May 14, 2009
So I locked horns with PhillyChief and John Evo, again, this time it was over the following comment from PhillyChief – who if I remember correctly – claims to be a scientifically-minded rationalist atheist:
Prayer helps no one but the one praying, providing a euphoria and calming effect, which could be comparable to ejaculating.
I felt that was an odd statement for a scientifically-minded rationalist to make, but was not surprised that it came from a sarcastic atheist who claims to be "almost always right", and so I replied,
How would you know? Where is that "demonstrable evidence" you're so fond of? Aside from being grossly unscientific, statements like the above appear contradictory alongside appeals to soft atheism as you've recently made on my site.
At this point John Evo chimed in, submitting some link that was presumably to some prayer study that he felt somehow qualified as the evidence I asked for, but some silly form asking for private information had to be completed in order to read it – and I don't want to join some organization just to refute what is really basic and obvious scientific misunderstanding. As I said multiple times in the thread, John Evo and PhillyChief are more than welcomed to cite any pertinent facts themselves.
To make a long story short, in our discussion I essentially challenged the scientific credibility of all prayer studies. Now I'm not normally a fan of absolute quantifiers, but in this case I will justify my use of the word all. In fact, if I were still screenwriting and had extra money sitting around, I would gladly offer a Randi-esque incentive of at least $10,000 to anyone who can prove the following arguments wrong: Philly's aforementioned comment is inherently unfalsifiable, and sans invocation of counterfactuals no scientifically reliable prayer study is possible.
Atheists and skeptics are rather fond of criticizing and ridiculing believers for relying on blind faith regarding their beliefs, yet I felt Philly and John Evo rely on nothing more than blind faith, anti-cl bias, or the typical knee-jerk misunderstanding of science we've rightly come to expect of a certain subset of believers. Of course sheer stupidity always remains an option, but I think both Philly and John Evo are intelligent guys, and I feel safe excluding that idea.
The "demonstrable evidence" part of my response refers to another thread where Philly moved the goalpost to "demonstrable in part" and had to equivocate pretty badly in order to save face after being called him out for an appeal to telepathy. If you've got the time, read that one, too. Much more than just another chance to roast one of my loudest critics, the errors in methodology Philly exemplified in that post actually relate directly to this one: Philly's aforementioned comment assumes there is demonstrable evidence regarding prayer studies – yet even if he were forced to equivocate again to "demonstrable in part" – he could not sustain his claim. That is the true basis of my aforementioned "10%" comment, and I can prove this beyond any reasonable doubt.
Misunderstanding me as he does on occasion, Philly replied:
No one KNOWS WITH ABSOLUTE CERTAINTY what anyone else is thinking. Here's an example: I swear I think you're a fucking asshole, cl. Do you know with absolute certainty that's what I'm thinking? …If you'd like to make a case that prayer isn't euphoric or calming, be my guest.
..to which I replied,
Why would I? Your statement is one-sided and only applies to a certain subset of people, and shows as much ignorance of sound science as the original statement cited. Point is, making claims that cannot be substantiated by evidence is a departure from rationalism, and with statements like these, you're a frequent flier. The 100% certainty argument is not the argument I'm accusing you of making. I'm accusing you of making a claim you can't even be 10% sure of. And what's this "reasonable degree?" Is that sort of like "demonstrable in part?"
Still misunderstanding me, Philly tries another angle:
Ok, then what percentage of Christians praying would I have to observe then? From your statement, all I can tell is less than all but more than 10%.
I'm fully convinced that Philly never understood this, but anyways, let's get to it.
False Argument #27: Prayer studies are scientifically credible – In order to even debate this, it seems we need working criteria of just what constitutes a credible study. First off, a credible study needs a falsifiable claim. I hinted at this much in the following comments to both of my detractors:
[Y]our statement suggests that you misunderstand the scientific method and falsifiability. …Can science reliably proceed sans a falsifiable claim? …My basis for discounting prayer studies is their inherently unscientific nature, and I've always thought such studies irrelevant in the misguided attempt to prove God. You're right that "the god thing is still unfalsifiable," and that's why you're completely out of your mind to claim that the prayer thing is falsifiable. Make sense yet? …Any scientist who claims God or prayer are amenable to the scientific method is not worth their paycheck.
Somewhere around or between those comments, John Evo makes the following statement, also an odd one for someone who claims to be a scientifically-minded rationalist. It also shows how little he's actually studied prayer studies:
If remote prayer studies had ever shown a positive impact outside of statistical fluctuations you would cite the study in any discussion about the value of prayer.
Really? Apparently John Evo assigns the same amount of credibility to telepathy as he does to prayer studies. I'll prove Evo wrong just as easily, as the two studies we're about to discuss both purported promising results – and you won't hear me claiming them as evidence for the value of prayer.
Now, it's never fun to argue in generalities, and that's why I challenged PhillyChief and John Evo to cite these studies they continually alluded to, yet for some reason kept shrouded in mystery not unlike Christians backed into a corner by an atheist in a discussion about evolution. As of the time of this post, citations have not been forthcoming – not even when I led by example and cited an actual study demonstrating that mere words can possibly function as effective placebos, which is only one of many potential confounders that inherently skew a certain subset of prayer studies (Thomas KB. (1987) General practice consultations: is there any point in being positive? Br Med J (Clin Res Ed). 294:1200-2. PubMed).
Now let's take a brief look at two more actual studies: The first from my own liberal backyard and published in Southern Medical Journal; the second from America's conservative heartland and published in the highly prestigious Archives of Internal Medicine.
In 1999, San Francisco cardiologist Dr. Randolph Byrd conducted a study on intercessory prayer (IP) where 393 CCU (coronary care unit) patients were randomly treated with either standard medical procedures, or standard medical procedures + daily IP from three to seven alleged "born-again Christians". The results were touted as promising, but I see so many red flags it feels more like I'm a matador in the middle of an intense bullfight on 10 hits of LSD.
Presuming them all honest, how do we know they weren't one of the majority whom the Bible describes as saying, "Lord, Lord" but not having faith? Humans can be dishonest too, you know. How do we know one or more of the "born-again Christians" wasn't actually an anti-Christian attempting to purposely skew the results? On the purely medical angle, newly-prescribed diuretics and newly-diagnosed cases of heart failure were two of Byrd's categories – and the prayer group showed fewer instances of both – but aren't diuretics routinely prescribed for heart failure, essentially making these categories measurements of the same thing? What an epistemological nightmare already, right?
Also published in 1999, at first glance the Harris Study seems methodologically superior to Byrd's, yet entails its own set of problems, and by its very nature remains vulnerable to the penultimate problems which cripple all prayer studies. The scenario was similar: 990 CCU patients in a Kansas City hospital received IP from teams of "Christian" intercessors on a completely randomized, controlled, double-blind, prospective basis, again with the control group receiving standard care and the prayer group standard care + IP.
Again, how do we know one or more of these "Christian" intercessors weren't among the many whom the Bible describes as saying, "Lord, Lord" but not having faith, or anti-Christians attempting to purposely skew the results? Further, the test was structured in a manner that reveals the researchers' own presuppositions about prayer as an agency: Requesting the intercessors to pray daily for 28 days suggests the conflation of prayer with magic. If it's true that "the Lord knows what we need even before we ask," of what value are a bunch of redundant prayers? How would we know negative results weren't God's just responses to foolish chatter, which is tantamount to lack of faith?
Now that we've briefly discussed specific difficulties with two actual studies that were both touted as promising, I'd like to discuss the general weaknesses that inescapably befall all prayer studies.
In fact, in the interest of brevity, let's start with the most crippling factor first: Regardless of how we structure our tests, we cannot reasonably control for or even quantify the confounding effects of extra-study prayer. That is, any methodological advantage we can conceive of cannot possibly eliminate or even quantify the subjects' exposure to other sources of prayer occurring outside the study, or to other types of spontaneous healing that may in fact exist. How can we know that any positive results in the prayer group were the result of the intercessors as opposed to other people praying in other places around the world that may or may not be focusing specifically on subject in the prayer group? It is entirely reasonable that other people besides the selected intercessors would be praying for the subjects in both groups – for example family members, friends, neighbors or members of their congregation – and I'm open to suggestions as to how we might eliminate or even quantify this massive problem. To those who might be wondering, "Wouldn't extra-study intercessors simply amount to a bigger sample group, effectively increasing the probability of positive results?" To them I would ask, "Is padding a study in favor of positive results scientifically credible?"
We've barely begun to discuss when and where the placebo effect becomes relevant, but that's another angle we could discuss, and frankly I see it as overkill. I've got some final objections to prayer studies in general, and then I'll conclude.
As C.S. Lewis so eloquently put it, prayer – by its very nature – is a request, not a compulsion. It logically follows that any request – also by its very nature – entails the possibility of being granted, rejected – or even postponed. The actual IP studies I've seen presume that God is like a magic genie who would, could or should simply grant any and all requests from any and all Joe's or Jane's off the street who claim to be Christian. What scientifically credible study begins with such debatable assumptions about the agency it seeks to reliably evaluate?
If someone in an hypothetical control group had cancer, and the cancer goes into remission, how do we know whether such resulted from IP vs. spontaneous remission? How do we know undiscovered phenomena analogous to spontaneous remission don't occur in other forms of cumulative or degenerative diseases like heart conditions, which were the subject in both of the aforementioned studies? How do we control for the fact that Kansas City likely dwarfs San Francisco in its population of born-again Christians, thus increasing the likelihood of confounding via extra-study prayer? How do we control for people who pray for all of those who suffer from disease and suffering? Do all prayer studies accurately control for the fact that every instance of a prayer study increases the likelihood of positive results by chance? How can we control for the fact that individual intercessors are going to have individual and varying degrees of faith? For these and many more reasons we've only begun to scrape the surface of, I submit that all prayer studies are inherently flawed and no prayer studies are scientifically credible.
Hence, False Argument #28 (That prayer has been proven or that studies with credible, promising results exist) and False Argument #29 (That prayer's inefficiency has been successfully demonstrated) both flow logically from False Argument #27, and here's the bottom line: Philly's aforementioned comment is inherently unfalsifiable, and no scientifically reliable prayer study is possible without the invocation of counterfactuals – because no matter how we structure the tests – it is impossible to successfully eliminate confounders and co-dependent variables.
I reminded PhillyChief and John Evo several times that the positive claimant retains the burden of proof. They made the positive, generic claim that prayer studies were credible, but failed to support that claim with evidence, hence that claim is irrational.
NOTE: In no way complimentary towards myself, John Evo has since changed his opinion, agreeing with me and challenging PhillyChief to take a more reasoned look at the evidence. "Ask yourself if you did any research at all when cl challenged us… You will find, as you continue your learning and growth, that the advice I gave in this message was well worth considering as it is said to you by someone who thinks highly of you. But take it as you will." John Evo to PhillyChief, May 16, 2009 6:07pm.