Hi all,
I have a dataset with treatment variable and choice variable. I am trying to understand the persistence of the choice variable using -cmxtmixlogit-. I am using -cmxtmixlogit- instead of -xtmlogit- because I have unbalanced choice set, e.g., sometimes, it is 2 available choices, sometimes it is 3.
Here is the code I have now
I want to check if there is any difference in persistence between the treatments, but because treatment is invariant with alternatives, I cannot do
I could do
I am just wondering if this regression makes any sense.
I cannot do separate regressions for the treatments, because of limitations about control variables.
I have a dataset with treatment variable and choice variable. I am trying to understand the persistence of the choice variable using -cmxtmixlogit-. I am using -cmxtmixlogit- instead of -xtmlogit- because I have unbalanced choice set, e.g., sometimes, it is 2 available choices, sometimes it is 3.
Here is the code I have now
Code:
cmxtmixlogit choice l.choice, casevars(i.treatment)
Code:
l.choice##i.treatment
Code:
cmxtmixlogit choice l.choice l.choice#i.treatment, casevars(i.treatment)
I cannot do separate regressions for the treatments, because of limitations about control variables.
