ABSTRACT
In Part I, it was shown how a variety of numerical measures and descriptive models could be used to summarise relationships in geographical data sets. The potential value of inferential procedures was also stressed, in that data from a single survey may be used to illuminate, or expand on, findings relevant to an unseen population which have been established in other contexts. The advantage of this is that survey data may be generalised or used to expand knowledge beyond the limited confines of the specific survey.
