Prayer & Promises
The Bible makes concrete promises about prayer that controlled studies of distant intercessory prayer have not consistently confirmed.
7 min readUpdated
The New Testament contains a number of passages promising that prayers will be answered. Several describe results that can be seen, which makes them possible to compare against evidence. Praying for someone else's recovery has been tested in controlled medical trials since the 1980s.
This page collects the promises, the studies, and the answers commonly given when the two do not line up.
The promises
| Citation | Verse |
|---|---|
| Matthew 7:7-8 | "Ask, and it will be given to you... For everyone who asks receives." |
| Matthew 17:20 | "If you have faith like a grain of mustard seed... nothing will be impossible for you." |
| Matthew 18:19 | "If two of you agree on earth about anything they ask, it will be done for them by my Father in heaven." |
| Matthew 21:22 | "Whatever you ask in prayer, you will receive, if you have faith." |
| Mark 11:22-24 | "Whatever you ask in prayer, believe that you have received it, and it will be yours." |
| John 14:13-14 | "Whatever you ask in my name, this I will do... If you ask me anything in my name, I will do it." |
| John 15:7 | "If you abide in me, and my words abide in you, ask whatever you wish, and it will be done for you." |
| John 16:23-24 | "Whatever you ask of the Father in my name, he will give it to you." |
| James 5:14-15 | "The prayer of faith will save the one who is sick, and the Lord will raise him up." |
| 1 John 5:14 | "If we ask anything according to his will he hears us." |
Most of these attach no limit to what may be asked — the words used are "whatever," "anything," and "everyone." Where a condition does appear it is faith, agreement between two people, or asking in Jesus's name. Only the last passage conditions the promise on the request matching God's will, which is the basis of a common objection below.
James is the most specific: it gives a procedure — call the elders, anoint with oil, pray in faith — and a physical result, which is concrete enough to test.
The studies
Researchers have tested prayer by randomly assigning patients to be prayed for or not, then comparing how each group does. Random assignment matters, because it rules out healthier patients simply attracting more prayer.
| Study | What it did | Patients | What it found |
|---|---|---|---|
| Byrd, 1988 | Coronary care patients, San Francisco General Hospital | 393 | 6 of 26 measures favored the prayer group; 20 did not |
| Harris and others, 1999 | Coronary care patients, Mid America Heart Institute | 990 | Small benefit on a new scoring system; no difference on Byrd's |
| Cha, Wirth and Lobo, 2001 | Prayer and pregnancy rates after IVF | 199 | Reported doubled pregnancy rate. Later withdrawn — see below |
| Benson and others, 2006 | Heart bypass patients at six hospitals, three groups | 1,802 | Complications in 52% prayed for, 51% not prayed for, 59% prayed for and told so |
| Masters and Spielmans, 2007 | Combined analysis of earlier studies | — | No detectable effect on health |
| Cochrane review, 2009 | Combined analysis of 10 randomized trials | 7,646 | Evidence does not support prayer either way |
The Benson study is the largest and most carefully built. It included a third group who were told they were being prayed for, and that is where the one notable difference appeared: they did slightly worse. The authors offered several possible explanations, including anxiety about being singled out as sick enough to need prayer.
Common objections
"God answers every prayer — yes, no, or wait"
Between them, those three cover every possible outcome. A recovery is a yes, a death is a no, and no change is a wait.
The difficulty is that nothing could ever count as a prayer going unanswered. Someone praying to a milk jug could keep score the same way and record a perfect success rate. That does not show the claim is false, but it does mean the perfect record is not evidence for it, because the record would look identical either way.
"God's timing is not our timing"
A prayer may not have been refused, only not answered yet.
Nothing rules this out. But no length of delay falls outside it — a week, a decade, and a lifetime are all equally consistent with the claim. It also sits awkwardly with the passages themselves, since Matthew 7:8 says everyone who asks receives, and Mark 11:24 tells the person praying to believe they have already received it.
"Prayers are only answered if they match God's will"
This comes from 1 John 5:14, which does add that condition.
The condition appears in that verse. It does not appear in Matthew 7:7-8, Matthew 21:22, Mark 11:24, John 14:13-14, John 15:7, or John 16:23-24, which say "whatever" and "anything" without qualification.
Reading the condition into all of them means the broader passages do not mean what they appear to say. It also changes what is being promised: if the answer is limited to what God had already decided to do, the outcome would have been the same without the prayer.
"Studies have shown that prayer works"
This normally refers to Randolph Byrd's 1988 study, which found better outcomes among patients who were prayed for.
Byrd tracked 26 separate measures. Six favored the prayer group and twenty did not. When that many measures are tested at once, a few positive results are expected by chance alone, and Byrd did not adjust for this. The summary score he reported was put together after the data came in rather than decided in advance.
A larger study by Harris and colleagues in 1999 set out to repeat the finding and found no difference on Byrd's own scoring system. Benson's 2006 study, with 1,802 patients and a measure chosen before the study began, found no difference at all.
The other study sometimes cited is Cha, Wirth and Lobo (2001), which reported that prayer doubled pregnancy rates after IVF. It should not be used. Rogerio Lobo asked for his name to be taken off it in 2004, saying he had not taken part in the research. Daniel Wirth was found to have been using a false identity and pleaded guilty to fraud charges in federal court the same year. The journal removed the paper from its website. Bruce Flamm's account in Skeptical Inquirer documents the case.
"The Bible forbids testing God"
Deuteronomy 6:16 says "You shall not put the LORD your God to the test," which Jesus quotes in Matthew 4:7.
Other passages do the opposite. Malachi 3:10 says "Test me in this." In 1 Kings 18, Elijah proposes a public contest between Yahweh and Baal and treats the result as proof. In Judges 6, Gideon asks for two physical signs to confirm what he has been told, and gets both.
The objection also cuts both ways, since it would rule out Byrd's study and the IVF study as well — the two most often cited as evidence that prayer works.
What the studies can and cannot show
These trials test whether being prayed for by strangers changes measurable medical outcomes. They cannot show whether prayer helps the person praying, whether God exists, or whether prayers are answered in ways that leave no trace in hospital records.
What they do test is the plainest reading of the promise that prayer heals the sick. Across the largest and best-designed studies, no such effect was found.