Computer simulation and statistical modelling: rivals or complements?
Thomas K. Burch, University of Victoria
Conventional statistical analysis and computer simulation exist as two separate research traditions within demography. The lack of better integration of these two streams of analysis is due in part to methodological confusion, in particular, the failure to recognise a crucial distinction – between the form of an analytic tool and the purpose for which it is used. It is the latter which determines the epistemological standing of a particular analysis. In light of this distinction, it is argued that statistical modelling and computer simulation tend toward different uses. Statistical modelling is largely oriented toward representing data; computer simulation is largely oriented toward representing real-world or theoretical systems. They are complementary modes of analysis. And both are products of the human mind, abstract representations of the real world. In this respect, but not others, they are epistemologically similar. Like all scientific knowledge, both are abstract, incomplete and provisional.