[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Inequality indices with complex survey design (Was: Problem to run Inequal7)
Dears Stephen P. Jenkins and Stas Kolenikov,...
My problem is not theoretical but computational (i dont have any
response of inequal7 and I need to use ctrl+alt+del to stop stata9 or
the program run all the life) - I have Stata 8.2 and 9.2. Im working
with a complex sample (National Household from Brasil - PNAD, itīs a
complex sample with cluster, stratum and PSU) but itīs not necessary
to calculate point estimates ( complex design itīs only important to
estimate variance or unpoint estimates).
I know a lillte bit about complex sample (to estimate variance in
general - I could use linearization (Taylor, Rao) or replication
(Bootstrap, Jackknife)) and some ado/do files to estimate
poverty/income distribution (povdeco, ineqerr, inequal, inequal2,
inequal7, sepov, ineqdeco or svy: svylorenz).
I do a ado file to run point estimate with Inequal - itīs ok but I
would like to estimate Theil(a) and because this I tried to use
inequal7. The syntax of inequal and inequal7 is very similiar but the
version that I have (Stata 8.2 and 9.2) - dont run inequal7 (the
problem itīs not with . or zero - my ado file change this value to
zero or next to zero).
Itīs not memory (when this happen - the programa dont run) and I have
the same problem when I try to eliminate unit with only one PSU (for
replication or linearation methods there are a theory that say to
estimate variance - the psu could not have with only one observation).
Thanks for the help...
* For searches and help try: