Effects of retroactive prayer on outcomes in patients with infection (2001)
ncbi.nlm.nih.gov
ncbi.nlm.nih.gov
But this one did: http://rsw.sagepub.com/content/17/2/174.short
Another interesting study: http://online.liebertpub.com/doi/abs/10.1089/107555304132380...
("No direct intervention effect on the primary outcomes was found. A marginally significant reduction in the amount of pain was observed in the intervention group compared to controls. The amount of concern for baseline problems at follow-up was significantly lower in the intervention group when stratified by subject's baseline degree of belief that their problem could be resolved. Prayer intervention appeared to effectively reduce the subject's level of concern only if the subject initially believed that the problem could be resolved. Those in the intervention group who did not believe in a possible resolution to their problem did not differ from controls. Better physical functioning was observed in the intervention group among those with a higher belief in prayer and surprisingly, better mental health scores were observed in the control group with lower belief in prayer scores.")
Think about it -- retroactive prayer? As in, the people got sick, either died or got better, and were then prayed for?
I mean, being an omnipotent deity is one thing, but effective retroactive prayer? Now that's a neat trick.
I suppose, given the existence of a deity that is able to heal people, then this would be a decent way to determine whether or not said deity was truly omnipotent.
Perhaps also 'omniscient': if your god is all-knowing, then it need not wait until you pray to act on your prayer.
Even assuming the study were perfect, the conclusion is not that retroactive prayer works, but that retroactive prayer in the context of conducting a scientific study works, since if you believe the deity being prayed to exists, you also have to believe that the deity knows it's part of a scientific study and not genuine spontaneous prayer.
Randomization should balance confounders like that: there should be on average an equal number of pre-prayed-for people in both the experimental and control groups.
(And even if all subjects had been prayed for, it seems reasonable that prayer would be additive and one would still expect to detect an average treatment effect.)
I wonder if their p-values are Bonferroni corrected?
That is, they ran three different experiments (that they reported on) and so the significance of their results should be concomitantly reduced.
Consider: if you examined a hundred different outcome measures between control and treatment groups in any given study you would be almost certain to find at least one that by chance alone was different at a p=0.05 level. The Bonferroni correction (or some variant of it--there are numerous similar approaches that go by different names) adjusts for this.
Failure to correct p-values for multiple experiments is one of the most common causes of spurious results in otherwise apparently well-designed studies.
Furthermore, they don't actually include the raw data, don't give a detailed description of their method of analysis ("The χ2 test was used to test for the significance of the results shown in the tables. As most of the continuous variables did not have a normal distribution, the Wilcoxon rank sum test was used for comparisons" is no-where near sufficient for me to reproduce their results) and base their "fever" indicator on multiple measures of which they take one.
Also, their control group has a maximum hospital stay of over twice that of their treatment group. How do the p-values vary if you look at large sub-sets of the data? That is, re-run the analysis with 10% of the data left out. Do the effects persist? This is a pretty routine test for robustness, and a claim with a prior as low as "Retroactive intersessory prayer is effective in reducing fever and hospital stay duration but not mortality" requires extremely compelling evidence to increase its plausibility to the point of being worth bothering with.
Not only is the conclusion questionable on statistical grounds, but also on theological ones: why would one expect a loving god to intervene to shorten trivialities on hospital stays and fever duration, but not morality?
The answer cannot be "God works in mysterious ways" because that completely obliterates any reason for believing that intersessory prayer will ever have any effect of any kind. It is equivalent to the claim that god is a black box that has a completely random and unpredictable response to stimuli (prayer) so we should study which stimuli have predictable outcomes.
What amuses me is that a lot of real research papers would flunk your criteria. To pick just one:
> Furthermore, they don't actually include the raw data,
So, do you see papers providing the raw data often? I don't.