[R] Heteroskedasticity in Tobit models

[Thomas Lumley]

On Tue, 25 Apr 2006, Alan Spearot wrote:

> Hello,
>
> I've had no luck finding an R package that has the ability to estimate a
> Tobit model allowing for heteroskedasticity (multiplicative, for example).
> Am I missing something in survReg?  Is there another package that I'm
> unaware of?  Is there an add-on package that will test for
> heteroskedasticity?

If you mean survreg() [rather than survReg(), which is in S-PLUS] then it 
can estimate models where the variance depends on a discrete covariate  by 
adding a strata() term to the model formula. For example:

> survreg(Surv(futime, fustat) ~ ecog.ps+strata(rx), data = ovarian,
+     dist = "weibull")
Call:
survreg(formula = Surv(futime, fustat) ~ ecog.ps + strata(rx),
     data = ovarian, dist = "weibull")

Coefficients:
(Intercept)     ecog.ps
   8.0159674  -0.5940253

Scale:
      rx=1      rx=2
1.2047759 0.5605876

Loglik(model)= -96.2   Loglik(intercept only)= -97.1
         Chisq= 1.68 on 1 degrees of freedom, p= 0.2
n= 26

is a weibull model with different variance depending on the value of rx.


 	-thomas

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle