Toward comparative polycentricity scores: Assessing variations in regional delineation and subcenter identification

Abstract

Urban spatial structure is increasingly characterized by polycentricity, the presence of multiple interconnected centers of similar size. Polycentricity indicators, which influence research and policies related to urban development, rely on three phases of analysis: (a) delineating regions, (b) identifying subcenters within these regions, and (c) operationalizing polycentricity; and each phase contains decision points for analysts. This paper argues that polycentricity methodologies should be thought of in terms of pathways, then systematically applies 15 such pathways to the case of German regional polycentricity and compares the results. Findings suggest that questions of polycentricity are more robustly measured by comparing across multiple regional delineation methods and selection of subcenters, then looking for signs of agreement or disagreement. When possible, constructing regions from larger areas through bottom-up methods tends to avoid the biases of administratively defined regions. When this is not possible, statistical approaches to subcenter identification can serve as a check to avoid forced selection of subcenters in poorly defined regions.