Are regions within a country really independent islands? Do economic relations and effects really have a homogenous, unique size across an entire country? These two assumptions are often imposed implicitly in empirical economic and social research. In his doctoral thesis, Norbert Schanne discusses how statistical methods can deviate from this unrealistic model structure through employing spatial patterns in both observable variables and presumed relations. Opportunities to improve our understanding of the economy as well as chances and perils in the application of such methods are demonstrated in a number of studies on aspects of regional labour market dynamics.