Robustness Report: A Bayesian Multiverse Analysis of the Water Tasting Experiment

The current Report assesses the robustness of the evidence for a preference of Chaudfontaine over tap water (Wagenmakers & Bartoš, 2025) through a Bayesian multiverse analysis. Varying data inclusion criteria and prior settings, we found that 93.8% (45/48) of the analysis paths resulted in evidence in favor of the hypothesis. At the same timeHowever, only one third (17/48) of the paths indicated strong evidence, all based on a small subset of the data. We therefore conclude that the claim that people prefer Chaudfontaine water over tap water lacks robust evidence in this particular experiment.