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

Re: st: -areg- versus -xtreg, fe- (revised)


From   "Clive Nicholas" <Clive.Nicholas@newcastle.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: -areg- versus -xtreg, fe- (revised)
Date   Thu, 3 Jun 2004 02:13:30 +0100 (BST)

Richard Williams replied:

> Whenever Stata doesn't do something it normally does, like allow robust
> standard errors, I always wonder if it is because (a) it is statistically
> inappropriate, or (b) nobody has gotten around to programming it yet.
>
> I don't know much about XT models, but I notice that some commands (e.g.
> -xtgee-) do allow for robust standard errors, if there is any way you can
> use them instead.

Thanks for this, Rich. I rather think it's (a). It normally is when I'm
involved. The good news is is that -xtgee, robust- does work for me; and I
can fit it for all of my models and get full statistics for everything.
Coefficients _and_ standard errors look sensible. Hoorah. But watch this:

note:  observations not equally spaced
       modal spacing is delta id = 1
       48 groups omitted from estimation
note:  some groups have fewer than 2 observations
       not possible to estimate correlations for those groups
       22 groups omitted from estimation

(...-corr(ar1) is specified here, default options otherwise: more groups
are taken out if I specify higher orders of autocorrelation, but I guess
that's unavoidable...)

Iteration 1: tolerance = 1.2783629
Iteration 2: tolerance = .11461104
Iteration 3: tolerance = .0154233
Iteration 4: tolerance = .00231007
Iteration 5: tolerance = .00035132
Iteration 6: tolerance = .00005355
Iteration 7: tolerance = 8.166e-06
Iteration 8: tolerance = 1.247e-06
Iteration 9: tolerance = 1.887e-07

Now I seem to recall that low tolerance = high multicollinearity = very
bad news for my models. Now I do have linear and quadratic terms in my
models, both of which are highly significant and are, naturally, highly
colinear with each other. But dropping either of these terms does not
affect the tolerance statistics, yet the fit plummets dramatically). So,
what to do?

CLIVE NICHOLAS        |t: 0(044)191 222 5969
Politics              |e: clive.nicholas@ncl.ac.uk
Newcastle University  |http://www.ncl.ac.uk/geps
*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index