01-21-2016 02:52 PM
Suppose to have a Dataset with the following variables Simulation (1 to 1000),Study (1 to 5), Treated (Dummy variable 0 not treated, 1 treated), Outcome (1 to 3), Outcome1 (Dummy variable 0 is not measured, 1 is measured), Outcome 2 (the same of Outcome1), Outcome3 (the same of Outcome1 and Outcome2), ID: identification of patient, Y: response variable.
Each studies has a sample size of
Study 1: 60:
Study 2: 80;
Study 3: 100;
Study 4: 120;
Study 5: 140.
I am trying to run this proc mixed:
proc mixed cl method=reml data=work.Datasim2;
by Simulation Study;
ods listing close;
model Y = Outcome1 Outcome2 Outcome3/ noint s cl;
random Treated*Outcome1 Treated*Outcome2 Treated*Outcome3/ s cl;
To analyse the effects of the Treatment between the two groups of patients (Treated = 1 and Not Treated = 0) for each Simulation and Study.
But it seams to not work properly this proc mixed as
1. The DF are to high;
2. i would like to obtain an effects of the Treatment in terms of mean difference so I would like to obtain negative estimates.
It is clear the question?