Hi,
I have stuck with simqi on a subsample after estsimp with multiply imputed datasets. The
datasets contain both males and females aged 20 and over from 1986 to 2002.
I am going to predict blood pressure (outcome variable) for a subsample of females aged
30-34 in 1986. I did estsimp on a subsample of females of all ages from 1986 to 2002, and
set independent variables at their means by setx for the subsample of females aged 30-34
in 1986. However, simqi derives predicted values not for these females but for a part of
males only (200 out of 90554 males of all ages in all years). I understand that
"if" command specified in estsimp or setx would be automatically considered in
simqi. I wonder if it occurs because simulated parameters were added to the first 1000
records in one of the multiply-imputed datasets opended befofre estsimp. But I cannot find
any solution. I would be absolutely grateful if you could please give me any idea.
Followig are simplified commands;
xi: estsimp reg bp i.age5 bmi i.exerc med i.year if sex==1, mi(out1 out2 out3 out4 out5)
setx mean if sex==1 & year==1986 & age5==30
simqi, genpv(bp)
(bp, blood pressure; age5, 5-year age groups; bmi, BMI; exerc, exercise (0/1); med,
probability of being medicated; year, survey year)
Thank you in advance for your help.
Sincerely yours,
Nayu
Nayu Ikeda
Visiting Scientist
INSTITUTE FOR HEALTH METRICS AND EVALUATION
University of Washington
Seattle, Washington, USA
--
Clarify mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.gking.harvard.edu/?info=clarify