Researchers have been measuring prayer for about sixty years. The studies fall into three categories, and the conclusions you can draw from them are completely different depending on which category you are reading. Most popular writing about prayer research confuses the categories, which is how you get the same studies cited as proof that prayer works miracles and as proof that prayer does nothing.

The categories are not exotic. Once you know which one a study belongs to, what it found tends to be unsurprising. Here is the map.

Category 1: Studies of what happens to the body when a person prays

These are the studies on contemplative prayer, repetitive prayer, and structured devotional practice as physiological events. The researcher attaches sensors to a person, asks them to pray for twenty minutes, and measures what changes.

What these studies have found, consistently, since Herbert Benson's foundational work in the 1970s:

The 2008 Harvard study by Dusek and Benson, published in PLoS ONE, traced more than 2,000 genes whose expression shifted in long-term practitioners of the relaxation response. The Italian cardiologists Bernardi and colleagues, in their 2001 BMJ paper, showed that reciting the Rosary at its natural pace slows breathing to six breaths per minute, the rate that maximizes cardiovascular resilience. Andrew Newberg's neuroimaging work has tracked the brain changes associated with deep contemplative prayer in Franciscan nuns and Tibetan monks.

This body of research is solid. It has been replicated across multiple institutions, decades, and types of contemplative practice. The effects are real. They are also not unique to prayer: meditation, yoga, deep breathing, and Tai Chi produce essentially the same physiological pattern. Prayer is one member of a larger family of practices that have these effects.

If someone asks whether prayer is "good for you" in a measurable bodily sense, this is the research that answers yes.

Category 2: Studies of what happens to a person's psychology and life over time

These look at the long-term effects of being a person who prays. They measure subjective well-being, depression rates, anxiety, sense of meaning, social connection, and life expectancy.

Harold Koenig at Duke University has spent four decades reviewing this literature. His 2012 systematic review in ISRN Psychiatry surveyed over 1,200 studies on religion, spirituality, and health. The pattern across the literature: regular prayer practice is associated with lower depression rates, lower anxiety, better recovery from illness, stronger immune markers, and modestly longer life expectancy.

The size of the effect varies by study and outcome. The direction is almost uniformly positive.

One important caveat. These are association studies, not controlled experiments. Researchers cannot randomly assign half the population to pray for thirty years. What they can do is find people who do pray regularly, find people who don't, and compare them. This means the effects may reflect prayer itself, but they may also reflect that people who pray regularly tend to belong to communities, have established habits, and report higher purpose, all of which independently improve health outcomes.

The honest reading: there is a strong correlation between regular prayer and better health. Whether the correlation is causal is harder to establish. Most researchers in the field think the answer is probably "partially," and that the social and habit-forming aspects of prayer practice carry a meaningful share of the effect.

Category 3: Studies of intercessory prayer

These ask a different question entirely. Does praying for someone, without telling them, change their medical outcomes?

The answer from the best evidence available is: no, in any way researchers can measure.

The most famous study is the 2006 Benson STEP trial in cardiac patients, published in the American Heart Journal. The largest, longest, and most carefully controlled study of intercessory prayer ever conducted, funded by the Templeton Foundation at a cost of $2.4 million. It found no difference in recovery between patients who were prayed for and patients who were not. Patients who were told for certain that strangers were praying for them did slightly worse, which the researchers attributed to performance anxiety.

The STEP trial was the largest. It was not the only one. Earlier studies (Byrd 1988, Harris 1999) found small positive effects that did not survive replication. Later studies (Astin and Astin 2000 meta-analysis, Roberts and Ahmed 2009 Cochrane Review) have consistently failed to find evidence that intercessory prayer changes medical outcomes.

This is the research that gets cited as "prayer doesn't work." That citation is accurate as far as it goes. Intercessory prayer, in the specific sense of praying for a stranger's medical recovery without telling them, has been studied extensively and has not produced measurable effects in the body of evidence that exists.

What this means in practice

The three categories of research tell three different stories. Most prayer is the first category: people praying themselves, for themselves, on a regular basis. The research on that is robust and positive. Most popular writing about prayer's "scientific evidence" cites the third category, where the research is robust and negative, and applies the conclusion to the first category, where it does not apply.

The simplest summary that does not distort the evidence:

These are three different findings about three different practices. The first two are good news. The third is what science has measured. The deeper question of whether prayer reaches God remains where it has always been, which is outside what science is designed to measure.

The fact that all three findings can be true at once is the part most popular coverage gets wrong.