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

st: RE: Xtprobit - comparing coefficients across groups

From   "Martin Weiss" <>
To   <>
Subject   st: RE: Xtprobit - comparing coefficients across groups
Date   Sun, 19 Jul 2009 10:18:31 +0200


The heterogeneity btw the groups could also be handled by


-----Original Message-----
[] On Behalf Of Stephen O'Neill
Sent: Sonntag, 19. Juli 2009 02:18
Subject: st: Xtprobit - comparing coefficients across groups

I am estimating a dynamic probit model using xtprobit following the approach
of Wooldridge (2005) "Simple solutions to the intitial conditions
problem...". My dataset contains observations for 1,200 individuals over 8
years. I wish to allow the coefficients in my model to differ across groups
(6 groups - I have a grouping variable). To do this I am currently
estimating the model separately for each group. [I had used interaction
terms originally but after reading Ai and Norton -"Interaction terms in
logit and probit models" I decided against this]. I now have 3 related

1) Is there an alternative method which does not require me to split my
2) Given that the coefficients are really (beta/sigma), how can I compare
the various coefficients (I presume I need to adjust by sigma_u)?
3) How can I test whether the coefficients differ across groups? [I also
plan to look at different two periods so this question will arise here also]

Thank you in advance for any help which you can offer,


*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index