Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

RE: st: Bug in stsplit and stptime - or me ?

From   "Jesper Lindhardsen" <>
To   <>
Subject   RE: st: Bug in stsplit and stptime - or me ?
Date   Mon, 10 Oct 2011 22:30:14 +0200

Hi Yulia,

Thanks for clarifying this issue, as I was puzzled whether this was a "legal" option, but as mentioned in the original post, this has been used in different settings.

Of note, the specific use of enter and origin was very deliberate (in the original dataset) in order to change the underlying analysis time after some data management. However, I gather a simple clonevar or rename of system variables can solve the problem. Or am I mistaken ??


-----Original Message-----
From: on behalf of Yulia Marchenko, StataCorp LP
Sent: Mon 10/10/2011 21:15
Subject: Re: st: Bug in stsplit and stptime - or me ?
Jesper Lindhardsen <> used -stset- to set his
data, issued some -st- commands, and then changed the -st- settings by
using -stset- again.  That would be fine except that, in the second
-stset- command, Jesper used system variables created by the previous
-stset- command such as _t, _t0, _d, etc.

You must not do that.  You can re-stset if you wish, but you must not
use variables _t, _t0, _d, etc.  We will modify -stset- to issue an
error message when system variables are used.

As a side note, Jesper used the -enter()- and -origin()- options
interchangeably in his code.  Although in his example it led to the
same results, this will not be true in general.  It is important to
remember that the -origin()- option specifies the onset of risk of
subjects and, thus, redefines the analysis time.  The -enter()- option
specifies when subjects enter a study which may be after they have
already been at risk of a failure.  The distinction between these
options is crucial for correct declaration of your survival data.

-- Yulia
*   For searches and help try:


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index