Claims about religion’s beneficial effects on physical and psychological health have received substantial attention in popular media, but empirical support for these claims is mixed. Many of these claims are tenuous because they fail to address basic methodological issues relating to construct validity, sampling methods or analytical problems. A more conceptual problem has to do with the near universal lack of atheist control samples. While many studies include samples of individuals classified as “low spirituality” or religious “nones”, these groups are heterogeneous and contain only a fraction of members who would be considered truly secular. We illustrate the importance of including an atheist control group whenever possible in the religiosity/spirituality and health research and discuss areas for further investigation.