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: Survival analysis and control variables

From   maarten buis <>
Subject   Re: st: Survival analysis and control variables
Date   Wed, 4 May 2011 14:48:59 +0200

On Wed, May 4, 2011 at 2:04 PM, D.W.Richards wrote:
>  I am trying control for wealth in relation the survival time of losses, not only survival
>  time in general.  In this case, does this equation seem logical?
> h(t)= ho(t) exp ( LossB1 + lossxageB2 + ageB3 + lossxwealthB4 +wealthB5)

This gets tricky. To make such decisions one really needs to be well
into what you
exactly want to measure, what the theory is, what the data looks like
and how it was
collected, etc. etc. At the very least I would center wealth and age
before making the
interaction terms. Right now B1 refers to the effect of loss for a
newly born baby (age
= 0) with no wealth (wealth = 0). I suspect that that is well outside
the range of your
data. You do not have to center at the mean, any meaningful value
within the range
of your data will do.

> Also do you know if it's possible to get a statistic like an "r-square" to test the
> difference between the above model and model below:
> h(t)= ho(t) exp ( LossB1 + lossxageB2 + ageB3)

I would just use -test-. So after you estimated the top model I would type

test lossxwealth wealth

To test the hypothesis that B4 and B5 are both simultaneously equal to
zero, in which
case you would end up with your second model.

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index