Argument from Petitionary Answered Prayer

Petitionary prayer is prayer that asks God for something specific: healing from a named illness, safe travel, provision, or deliverance from danger. Unlike prayers of praise or confession, petitionary prayer makes a request whose fulfillment or non-fulfillment can in principle be observed. This is what makes it, unlike much religious experience, at least partially amenable to empirical investigation, and it is why researchers from biblical scholars to cardiologists have tried to study it directly.

The argument from answered prayer proposes that when specific petitions are followed by outcomes that seem improbable given ordinary causal processes, this constitutes evidence, however modest, that a personal agent is responding to the request. It is important to be clear from the outset that this is presented here as a weak, cumulative argument, one among many considerations in a larger cumulative case, not a standalone proof of God's existence.

The argument's plausibility depends heavily on how one weighs three bodies of evidence: large collections of testimonial reports, a smaller number of field studies attempting direct physiological measurement, and randomized controlled trials designed to eliminate confounds. These three bodies of evidence, as we will see, do not point in the same direction, and a scholarly treatment must hold that tension honestly rather than resolve it by ignoring inconvenient data.

Craig Keener, a respected New Testament scholar at Asbury Theological Seminary, undertook an unusual project in his two-volume Miracles: The Credibility of the New Testament Accounts (2011). Motivated initially by a desire to defend the historical plausibility of the miracle accounts in the Gospels and Acts against Humean skepticism, Keener assembled hundreds of contemporary reports of healings following prayer, drawn from interviews, missionary records, and in some cases prior medical documentation.

Keener is careful to note the limitations of his method: he is a historian and biblical scholar, not a clinician conducting trials, and most of his cases rely on testimony that cannot be independently re-verified decades later. Nevertheless, he argues that a subset of cases, particularly those involving physician-documented diagnoses of conditions such as blindness, paralysis, or advanced organic disease followed by prompt and complete resolution, are difficult to explain by ordinary means, and that dismissing all such testimony a priori reflects a philosophical commitment to naturalism rather than a neutral historical method.

Critics reply that testimonial evidence, even when it includes some documentation, remains vulnerable to selective memory, retrospective reinterpretation, and the absence of a comparison group of people who prayed and were not healed. Keener's compilation is valuable as a catalogue of candidate cases worth investigating, but by itself it cannot establish causation; it raises the question rather than settling it.

Candy Gunther Brown, a historian of American religion at Indiana University, has taken a further step by attempting direct physiological measurement in the field. In a widely discussed 2010 study published in the Southern Medical Journal, Brown and colleagues used pulse oximetry and standardized audiometric and vision tests to measure hearing and vision in Mozambican villagers before and immediately after proximal intercessory prayer (prayer offered with hands laid on the person), reporting statistically significant improvements in several participants.

Brown's later book Testing Prayer: Science and Healing (2012) situates this fieldwork within a broader, careful discussion of the methodological challenges of studying prayer scientifically, including the difficulty of blinding, the absence of a no-prayer control group in the Mozambique study, and the problem of measurement reliability in field conditions with limited equipment. Brown herself is notably measured in her conclusions, presenting the data as suggestive and worth further controlled investigation rather than as proof of divine healing.

The Mozambique study illustrates both the promise and the limits of field research on prayer: it moves beyond pure testimony to instrument-based measurement, which is a methodological improvement, but the absence of randomization and control groups means regression to the mean, measurement error, and expectation effects on the part of examiners cannot be excluded. It remains one data point among many that call for a larger, better-controlled follow-up that, to date, has not been conducted at comparable scale.

The most rigorous test of intercessory prayer to date is the Study of the Therapeutic Effects of Intercessory Prayer (STEP), led by Herbert Benson and colleagues and published in the American Heart Journal in 2006. The trial randomized over 1,800 cardiac bypass surgery patients into three groups: those prayed for without their knowledge, those not prayed for without their knowledge, and those prayed for who were told with certainty they would be prayed for. The trial found no significant difference in complication rates between the uncertain groups, and, strikingly, a higher rate of complications among patients who knew with certainty they were being prayed for.

STEP was a methodologically strong trial: large sample size, randomization, and blinding of patients (where feasible) to reduce placebo-type confounds. Its null and slightly negative results are widely regarded in the medical and philosophical literature as the strongest empirical evidence against any straightforward efficacy claim for petitionary prayer, and any honest treatment of this argument must take that result seriously rather than explain it away reflexively.

Earlier and smaller trials, including the MANTRA II trial on cardiac angioplasty patients led by Mitchell Krucoff and published in The Lancet in 2005, similarly failed to find significant benefits from intercessory prayer. Taken together, the controlled-trial literature offers a consistent pattern: whatever is happening in the anecdotal and field-study reports, it does not manifest as a reliable, replicable, mechanism-like effect under experimental conditions designed to detect one.

Several well-understood psychological and statistical phenomena can generate a steady stream of apparently answered prayers without any causal contribution from prayer itself. Regression to the mean predicts that unusually severe symptoms will often improve on their own, simply because extreme measurements tend to be followed by less extreme ones; a person who prays during a symptom flare is likely to see improvement regardless of prayer's efficacy.

Selection and reporting effects compound this: a person whose prayer for healing goes unanswered rarely publicizes the failure, while a single striking recovery becomes a widely retold testimony. Daniel Kahneman's work on cognitive biases, particularly confirmation bias and the illusion of causal narrative, explains why humans reliably overestimate the frequency and significance of hits while discounting misses. Given the enormous number of petitionary prayers offered daily worldwide, base-rate considerations alone predict that some strikingly improbable-seeming recoveries will occur by chance.

Medicine also documents rare cases of spontaneous remission in serious diseases, including certain cancers, entirely independent of religious practice, which shows that dramatic, medically surprising recoveries are not by themselves evidence of a supernatural cause. A scholarly defender of this argument must concede that these mechanisms plausibly account for the great majority of reported cases, reserving the argument's remaining weight only for a small residue of cases involving irreversible organic pathology reversing suddenly, an evidential residue that is itself difficult to verify retrospectively and vulnerable to charges of original misdiagnosis.

Given this mixed picture, what should a careful thinker conclude? The controlled-trial evidence weighs against treating petitionary prayer as a reliable, testable mechanism of divine action, and this should be stated plainly rather than minimized. At the same time, a null result in a mechanistic clinical trial does not straightforwardly refute a theological picture in which God acts as a free personal agent responding to relationship rather than as an impersonal force triggered predictably by ritual, a point Richard Swinburne has pressed in his broader cumulative case for theism.

This asymmetry, however, comes at a cost: if a proponent's account of God's action is compatible with both positive and null trial results, then the trial results cannot straightforwardly count as confirming evidence either, which limits how much weight the argument can bear even for the believer. What remains is a small residue of individually striking, hard-to-explain anecdotal cases, weighed against a strong base of statistical and psychological explanations and a discouraging controlled-trial record.

The most defensible conclusion is that petitionary prayer functions, at best, as a very weak corroborating consideration within a larger cumulative case for theism built primarily on other grounds, such as cosmological, fine-tuning, or moral arguments, rather than as independent evidence capable of shifting a skeptic's credence on its own. Presented honestly, with its strongest counter-evidence in view, the argument retains modest interest without overreaching its actual evidential force.