in the norwegian context it is for instance e.g. hard to tell what is common between municipalities varying in population from 200 to more than 510000 it is a pure statistical observation that small municipalities will tend to have extreme values on indicators. combined with very varying area (between 6 to 9000 sg. km) and differences in population distribution, such a geographical basis cause problems for mapping-and for regional analysis. a thematic map-and a regional analysis-will contain a certain level of "statistical noise:, much of the observed regional variations and patterns may be a function of the structure of the geographical units, while, real variation are somewhat hidden. but still we too often use these non-comparable units as a basis for ranking based on a selection of indicators