R help - May 2017 - p-value=0 running log-rank test

Dear all,

I have a question concerning the p-value. When running log-rank test I get
a p-value = 0.
What is it mean? Can this be true? Why aren?t there decimal points? Is
there a way to find out the exact p-value?

Here is the output:
> survdiff(Surv(tempo2,status)~tphist, data=base,rho=0)
Call:
survdiff(formula = Surv(tempo2, status) ~ tphist, data = base,
    rho = 0)

           N Observed Expected (O-E)^2/E (O-E)^2/V
tphist=1 513       19   40.879    11.710  9.00e+01
tphist=2  49        4    3.892     0.003  3.27e-03
tphist=3  23        9    1.686    31.717  3.29e+01
tphist=4  18       15    0.543   385.172  3.91e+02

 Chisq= 430  on 3 degrees of freedom, p= 0

Thank you in advance.

Anne.

	[[alternative HTML version deleted]]

This is not an R question.

Your question indicates that you really need to learn some statistics.

To answer the very last part:

 > pchisq(430,3,lower=FALSE)
[1] 7.020486e-93

And if that is not 0 to all intents and purposes, then God help us all.

cheers,

Rolf Turner

On 19/05/17 12:48, Anne Karin da Mota Borges wrote:
> Dear all,
>
> I have a question concerning the p-value. When running log-rank test I get
> a p-value = 0.
> What is it mean? Can this be true? Why aren?t there decimal points? Is
> there a way to find out the exact p-value?
>
> Here is the output:
>
>> survdiff(Surv(tempo2,status)~tphist, data=base,rho=0)
> Call:
> survdiff(formula = Surv(tempo2, status) ~ tphist, data = base,
>     rho = 0)
>
>            N Observed Expected (O-E)^2/E (O-E)^2/V
> tphist=1 513       19   40.879    11.710  9.00e+01
> tphist=2  49        4    3.892     0.003  3.27e-03
> tphist=3  23        9    1.686    31.717  3.29e+01
> tphist=4  18       15    0.543   385.172  3.91e+02
>
>  Chisq= 430  on 3 degrees of freedom, p= 0
>
> Thank you in advance.

This is a minor problem with the print method for "survdiff" objects. 
It
acts like it computes the p-value with
   1 - pchisq( chisq, df)
instead of
    pchisq( chisq, df, lower.tail = FALSE)
The former will give zero when the latter gives a number less than about
10^-16.

Most people would accept 0 instead of 10^-16, but the 6-sigma folks worry
about probabilities of 10^-10 so perhaps it is worth making the easy change.


Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Thu, May 18, 2017 at 10:14 PM, Rolf Turner <r.turner at auckland.ac.nz>
wrote:
>
> This is not an R question.
>
> Your question indicates that you really need to learn some statistics.
>
> To answer the very last part:
>
> > pchisq(430,3,lower=FALSE)
> [1] 7.020486e-93
>
> And if that is not 0 to all intents and purposes, then God help us all.
>
> cheers,
>
> Rolf Turner
>
> On 19/05/17 12:48, Anne Karin da Mota Borges wrote:
>
>> Dear all,
>>
>> I have a question concerning the p-value. When running log-rank test I
get
>> a p-value = 0.
>> What is it mean? Can this be true? Why aren?t there decimal points? Is
>> there a way to find out the exact p-value?
>>
>> Here is the output:
>>
>> survdiff(Surv(tempo2,status)~tphist, data=base,rho=0)
>>>
>> Call:
>> survdiff(formula = Surv(tempo2, status) ~ tphist, data = base,
>>     rho = 0)
>>
>>            N Observed Expected (O-E)^2/E (O-E)^2/V
>> tphist=1 513       19   40.879    11.710  9.00e+01
>> tphist=2  49        4    3.892     0.003  3.27e-03
>> tphist=3  23        9    1.686    31.717  3.29e+01
>> tphist=4  18       15    0.543   385.172  3.91e+02
>>
>>  Chisq= 430  on 3 degrees of freedom, p= 0
>>
>> Thank you in advance.
>>
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posti
> ng-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]