Multiverse Analysis for Reporting Robustness to Analytical Flexibility - Love Replications Week 2026

Multiverse analysis is a comprehensive and systematic approach to reporting the robustness of results across the defensible “garden of forking paths.” It does so by systematically specifying and computing all defensible combinations of analytic decisions applied to a single dataset, and reporting the result obtained from each. By explicitly acknowledging analytical uncertainty, multiverse analysis increases transparency and reduces the risk of overstating robustness and reproducibility. These slides accompanied a talk on multiverse analysis at Love Replications Week, 2nd March 2026.