[R] survreg() provides same results with different distirbutions for left censored data

[Thomas Lumley]

On Thu, Dec 8, 2011 at 3:55 AM, ifejobi <ifejobi at gmail.com> wrote:
> Hello,
>
> I'm working with some left censored survival data using accelerated failure
> time models.  I am interested in fitting different distributions to the data
> but seem to be getting the same results from the model fit using survreg
> regardless of the assumed distribution.
>
> These two codes seem to provide the same results:
>
> aft.gaussian <- summary(survreg(Surv(y,y>0),type="left")~
> group),dist="gaussian")
>
> aft.weibull <- summary(survreg(Surv(y,y>0),type="left")~
> group),dist="weibull")
>

That's because you're fitting the same model: the dist= argument needs
to be inside the parentheses for survreg().

   -thomas

-- 
Thomas Lumley
Professor of Biostatistics
University of Auckland