Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# Re: st: mlogtest after mlogit

 From Maarten Buis To statalist@hsphsun2.harvard.edu Subject Re: st: mlogtest after mlogit Date Mon, 24 Oct 2011 14:50:56 +0200

```On Mon, Oct 24, 2011 at 11:59 AM, Chiara Mussida wrote:
> I run I mlogit with a dependent variables for 9 outcome
> categories. <snip> my dependent variable, transition, is
> built as follows:
> ge transition=0
> replace transition=1 if UE==1
> replace transition=2 if UN==1
> replace transition=3 if UU==1
> replace transition=4 if EU==1
> replace transition=5 if EN==1
> replace transition=6 if EE==1
> replace transition=7 if NE==1
> replace transition=8 if NU==1
> replace transition=9 if NN==1
>
> where each outcome is a transition or permanence in each labour market
> state (namely, Unemployment, Employment, Non labour force)

Sounds to me like you want to run a Markov model. The simplest version
of which requires two variables: origin and destination. Origin is the
previous state and destination is the current state: so both can take
"values" U, E, or N (represented by appropriate numerical values, e.g.
1, 2, 3). After that you run three -mlogit- models (assuming you have
one explanatory variable imaginatively called x):
mlogit destination x if origin == 1
mlogit destination x if origin == 2
mlogit destination x if origin == 3

So you are modeling the probabilities of each destination state
conditional on the three possible origin states.

Things (as always) get more complicated when you want to control for
unobserved heterogeneity. However, whether or not one really wants to
control for unobserved heterogeneity, and if so, what the causal
effect is one is really after, i.e. what (thought) experiment you
would like to execute, is a remarkably subtle problem. See e.g.:
Robert D. Mare (2011) "Introduction to symposium on unmeasured
heterogeneity in school transition models" Research in Social
Stratification and Mobility, 29(3): 239-245.

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/
```