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

st: On the interpretation of sts list and ltable in duration analysis


From   Gijs Dekkers <gd@plan.be>
To   statalist@hsphsun2.harvard.edu
Subject   st: On the interpretation of sts list and ltable in duration analysis
Date   Thu, 31 Mar 2005 11:59:17 +0200

Dear all,

I am applying the various lessons of the Essex summer school course on spell duration data (by Stephen Jenkins) to my own discrete paneldata set.
My problem is the following: Jenkins applies the commands 'sts list' and 'ltable (noadjust)' in his lesson 4, and gets exactly the same output.

Unfortunately, in my case, this is not so. My question therefore is: how come that both commands give different results, where they do not in Jenkins' case? Am I doing something wrong? How should I interpret the results of both commands?

I have a simple discrete dataset of the following form:
id year (called 'jaar2') poor
1 1 0
1 2 0
1 3 1
2 2 0
2 3 1
2 4 1
and so forth.

Below follows the log file: I start with 'stset'-ting this data set, then I apply 'sts list', followed by 'ltable' including noadjust.

Here's what I have figured out so far: the number at risk (beg. total) is the number of individuals in 'sts list' whereas it is the number of observations in 'ltable' (in my case, that is, in Jenkins's course it is exactly the same thing). Moreover, those who fail (become poor) or net lost (censored) in 'sts list' add up to 'deaths' in 'ltable', which immediately raises the question on the meaning of 'lost' in ltable.

I would appreciate some help with this, as I am not an experienced user of stata.

Gijs

LOG-file:
------------------------------------------------------------------------------------------------
log: C:\usr\STATA\data\hazard.log
log type: text
opened on: 31 Mar 2005, 11:12:37

. use poverty.dta

. * bron: jenkins, lesson 3, page 6
. gen jaar2=jaar-1994 if (jaar<2003)

. stset jaar2, failure(poor) id(whhid)

id: whhid
failure event: poor != 0 & poor < .
obs. time interval: (jaar2[_n-1], jaar2]
exit on or before: failure

------------------------------------------------------------------------------
17280 total obs.
2182 obs. end on or before enter()
2004 obs. begin on or after (first) failure
------------------------------------------------------------------------------
13094 obs. remaining, representing
3241 subjects
454 failures in single failure-per-subject data
17469 total analysis time at risk, at risk from t = 0
earliest observed entry t = 0
last observed exit t = 8

. sts list

failure _d: poor
analysis time _t: jaar2
id: whhid

Beg. Net Survivor Std.
Time Total Fail Lost Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
1 3241 109 162 0.9664 0.0032 0.9596 0.9720
2 2970 87 226 0.9381 0.0043 0.9291 0.9459
3 2657 72 214 0.9126 0.0051 0.9021 0.9221
4 2371 59 404 0.8899 0.0058 0.8780 0.9007
5 1908 26 177 0.8778 0.0062 0.8651 0.8894
6 1705 27 212 0.8639 0.0066 0.8503 0.8763
7 1466 42 273 0.8392 0.0075 0.8239 0.8532
8 1151 32 1119 0.8158 0.0083 0.7989 0.8315
-------------------------------------------------------------------------------

.
end of do-file

. do "C:\windows\TEMP\STD040000.tmp"

. ltable jaar2 poor, noadjust

Beg. Std.
Interval Total Deaths Lost Survival Error [95% Conf. Int.]
-------------------------------------------------------------------------------
0 1 17280 217 1965 0.9874 0.0008 0.9857 0.9890
1 2 15098 231 1934 0.9723 0.0013 0.9697 0.9748
2 3 12933 240 2072 0.9543 0.0017 0.9508 0.9575
3 4 10621 225 1927 0.9341 0.0021 0.9297 0.9381
4 5 8469 208 1979 0.9111 0.0026 0.9059 0.9161
5 6 6282 145 1587 0.8901 0.0031 0.8839 0.8960
6 7 4550 144 1485 0.8619 0.0038 0.8543 0.8692
7 8 2921 155 1372 0.8162 0.0051 0.8060 0.8259
8 9 1394 141 1253 0.7336 0.0080 0.7176 0.7490
-------------------------------------------------------------------------------

.
end of do-file


--
Dr. Gijs Dekkers
Federaal Planbureau
Algemene Directie
Kunstlaan 47-49
B 1000 Brussel
++32/(0)2/5077413
fax 7373






**********************************************************************
Disclaimer: This e-mail may contain confidential information
which is intended only for the use of the recipient(s) named above.
If you have received this communication in error, please
notify the sender immediately and delete this e-mail from
your system.
Please note that e-mail messages cannot be considered as official
information from the Federal Planning Bureau.
**********************************************************************
*
* 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