st: Markov model

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- Youngsoo Yu wrote me privately:
> I found your name and messages on STATA Q/A pages. Can I ask you a
> favor? Could you let me know how I can compute stata code for Markov
> models?

The convention is not to ask questions to individual members of the 
statalist privately. Instead you should sent your questions directly
to the statalist. The reasons for this are explained here:
<http://www.stata.com/support/faqs/res/statalist.html#private>

> I have a dichotomous variable, human rights success(1) and failure(0).
> And the transition to and from each other follows a first-order Markov 
> process.
>
> Matrix of transition is like this.
> 
> [  P(Rt=F)  ]       =    [   Pff  Psf  ]  [P(Rt-1 =F)   ]
> [  P(Rt=S)  ]            [   Pfs  Pss  ]  [P(Rt-1= S)   ]
>
> I assume that the probability of transition from failure to success is
> a linear function on exogenous variables. Pfs(t) = F(Xt-1*b)
> 
> I'm sure what I'm doing with Markov transition and understand the 
> equations. Unfortunately, I don't know how I can estimate the 
> coefficeints of variables with STATA.

A straightforward Markov model is very easy to estimate: Just estimate
one logit model for all observations whose previous observation was 
a failure and another logit model for all observations whose previous
observation was a success. 

The hard part comes when you believe that you need to control for 
unobserved heterogeneity. My take on that is to just not do that, but
others (and other disciplines) disagree. In that case you'll probably
need to do some programming. If you want to go that road then this 
special issue of the stata Journal on maximum simulated likelihood 
may be useful:
<http://www.stata-journal.com/sj6-2.html>

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/