Scientific studies on intercessory prayer show it has no measurable effect on healing outcomes — prayer simply doesn't work.

Why people find it persuasive

Large controlled studies (most famously the 2006 STEP study on cardiac surgery patients) exist specifically testing whether being prayed for by strangers affects recovery, and headline results reported no significant positive effect (and in one arm, a negative one), which sounds like a direct empirical refutation of a specific religious claim.

Important distinctions

• • Prayer as a testable mechanical intervention analogous to a drug vs. prayer as personal communication with a free, personal God who is not obligated to respond in a statistically detectable uniform pattern.
• • Third-party intercessory prayer by strangers for anonymized subjects vs. the relational, communal, and covenantal prayer contexts described in Scripture.
• • 'No detectable statistical effect in this specific study design' vs. 'prayer categorically does nothing, ever.'
• • Physical healing as the sole purpose of prayer vs. prayer's broader purposes (relationship, submission, comfort, transformation) in Christian theology.

Christian response

• • Look honestly at what the STEP study and similar studies actually tested: whether prayer by strangers, for anonymized patients who may or may not have known they were being prayed for, following a specific fixed protocol, produces a statistically detectable physiological effect on a specific outcome measure — this is a much narrower and more mechanistic model of prayer than what most theistic traditions actually teach or practice, and a null result in that narrow paradigm does not straightforwardly generalize to 'prayer does nothing' in the broader relational sense Scripture describes.
• • Point out the significant methodological and theological problem of treating God as a variable to be manipulated and measured like a drug in a double-blind trial: Christian theology holds that God is a free personal agent who responds (or does not) according to his own wisdom and purposes, not a mechanical force with a fixed dose-response curve — this is not merely a convenient escape hatch, since Scripture itself explicitly criticizes exactly the mindset of trying to compel or test God for a guaranteed sign or result (Jesus' rebuke to those seeking a sign in Matthew 16, or Deuteronomy 6:16's 'do not put the Lord your God to the test').
• • Note the actual body of research on prayer and health is considerably more mixed than 'prayer doesn't work' headlines suggest, including some studies showing positive correlations for personal or communal (as opposed to third-party anonymous) prayer and religious practice on outcomes like recovery time, mental health, and longevity, and that this remains a genuinely contested area of behavioral medicine research rather than a settled negative finding.
• • Distinguish the broader purpose of Christian prayer from a narrow healing-request-and-measurable-result model: Scripture presents prayer primarily as relational communication, worship, confession, and alignment of the believer's will with God's, with physical healing as one possible but never guaranteed outcome (even Paul reports his own request for healing being denied in 2 Corinthians 12:7-9, with a theological reason given rather than a scientific mechanism), so 'this study found no detectable healing effect' does not touch the primary claims Christian teaching actually makes about what prayer accomplishes.
• • Acknowledge honestly that this remains one of the genuinely difficult and unresolved areas where a supernaturalist worldview and empirical measurement are hard to fully reconcile, and that Christians should resist both overclaiming prayer as a reliably testable mechanism (which sets up predictable disappointment and false claims) and dismissing the studies outright without engaging their actual methodology and limits.

Christian counterresponse

Theists respond this looks unfalsifiable only if one first assumes prayer was supposed to function like a mechanical variable with a predictable dose-response signature; given that Christian teaching independently and consistently (not just after the fact) describes God as a free personal agent whose responses are not guaranteed or mechanically triggered, a null result in a study designed around exactly that mechanical assumption is unsurprising on Christian premises rather than being an ad hoc rescue invented only when the data came back negative.