Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

st: Choice among different heckprob models


From   Marco Sangiorgi <sangiorgi.marco.13@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Choice among different heckprob models
Date   Wed, 9 Nov 2011 16:02:45 +0100

Hi,

First of all please sorry for the bad layaout of the previous post, I
had some problem with plain text settings.

Here it is my concern. I'm running a probit model with sample
selection (-heckprob-). I've built three different models:

1) full model:   y = x1 + x2 + x3
2) model 2:      y = x1 + x2
3) model 3:      y = x1      + x3

Note that model 2 and model 3 are nested in the full model.  Now I
have to select the model that better fits the data. Here it is the
procedure I've followed:

a- In order to choose the model that better fits the data between the
full model and model 2, I've tested the significance of x3 in the full
model   [results: x3 is highly significant, that suggests to not
exclude it from the model]

b- Then I did the same in order to choose between the full model and
model 3 (that is, I've tested the significance of x2 in the full
model)   [results: x2 is highly significant, that suggests to not
exclude it from the model]

c- Since I found that x3 and x2 were both significant, I conclude that
the full model fits the data better than model 2 and model 3

Is it a correct procedure? Should I have to conduct different tests
for each equation (outcome and selection) separately?

Thanks,


Marco Sangiorgi
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index