Bookmark and Share

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

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: statalist-digest V4 #4681

From   "David Roodman (" <>
To   "" <>
Subject   st: RE: statalist-digest V4 #4681
Date   Thu, 25 Oct 2012 14:16:42 +0000

	In principle, you can do fixed effects in cmp by putting in group dummies. I think the only limitation would be if you have so many groups that the dummies overload cmp. Also, cmp now does random effects too.

-----Original Message-----
Date: Wed, 24 Oct 2012 10:15:36 -0400
From: Kirk Geale <>
Subject: st: Accounting for panel nature in a mixed process IV model?


I have a model with first stage that is an ordered probit, and a
second stage that is a probit. For this I would use a routine like
- -cmp-.  This allows me to use mixed dependant data types as well as
IV's.  But this data also has a panel structure.  I can use something
like the linear -xtivreg-, but then I cannot take into account the
mixed processes that are present.  As far as I am aware, there is no
routine that allows for mixed processes, IV's, and fixed effects.  I
was thinking about centering the relevant variables before using
- -cmp-, but does de-meaning only apply to linear fixed effects?  Any
other thoughts on how I can incorporate the full suite of mixed data
types, fixed effects, and IV's?

As a side note, can one de-mean variables in an ordinary nonparametric
kernel regression to take into account of the panel nature of data?

Kirk Geale

- --
Kirk Geale
M.Sc. Economics
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index