
Model Validation 431
11.6 Validating the independence assumption
11.6.1 Caveat on residual plots
The residual plots described in Sections 11.2–11.5 above a re useful for d etecting misspecification
of the intensity in a fitted Poisson proce ss model. They are not very useful for checking the inde-
pendence assumption of the Poisson process (assumptio n (PP3) on page 300).
An extreme example is provided by the cells dataset which gives the locations of the c entres
of biological ce lls observed in a histological section [572]. The four-pa nel diagnostic plot for a
unifor m Poisson process fitted to the cells data, in Figure 11 .13, shows no evidence against ...