Choice in Classification: The Subjectivity Ofobjective Methods¹

A number of recent studies have presented classifications of geographical data using numerical methods which have been adopted from other scientific disciplines. Before a classification can begin, however, the investigator must make three sets of decisions regarding the exact process to be employed. Some of his choices are predetermined by the nature of the data and the research design but others may be taken on subjective grounds alone. Computational examples demonstrate the effects of third stage choices on the results of two classifications. The results are analogous to the general geographical problem of regional cores and zones of transition. All methods extract virtually the same core but choice can govern the classification of units within the zone of transition. Classification is thus basically a subjective process, despite the apparent objectivity of the methods employed.