Which Robustness Value Should You Report for an Indirect Effect? Sensitivity Analysis for Mediation Across Tests

Simulation code, simulation results, and the worked-example script for the article “Which Robustness Value Should You Report for an Indirect Effect?” Simulation 1 (normal errors, 13,489 data sets) and Simulation 2 (heteroskedastic errors, 3,494 data sets) compare robustness values for the joint significance, Sobel, Monte Carlo, and percentile bootstrap tests of an indirect effect. mc_analytic_check.R computes the exact Monte Carlo robustness values (Proposition 4). The example uses the economic stress data distributed with Hayes (2022) and the R package processR; the data file is not included. Requires R and RobustMediate 0.3.1 (https://github.com/subirhait/RobustMediate).