Hi there,
I have created two piecewise graphs from linear mixed effects models. I now want to overlay them onto one graph.
Outcome is A1C.
Unit of measurement is months (pre = -12 to 0 months; post = 0 to 12 months).
Exposure is group (group==0 is the control group; group==1 is the intervention group). Study subjects only received the intervention at "post" (i.e. 0-12 months).
I get the following marginsplot graphs from the following code:
mixed a1c c.pre c.post i.group c.post#i.group if group==0 || id: , cov(unstr) reml
margins, at(pre=-12 post=0) at(pre=0 post=0) at(pre=12 post=12) vsquish
marginsplot, recast(line) ciopt(color(%20)) recastci(rarea)
Array
mixed a1c c.pre c.post i.group c.post#i.group if group==1 || id: , cov(unstr) reml
margins, at(pre=-12 post=0) at(pre=0 post=0) at(pre=12 post=12) vsquish
marginsplot, recast(line) ciopt(color(%20)) recastci(rarea)
Array
I want to combine the two graphs, so I used the following code:
mixed a1c c.pre c.post i.group c.post#i.group if group==0 || id: , cov(unstr) reml
margins, at(pre=-12 post=0) at(pre=0 post=0) at(pre=12 post=12) saving(file1, replace) vsquish
mixed a1c c.pre c.post i.group c.post#i.group if group==1 || id: , cov(unstr) reml
margins, at(pre=-12 post=0) at(pre=0 post=0) at(pre=12 post=12) saving(file2, replace) vsquish
combomarginsplot file1 file2, labels("Control" "Intervention")
Unfortunately, I get a graph that looks like this:
Array
How do I tell combomarginsplot that I want to plot the outcome against the original x-variable, rather than (File 1 vs. File 2)?
Thanks in advance!
Joseph Leung
Clinician Investigator Program PGY-7
University of British Columbia
Related Posts with Problem with combomarginsplot
Can I use diff-in-diff for my research setting?I want to use the legislation adoption at the state level to perform a diff-in-diff in my panel data…
How to perform this propensity score matchI have a panel data and want to examine the relation between the passage of legislation (LawDummy, i…
how to create equal sample size for all the regressionsHi, I have a panel data and will need to run several regressions that involve different dependent an…
Time invariant variables fixed effect modelHi, I have a panel dataset and I intend to run a fixed-effect model. I am having issues with one va…
Propensity score weighted and block balanceHi, I´m trying to use pscore weighted, so I'm using the command without matching, and having problem…
Subscribe to:
Post Comments (Atom)
0 Response to Problem with combomarginsplot
Post a Comment