I’m using the following code to weight the 2020 ASEC-CPS, to check that I weighted the data correctly. I compared population total to the 1-year ACS total; and it was completely off. Did I weight the data incorrectly? Or should I expect a difference? If so, how much of a difference is okay?
data_rewgt ← svrepdesign(data = data, weight = ~ASECWT, repweights = ‘REPWTP[0-9]+’, type = ‘JK1’, scale = 4/160, rscales = rep(1, 160), mse = TRUE)
What population totals did you calculate using ASEC and ACS, respectively? They should be very close. The code you posted looks correct to me, based on this thread.
I grouped age by 12-17, 18-64, and 65+. I got: 19,267,156. This is for Illinois, the 1-year 2019 ACS is 12,671,821. The estimates are very different.
For context, I am only using IPUMS-CPS, ASEC 2020. I don’t expect an exact match but these estimates are off by a lot.
Using the IPUMS online data analysis system, here are the totals I calculated for Illinois:
2019 1-year ACS:
They are very similar. Your R code for svrepdesign() looks correct, so I’m guessing the issue is somewhere else in your code.
I found the issue! Thank you!