DEMPOP
1
Hi Joe,
I am looking to compute median income for different racial groups from the ACS, while using svy. Stata doesn’t seem to allow me to do:
svy: bysort race: sum inctot, de
or
bysort race: sum inctot [pweight=perwt], de
How do you recommend I get medians while using the svy command? Thanks!
Ps- my apologies for contacting you directly, but I am getting pretty anxious.
findit epctile
* should lead to my webpage
epctile inctot, svy percentile(50) over(race)
HTH, Stas
DEMPOP
3
Thanks for the epctile tip. I used it, and I got:
Mata run-time error
r(3499);
Is there a bug?
There is no Mata code in epctile.ado, so this must be some sort of version incompatibility. Can you post the full output?
DEMPOP
5
Sure,
epctile inctot, svy percentile(50) over(asian1g)
(running mean on estimation sample)
Mata run-time error
r(3499);
Thanks again
OK, didn’t help :-\
set trace on
set tracedepth 2
which epctile
epctile again
Send to the address in the header of the epctile.ado file which would show up when you “which” it.
Stas
DEMPOP
7
When I “Which” it, I get:
c:\ado\plus\e\epctile.ado
*! percentile as estimation procedure
Where shall I send it exactly?
DEMPOP
8
Also, it seems to think that “again” is a variable, it says,
variable again not found.
Hm. This is probably a REALLY old version of the file, then.
skolenik at gmail dut com
I have been able to update the .ado file on my website, so try
net from http://members.socket.net/~skolenik/s…
net d epctile
net install epctile, replace
It should show something like
. which epctile
c:\ado\plus\e\epctile.ado
*! percentile as estimation procedure v.1.4 by Stas Kolenikov, skolenik at gmail dot com
If it still does not work, I’ll ask for more detailed debug info.