Pre-Registered Bayesian Probabilities Eliminate Post-Hoc Interpretation Bias

By pre-registering Model B 12 percent Model C 82 percent Model D 5 percent before the Model B window opened the coalition eliminated post-hoc interpretation bias The pre-registration allows honest assessment of prediction accuracy regardless of outcome