[R] How to fit a Tobit model with observations censored at different values

[Terry Therneau]

> I am a new user of R. I have a dataset with a dependent variable (DV) censored
> at different values. The dataset looks like, 
conditions .....    IDV1 IDV2    DV
1                    2      4      89
1                    6      6      75
1                    4      5      0  ( DV<=70)
...
>I do not know how to ajust the code in the VGAM pacakage. I would like to get
> some some suggestions. Thank you very much.

  I do not know how to adjust the VGAM package, but it is very easy to fit Tobit 
models using survreg.  Let DV1 = the value of DV or the lower threshold and DV2 
= 1 if the value is exact and 0 if it is censored.  So for the above data
    DV1   DV2
    89     1
    75     1
    70     0
    
    Then the following will fit a Tobit regression
    
survreg(Surv(DV1, DV2, type='left') ~ x1 + x2 + ..., data=mydata,
    	       dist='gaussian')
    	       
    Terry Therneau