Hi (again),
(Thanks for all your replies!)
Is it possible to get help compacting the censuses?
I have difficulties accessing sufficient memory capacity to work with the large populations. At the same time we are interested in the prevalence of the unusual combinations of characteristics so that we don’t want to work on subsamples.
I can do the analyses we need with a dataset with frequencies of individuals observed with different combinations of values across 12 variables. (See, e.g., the Stata command ‘contract’.)
We need such tables for 29 historical censuses and 65 censuses from Europe and North America.
I’d be happy to prepare code if that would be any help. If it’s possible for a fee that would also be of interest.
All the best,
Stefan