No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • HELP: How to get predicted dependent variables after a xtmixed regression (not overall margins!)

    Hi all,
    in a group of alcoholic patients (segregated by Sex), I've measured the association between mean brain volumes (dependent variables) and alcohol dosage (i.e., alc_sqrt), while controlling for the effect of head size (i.e., ICV), age and education. I've used a mixed effect model with Site as the cluster variable because data have been pooled from several research Sites.

    Here is the code:

    foreach var of varlist LeftCerebellumWhiteMatter - CC_Anterior {
    bysort Sex: xtmixed `var' ICV Age Education alc_sqrt || Site:

    For the brain regions where I've got a significant association with alcohol dosage, I would like to make two-way scatter plots to illustrate that association graphically. BUT I want to use another software. So what I do need are the mean brain volumes for each participant adjusted for ICV, age, education. Does it make sense to you?

    I replicate the code but not including alcohol dosage

    foreach var of varlist ... {
    bysort Sex: xtmixed `var' ICV Age Education || Site:

    Now, How can I estimate the adjusted mean brain volumes for each participant?

    Thanks in advance for your help!
    Last edited by Maria Gloria Rossetti; 12 Jul 2018, 09:21.

  • #2
    -help predict-


    • #3