Reproducibility Package for "From Statistical Significance to Practical Relevance: A Monte Carlo Comparison of NHST, TOST, Second-Generation p-Values, REACT, and Bayesian HDI-ROPE"

This reproducibility package accompanies the study “From Statistical Significance to Practical Relevance: A Monte Carlo Comparison of NHST, TOST, Second-Generation p-Values, REACT, and Bayesian HDI-ROPE.”
The package contains the canonical R implementation, simulation-design files, deterministic random seeds, validated aggregate simulation outputs, quality-control reports, publication table and figure source data, manuscript-facing assets, and R session information.
The study compares five inferential procedures: null-hypothesis significance testing (NHST), two one-sided tests (TOST), second-generation p-values (SGPV), REACT, and Bayesian HDI-ROPE. The reference Monte Carlo design considers six sample sizes, three practical-relevance thresholds, 29 true mean differences, and 20,000 replications per data-generating scenario, together with sensitivity analyses under heteroscedastic normal, heavy-tailed, and skewed distributions.
Raw scenario-level checkpoint files are not included in this compact deposit because they can be deterministically regenerated from the supplied code and random seeds. The aggregate outputs required to reproduce and verify the reported tables, figures, operating characteristics, sensitivity analyses, and quality-control results are included.