Can't hurt.
So... does anyone know how to produce summary statistics, including means and modes, in STATA while using p-weights? I'm fairly decent with my chosen stats package but this is getting into that arcane realm that only a handful of people (who probably wear orange robes and live in caves in the Bay Area) are actually familiar with.
posted by Drek | 12:50 PM
4 Comments:
I did this a lot in a previous paper: append "[pw=WEIGHTVARIABLE]" to your command. Do that without the quotes, but with the brackets.
Ah, but you see, that's the trick. The "sum" command and most of its bretheren won't work with pweights- only fweights, aweights, and iweights.
Ah, I forgot about that. I'll poke around, as I seem to remember finding some way to get good summary statistics with pweights a while back.
I think you get the same results as what you would think pweights should produce if you use the iweight/aweight instead. There is a Stata FAQ on this topic (why can't you use pweights with summarize?). You can also use the svymean command.
Post a Comment
<< Home