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

st: RE: random order after sub-sorting


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: random order after sub-sorting
Date   Fri, 12 Aug 2005 17:32:15 +0100

It is difficult to comment on Stata's behaviour 
here as you do not give information on precisely 
what commands you used. 

In general, if you type 

sort husband 

any solution in which -husband- is in the 
correct order is possible. If you type 

sort husband wife 

then the solution will be constrained 
by the order of -wife-. It is also possible
that 

sort husband, stable 

will be useful for your purposes. It is not 
clear which part of this you already know. 

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

D.C.J. Vissers
 
> I have a dataset of nearly 3000 couples. One record consists 
> of data of
> both the man and the woman. There are some polygamous marriages, so 1
> man can have 2 or 3 women instead of 1 woman. In this case, 
> there are 2
> or 3 records with the same data for the man and different data for the
> women.
> For my analyses of men, I only want 1 record of polygamous men.
> Therefore, I create duplicates based on idnumbers. Before this
> duplication I sorted the data on idnumbers and 1 variable of the men,
> namely mobility status. This status is the same for 1 man 
> with different
> wives.However, the data of the corresponding wifes will be 
> different. I
> noticed that Stata randomly selects the order of the 
> different wives in
> polygamous men. See also the following example:
> 
> Attempt 1:
> data of man 1 - wife 1
> data of man 1 - wife 2
> data of man 1 - wife 3
> 
> Attempt 2:
> data of man 1 - wife 3
> data of man 1 - wife 1
> data of man 1 - wife 2
> 
> I kept the first record of every duplicate. This means that 
> in attempt 1
> Stata kept wife 1, and in attempt 2 wife 3. This might not give big
> problems, but after this I make another selection. I also need to know
> the mobility status of women. Suppose, this status is known 
> for wife 1 &
> 2, but not known (missing) for wife 3. This means that I will find
> different numbers in attempt 2. And that is also what 
> happened. I found
> small differences in different attempts (number of records: 
> 1519 vs 1517
> vs 1512).
> 
> My questions:
> 1. Why does Stata randomly order records after a certain sorting?
> 2. How can I make sure that every time the same selection is made?

*
*   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