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

st: re: data management

From   Kit Baum <>
Subject   st: re: data management
Date   Thu, 27 Mar 2008 21:07:29 -0400

Sergiy said

> Raphael said
> I have 100 subjects with 6 obs each. I would like to create a new
> variable containing 1s for the first three obs and 2s for the
> remaining obs for each id.
> bysort id: gen x = cond( _n < 4, 1, 2)

and this raises the question, is _bysort_ stable? Because if it is
not, then not the first three of Raphael's observations get 1s, but
three random of each six get 1s. Stata's manual is silent about it,
but it seems to be the case. Quote:

"sort specifies that if the data are not already sorted by varlist, by
sort them."

Does anybody have more precise info? Is bysort stable?

This can be dealt with, presuming Raphael has an obs number for each reading on a subject, by doing

bysort id (obs): gen x ...

which will ensure that the observations for each ID are ordered 1,2,...,6.

Kit Baum, Boston College Economics and DIW Berlin
An Introduction to Modern Econometrics Using Stata:

* For searches and help try:

© Copyright 1996–2019 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index