Measuring Multidimensional Social Identification Using a Conjoint Experiment

Abstract

The concept of social identity has gained prominence among empirical social scientists in the recent past. Yet, its measurement remains elusive. Despite a multiplicity of approaches, no measure currently exists that provides a holistic, multidimensional, and relative assessment of subgroups and individuals’ (hidden) identification with different social groups at reasonable costs. This unfortunate circumstance complicates a comparable, realistic, and integrated assessment of social identity that adheres to its theoretical foundations. Confronting this gap, we propose a method that combines recent advances in conjoint analysis with the advantages of Bayesian hierarchical modeling. Our method offers theoretically grounded and reliable estimates of identification with any social group, along with its relative importance for overall social identity. Through an original survey conducted in Germany, we demonstrate the measure’s utility and validity with respect to partisan and national identity. We suggest potential applications of our measure and discuss implications for future work on social identity.

Keywords: Social Identity, Conjoint, Preferences, Bayesian, Hierarchical Modeling, Measurement