Very well said.
Nick
[email protected]
Maarten buis
--- On Thu, 5/11/09, Tim Waring wrote:
> One final point on the meaning of fixed effects - there
> seems to be some confusing differences in how people use the
> term. The safest meaning is that used by
> statisticians, but there are many other uses. It is
> probably good that everyone know what those uses are, and
> Andrew Gelman provides a nice summary of the different
> meanings that people use for "fixed effects":
I disagree, on the Statalist it would be safest to follow the
conventions in official Stata, which would seem to be that
fixed effects refers to the model that only uses information
within a level rather than to the non-random coefficients in
a random effects model. For example, if you want to estimate
such a model, you would typically have to specify the -fe-
option within an -xt- command, where -fe- stands for fixed
effects.
A more general lesson is that we are a multi-disciplinary
list, so we should not assume that the terminology (and
literature) that is very well known within our own
discipline is also known by the rest of the list. The one
thing that we do have in common is Stata, so that should
be our point of reference.
*
* 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/
