i think you should probably do this all in R, but what
clarify does is to
create 1000 sims using however many datasets you have. it combines them
for you, and assumes you only need 1000 sims of your QofI.
We did do it all in R, but have not been able to calculate the robust
standard errors for a logit model in R. Hence, this last minute attempt
at simulating results through Stata.
you can
increase the number of sims but don't need to
only because you have
multiple imputations.
Gary
Ok. So I can use sims[5000] to increase the number of simulations, but
it is not necessary.
Thanks,
Nirmala
On Fri, 9 May 2003, Nirmala Ravishankar wrote:
I am trying to estimate the simulated betas for a
logit regression with 5
imputed datasets. Instead of getting 5000 simulated values (1000 for each
dataset, which is the default), I only get 1000 in total. Here are the
steps I followed:
use "C:\imp1.dta", clear
estsimp logit winner racebl racehi raceas inchi incmed, robust
mi(C:\imp1.dta, C:\imp2.dta, C:\imp3.dta, C:\imp4.dta, C:\imp5.dta)
At this stage, when I do "sum", it lists the newly created variables along
with the variables in imp1.dta, and there are only a 1000 observations.
Ultimately, I would like to "rbind" all the coefficients from the 5
datasets, and export them to R. How do I do that?
Thanks,
Nirmala
**********************************************************************
Nirmala Ravishankar
First Year Graduate Student,
Government Department,
Harvard University
_______________________________________________
gov2001-l mailing list
gov2001-l(a)fas.harvard.edu
http://www.fas.harvard.edu/mailman/listinfo/gov2001-l