Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: missing standard errors when using nl command


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: missing standard errors when using nl command
Date   Tue, 16 Sep 2003 12:01:18 +0100

stephen hynes (modulo deprecated HTML): 

> I am using the nl command to fit a nonlinear 
> function to my dependand variable by least squares. 
> I have a dataset with 6 years of observations on approx. 
> 5000 individuals. I am carrying out a variance covariance 
> analysis using the minimum distance method and am trying 
> to fit an appropriate econometric specification to my error 
> component model (for Irish earnings). When I run my model 
> specification using nl I get no results back for my standard 
> errors, only dots. Can anyone tell me why this may be so. 
> Are the number of years of observations too small?

Difficult to diagnose at a distance, especially 
with no knowledge on my part of this class of models
or your data set. 

Missing standard errors can arise for all sorts 
of reasons. I would tend to guess collinearity 
or convergence problems. Whenever I've used it 
I've found -nl- unable to produce miracles if (in 
retrospect) I made silly guesses initialising parameters. It 
often needs a good hint about where to start. 

Nick
n.j.cox@durham.ac.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/



© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index