To believe or not to believe Open science and replication in the psychology of religion

Over the last decades, scholars in the psychology of religion have been trying to understand the origin, the function, and the consequences of the fascinating phenomenon that religion is. In the current dissertation, we aimed to contribute to the scientific inquiry of religion, not by proposing new theories or adding hypotheses, but by rigorously testing influential existing ones. Specifically, we tried to add to previous research by reexamining previously reported effects while (1) including large and diverse samples, (2) applying open science practices and Bayesian modeling techniques (3) conducting replications of key effects plus potential alternative explanations or effect (e.g., correlational instead of experimental effects), (4) critically assessing and visualizing patterns in the raw data, and (5) applying new tools to assess robustness (e.g., a many-analysts approach, analysis blinding).
Based on the findings in this dissertation, the –perhaps unsurprising– conclusion we can draw is: some effects are and some aren’t replicable. We found that religiosity indeed seems positively related to self-reported well-being; religiosity seems predictive of the tendency to make post-mortem continuity judgments of psychological states (e.g., love); religiosity seems related to credibility ratings for gobbledegook statements, and a reduced relative difference for those from a scientist compared to a spiritual guru. At the same time, we obtained convincing evidence for the absence of other effects: an experimental attenuation of personal control does not seem to activate a compensatory mechanism of belief in a controlling God and neural markers of cognitive conflict and error processing do not seem to be associated with religiosity.