Abstract
The use of areal units as units of analysis is widespread in the social sciences. Typically, however, studies employing areal units gloss over the way in which the units were chosen. After illustrating how different theoretical and policy conclusions can be drawn from different lattices, it is asserted that the problem is not methodological but epistemological. “True” relationships are not to be found due to the lack of correlation between proximity and closeness. Moreover, judgement must enter into the decisions to employ a particular lattice. Thus, we need to understand the prejudices that must underlie any attempt to carve up the world. Several guidelines are presented to aid the researchers.