It’s about Fertility, stupid! The Evolutionary Adaptivity of Religion
Religiosity (defined as behavior towards superempircal agents) is today clearly adaptive: Members of competitive religious communities are building stronger families with more offspring worldwide as their secular neighbours of the same education and income levels. This is observable in empirical studies, censusses worldwide, as well as in case studies (i.e. Amish, Hutterites, Mormons, Orthodox Jews). In contrast, non-religious populations and those religious communities who do not build and support families inevitably succumb to cultural evolution (i.e. late Greek and Roman Polytheism, Gnostic groups, the Shakers) and are replaced by demographically successful religious competitors. Please note: We found many religious traditions that were able to attain high levels of fertility throughout the generations. But in sharp empirical contrast, we didn't find a single non-religious community, movement or population that was able to retain at least replacement level (two births per woman) for a century!
Both genders are exercising evolutionary choices: Women more often favoring stronger marital and communal security (as in strict, monotheistic religions) and offering privileged roles to men in return - whereas the latter tending in higher percentages toward hedonistic, poly- or atheistic worldviews. Even celibate institutions (offering services like policing rule-abiding behavior, strengthening communities and families, child care, education, healing and more) can be adequately described as religious helpers-at-the-nest swapped between family networks (i.e. by Berman et al. 2007).
Survival benefits of religiosity (like health, welfare, security and more) might also apply. But from a biological (and biocultural) perspective, the strong intra-generational "reproductive advantage of religion(s)" (F.A. von Hayek) is setting the decisive benchmark of evolutionary success.