45th Congress of the European Regional Science Association: "Land Use and Water Management in a Sustainable Network Society", 23-27 August 2005, Amsterdam, The Netherlands
There is a very well known fundamental problem in spatial data analysis namely that all results of quantitative methods are potencially influenced by the way of spatial delimitation. This problem is mostly called modifiable areal unit problem (MAUP). However, beside the rich tradition in the empirical spatial data analysis, the effect of MAUP on putting forward and testing a theory and the effect on model-building is an issue rarely investigated. The MAUP creates the need for the investigation of the connection between theories and data and the micro-macro dualism. My paper presents the epistemological background of the problem and gives illustrations of the negative consequences of ignoring them in regional macroeconomics.