From Maarten buis To qglian@fudan.edu.cn Subject st: Re: More information about the pathreg2 command in path analysis Date Sun, 25 Mar 2007 16:52:31 +0100 (BST)

```--- Lian,Qiguo <qglian@fudan.edu.cn> wrote:
> Dear Prof.Maarten buis,
> I am a graduate student of Fudan university, Shanghai,China. And my
> major is Epidemiology.

You're in good company, because I am also a graduate student.

> My first question:
> The example in http://www.ats.ucla.edu/stat/Stata/faq/pathreg.htm
> shows us how to use the pathreg, my model,however,is different with
> it in some way. I have two
> intermediate exogenous_vars(agressive and depressive) in the Read
> position. How to use the pathreg? Should I apply the pathreg for the
> two intermediate exogenous_vars respectively?

For simplicity assume that the other variable names from the ATS
example remain unchanged. In that case you would type:

pathreg (agressive ses female) /*
*/ (depressive ses female) /*
*/ (write ses agressive depressive female)

( The use /* and */ is explained in
http://home.fsw.vu.nl/m.buis/stata/exampleFAQ.html#/**/ )

> My second question:
> The pathreg2 command has the power to estimate the log likelihood. I
> am confused what does the  log likelihood stand for because I want to
> estimate the whole model,but the pathreg has to run several times to
> get the result for the model and you can get a goodness fit index
> each time. I want to know how to get the whole model's log likelihood
> using the pathreg2 command written by you.

Based on my intuition obtained from similar models I think you can just
add the log likelihoods from the different model (i.e. multiply the
likelihoods), to get the (log) likelihood for the entire model. This is
what -pathreg2- does. However, this is just my intuition. I shared that
intuition on the stalist so the person asking the question could have a
concrete lead to investigate. I also posted that program to show how
that person could implement it, if my intuition turned out to be
correct.

Right now I don't have time to investigate this further. I forwarded
this mail to the statalist, in case somebody else knows the answer, or
has time to investigate.

I hope that this is at least some help,
Maarten

-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------

___________________________________________________________
New Yahoo! Mail is the ultimate force in competitive emailing. Find out more at the Yahoo! Mail Championships. Plus: play games and win prizes.
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```