Operational criteria for context-dependent choice of statistical school

Determine operational decision criteria for selecting an appropriate normative statistical framework under ambiguous research contexts, including whether to elicit and use expert priors for subjective Bayesian analysis when expert competence is uncertain, and whether a given number of replications (e.g., 100) suffices to justify adopting frequentist methods.

Background

The essay advocates a context-dependent approach to choosing between statistical schools but acknowledges unresolved practicalities in ambiguous cases. Specifically, when expert quality is uncertain, it is unclear whether subjective Bayesian analysis via prior elicitation is warranted.

The authors also highlight uncertainty about sample-size thresholds that justify frequentist methods in domains where replications are typically scarce, underscoring the need for concrete, context-sensitive decision rules.