The Power of Prayer
In the March/April Skeptical Inquirer Tessman and Tessman (2000) suggest that the results of the study by Harris and others (1999) of intercessory prayer can be “well explained by chance.” So we are invited to conclude that there was in fact nothing significant going on: neither God nor anyone else was biasing the outcome in favor of the prayer group. But they spoke too soon. For the truth is that hidden in the details of the original paper is statistical evidence that there was indeed some kind of intercession taking place.
In November 1999 a long report of this study appeared in New Scientist magazine, and I was provoked enough by it to take a closer look. Here follows a copy of the letter I had occasion to write to the editor a few weeks later:
“Prayers can help patients recover even when they don't know people are praying for them, says a provocative new study" (New Scientist, 13 November, p. 24). But it turns out that hidden in the original paper that describes this study (Archives of Internal Medicine, 159, 2273) is a result that is even more curious: namely, that prayer works best of all before it has been started!
The paper reports the following facts without comment. “1,013 patients were randomized, 484 to the prayer group and 529 to the usual care group. After subsequent removal of those patients who spent less than twenty-four hours in the Coronary Care Unit [because it took a day to get the prayers up and running], 524 remained in the usual care group and 466 in the prayer group.”
But look at these figures. While 18 of 484 patients who were going to be prayed for got better within 24 hours, only 5 of 529 patients who were not going to be prayed for did so. This difference is significant at p<.001 by a chi-squared test. By contrast the main effect reported in the paper, namely that among patients who stayed more than 24 hours those who were prayed for had better clinical outcomes than those who weren't prayed for, only just makes it to the p<.04 level.
So it seems that either the study has come up with strong evidence of prayer producing backward causation of recovery, or else that there was something wrong somewhere with the way the study was conducted (e.g., that, despite the claim of randomization, less sick patients were in fact being assigned to the prayer group). Readers should take their pick of these two interesting alternatives.
I sent a copy of my letter to the first author of the paper, William S. Harris. He acknowledged, in a courteous reply, that these numbers were somewhat puzzling. But he moved to head off my veiled accusation (that someone involved with the study had deliberately tried to rig the results by assigning less sick patients to the prayer group) by reminding me that I was making an unwarranted assumption: namely, that those who left the unit within twenty-four hours had indeed got better-when in fact they might have died (this is something he said he had not yet had time to look into). Point taken. If it should turn out that patients who were assigned to the to-be-prayed for group were actually significantly more likely to die within twenty-four hours, the implications of this study would surely be more interesting still.
- Harris, W.S., M. Gowda, J.W. Kolb, C.P. Strychacz, J.L. Vacek, P.G. Jones, A. Forker, J.H. O'Keefe, and B.D. McCallister. 1999. A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Archives of Internal Medicine 159: 2273-2278.
- Tessman, Irwin, and Jack Tessman. 2000. Efficacy of prayer: A critical examination of claims. Skeptical Inquirer 24(2): 31-33, March/April.