EARNWT, what kind of weight should I apply?

Thank you so much!


The Poverty variables are only available as part of the ASEC sample and should therefore use the WTSUPP weight variable. These poverty variables are constructed using the annual income and earnings figures reported as part of the ASEC supplement questionnaire, so you do not have to worry about needing to use the EARNWT weight variable. If you did happen to need to use a variable from the earnings supplement, you would want to use the EARNWT weight variable, as only individuals who participated in the Earner Study questions would have valid data.

I hope this helps.