Multilevel logit model command

[Amber Pong]

I have a sample dataset where individuals were randomized into a clinic and given placebo or drug. The outcome is binary, whether their symptoms reduced or not after being given the drug or placebo. Symptom is 1 = Yes, 0= No and drug is 1= drug and 0=placebo. There are five different clinic, agegroup is 1=younger, 2=middle, 3=older, and sex is 1=male, 2=female. I’m trying to run a multilevel logit model with agegroup and sex as covariates.

clear
input int(symptom drug clinic agegroup sex)
1 1 2 3 1
1 0 1 2 1
1 0 5 1 2
0 0 3 1 1
0 1 4 3 2
1 1 1 2 1
0 1 2 2 2
1 0 4 3 2
0 0 3 1 1

[Ken Chui]

Try help melogit

[Amber Pong]

Thanks for the guidance Ken Chui. I looked at help melogit and I'm wondering if I have the syntax right

 melogit symptoms drug agegroup sex || clinic

[Joseph Coveney]

. . . I looked at help melogit and I'm wondering if I have the syntax right

You'll probably want the additions shown below in red.

melogit symptoms i.(drug agegroup sex) || clinic:

First, you're missing a colon after the random effects variable.

In addition, all of your predictors are categorical, and only the drug treatment predictor is coded zero-one. And age group has three categories.

help factor variables

You do have far more data than what you show, right?

[Amber Pong]

Yes there are more data than what I showed. Thank you Joseph