Exoplaneteers Keep Overestimating Sigma Significances

Frequentist and Bayesian statistics in exoplanetary atmospheric studies.

The preprint identifies a drift within the current paradigm of exoplanetary atmospheric characterization. Specifically, it addresses the misapplication of a statistical conversion (Sellke et al. 2001) by exoplaneteers. The authors argue that while the conversion itself is valid, its frequent misuse leads to an overestimation of significance levels, inflating confidence in detections. This practice represents a drift from rigorous Bayesian statistical interpretation, not a full-blown crisis or revolution, as the core methodology of Bayesian model comparison remains accepted. The highlighted misapplication of the conversion necessitates clarification and standardization within the field, making 'Model Drift' the most accurate classification. There is however also an argument to be made for 'Model Crisis' since the authors are questioning the validity of some highly-cited results in the field (eg. detection of DMS around K2-18b). It's therefore plausible that community-wide adoption of the authors recommendations triggers a model crisis event in the near future (or could already be interpreted as one at present).