Selecting the Optimal Number of Archetypes in Environmental and Natural Resource Governance: Procedure and Metrics Based on Formal Concept Analysis

Abstract

Research on the governance of environmental and natural resources increasingly employs archetype analysis, but still requires some methodological advances. One gap we address are rigorous criteria and metrics for selecting the appropriate number of archetypes. Another gap is that some commonly combined methods, e.g., formal concept analysis (FCA), have difficulties in treating outcome attributes differently than explaining attributes. Addressing both gaps is important, since archetype analysis aims to identify a suite of different attribute configurations that explain one outcome of interest (equifinality), and different archetypes should be combined like building-blocks when explaining a single case (modularity). This paper thus proposes a general, orderly and replicable procedure to enhance standard FCA to treat outcome attributes independently, and to choose the appropriate number of archetypes based on the metrics that compare different archetypes and suites of archetypes. Some steps are computational and can be done with conventional software packages, while other steps are qualitative. We illustrate the procedure by replicating the data analysis part of previously published archetype analyses of environmental and natural resource governance. This admits to delve into the particularities of each step, show how to circumvent possible pitfalls, and assess the strengths and limitations of the procedure.