Jean Marie
Thanks.
I would recommend, however, that the axes only show the range of values
greater than -tmin- when this option is specified.
An example where the -tmin- option would be very useful is when age is
being used to denote the time variable and the failure events do not
occur until middle age or later. For example, if we are plotting
cumulative coronary heart disease (CHD) morbidity as a function of age,
the number of events before age 30 is likely to be zero. In this case
it makes sense to start the plot at age 30 rather than at age zero,
which is the default value for sts graph. Having the x-axis extend back
to zero in this graph generates a lot of blank paper and reduces the
plot detail for ages when CHD actually occurs.
Bill Dupont
-----Original Message-----
From: Jean Marie Linhart, StataCorp LP [mailto:[email protected]]
Sent: Monday, January 26, 2004 2:26 PM
To: [email protected]
Subject: Re: sts graph bug
Bill Dupont and Roger Harbord have recently discussed
-sts graph- with the -tmin- option.
(e.g.
. sysuse cancer, clear
. stset age, failure(died)
. sts graph
. sts graph, tmin(55)
)
As they have discussed, currently the survival graph starts at the
origin (_t = 0) with the -tmin- option, and shows a survival function
equal to one (or -per- if the -per option is specified) until time
-tmin-. At time -tmin- it then drops as if all the failures preceding
time -tmin- occurred at time -tmin- and then proceeds to plot the
remainder of the survival function. This is misleading and should be
considered a bug -- the axes should show the range of the entire
survival curve but the actual plot should start at the time specified
via -tmin-. I have already changed -sts graph- to do this, and I expect
this fix to be available in the next ado-file update.
--Jean Marie
[email protected]
*
* 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/
*
* 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/
