Comparing Two Worlds. Theoretical and Empirical Assumptions in Discriminant Validity Testing
(A2021-92931)

AUTHORS

Thomas Niemand, Clausthal University of Technology, Institute of Management and Economics, Department of Market Research; Robert Mai, Grenoble Ecole de Management

ABSTRACT

Discriminant validity testing has become a standard procedure for checking as to whether fac-tors or latent variables overlap. Two approaches, the Fornell-Larcker criterion (FLC) and the Heterotrait-monotrait ratio (HTMT), have gained considerable attention in marketing research, but which pursue different conceptual assumptions. The recently proposed HTMT approach adopts an empirical route, whereas FLC follows a theoretical route. We propose that ignoring this distinction provokes misleading conclusions, such that the seeming superiority of HTMT is a result of the empirical assumptions about acceptable factor correlations. This research pin-points that both approaches’ performance strongly depends on the assumptions made. Account-ing for this so-far ignored aspect overturns previous findings, in that FLC outperforms HTMT when pursuing a conservative approach where critical correlations are relatively low (.7), while the opposite is true for stronger acceptable correlations.