STATS ARTICLES 2006
Does Faith Affect Health?
March 27 2006
Rebecca Goldin Ph.D.
A recent spate of stories has suggested that studies of faith-based healing are prompting scientific debate. Last week, the Washington Post had a front-page article headlined “Researchers Look at Prayer and Healing; Conclusions and Premises Debated as Big Study's Release Nears.” But the Post and other news organizations (such as Newsweek) are confusing scientific questions about the interaction of religion and health, with whether prayer itself affects healing. The debate here is not scientific, and not controversial in the scientific community.
There are three rough categories of inquiry, depending on who is praying for whom:
- Do our own prayers affect our own health? How does the health of someone who prays compare with that of someone who doesn’t pray – or someone who has emotional support that isn’t based in religion?
- Do prayers affect the health of someone else, such as a friend or relative? In this scenario, a sick person knows that someone is praying for him or her. Are there perceived benefits compared with those for whom no-one prays? What about a comparison with people who have a supportive community but no one actually praying for them?
- Is there an effect on health from prayers said without the knowledge of the person being prayed for?. How does this compare with others treated equally in every other respect save that of having prayers said, anonymously, on their behalf?
The last question has a very different nature than the first two, yet it has been thrown into the first group as if it were equally scientific and equally controversial.
It is not. There are scientific debates about the role of religion in medicine – and these range from whether it is appropriate or helpful for doctors to discuss religion with their patients, to whether religion has a calming effect that can easy pain and suffering, or reduce the chance of a heart attack.
There is no serious scientific debate about the role of prayers at the Western wall for patients in New York, or the effect of online prayer groups for cancer patients (unconnected, except by prayer and possibly six degrees of separation, from those praying).
The first two questions stem from a well-known “religiosity effect” on health: religious people are likely to live longer than non-religious people. But no one can prove whether this is because healthier people are drawn to religion, whether religious practice correlates with healthier living styles (like better diet or fewer indulgences), whether practicing religion itself reduces health risks through biological mechanisms (such as lowering heart rates and reducing stress) or whether there is divine intervention. And when it comes to the divine part, the nature of the charge – that God has something to do with it – can never be validated or refuted.
The other explanations are a subject of scientific discussion: Does praying help relieve stress? Does church (or other organized religion) help keep people connected to their communities, leading to better health care or healthier habits? These questions are important and scientific – as scientific as asking whether art therapy helps reduce pain of terminally ill cancer patients, or whether massage therapy reduces the likelihood of repeat heart attacks.
And when taxpayers find that National Institutes of Health have spent 3.5 million dollars on questions of faith and healing, they should be assured that it is the interaction of religious practice (and all it comes with) and healing that is being studied, not the effect of religion itself on healing.
The popular and unscientific belief that strangers’ praying for the sick makes a difference is not the subject of scientific controversy, despite recent press coverage. In this case, unlike examining the influence of religious practice among sick people or their family members, it is very easy to set up a scientific experiment to evaluate the hypothesis. Take two groups of people who are comparable in every sort of way, and have one group prayed for and the other not, in a double-blind environment (even those who interact with the patients do not know who is prayed for and who not), and then compare results. This can, and has, been studied. The results: Nada - unless you count studies whose premises and methods are easily dismantled by serious researchers.
Then come the explanations (from believers of faith-based healing) for why a measurable effect was not observed: there was too much “noise”, i.e. too many prayers already being said for everyone, such that the effect of extra prayers couldn’t be measured. Another excuse: the prayers were not said loud enough, or close enough, or the “energy” wasn’t “channeled” properly. The list goes on and on. And because you can never prove the null hypothesis – you can never prove there isn’t an effect – people continue to attempt to show a relationship. No doubt, by chance alone, there will someday be a study that finds such an effect – bias aside. That’s because statistics has a built-in possibility for error.
The statistical gold standard for a “result” is the p-value. This is a measure of how likely an outcome can occur by chance alone. A p-value of .05 means that there is a five percent chance that a certain outcome (like ‘prayed-for’ people healing better) would be observed by chance alone. The value p=.05 is the limit at which researchers say an effect is observed. There is a small chance that, in any study, we will observe that prayers by distant well-wishers have a measurable impact on patients’ health (though the phenomenon is due to chance alone). If there are enough studies trying to evaluate this, one will surface with such a result. This occurs for the same reason that, for every thirty-seven fools who lose money at roulette, there is one winner who walks away saying he knew how to play well.
To call the role of praying in healing “controversial” in the scientific community is to give it far too much credit. There is little if any scientific evidence that the prayers themselves are making a difference. Let’s keep science in the realm of science – that is a realm where there are measurable effects, and plausible mechanisms.