[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Obtaining variable-specific p-value from -estat phtest-

From   "Peter Jepsen" <PJ@DCE.AU.DK>
To   <>
Subject   st: Obtaining variable-specific p-value from -estat phtest-
Date   Wed, 10 Oct 2007 12:22:56 +0200

Dear Carlo,

Oh, sorry. I should have explained more carefully. I realize that the
p-value is shown when I run my example. But I am doing a study where I
repeat a Cox-model 10,000 times, each time drawing matched pairs at
random from a large population. My model includes an exposure and
several confounders. I am using -postfile- to generate a new dataset
that contains the coefficient for my exposure from each of the 10,000
runs, but I would also like it to contain the p-value for the test of
proportional hazards for my exposure. I am less concerned with
proportional hazards for my confounders. Therefore, I need to obtain the
p-value for 'exposure' in my example below in a more automatic manner.
And, as I wrote, -estat phtest- leaves behind only the chi2 and df for
the GLOBAL test.

Best regards,

-----Oprindelig meddelelse-----
Fra: Carlo Lazzaro [] 
Sendt: 10. oktober 2007 11:37
Cc: Peter Jepsen
Emne: R: Date: Wed, 10 Oct 2007 10:41:19 +0200

Peter wrote:

Dear Statalisters,

I would like to obtain the p-value (or chi2) for the test of
hazard assumption for the variable 'exposed'. -return list- gives me the
chi2 for the global test, but not for the individual variables in the
An example:

sysuse cancer, clear
stset studytime, fail(died)
gen exposed=drug==1
stcox exposed age, sca(sca*) sch(sch*)
estat phtest, detail

I can obtain the rho for 'exposed' with -corr sca1 _t-, but I don't know
to get the chi2. I have tried reading the source code with - viewsource
stphtest.ado- but I couldn't make it out. Can anybody help me?

Best regards,

Dear Peter,

I have tried to run your example and as far as - estat phtest, detail -
concerned, I have obtained what follows:

      Test of proportional hazards assumption

      Time:  Time
                  |       rho            chi2       df       Prob>chi2
      exposed     |     -0.00949         0.00        1         0.9603
      age         |     -0.11758         0.42        1         0.5168
      global test |                      0.43        2         0.8064

Is it what you were looking for?

Kind Regards,


*   For searches and help try:

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