My question is about the sample size of the FSS. I’m trying to get my summary statistics to match the ones put out by the ERS in its annual summary, however, my sample size for FSS respondents seems to be about 5,000 people smaller than theirs for every year (I’m looking at only years since 2006).
Not sure if there are households IPUMS doesn’t let users download or if the ERS somehow inflates their sample size? It’s causing all my summary stats to be about 1 or 2 pp off from the official stats on food security.
This does sound a bit puzzling. Without a link to the exact tables and figures you are comparing your results to, it is difficult to exactly identify the source of the discrepancy. From what you’ve said here, it sounds like the issue is likely related to a universe definition or some other filter used when defining the key variables used to construct the figures. If the issue still persists after checking these details, feel free to email ipums@umn.edu with links and more detail about the tables you are trying to reproduce.