Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Pls help me! question regarding panel fixed effect model

From   "Clive Nicholas" <>
Subject   Re: st: Pls help me! question regarding panel fixed effect model
Date   Sun, 20 May 2007 07:32:07 +0100

On 20/05/07, <> wrote:

I put questons regarding panel fixed effect
estimation, but did not get any reply. Can anybody pls help me?
As there are around 1200 Statalisters, it's very unlikely that
everybody will ignore your perfectly reasonable questions. But do also
remember that it's the weekend.

# I want to incorporate time effects in fixed effect model. How can I
do this
in stata? What command should I use?
By entering time-dummies directly into your model. See below.

# There are 3 values for  R square...within, between and overall. Which
should be used? The value of R square is very low when I use fe, but it
high in case of pooled OLS. How can I  know about the R square value
for the
variables and for the country or time effects?
You would look at both the 'within' and the 'overall' R^2 statistics.
Remember that 'within' refers to the observation of effects over time
_within_ units; 'between' refers to the observation of effects
_between_ units at any one point in time. Any decent econometric
textbook would tell you this.

# If I use cluster option, does it correct for serial correlation and
heteroskedasticity? I need to correct both in my estimation. What does
group correlation mean?
Heteroskedasticity. As for the intra-group correlation, this link may
help to explain:

As for which model you should use, to achieve what you need, why not
use Baum, Schaffer and Stillman's -ivreg2-, downloadable from SSC?
Then you could do something like this:

. ivreg2 y x1 x2 d1 d2 d3, bw(2) robust cluster(group) small

where the d-variables are your time-dummies. This fits an OLS
fixe-effect model with Newey-West standard errors (although using both
the -robust- and -cluster()- options might be overdoing things). I'm
sure Kit, Mark and/or Steve will add further comments on this should
they feel the need.

Clive Nicholas
*   For searches and help try:

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