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

st: RE: variable for birth order


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: variable for birth order
Date   Fri, 25 Nov 2005 14:29:41 -0000

I don't know any general rule from demography
or medicine. My wild guess based on no evidence
whatsoever is that demographic or medical expertise
wouldn't help here anyway. That is, if you 
have information on order of birth, it should be 
used; if you don't, there is no magic that will
make up for that lack. 

Other things being equal, what would drive my 
analysis if this were my problem is knowing that
statistical techniques that feed on ranks as ranks 
presume that (apparently) equal values are assigned
equal ranks and that the sum of the ranks is preserved. 
This is the default of -egen, rank()- anyway.

That doesn't mean that ties don't need corrections 
for some kinds of nonparametric tests 
that may or may not be offered by the programs you
use. 

There is an argument for using -egen, rank() unique- 
if you want to shake data points apart on some kinds of
graphs, but unless you have a very small dataset this is 
probably not an issue. 

Nick 
n.j.cox@durham.ac.uk 

Kemal Aslan

> I have information on the age of children, but for twins and 
> triples I do
> not have information on who comes the first tothe world. I want to use
> egen-rank to create to a birth order for each child. Given 
> the incomplete
> information set, which option of egen-rank should I use? Do 
> you know the
> general rule to create birth order when the information is 
> incomplete as
> described above in the demography and medicine literature?

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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   |   What's new   |   Site index