I’m trying to download annual poverty CPS data from 1990-2013. The data needs to be stratified by race/ethnicity - specifically non-Hispanic black, non-Hispanic white, and Hispanic. I would like the data at the county level if possible. What variables do I use to do this? Thanks!
IPUMS-CPS provides microdata, which means the data available to download include one observation for each individual respondent. You could then calculate annual poverty levels for specific subgroups and geographies by manipulating this microdata in a statistical package such as SAS, STATA, or SPSS. Keep in mind that you will need to weight your sample data to produce estimates representative of the population.
For the race/ethnicity variables, you will want to include RACE and HISPAN in your data extract. As for geography, it is possible to create county-level statistics by using the COUNTY variable; however, there are two drawbacks. First, due to confidentiality protections, more than half of CPS households are located in an unidentified county. Second, the COUNTY variable is not available prior to 1996. As an alternative, you can identify the state of residence for all sample households.
Finally, you have two choices for determining a respondent’s poverty status: POVERTY and OFFPOV. The variable POVERTY is the original poverty status variable found in the CPS PUMS file. In contrast, OFFPOV is the official poverty status that should be used in order to replicate official poverty rates. You can find a detailed explanation of the differences between these two variables in this User Note.
Hope this helps.