Applying and assessing large-N QCA: Causality and robustness from a Critical Realist perspective

Applying qualitative comparative analysis (QCA) to large Ns relaxes
researchers’ case-based knowledge. This is problematic because causality in
QCA is inferred from a dialogue between empirical, theoretical, and casebased
knowledge. The lack of case-based knowledge may be remedied by
various robustness tests. However, being a case-based method, QCA is
designed to be sensitive to such tests, meaning that also large-N QCA
robustness tests must be evaluated against substantive knowledge. This
article connects QCA’s substantive-interpretation approach of causality to
critical realism. From that perspective, it identifies relevant robustness tests
and applies them to a real-data large-N QCA study. Robustness test findings
are visualized in a robustness table, and this article develops criteria to
substantively interpret them. The robustness table is introduced as a tool to
substantiate the validity of causal claims in large-N QCA studies.