[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: AW: sample size

From   "Martin Weiss" <[email protected]>
To   <[email protected]>
Subject   st: AW: sample size
Date   Tue, 29 Sep 2009 14:48:08 +0200


This question is not so much about the size of the sample than about the
members that are or are not included by -ivreg2-: If it was all about the
size, you could simply say -in 1/2122- and be done...

By conditioning on -e(sample)- from the -ivreg2- estimation, you tell
-regress- to use the observations that -ivreg2- chose to use for whatever
reasons, and you would need a very good explanation for that...

You could make -e(sample)- permanent by saying something like -gen byte
inivreg=e(sample)- and then -summarize- or -inspect- the resulting dummy to
look for patterns that promote or discourage inclusion...


-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Nirina F
Gesendet: Dienstag, 29. September 2009 14:36
An: [email protected]
Betreff: st: sample size

hello all,

When I run ivreg2, my sample size dropped from 2445 to 2122.
What is the logic behind this sample drop?
Also, I restricted the plain reg with if e(sample) so that I have now
2122 in my OLS but is it ok to do that? I mean what is the meaning
econometrically of restricting the sample size to be the same?

Thank you in advance fo rtaking the time to respond.
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index