Weibull (frailty) model for survival data

[Tristan Lorino <tristan.lorino@lcpc.fr>]

Hi,

I   have  one pavement cracking measure on several sections of several
roads. I would like to use a (parametric) Weibull model for
survival data. My questions are:
1.  once fitted the  model  (without  frailty),  how can I study the
goodness-of-fit?  S-Plus  Guide  to  Statistics  vol.2 (p. 437) does not
develop this point...
2.   once  fitted  the model with frailty (frailty represents the road
effect), how can I test the role of the
frailty,  e.g.  how  can  I be sure that introducing a frailty term is
necessary?
3.  in presence of frailty, how can I study the goodness-of-fit?

Thanks for your help,
Tristan Lorino
------------------------------------------
Laboratoire Central des Ponts et Chaussées
[Division ESAR - Section AGR]
Route de Bouaye
BP 4129
44341 Bouguenais Cedex
Tél. 33 (0)2 40 84 55 95