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

Re: st: Problem with -est table- after -xtmixed-

From (Roberto G. Gutierrez, StataCorp)
Subject   Re: st: Problem with -est table- after -xtmixed-
Date   Tue, 16 Aug 2005 17:04:30 -0500

Miguel Foguel <> asks:

> I am estimating a model using -xtmixed- and ols (for comparison purposes).
> The problem I have is the following: When I try to display the results for
> both models using -est table- (in fact -estout-) the results are not matched
> by the name of the variables. This seem to be caused by the fact that a
> strange "y" is showing up above the name of the variables in the display of
> the -xtmixed- model.  As a result the display from ols includes an
> underscore sign above the variables' names too.  As an example I give the
> following very simple code so that you can see what's going on.

> set obs 100
> gen x1 = uniform()
> gen eps = invnorm(uniform())
> gen y = 1 + x1 + eps
> reg y x1
> estimates store ols
> xtmixed y x1, mle
> estimates store xtm
> estimates table *

> Note that the variables are the same but the display from -est table- is not
> matched. Could anyone tell me how to sort this problem out? Many thanks in
> advance.


  . estimates table *, equations(y=1)

which forces Stata to match up the first equations and label them with the
name of the dependent variable, -y-.

-xtmixed- labels the first equation, the fixed portion, of the model with the
name of the dependent variable, as most Stata commands do.  -regress-, on the
other hand, is an older command that does not label the equation.  When
displaying coefficients from -regress-, -estimates table- uses an underscore
to denote the lack of such a label.

What I suggest above works around this.

*   For searches and help try:

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