Dealing with Descriptive Heterogeneity in Fuzzy-Set Qualitative Comparative Analysis (FsQCA)

This Paper aims to clarify fundamental aspects of the process of assigning fuzzy score to a condition while considering descriptive heterogeneity in fuzzy-set Qualitative Comparative Analysis (fsQCA). The QCA method is a theory-centered technique with a focus on testing conditions and determining which conjunctions of conditions lead to the outcome. Theoretical knowledge plays an important role in relation to representing conditions in fuzzy terms. The role of the theory-based calibration is justified by the scope of identifying coherent theoretical solutions. Consequently, descriptive heterogeneity, expressed by quantitative indicators, can be seen as optional in assigning final fuzzy score to conditions: it is not an intrinsic and necessary defining attribute of a concept but instead merely contributes to its description. In order to unite theoretical data and quantitative indicators, we refer to complex conditions called Fuzzy Multiple Attribute Condition (FMAC). Linked to this, we propose a new strategy that conceives the descriptive heterogeneity as a new non-idiosyncratic tuple to aggregate to each FMAC’s attribute. The minimal and necessary elements that determine the qualitative status of a case are derived by theoretical knowledge. Meanwhile, the non-idiosyncratic elements that determine how a case inside a set is similar to its ideal type are determined by specific empirical measurements. The empirical measurements are not necessary to determine the qualitative membership, hence, the ontological relationship between empirical measurement and theoretical definitional elements of concept is based on family resemblance. After defining what family resemblance is and explaining why it can be used in defining the descriptive heterogeneity in QCA, we employ an empirical example in order to extrapolate axiomatic properties able to determine fundamental principles of family resemblance aggregation. Finally, we identify an operator based on a Quadratic Mean that is able to aggregate fuzzy scores based on family resemblance.