Hi Mr Bloem,
You previously (in March 2017) answered a question about erroneous 1920 census complete-count data variables DWELLING and DWSEQ: Unreliable DWELLING and DWSEQ in 1920 complete data?
It seems these variables are still erroneous. Downloading the 100% 1920 data with serial and dwelling selected show many, many family “serials” at each “dwelling”, consistent with what the previous user found. I was hoping to inquire as to whether a correction for this was on the horizon, as I am considering a project for which this information is important.