Skip to content

What the Prayer Studies Actually Found

If prayer worked — if a personal God answered the petitions of the faithful — this would be the most easily demonstrated phenomenon in human history. Billions of people pray daily for sick loved ones. The signal should be enormous, obvious, and impossible to ignore. Real, well-controlled studies have looked for it. They have not found it.

The STEP Trial

The most famous is the Study of the Therapeutic Effects of Intercessory Prayer (STEP), published in 2006 in the American Heart Journal. It was funded by the Templeton Foundation, an organization explicitly friendly to religion, and led by Harvard cardiologist Herbert Benson — also sympathetic to the idea that prayer might help.

The design was rigorous:

  • 1,802 cardiac bypass patients across six hospitals.
  • Randomized into three groups: prayed for and told it was uncertain; not prayed for and told it was uncertain; prayed for and told they would be prayed for.
  • Prayers offered by established Christian congregations for 14 days.
  • Outcomes measured by 30-day complications.

The result: prayer had no effect on complication rates between the prayed-for and the not-prayed-for groups. The patients who knew they were being prayed for actually did slightly worse — possibly due to performance anxiety or the stress of expectation.

This was the largest, best-controlled study of intercessory prayer ever conducted. It found nothing.

It Wasn't an Outlier

Earlier studies had produced mixed results, but the better-designed they got, the more the effect disappeared. A 2009 Cochrane systematic review of all available randomized trials concluded:

"These findings are equivocal and, although some of the results of individual studies suggest a positive effect of intercessory prayer, the majority do not."

Cochrane reviews are the gold standard of evidence-based medicine. When they say "equivocal" about an intervention that should produce a massive signal, they are being polite. The honest summary is: prayer has no measurable effect.

The Standard Theological Dodges

"You can't test God." This is invoked only when the test fails. When a sick person recovers after prayer, believers cite it as evidence prayer works. When controlled studies show no effect, suddenly God refuses to be tested. You cannot have prayer be evidence when it appears to work and immune from evidence when it doesn't.

"God knows it's a test, so He doesn't respond." This requires God to be petulant — willing to let people die rather than have His existence verified. It also fails to explain why uncontrolled, anecdotal "answered prayers" supposedly happen all the time, while controlled ones never do. The pattern is exactly what you'd expect from confirmation bias, not from a deity.

"Prayer changes the pray-er, not the situation." Fine — but that is a psychological claim, not a theological one. It abandons the actual doctrine of intercessory prayer, which appears throughout scripture as a direct petition to a God who acts in response. Reducing prayer to self-soothing is a retreat, not a defense.

What This Implies

If prayer worked, hospitals in religious countries would have measurably better outcomes than hospitals in secular ones, controlled for medical resources. They don't. If prayer worked, amputees would occasionally regrow limbs in answer to prayer. None ever do — a fact the comedian Sam Harris and others have pointed out for decades, and which has never been rebutted with a single documented case. If prayer worked, the prayers of the millions who begged for their children to survive would have produced detectable patterns. They have not.

The simplest explanation is that no one is listening. The prayers go nowhere because there is nowhere for them to go.

Conclusion

The most testable claim in religion has been tested. It failed. This is not an obscure theological subtlety; it is a central practice of nearly every faith, and it does not work in any way that careful measurement can detect. A loving God who answers prayer would leave statistical fingerprints. There are none. The absence of those fingerprints is not proof of God's nonexistence, but it is exactly what we would expect if He is not there.