Discussion on ''Nonparanormal Adjusted Marginal Inference'' by Susanne Dandl and Torsten Hothorn.

We comment on the work of Dandl and Hothorn, who propose Nonparanormal Adjusted Marginal Inference, a covariate-adjusted method for estimating marginal treatment effects in randomized studies. We discuss their likelihood-based inferential strategy for parameterizing and estimating marginal contrasts and compare this fully parametric approach with established semiparametric methods. We then critically assess its robustness and practical suitability for routine use in clinical trials.