# Question for statisticians about new prayer study. **URL:** \ **Category:** Great Debates\ **Created:** [March 31, 2006, 5:21pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721 "2006-03-31T17:21:40Z")\ **Posts on this page:** 20\ **Page:** 3
**Author:** ![Thudlow\_Boink](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/thudlow_boink/32/320_2.png) [@Thudlow\_Boink](https://boards.straightdope.com/u/Thudlow_Boink)\ **Post date:** [March 31, 2006, 9:28pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/41 "2006-03-31T21:28:17Z")
> [@Diogenes the Cynic](#): > > That’s exactly what I’m saying. Everyone recovered at a perfectly ordinary rate. There was no extraordinary effect on anybody. The “rocks” rolled downhill at just the speed gravity would predict and no faster. Forget for a moment that it’s prayer we’re talking about; the same idea would apply to any procedure, drug, or medical treatment that claims to help—not to produce extraordinary effects, but to make ordinary effects more probable or more effective. Suppose, for example, you have one group given megadoses of Vitamin Q, and a control group given no Vitamin Q. And in each group, some people recover and some don’t (or, some experience complications and some don’t or whatever). But suppose the recovery rate is 80% in the Vitamin Q group and 30% in the control group. This would be evidence that Vitamin Q has some real, beneficial effect, even though no one individual shows anything out of the ordinary. Or suppose everybody recovers, but the time for recovery can take anywhere from 10 to 90 days. Within the group given Vitamin Q, some recover in as few as 10 days and some take as long as 90, but the average recovery time is 20 days. In the control group, recovery times also range from 10 to 90 days, but their average is 70 days. Again, no individual case is extraordinary, but the data as a whole suggest a beneficial effect from the Vitamin Q. ---
**Author:** ![Diogenes\_the\_Cynic](https://avatars.discourse-cdn.com/v4/letter/d/e99b99/32.png) [@Diogenes\_the\_Cynic](https://boards.straightdope.com/u/Diogenes_the_Cynic)\ **Post date:** [March 31, 2006, 9:28pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/42 "2006-03-31T21:28:51Z")
> [@bup](#): > > We can’t prove that there has ever been a heart surgery where the healer wasn’t prayed for. You also can’t prove that it’s ever rained without somebody doing a rain dance. As **Voyager** said, the defaut assumption is that there is that ritualistic chanting should have no effect on the healing rates of the human body. This study shows that there is no reason to move off of that assumption. Only one side has a burden here. ---
**Author:** ![Diogenes\_the\_Cynic](https://avatars.discourse-cdn.com/v4/letter/d/e99b99/32.png) [@Diogenes\_the\_Cynic](https://boards.straightdope.com/u/Diogenes_the_Cynic)\ **Post date:** [March 31, 2006, 9:31pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/43 "2006-03-31T21:31:40Z")
> [@Thudlow Boink](#): > > Forget for a moment that it’s prayer we’re talking about; the same idea would apply to any procedure, drug, or medical treatment that claims to help—not to produce extraordinary effects, but to make ordinary effects more probable or more effective. > > Suppose, for example, you have one group given megadoses of Vitamin Q, and a control group given no Vitamin Q. And in each group, some people recover and some don’t (or, some experience complications and some don’t or whatever). But suppose the recovery rate is 80% in the Vitamin Q group and 30% in the control group. This would be evidence that Vitamin Q has some real, beneficial effect, even though no one individual shows anything out of the ordinary. > > Or suppose everybody recovers, but the time for recovery can take anywhere from 10 to 90 days. Within the group given Vitamin Q, some recover in as few as 10 days and some take as long as 90, but the average recovery time is 20 days. In the control group, recovery times also range from 10 to 90 days, but their average is 70 days. Again, no individual case is extraordinary, but the data as a whole suggest a beneficial effect from the Vitamin Q. Yes but there was no similar result in this case. The Vitamin Q people did no better than the control group (in fact, did slightly worse). ---
**Author:** ![Thudlow\_Boink](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/thudlow_boink/32/320_2.png) [@Thudlow\_Boink](https://boards.straightdope.com/u/Thudlow_Boink)\ **Post date:** [March 31, 2006, 9:36pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/44 "2006-03-31T21:36:01Z")
> [@Diogenes the Cynic](#): > > Yes but there was no similar result in this case. The Vitamin Q people did no better than the control group (in fact, did slightly worse). Right. Whatever kind of effect they were looking for, ordinary or extraordinary, they didn’t find any. ---
**Author:** ![Diogenes\_the\_Cynic](https://avatars.discourse-cdn.com/v4/letter/d/e99b99/32.png) [@Diogenes\_the\_Cynic](https://boards.straightdope.com/u/Diogenes_the_Cynic)\ **Post date:** [March 31, 2006, 9:36pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/45 "2006-03-31T21:36:45Z")
> [@Voyager](#): > > So is any prayer for something not tempting God? (Prayers thanking god wouldn’t fall under this.) Luke (quoting Deuteronomy) actually uses a word that’s better translated as “test thoroughly.” Sometimes you see that verse translated as “You will not put God to the test,” i.e. you’re not supposed to ask him to prove himself. My question is, why ever not? Why is it not kosher to demand proof? If you ask me, that sounds kind of weaselly on God’s part. ---
**Author:** ![Gangster\_Octopus](https://avatars.discourse-cdn.com/v4/letter/g/a8b319/32.png) [@Gangster\_Octopus](https://boards.straightdope.com/u/Gangster_Octopus)\ **Post date:** [March 31, 2006, 9:57pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/46 "2006-03-31T21:57:01Z")
I think **Apos** pretty much hit the nail on the head. The experiment is ill-defined. In strictly anlytical terms the study is very narrow, in the sense that under the conditions and assumptions of the study, prayer has been shown to have no effect. The study does not show prayer has no effect, ever. But there is nothing to prove that it does have an effect either, which I guess is **Dio’s** point. I don’t think many of us who are debating the study are arguing that prayer works,I think we are just ponting out the limits of the conclusion. From a scientific standpoint, since there is no theorized reason why prayer should help (except maybe in a mental health sense), the default supposition should be, “Prayer doesn’t do anything, as far as we know”. Even if the study is narrow, that doesn’t mean the standing “truth” is “Prayer helps”. That hypothesis cannot be the default waiting to be disproved, and has clearly been shown to not be true, under the assumptions of this study. ---
**Author:** ![David\_Simmons](https://avatars.discourse-cdn.com/v4/letter/d/9de053/32.png) [@David\_Simmons](https://boards.straightdope.com/u/David_Simmons)\ **Post date:** [March 31, 2006, 10:06pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/47 "2006-03-31T22:06:26Z")
> [@Apos](#): > > So, Shodan, are you happy that government paid millions of dollars for a study that: The study was mainly funded by the [John Templeton Foundation.](http://www.templeton.org/) ---
**Author:** ![bup](https://avatars.discourse-cdn.com/v4/letter/b/6bbea6/32.png) [@bup](https://boards.straightdope.com/u/bup)\ **Post date:** [March 31, 2006, 10:17pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/48 "2006-03-31T22:17:54Z")
The ‘Vitamin Q’ analogy is lacking. To make it more like the experiment, we have to then say, “We don’t know if the control group also received Vitamin Q; it may have been in the food they were eating, and more than it was in the food the people who were given Vitamin Q got.” I am willing to say that it has rained without someone doing a rain dance because there is geological evidence of rain before there were people. Will everybody own up to the fact they didn’t control the control group? They didn’t tell friends and relatives not to pray because “that would have been unethical” (see the article). So you can’t draw any conclusions from this study. ---
**Author:** ![Diogenes\_the\_Cynic](https://avatars.discourse-cdn.com/v4/letter/d/e99b99/32.png) [@Diogenes\_the\_Cynic](https://boards.straightdope.com/u/Diogenes_the_Cynic)\ **Post date:** [March 31, 2006, 10:39pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/49 "2006-03-31T22:39:03Z")
How do you know that someone on another planet wasn’t doing a rain dance before there were humans? Like I said, there is no theoretical reason why prayer should promote physical healing so the assumption is that it does not. This is no more unreasonable than assuming that dancing can’t make it rain. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 31, 2006, 10:41pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/50 "2006-03-31T22:41:53Z")
> [@Diogenes the Cynic](#): > > Luke (quoting Deuteronomy) actually uses a word that’s better translated as “test thoroughly.” Sometimes you see that verse translated as “You will not put God to the test,” i.e. you’re not supposed to ask him to prove himself. > > My question is, why ever not? Why is it not kosher to demand proof? If you ask me, that sounds kind of weaselly on God’s part. Except of course that God got pissed off at Moses for not putting him to the test when he smote the rock to produce water. So it sounds more like “pay no attention to the priest behind the curtain” to me. Not that we should ever expect consistency from religion. Anyhow, isn’t asking for something putting god to the test, since if you get it you’ll believe more in god? That’s why I distinguished that case from prayers of thanks which can’t be interpreted as putting god to the test. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 31, 2006, 10:45pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/51 "2006-03-31T22:45:13Z")
> [@bup](#): > > The ‘Vitamin Q’ analogy is lacking. > > To make it more like the experiment, we have to then say, “We don’t know if the control group also received Vitamin Q; it may have been in the food they were eating, and more than it was in the food the people who were given Vitamin Q got.” But if Vitamin Q is newly discovered, it might be in other food, or there might be unknown metabolic pathways that synthesize vitamin Q in the body. > [@](#): > > So you can’t draw any conclusions from this study. Sure you can - the ones I mentioned. No study is ever the last word, which is why enough details get published to allow others to reproduce it. ---
**Author:** ![Loopydude](https://avatars.discourse-cdn.com/v4/letter/l/e56c9b/32.png) [@Loopydude](https://boards.straightdope.com/u/Loopydude)\ **Post date:** [March 31, 2006, 10:46pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/52 "2006-03-31T22:46:16Z")
I think for the null result (seen between the control and one experimental group) to somehow be generated by prayer contamination, you’d have to evoke a mechanism by which uncontrolled prayer rather nicely cancelled the difference one would have seen if the result deviated significantly from the null hypothesis. To assert otherwise is to assume prayer does something to begin with, perhaps only if done in secret, or only if done by the individual, or a host of other possibilities that fine-tune this cancelling. I’m not sure how such assumptions can be justified. I must admit being baffled by the result showing decreased benefit. To me that says there’s likely some systematic error here, though I rather doubt prayer contamination is the cause. ---
**Author:** ![bup](https://avatars.discourse-cdn.com/v4/letter/b/6bbea6/32.png) [@bup](https://boards.straightdope.com/u/bup)\ **Post date:** [March 31, 2006, 10:47pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/53 "2006-03-31T22:47:43Z")
> [@Diogenes the Cynic](#): > > Like I said, there is no theoretical reason why prayer should promote physical healing so the assumption is that it does not. This is no more unreasonable than assuming that dancing can’t make it rain. I follow that. You’re ready to dismiss the claim out of hand. You’re missing the criticism, though, which is a valid one. It is plausible, and altogether likely, that every person in that study got prayed for. It’s a hell of a lot more common than rain dancing. The default assumption was prayer doesn’t work before the study. The default assumption afterward is that it does nothing, _but not because the study demonstrated anything. They didn’t even figure out how to test the hypothesis cleanly._ ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 31, 2006, 10:50pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/54 "2006-03-31T22:50:08Z")
> [@Apos](#): > > At the very most, this study balances out sloppy studies that claimed tha prayer had an effect. But it’s still just a crime that it ever existed, and I think the real question is what idiot greenlighted it. I mean aside from its subject matter, it was a really good study, with tons of good funding… that had absolutely no purpose whatsoever and no chance of producing a meanginful result. I mean, here’s a study that can only exist if the greenlighter has a prior religious belief in prayer (because otherwise there’s no surface plausibility to justify the study, no mechanism that can be tested), would only be done if they want to prove something about prayer… and then when it is done, they discount the results. What the hell was the point? Given the number of credulous people in the world today, and the amount of crap that people believe, I think this was a very good study. I just read about a study on chiropractic in England, where it was found not only did chiropractic not heal any diseases, it didn’t even help back pain. I assume the PI didn’t believe in chiropractic. We could use more studies about these “herbal remedies” now on the shelves. Yeah the study isn’t perfect - none are. But now the nuts can’t claim that science is scared to test their claims of prayer working. And maybe someone will be convinced. ---
**Author:** ![bup](https://avatars.discourse-cdn.com/v4/letter/b/6bbea6/32.png) [@bup](https://boards.straightdope.com/u/bup)\ **Post date:** [March 31, 2006, 10:55pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/55 "2006-03-31T22:55:18Z")
> [@Voyager](#): > > Sure you can [draw conclusions from the study]- the ones I mentioned. No study is ever the last word, which is why enough details get published to allow others to reproduce it. You’re wrong about that. You had a null hypothesis, and a flawed experiment, so you’re claiming the study demonstrated you can’t reject the null. It did no such thing. There was no reason to reject the null hypthesis before the study, and because the study was done sloppily, it didn’t demonstrate anything. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 31, 2006, 11:24pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/56 "2006-03-31T23:24:58Z")
> [@bup](#): > > You’re wrong about that. You had a null hypothesis, and a flawed experiment, so you’re claiming the study demonstrated you can’t reject the null. It did no such thing. > > There was no reason to reject the null hypthesis before the study, and because the study was done sloppily, it didn’t demonstrate anything. I never said that the study demonstrated that you can’t reject the null. What I did say (in a garbled way) was that the study showed there was no reason to reject the null hypothesis. I did say this kind of study never can disprove anything. And I dispute the notion that the experiment was done sloppily - at least for the reason you mention. all experiments are bound by the limits of the practical. I’d assume that the paper, in the Conclusions section, mentions the impossibility of preventing anyone from praying for the control population. I don’t see anyway of getting around this. You mentioned praying for rats. But a negative conclusion might just mean God doesn’t care about rats. It might mean that prayers for rats can’t be sincere. The experiment is certainly not conclusive - but it is interesting, and worthwhile. ---
**Author:** ![Apos](https://avatars.discourse-cdn.com/v4/letter/a/838e76/32.png) [@Apos](https://boards.straightdope.com/u/Apos)\ **Post date:** [March 31, 2006, 11:39pm UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/57 "2006-03-31T23:39:53Z")
My point is simply this. Given the excuses that are being made after the fact as to why the experiment doesn’t prove anything about prayer, the people who suggested the study were litterally pissing 2 million dollars down the drain. If they were ready with those objections for any negative result, then the entire study was pointless, and they should have known that from the start. The exact same excuses would invalidate any positive result as well. The experiment was not sloppy. It was very good: it’s methodology was sound, it was well funded, the assignments were well done. The problem is the underlying element of study: something not only intangible and unmeasurable (making any sort of consistent “dose/effect” measure impossible) but without any underlying mechanism to make the study plausible in the first place (in the abscence of any plausible effects to explain). We are left with: nothing to explain, and no way or willingness to explain it. In other words, a complete boondoggle. ---
**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\ **Post date:** [April 1, 2006, 12:20am UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/58 "2006-04-01T00:20:33Z")
From a scientific point of view, I’d say this test was a good one. But from a philosophical viewpoint, I don’t think this question can ever be resolved. Since we can’t detect the presence/absence of a prayer for or by any one individual, how do we know that someone, somewhere in the world, isn’t praying for the _opposite_ of what someone else is? If someone prays that subject X dies, and someone else prays that he lives, do these cancel each other out? Is a general prayer for “all souls everywhere” as powerful for one subject as a single prayer for that subject only from ten people? Suppose (just being a Devil’s Advocate here) that 10 anti-prayers (die, sickie!) were said for one group, exactly balancing out the 10 pro-prayers (get well, sickie!) for the same subject. Then we could get exactly the same outcome as this test. Preposterous? Sure. Impossible? No. I like **Bup’s** idea that a bunch of mice, who no one knows exist, be the subjects of a prayed-for and not-prayed for group. But even that has a flaw – what if someone, somewhere in the world, is praying, “God, let all of your creatures live a long life.” Whoops – the mice have just been prayed for. And what if the prayer is said by someone like a Mother Theresa? Would her prayers have more weight than a skid row junkie’s? If you can’t detect it and can’t measure it, you can’t handle it. I just don’t see any way around the problem. ---
**Author:** ![Loopydude](https://avatars.discourse-cdn.com/v4/letter/l/e56c9b/32.png) [@Loopydude](https://boards.straightdope.com/u/Loopydude)\ **Post date:** [April 1, 2006, 1:17am UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/59 "2006-04-01T01:17:21Z")
I still say the fact they did see an effect in the group who knew they were being prayed for likely reveals an uncontrolled flaw in the study. The investigators suggest it might be some kind of performance anxiety, but don’t really seem to have a good explanation. My guess is there is a naturalistic explanation, though, and that alone would trash the whole effort. ---
**Author:** ![Musicat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/musicat/32/20189_2.png) [@Musicat](https://boards.straightdope.com/u/Musicat)\ **Post date:** [April 1, 2006, 1:27am UTC](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721/60 "2006-04-01T01:27:22Z")
> [@Loopydude](#): > > I still say the fact they did see an effect in the group who knew they were being prayed for likely reveals an uncontrolled flaw in the study. The investigators suggest it might be some kind of performance anxiety, but don’t really seem to have a good explanation. My guess is there is a naturalistic explanation, though, and that alone would trash the whole effort. I suspect it is natural variation. Heads won’t come up exactly 500 times every 1000 coin tosses, either. Do enough studies, and you will find some small differences in some samples. They already said it wasn’t statistically significant. But imagine what would have happened if the difference went the OTHER way. “Study shows 8% improvement in prayed-for patients. Prayer works!” [Previous page](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721.md?page=2) [Next page](https://boards.straightdope.com/t/question-for-statisticians-about-new-prayer-study/350721.md?page=4)