
152 Spatial Point Patterns: Methodology and Applications with R
experiments, a line ar model Y = X
β
+
ε
is confounded (in the context of a particular experimen t)
if the columns of the design matrix X are not linearly indepen dent, so that th e parameter vector
β
is not identifiable. Bartlett’s examples show th a t a point pr ocess model involving both spatial
inhomogeneity and interpoint interaction may be conf ounded, that is, unidentifiable, given only a
single r ealisation of the spa tial poin t pro cess.
The potential for confounding spatial inhomogeneity a nd interpoint inte raction is important in
the interpretation of summary statistics such as