Direction and specification in a moderated mediation analysis of interprofessional education: a commentary on He et...

He et al. (2026) report that agreeableness and openness are indirectly associated with emotional intelligence (EI) through perceived group cohesion (GC) and perceived team emotional intelligence (TEI) among 344 health professions students completing an interprofessional education (IPE) course, and that attitudes toward IPE (ATIPE) moderate the GC→TEI path. This commentary raises three concerns. First, the estimated interaction is significantly negative in both models: the conditional association between GC and TEI, and the serial indirect effect, are strongest at low ATIPE and weakest at high ATIPE, as the Results correctly state. The Discussion nevertheless describes favorable attitudes as enhancing GC's impact on TEI (a directional contradiction), and the practical recommendations speak of maximizing and amplifying these effects in language that at minimum requires clarification. Second, assuming the supplementary appendices report conventional full-model F statistics for the analysis sample of 344, the R² and F values are numerically consistent with grade and discipline each entering the models as a single predictor, and cannot be reconciled with the stated dummy coding of a sample spanning seven disciplines; and the conditional effects are reproducible only under an unreported mean-centering of ATIPE. Third, statistical support for the TEI-only indirect effect is specification-dependent (present in Model 6 but not in Model 91), and the wording offered for the latter cannot explain the contrast between the two models. I outline why each point matters and which clarifications would resolve them.