[R] nesting in CoxPH with survival package

Bert Gunter gunter.berton at gene.com
Sat Jan 12 22:54:08 CET 2013


You need to get local statistical help. What you are doing makes no
sense. See inline below.

-- Bert

On Sat, Jan 12, 2013 at 1:03 PM, David Winsemius <dwinsemius at comcast.net> wrote:
> On Jan 11, 2013, at 5:35 PM, Katie Anweiler wrote:
>> Hello all,
>> I am trying to understand how to specify nested factors when using
>> coxph(), and if it is appropriate to nest these factors in my
>> situation.
>> In the simplest form, I am testing two different temperatures, with
>> each temperature being performed twice in different experimental
>> periods (e.g. Temp5 performed in Period A and C, Temp4 performed in
>> Period B and D)

Period is confounded with temperature. That is the source of the
singularity. in the message received below.

You can estimate the C-A and the D-B differences.
As I said, get statistical help. These are not R questions.

-- Bert

>> I am trying to see if survival time is affected by the treatment
>> temperature.  To do this I am using temperature and experimental
>> period nested within temperature as factors.
>>> LOEtempmod.5days=coxph(LOE.stable.5days~Temp+Temp/Period,data=goodstable)
>> Warning message:
>> In coxph(LOE.stable.5days ~ Temp + Temp/Period,  :
>>  X matrix deemed to be singular; variable 2 5 6 7
>> 1. Is this an appropriate way of nesting?
> Have you looked at the coxme package?
> http://cran.r-project.org/web/packages/coxme/index.html
>> 2. Can this error message be ignored?
> Sometimes R packages correctly drop variables that are exactly collinear:
> other times the correct solution is not clear. I would think the answer in
> this case would be "no", but do not have a lot to go on at this point.
> --
> David Winsemius, MD
> Alameda, CA, USA
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.


Bert Gunter
Genentech Nonclinical Biostatistics

Internal Contact Info:
Phone: 467-7374

More information about the R-help mailing list