Two large analyses conducted decades apart, using different populations and different methods, arrived at the same broad conclusion: people who attend religious services regularly live substantially longer, on average, than those who do not. The gap is large enough to have survived several attempts to explain it away, and the explanation that consistently fails to account for it is the one most people reach for first.
The finding is not about faith, or not in any way the research can measure. The part of religious life that the data keeps connecting to longer survival is not prayer, not belief, and not the content of what happens inside a religious service. It is the act of showing up, regularly, in a room with other people.
What two large analyses found
The figure most often cited in this literature comes from a 1999 study published in Demography, using nationally representative data from the National Health Interview Survey linked to mortality records. That study found a seven-year difference in life expectancy at age twenty between people who never attended religious services and those who attended more than once per week. People who never attended faced roughly 1.87 times the risk of death during the follow-up period compared to frequent attenders.
A more recent and methodologically careful analysis arrived at comparable results. Shanshan Li and colleagues, including epidemiologist Tyler VanderWeele of Harvard T.H. Chan School of Public Health, published a study in JAMA Internal Medicine in 2016 that followed 74,534 women enrolled in the Nurses’ Health Study for sixteen years. Women who attended religious services more than once per week had a 33 percent lower risk of all-cause mortality compared to women who never attended. The difference held across both cardiovascular disease and cancer mortality.
These are not small effects. A 33 percent reduction in mortality risk over sixteen years, in a sample of more than seventy thousand women, does not dissolve under standard statistical scrutiny. The Hummer et al. analysis used a nationally representative sample of men and women and adjusted for a wide range of sociodemographic and behavioral variables. Both papers acknowledge limitations. Neither collapses under methodological pressure.
The limits of what observational data can show
Observational studies of this kind cannot establish causation, and it is worth stating that plainly before going further. People who attend religious services regularly differ from those who do not in ways that are difficult to fully disentangle statistically: family background, baseline health, existing social networks, other behavioral habits that accompany a religious life. The Li et al. study attempted to address reverse causation (the concern that healthier people may simply be more able to attend) by tracking the relationship over sixteen years and applying statistical methods designed to reduce that bias, and the association persisted across the follow-up period. But the finding remains correlational, and the research cannot rule out every possible confound.
The Nurses’ Health Study enrolled only women, which limits how far those results can be generalized to the full population. The analysis is now more than twenty-five years old, drawn from a period when American religious participation looked different from today in both scale and composition. The consistent pattern across datasets and decades is meaningful; the specific figures should not be read as precise universal values. They are findings from particular samples, at particular historical moments.
Why prayer alone does not account for it
In a 2017 paper for Current Directions in Psychological Science, VanderWeele reviewed the accumulated evidence on which aspects of religious life appear to drive the health associations. His conclusion was pointed: the associations between religious participation and health and flourishing are substantially stronger for communal religious participation than for either spiritual-religious identity or private religious practices.
Private practice, including prayer and scriptural reading, showed little independent association with mortality after accounting for service attendance and other variables. Identifying with a religious tradition without regularly attending services was similarly unassociated with the survival advantage. The measurable protective relationship is with attendance specifically, with the regular, embodied practice of being present with others in a structured communal setting.
This does not speak to the theological or personal value of prayer. That is a different question entirely, and one the research does not touch. What the research addresses is narrower: among the measurable dimensions of religious life, which ones predict longer survival? The answer, across multiple datasets, is communal attendance.
What appears to be doing the work
If not prayer, then what?
VanderWeele’s review identifies several proposed pathways: higher rates of social support among frequent attenders, lower rates of smoking, lower rates of depression, and a greater sense of meaning or purpose in life. These mechanisms are not mutually exclusive and probably operate together in some combination.
Social support is the most intuitive of the candidate explanations, and there is reasonable evidence for it. But what stands out in VanderWeele’s research is that religious service attendance predicted survival beyond what standard social connection variables — including marriage, friendship networks, and family contact — could fully account for. The attendance effect appears to exceed what equivalent social connection outside a religious community can fully account for.
That gap is what makes the finding genuinely interesting rather than simply confirming that social connection matters. Something specific may be happening in the structure of congregational life, whether that is the intergenerational character of religious communities, the regularity of gathering around shared ritual, the explicit orientation toward something larger than the individual, or some combination of those elements, that does not appear to be replicated by comparable amounts of secular socializing. These are plausible candidate explanations, not established mechanisms. VanderWeele’s review is careful to note that social support alone explains only a small portion of the measured effect, which means the rest remains, at present, incompletely understood.
What the research cannot say
These studies do not argue that people should attend religious services in order to live longer. That framing misunderstands what observational epidemiology can support. The findings identify a pattern in population-level data; they do not constitute a prescription for individuals.
What the pattern may be pointing toward, at a more structural level, is something about the difficulty of replicating what longstanding, multi-generational, obligation-structured community provides in terms of social integration. Religious services are one of the few remaining social institutions in American life that gather people of different ages, family relationships, and circumstances in the same room, around a shared practice, on a predictable schedule, across decades. That organizational feature, rather than any specifically religious content, may be what the mortality data is partly tracking.
Whether comparable health effects can be produced by non-religious institutions that provide similar regularity and social integration is a question the research has not directly tested. What it has tested, in multiple large samples over several decades, is whether those who participate regularly in religious communities live longer. The answer continues to be yes, with an effect size that demands a serious explanation rather than a quick dismissal. The part of religious life most responsible for that answer appears to be the room full of other people, rather than what is said or believed inside it.
We are writers and parents, not clinicians or epidemiologists. This article is a careful reading of the available research, not medical guidance.
