Statalist


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

st: How do I calculate prvalues for specific time points in my survival analysis?


From   "Stephen P. Jenkins" <stephenj@essex.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: How do I calculate prvalues for specific time points in my survival analysis?
Date   Thu, 10 Dec 2009 00:01:35 -0000

++++++++++++++++++++++++++++++++++++++++++++++++++++++

Date: Tue, 08 Dec 2009 16:20:10 -0700
From: Laura Templeton <ltempleton@ualberta.ca>
Subject: st: How do I calculate prvalues for specific time points
in my survival analysis?

Hello,

I am using discrete-time survival analysis to inspect the effects
of a 
host of background variables on employment state. I have 4 years
of data 
which has been broken down into person/month units. Event
completion is 
determined at each month with those finding appropriate
employment 
dropping from further analysis (using stset command).  Depending
on when 
a person is dropped from analysis (i.e., event
completion=suitable 
employment), I can have up to 48 time points per person.

My question:

I would like to be able to talk about the predicted probability
of white 
men, white women, non-white men and non-white women finding
suitable 
employment at specific time points (end of year 1, end of year 2,
etc) 
and am having trouble with the code.

So far I am successfully able to report the predicted probability
of 
finding appropriate employment for each additional month in the
labour 
market with this code:

FOR WHITE MEN
prvalue, x(female 0 vismin 0) rest(mean)

FOR WHITE WOMEN
prvalue, x(female 1 vismin 0) rest(mean)

However, what I really want to know is: what is the predicted 
probability of being appropriately employed by the end of the
first year 
(months==12) or by the end of the 2nd year (months==24) for white
men? 
Or white women? etc.

Any tips on how I would incorporate the time element into my
syntax is 
most appreciated!

Laura

++++++++++++++++++++++++++++++++++++++++++++++++++++++

I think you need first to clarify in your own mind whether you
want conditional or unconditional probabilities -- e.g. Pr(gets
job in month 12, given no job in months 1 to 11 and
characteristics X), versus unconditional Pr(has job in month 12,
given characteristics X). The former is the discrete time hazard
rate.

Have a look at the Lesson on discrete time survival analysis
associated with my on-line Stata course (URL below signature). It
has examples of how to create discrete time hazard rates,
survival probabilities (and their complement). They involve use
of -predict- after estimation, and then some basic manipulations.


Stephen
-------------------------------------
Professor Stephen P. Jenkins <stephenj@essex.ac.uk>
Institute for Social and Economic Research (ISER)
University of Essex, Colchester CO4 3SQ, UK
Tel: +44(0)1206 873374. Fax: +44(0)1206 873151
http://www.iser.essex.ac.uk 
Survival Analysis using Stata:
http://www.iser.essex.ac.uk/iser/teaching/module-ec968
Downloadable papers and software:
http://ideas.repec.org/e/pje7.html



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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