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

st: RE: Identifying regions within a cross-section with Stata

From   "Nick Cox" <>
To   <>
Subject   st: RE: Identifying regions within a cross-section with Stata
Date   Wed, 10 Oct 2007 18:09:12 +0100

There are lots of ways to think about this. 
Literature uses many different terminologies, 
breaks, changes, switches, whatever. 

My own prejudice is that you should publish 
(make public) your graph. If the evidence
is convincing, people will believe you, 
and not otherwise. A test is then just ornamental at best. 

There are many groups who regard that kind of 
procedure as subjective, impressionistic and 
even unscientific. 

But even those who want to attach a P-value 
to everything would perhaps be squeamish about
hypotheses based on looking at the data. If 
you say the idea came out of nowhere as a matter
of prior theory, then that makes everything all 
right, supposedly. 


Erasmo Giambona
> I have a cross section of firm-level data. If I plot the data
> according to firm debt and growth opportunities, it seems that the
> data is separating into 2 distinct regions. In the first region there
> are firms with very low debt and high growth and in the second region
> firms with very high debt and low growth. Is there a test to check
> whether indeed there is separation in the data? Is there a way to
> handle this with inregression?

*   For searches and help try:

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