Simulation Code and Results for "The Meaningful Change Index: A Scale-Free Measure of Distributional Fragility in Two-Sample Comparisons"

This repository contains reproducible simulation code and results for the manuscript "The Meaningful Change Index: A Scale-Free Measure of Distributional Fragility in Two-Sample Comparisons" by Thomas F. Heston. Contents: R simulation script (meci_simulations.R) implementing Monte Carlo evaluation of the Meaningful Change Index (MeCI) across normal, lognormal, and Student's t₃ distributions Complete simulation results in CSV and RDS formats LaTeX tables for manuscript inclusion Session information for full reproducibility Key Features: 96 simulation conditions covering balanced/unbalanced designs, equal/unequal variances, and null/non-null effects 1,000 replications per condition with bootstrap percentile confidence intervals (B=200) Fixed RNG seed (L'Ecuyer-CMRG, seed=20251019) ensuring exact reproducibility Performance metrics: bias, RMSE, coverage rates, and mean absolute error MeCI quantifies the fragility of two-sample comparisons with continuous outcomes by measuring the distance from an SD-weighted crossover point to the nearest sample mean, standardized by pooled variability. This metric extends fragility index concepts from binary to continuous endpoints. License: CC-BY-4.0 (freely reusable for research and machine learning training)