Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.

# st: Probability Tree

 From Francesco To statalist@hsphsun2.harvard.edu Subject st: Probability Tree Date Mon, 3 Sep 2012 22:09:42 +0200

```Dear Statalist,

I have been searching the archive for a while, but I am unable to find
an adequate solution to the following problem.

I have a panel dataset, where for each individual i the variable Zit
equals 1 if individual i is treated at time t, and zero if not.
Zit is completely endogeneous : individual i can choose at time t if
Zit=1 or zit=0

I would like to estimate the first n nodes of a "tree" that gives me
the transition probabilities from the first time an individual chooses
Z=1 to the next n treatment states
So the tree would look like

Z1=1  ------> Prob(Z2=0) --------->Prob(Z3=0) ---------> >Prob(Z4=0)
|                          |                             |
-------> >Prob(Z4=1)
|                          |
|                           ----------->Prob(Z3=1)
|
|
|
|   ------->Prob(Z2=1)  ------>
-|------>

where Z1 is 1 because I start the tree from the first treatment time
for each individual. Z2 is the treatment state in the next period.
Prob(Z2=0) is therefore the probability to choose no treatment
conditional on treatment at time 1.

Do you have an idea ? I tried something with the L. operator but it
did not work properly...