In this paper we briefly review Tversky's main criticisms of spatial m
odels of similarity. We then recapitulate the contrast model that was
proposed to remedy for the shortcomings of spatial models. Next, four
data-analytic methods that are based on the contrast model are describ
ed and illustrated with analyses of a small data set; special attentio
n is given to the interrelations of the models on which the methods ar
e based. Finally, the four data-analytic methods are checked against T
versky's criticisms of geometric models of similarity and against rece
nt cognitive objections to the contrast model.