The primary objective of this paper is to revisit a number of empirical modelling activities which are often characterized as data mining, in an attempt to distinguish between the problematic and the non-problematic cases. The key for this distinction is provided by the notion of...