Re: st: Small Sample Statistics

[Michael Hanson <mshanson@mac.com>]

On Mar 8, 2010, at 1:24 PM, odongo kodongo wrote:

> I'm conducting a multiple equation regression (system) through GMM.  
> The
> following few questions arise:
> 1. Can I be able to obtain small sample test statistics (t) rather  
> than z?

	Many e-class commands, such as -ivregress gmm- and -xtabond2-,  
support the -small- option.  You'll need to look at the -help- for  
your specific (but unnamed) command to see if it too supports this  
option.


> 2. Is it possible to determine the following measures of fit for each
> equation in the system? R squared? RMSE?

	Depends on the way you specify your model (easy, for example, in a  
model that can be re-written in a recursive form), and the Stata  
commands you use.

	In both cases, you really need to provide more information  
(including, most importantly, the exact Stata commands you are using),  
if you expect to get more specific help.  Please (re-)read the  
Statalist FAQ for advice on how to ask questions that will better your  
chances of getting the intended response.  The link to the FAQ is at  
the end of every message posted to Statalist (including this one.)

Best,
Mike

*
*   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/