Hello I have performed a latent class analysis
This was run on a dataset which consists of procedure == 1 (treatment) and procedure ==0 (control) variable, which have been weighted using IPTW and have pweights
I want to find out how the -earlyimprovers CLASS 1 - varied to the -late improvers CLASS 3- for both treatment and control; procedure ==1 and procedure == 0
I then proceeded to plot graphs:
This plots the data for both treatment and control
1. Is there a way to plot the graphs for PROCEDURE ==1 treatment only
and overlaying the same graph for procedure==0 control
I understand the ... -if- command, eg
if procedure== 1
or
if procedure == 0
does not work with -gsem-
2. Is there a way to take into consideration the already produced -weights- ?
Code:
gsem (q1 q2 q3 <-, ologit) (C <- preq1 preq2 preq3), lclass(C 3)
I want to find out how the -earlyimprovers CLASS 1 - varied to the -late improvers CLASS 3- for both treatment and control; procedure ==1 and procedure == 0
I then proceeded to plot graphs:
Code:
//Class 1 (this consist of both treatment and control)
margins, predict(outcomeq1) class(1) ///
predict(outcomeq1) class(1)
predict(outcomeq1) class(1)
marginsplot, recast(line) title ("Class 1" title("") ///
xlabel(1 "q1" 2 "Pain" 3 "kneel" 4 "walk", angle(45)) title ("") ylabel(0(25)125) name(class1)
1. Is there a way to plot the graphs for PROCEDURE ==1 treatment only
and overlaying the same graph for procedure==0 control
I understand the ... -if- command, eg
if procedure== 1
or
if procedure == 0
does not work with -gsem-
2. Is there a way to take into consideration the already produced -weights- ?
