===== Methodological Overreach =====

'''Methodological overreach''' refers to the inappropriate extension of a study’s methodology beyond what the data or design can reliably support.

It occurs when researchers:
  * Make claims that exceed the limitations of the study design (e.g., causal claims from observational data).
  * Apply complex statistical or machine learning tools without adequate justification or validation.
  * Generalize findings to populations or clinical settings not represented in the data.
  * Use sophisticated methods (e.g., SHAP, deep learning) to add perceived value without improving scientific rigor.

==== Examples in clinical research ====
  * Using a cross-sectional dataset to build a tool for **longitudinal prediction**.
  * Claiming **preventive or therapeutic impact** from an observational model.
  * Presenting **model performance metrics** (like AUC) as proof of clinical utility.

==== Consequences ====
  * Misleading conclusions
  * False sense of confidence in tools or interventions
  * Poor translation into real-world clinical practice

'''In summary:''' methodological overreach undermines scientific integrity by overstating what a study can truly demonstrate.