> "And if not, then I suppose I've made my point."

You made your point, though I still don't really understand why it was worth making for the discussion at hand.

> "I'm not sure if this is actually what you're saying,"

It wasn't. I thought you were trying to extrapolate from pure induction to all epistemology. Within the limits of induction as a means of knowing, you are of course correct, though induction is such a limited means of knowing (even the sciences rarely use it by itself) that working within those limits is pretty meaningless.

> "we don't presently have a good reason to believe immaterial things exist at all."

This isn't even close to true. It doesn't make your original point wrong, though.

> "Taking the contrary position, that prayer has no effect, doesn't require any faith on my part—it only requires citation to the fact that prayer doesn't appear to do anything."

That is only the case if it is also true that anything that cannot be proved by measurement can be assumed to be false. This is logical positivism (not Ockham's Razor) and it is not true; or, at least, its truth is an article of faith, since it cannot be measured and is not borne out by statistics. But that doesn't make your original point wrong, either.

If something can't be tested by measurement, either by current tools or in principle, investigators move on to other means of knowing (math, for example). In this case, though, our theology backs up your science; God doesn't grant wishes like a cosmic genie, so we really shouldn't expect to see wishes being granted. We believe (as an article of faith) that he CAN, but this would not be measured with statistics, but rather on a case-by-case basis, on which we have no data and no means by which to acquire any.

I'm not sure if any of this was intended to counter my actual argument, which is that your friend's recovery is not evidence of God answering prayer. If so, I don't see how. And if not, then I suppose I've made my point.

> "At that point, the conversation isn't about prayer studies anymore—it's about why faith isn't a substitute for evidence when it comes to acquiring knowledge." > In some arenas that is true, including this one. If I say that something can be measured empirically, and I further say that I can actually measure it (i.e., I offer the measurement as evidence), I had better be able to actually measure it according to the rules of how such things are measured (empirical induction), otherwise I'm kind of an idiot. But such a limited means of knowing has very little value apart from trivia, at least in fields outside of medicine and technology.
In other arenas, the conversation is about why evidence isn't a reliable means of acquiring knowledge about things that can't be measured evidentially. In which case, the discussion becomes about not whether one believes through faith or not, but in which object one places one's faith.

I'm not sure if this is actually what you're saying, but it seems like you might be saying that whether prayer works can't be verified empirically, so we're both just operating on faith—as in, you have faith that God answers prayer and I have faith he doesn't.

I don't know if this is what you meant, but if so, it doesn't work this way. Praying has no apparent effect—we've established this statistically. So we're left with one of two potential results—either prayer has no effect at all, or it has invisible/immeasurable effects.

We resolve this question by way of Occam's Razor, not faith. What's the difference between an invisible/immeasurable effect and no effect at all?

At the end of the day, the only fact I need to carry my burden to show that prayer doesn't do anything is to show that it doesn't appear to do anything. As between the following competing hypotheses:

1. Prayer has effects, but they aren't apparent to us.
2. Prayer has no effect.

(2) is the stronger answer based on form alone, and requires no faith to accept. By contrast, (1) depends entirely on faith. All I'm saying is that the world is the way it appears to be. This is always the best assumption unless you have specific evidence to indicate otherwise.

For that reason, this entire discussion is underpinned by the fact that we don't presently have a good reason to believe immaterial things exist at all. So far as we can tell, everything that in fact exists has measurable impacts on the world—in a sense, that's effectively what it means to exist.

So when you come along with a notion that prayer does some ill-defined immeasurable thing, you've effectively conceded the argument. Our experience teaches us that when something exists, it more or less necessarily appears to exist.

Obviously, for any particular phenomenon, there's always the possibility that we would need more suitable investigative tools—I realize the ancient Greeks weren't going to be able to detect dark matter, for instance. But what you're talking about is something effectively immeasurable in principle. Unless you think there's a reason to expect that future people will have developed tools that can measure the efficacy of prayer, then our investigative tools aren't the issue here.

At the end of the day, taking the position that prayer has effects requires you to adopt a purely faith-based position that it's doing something even though it appears not to be. Taking the contrary position, that prayer has no effect, doesn't require any faith on my part—it only requires citation to the fact that prayer doesn't appear to do anything.

> "But if you define your version of God in such a way that praying to him for things is expected to correlate with getting them at a greater-than-otherwise-expected rate, these studies indicate that your god doesn't exist."

That's a different syllogism, as indicated by "define your version of God." So:

1) If God exists, my prayer will be answered [in a way that is empirically measurable, etc].
2) my prayer was not answered
3) therefore God does not exist.

This one is a valid syllogism. Christian orthodoxy does not teach this, however.

> "At that point, the conversation isn't about prayer studies anymore—it's about why faith isn't a substitute for evidence when it comes to acquiring knowledge."

In some arenas that is true, including this one. If I say that something can be measured empirically, and I further say that I can actually measure it (i.e., I offer the measurement as evidence), I had better be able to actually measure it according to the rules of how such things are measured (empirical induction), otherwise I'm kind of an idiot. But such a limited means of knowing has very little value apart from trivia, at least in fields outside of medicine and technology.

In other arenas, the conversation is about why evidence isn't a reliable means of acquiring knowledge about things that can't be measured evidentially. In which case, the discussion becomes about not whether one believes through faith or not, but in which object one places one's faith.


And, just for fun, the bunny trail.

"Example: If Y is true, then X will happen when I do A. X happened when I did A. Therefore, Y is true. Deductively, this is obviously invalid. But it's exactly what's going on in the background behind 'If I do A, X will happen.' "

That's not how I understand what is happening. In your example, filling in the blanks might yield the following:

"If alchemical theory (Y) is true, [procedure] (A) will result in working gunpowder (X)." However, alchemical theory was not true, despite its ability to produce gunpowder, alcohol, etc. Induction is good at measuring results, not in testing/confirming the theory behind those results. Here's what I understand is happening:

1) Under condition A, X will occur. (Hypothesis/conclusion).
2) Condition A is true. (the experiment).
3) X occurred (the result).

The validity of the syllogism is then offered as evidence of the soundness of the first premise (the hypothesis). In deductive reasoning, this would not be allowed; there is no relation between soundness and validity in deduction. Since all three premises will be true, there is a sense that the consequent (3) is, in fact, affirmed. However, a fallacious "affirming the consequent" would entail assuming that the experiment was sound BECAUSE it produced the expected result, as opposed to because of the procedure it followed. The scientific method does not allow this; the result of an experiment is only admissible if it follows the procedure described (A must precede X).

> That's not how induction works. That's not induction, it's logical positivism, which is not the same thing. Induction works by affirming the antecedent (I hypothesize that if I do A, X will happen. I did A; X happened). But that's a side discussion.

I agree it's a side discussion so I don't want to overly belabor the point, but the inductive formulation of affirming the antecedent goes hand-in-hand with the deductive formulation of affirming the consequent—which is why it only works as an inductive argument instead of a deductive one. Let's try to look at this holistically:

Example: If Y is true, then X will happen when I do A. X happened when I did A. Therefore, Y is true.

Deductively, this is obviously invalid. But it's exactly what's going on in the background behind "If I do A, X will happen." The reason I expect X when I do A is because if I've got my ultimate conclusion right—that is, if Y is true—then it should follow that X will happen when I do A.

Ultimately, this is all a bit off to the side of the real issue, which is whether we have any reason to believe that praying for things has an effect on the world.

> And so I don't cite it as though it were evidence for my position.

Except that this entire discussion effectively began as a result of you doing precisely that. I stated that, statistically speaking, prayer has no real-world effect. You attempted to counter with an anecdotal example of someone making a medical recovery. I then demonstrated why it would be confirmation bias for you to claim that as a counterexample.

> The efficacy of prayer (or lack thereof) is untestable by induction due to an absence of experimental controls. So if that is really the totality of your argument, I suppose I would concede it, but I don't see what significance such an argument could possibly have.

It's not that the efficacy of prayer is untestable. It's that prayer has no apparent efficacy when tested. If you personally define your version of God in such a way that praying to him for things isn't expected to correlate with getting them at a greater-than-otherwise-expected rate, then this fact doesn't mean much to you. But if you define your version of God in such a way that praying to him for things is expected to correlate with getting them at a greater-than-otherwise-expected rate, these studies indicate that your god doesn't exist.

When put to the question in this way, you appear to be in the first camp. That's fine, but you don't get to take that position in one sentence and then start talking about how people making medical recoveries show that prayer works in the next—that would be confirmation bias, and would bring us essentially full circle to where we started.

> I suppose if I wanted to play devil's advocate...

You've described pretty much exactly how the efficacy of prayer is tested. But I don't think your concern with the data for groups 3 and 4 holds. For one thing, we can assume that the people in groups 3 and 4 aren't being prayed for at a 100% rate, and we know the people in groups 1 and 2 are. For another, these studies tend to be restricted to groups with similar medical conditions/prognoses—as such, similarity in distribution as between sets 1/2 v. 3/4 indicates that the prayers are simply superfluous.

To add on, there's another way of looking at the problem—studies have found a statistically significant difference between people who are told they're being prayed for and those who are not, irrespective of anyone's actual prayers. Importantly, this difference appears even when they're actually all being prayed for, or when the group told nothing is prayed for and the other group is lied to and not prayed for.

Stated another way, prayer studies indicate that praying for something to happen does not impact the likelihood that it will. You might still take the position that prayer has some sort of immeasurable impact, although I'm not sure what the difference would be between an immeasurable impact and a non-impact.

At the end of the day, literally 100% of the purported evidence that prayer "works" takes the form of "So-and-so prayed for X to happen and X happened." What I've demonstrated in this exchange is that none of that is actually evidence. As such, there is no evidence.

Again, nothing I've said precludes you from taking the position that prayer "just doesn't work that way," such that the whole project of even asking whether praying for things correlates with getting them is fundamentally misguided. A fair number of Christians take this position. The downside is that whatever it is you do believe prayer accomplishes, you're left with no means of acquiring evidence that it does so because we have no ability to measure "immeasurable impacts."

At that point, the conversation isn't about prayer studies anymore—it's about why faith isn't a substitute for evidence when it comes to acquiring knowledge.

> "affirming the consequent may be a fallacy in deductive arguments, but it's the bread and butter of how induction works."

That's not how induction works. That's not induction, it's logical positivism, which is not the same thing. Induction works by affirming the antecedent (I hypothesize that if I do A, X will happen. I did A; X happened). But that's a side discussion.

> "Instead, I'm arguing for one conclusion and one conclusion only—that instances of people getting what they prayed for are not evidence that prayer has a real-world effect, and people who claim them as such are guilty of confirmation bias."

I agree that's what you're arguing.

> "As such, when people do get things they pray for, it isn't evidence for your position, so you don't get to cite to it as though it were. That's the totality of my argument."

And so I don't cite it as though it were evidence for my position. Remember that I did start out by saying that was a bad argument. The efficacy of prayer (or lack thereof) is untestable by induction due to an absence of experimental controls. So if that is really the totality of your argument, I suppose I would concede it, but I don't see what significance such an argument could possibly have.

I suppose if I wanted to play devil's advocate, I could say something like this: in order to conduct an inductive prayer experiment, you need four sets of data:

1) those who prayed/were prayed for and received
2) those who prayed/were prayed for and did not receive
3) those who did not pray/were not prayed for and received
4) those who did not pray/were not prayed for and did not receive

Data across the general population shows that there is no statistically significant difference between sets 1 & 3 or sets 2 & 4 (that is the basis of your argument, if I understand correctly). Due to the lack of experimental controls, however, it is not really certain that sets 3/4 were not prayed for; there's no such thing as a "Faraday Cage" for prayer. (Remember, you're testing real-world effect of prayer as a general mechanism, not just prayer for the self). Therefore, data for 3/4 does not actually exist, meaning that it cannot be compared to 1/2; therefore you don't actually have any evidence of lack of effect across the population.

That's still a bad argument, but it does reveal that the confirmation bias can go both ways. You didn't pray, but maybe your grandma did—how do you know that wasn't a factor? You can't replicate the circumstances and try again (which is another important component in inductive reasoning).

I don't mind conceding your "one conclusion" because it's a trivial point, but you don't get to pretend to be doing induction (and thus be immune from deductive fallacies) unless you're actually following the methods of induction.

> It seems that you are basing your position off of an argument for the existence of God from the efficacy of prayer. ... I'm inclined to see the culprit as correlative fallacy rather than confirmation bias ("What I wanted happened after I prayed, therefore it happened because I prayed"), but the ultimate failure is the same.

Confirmation bias essentially is the correlative fallacy applied in a selective fashion—e.g., whenever praying for something correlates with getting it, I will draw causal inferences, but when praying for something fails to correlate with getting it, I will not draw causal inferences. I agree with you that the ultimate failure is the same—a faulty attribution of causation.

But your formulation of the argument I'm addressing is a bit different than how I'd describe it. What I'm really looking at is the following, drawn from your initial story about your friend:

1. I prayed that something would happen; and
2. The thing I prayed for actually happened; therefore
3. The thing I prayed for happened because I prayed for it.

Whether or not this is an argument you would make after having really considered it, it's implicit in your act of telling the story about your friend in response to my assertion that statistically, prayer has no real-world effect—you clearly meant the story as an example of a time when praying for something caused it to happen, and my central point is that we have no reason to think that's what it was.

> You said, "Answered prayer can't be evidence that YHWH exists unless unanswered prayer is evidence that YHWH doesn't exist." This is wrong. The syllogistic form of that statement is:
- If God does not exist, my prayer will not be answered. My prayer was not answered. Therefore God does not exist. (Q>P)
This is a logical fallacy called "affirming the consequent." Instead, what actually happened when prayer is not answered is this:
-If God does not exist, my prayer will not be answered. My prayer was not answered. No conclusion is possible ( = we don't know if God exists or not)."

I don't think that's really a fair characterization of my argument, which is about the evidentiary value of getting what you pray for and not whether God exists. I explained this a few paragraphs ago so I'll hold off going back over it here.

But in any event, bear in mind what we're talking about here—affirming the consequent may be a fallacy in deductive arguments, but it's the bread and butter of how induction works. Essentially all of science is premised on affirming the consequent.

The key is to repeatedly affirm as many different consequents as possible, without any failures. For instance, if humans have a common ancestor with modern apes, we should have largely identical DNA. We do have largely identical DNA, and thus our hypothesis remains in contention. So we add on additional consequents—for instance, we should also be able to find evidence of extinct humanoids closer to us than modern apes but still distinctly non-homo-sapien. We find this as well, and so forth and so on.

It's all a giant string of affirming the consequent until it becomes mind-bogglingly unlikely that we would have been able to affirm the consequent every single time if the conditional itself weren't true. Just because it's a fallacy in deductive logic doesn't mean it isn't exceedingly useful in figuring out how the world works. "Does prayer have any measurable effect in the real world?" is a fundamentally inductive, scientific question.

What you seem to be arguing—and I don't disagree with this—is that there could well still be at least one god out there even if prayer has no measurable effect in the real world. The only gods that prayer studies require us to jettison are those defined in such a way that we should expect them to grant requests—if not predictably, then at least in some statistically significant way. If your particular conception of God doesn't lead you to expect that prayer would have any measurable effects, then the fact that prayer has no measurable effects isn't an issue for you.

> I'm not convinced those two options are the only ones, or that they reduce to the same thing, but that doesn‘t matter because you found the right one: our asking for something has no impact on whether or not we will receive it. The purpose of prayer is not to motivate God to do something. God does what God will do according to the will of God, which is not contingent on anything that anyone else does. This is a corollary of a divine attribute called Aseity. Christians who know their theology should already affirm this.

I want to bring this back to how we got started, because I think you're trying to have a different argument than I am. For purposes of this conversation thread, I'm not making any arguments designed to establish that God doesn't exist. I'm also not making any arguments designed to establish that prayer doesn't have real-world effects, although I am taking as established that prayer has no measurable effects on the basis of the prayer studies others have done. Instead, I'm arguing for one conclusion and one conclusion only—that instances of people getting what they prayed for are not evidence that prayer has a real-world effect, and people who claim them as such are guilty of confirmation bias.

It's not that I think praying for things should correlate with actually getting them—it's that every argument of the form "I know God answers prayer because [insert example of person getting what they prayed for]" necessarily assumes that praying for things correlates with actually getting them. All I'm doing is attacking that assumption, because we've statistically established it to be false.

This is what I mean when I say that getting what you pray for can't be evidence that God answers prayer unless not getting what you pray for is evidence that God doesn't answer prayer. I don't think this is an issue for you personally, as you don't define your god in such a way as to result in any expectation that praying for things will cause you to get them. But look at what your position entails—if praying for things doesn't impact whether you'll get them, then your position doesn't predict that people will get things they pray for. As such, when people do get things they pray for, it isn't evidence for your position so you don't get to cite to it as though it were. That's the totality of my argument.

> Why do we pray if not to motivate God to action? Why does God not make his existence self-evident (in this case by answering prayers?) What is the significance of God hearing and acknowledging our prayers if he does not intend to respond? Theology can answer all of these (some more clearly than others), but these discussions are really only apprehensible after divine existence is established; you can't really debate the character and behavior of something that doesn't exist.

I'm sure you can appreciate the irony in that last sentence—for the last several thousand years, collective humanity has been embroiled in countless wars and debates over the character, behavior, proclivities, intentions, desires, and dictates of things that don't exist. We might disagree about whether your particular conception of God belongs in the list of things that don't exist, but the larger point stands either way—you don't believe in most of the gods people have believed in any more than I do, and the same statement would be true of anyone who has ever lived.

God of Mind, it seems that you are basing your position off of an argument for the existence of God from the efficacy of prayer. It goes something like this:

1. If God does not exist, my prayer will not be answered.
2. My prayer was answered.
3. Therefore God exists. (!Q>!P)

This is an inadequate argument for a host of reasons, all of which involve #2 being impossible to prove or verify. But I'm inclined to see the culprit as correlative fallacy rather than confirmation bias ("What I wanted happened after I prayed, therefore it happened because I prayed"), but the ultimate failure is the same. The case in which #2 CAN be proved is if the "answer" involves something so astoundingly coincidental and/or something that our current understanding of nature considers impossible, such that Ockham's Razor indicates that the simplest answer is divine intervention. But this is not longer the argument from efficacy of prayer, but rather the argument from miracles, which is a different discussion, and which your issue does not address.

You said, "Answered prayer can't be evidence that YHWH exists unless unanswered prayer is evidence that YHWH doesn't exist." This is wrong. The syllogistic form of that statement is:

1. If God does not exist, my prayer will not be answered.
2. My prayer was not answered.
3. Therefore God does not exist. (Q>P)

This is a logical fallacy called "affirming the consequent." Instead, what actually happened when prayer is not answered is this:

1. If God does not exist,my prayer will not be answered.
2. My prayer was not answered.
3. No conclusion is possible ( = we don't know if God exists or not).

The difficult in making this statement is that you have to prove its first premise. When Christians say that "no is still an answer," they aren't trying to prove premise 2 of the argument from the efficacy of prayer, they're refuting this premise (i.e., providing a [legitimate] reason other than nonexistence for non-answer). You yourself agree with this; you said specifically "I'm not saying that YHWH must answer all prayers or else he can't answer any prayers." This means that it's possible that God can exist but not answer prayer.

You said, "Either God doesn't exist, or he does, but we have no idea whether he answers prayers." This is pretty much true, though it would be more proper to say, "We have no idea whether He will answer any specific prayer," since one would need only ONE example (not a statistical majority, or even a statistically significant minority) to prove that he "answers prayer" (meaning "grants requests") in general. The Bible records several examples of answered prayers, and since the same Christians who believe that God does answer prayers believe that the Bible is the accurate record of the activity of God, it is not inconsistent for them to believe that God DOES answer prayer, though this gives them no assurance that he will answer any given (or any at all) prayer of THEIRS. Past performance is no guarantee of future results. Most Christians will affirm that they have no way of knowing whether or not God will grant a particular request, and most of the ones that won't affirm that are operating under faulty theology that I have no desire to defend.

"I'm assuming your theological position is that this isn't how prayer worlds, but the end result of you taking that position is that you also can't know if God answers prayer because your hypothesis makes no predictions." We know that God answers prayer didactically, not empirically. Causation (of any kind) can't be measure empirically without fully isolating variables and replicating results. Revelation ( = being told by God) is the only way we know ANYTHING about what God is like or how God acts. Generally, when we affirm something as an "answer to prayer," this is not on the basis of an absence of physical/biological efficient causes, but on the belief that God works by means of those causes. In other words, we believe that prayers are answered ONLY because we first believed in a God who answers prayer.

If the efficacy of prayer were the only argument for the existence of God, people who wanted to believe in God would have a pretty bad time of it. But it isn't. If your objective was to force Christians to admit that there is no assurance of answered prayer, fine, because we were doing that anyway. If it was to prove that the argument from the efficacy of prayer is invalid, well, technically it isn't. If anyone could manage to prove that even ONE indecent, ever, in the history of time, occurred as an answer to prayer, and NOT from some other cause, it would prove that God exists (or existed at that point in time). I have not idea how one could possibly go about proving this, however, so I will admit that the argument, while technically valid, is practically useless.

But you also said, "In order for you to avoid the conclusion that God doesn't exist, you need to take one of two positions: Either God ignores prayers (by which I mean you praying for something doesn't impact whether you'll get it, not that he's not listening to you), or he intentionally answers some of them and not others at a rate he specifically calculated to make it look like he's ignoring prayers. You can't take the position that God answers prayers at a rate other than what we'd expect if he weren't there, because the numbers are in on that one."

I'm not convinced those two options are the only ones, or that they reduce to the same thing, but that doesn‘t matter because you found the right one: our asking for something has no impact on whether or not we will receive it. The purpose of prayer is not to motivate God to do something. God does what God will do according to the will of God, which is not contingent on anything that anyone else does. This is a corollary of a divine attribute called Aseity. Christians who know their theology should already affirm this.

So if you were trying to produce a defeater for Christian theology, this isn't one. I would have given the same answer if you'd just asked, "How does prayer work?" But the impression I get is that you think the LACK of answer to some/many/most prayer is significant of something. You knew that what I said was a possible answer (I know you did, because you spelled it out), but it isn't a satisfying answer.

This is a variant of a problem I encounter all the time. Initially, at some point, we receive a description of God and what he is like. We hear that he is powerful, kind, loving, merciful, cares for us, answers prayers, etc. We hear this and we get an idea of what we can expect to experience in light of such a God. Then we go out and experience life and none of what we expected happens. At this point, we have a choice to make. Either the definition we received of God was wrong, or our ideas about what that description meant was wrong, and one of the two must be abandoned. The true disciple will abandon their conceptions and try to develop a better understanding of the God of whom they have been told. Everyone else will look for a new god who will either give them what they want or, as a consolation prize, at least fall into line with their self-generated conceptions.


As I hear it, your conception of "God answers prayer" is "people who pray for things [would] get them at a rate better than random chance would predict." You KNOW that this is not how a Christian understands "God answers prayer," because you SPECIFICALLY SAY that you know this ("I'm assuming your theological position is that this isn't how prayer works"), but here it is anyway. You have also observed that this conception does not bear out in reality (of course it doesn't). So now this is the question you need to ask: What use do you have for a God who will not give you things you ask him for?

If your answer turns out to be "none at all," than nothing I (or anything in Christianity) can say can help you. We do not serve God because we get things from him. God cures our sins and makes us like him, and that has nothing to do with answering our prayers (unless that is what we are praying for, which it should be, and note that these things can't be empirically measured). If the answer is anything else, however, this issue is really a technicality. Why do we pray if not to motivate God to action? Why does God not make his existence self-evident (in this case by answering prayers?) What is the significance of God hearing and acknowledging our prayers if he does not intend to respond? Theology can answer all of these (some more clearly than others), but these discussions are really only apprehensible after divine existence is established; you can't really debate the character and behavior of something that doesn't exist.

Remember:

This is a fallacy:
1. If God does not exist, my prayer will not be answered.
2. My prayer was not answered.
3. Therefore God does not exist.

This is not a fallacy:
1. If God does not exist, my prayer will not be answered.
2. My prayer was answered.
3. Therefore God exists.

This is not a fallacy either:
1. If my prayer is not answered, God does not exist.
2. My prayer was not answered.
3. Therefore God does not exist.

You can't prove that last one metaphysically, empirically, or theologically, but you can easily state it heuristically: "If my prayer is not answered, I don't want anything to do with the God who wouldn't answer it."

I'll start by addressing the first five full paragraphs—stated as concisely as possible, you're saying that the kinds of "prayer study" statistics I'm talking about aren't an appropriate means of determining whether prayers are actually being heard or answered by a deity.

I believe you're right, to an extent, but in a different way than you seem to think. You're right in the sense that a statistical study can't conclusively tell us whether God is or isn't answering prayer—all it can do is tell us whether there's something going on that a purely secular understanding can't account for. If, for instance, sick people who prayed to get better were more likely to get better than sick people who didn't pray, atheists like me would have a lot of trouble explaining this without invoking the idea of a god who answers prayers. All we could do is throw up our hands and say "Either God really does answer prayer, or I have no idea what's going on."

But in the real world, we find that praying for things doesn't appear to impact whether you get them at all. Like I said, this doesn't rule out the idea of God answering prayer. But it also doesn't give us a reason to suspect anything's going on in the first place, and we remain able to explain the world without invoking any supernatural beings. Why invent a prayer-answering God when the world you see around you is adequately explained without one?

> You asked for a definition of God so that we'd expect people who pray for things to get them at a rate better than random chance would predict. Your mind is set in mathematical analysis of experimental data, prediction rates, and controlled studies with control groups, placebo effects, and confirmation bias. You're applying the wrong parameters in the wrong arenas. Let me say it this way: A musician and a physicist studying Beethoven’s 3rd Symphony will bring radically different observations to the table, and it's completely illegitimate for either of them to apply their analytical perspectives in the other's domain. Your rate calculation logic and control groups studies are the physicist view; my observations and evidence are the musician's frame of reference. Both have their legitimacies.

You're misunderstanding my point. I'm not saying that God needs to be defined so that we'd expect people who pray for things to get them at a rate better than chance would predict. I'm saying UNLESS you define God in such a way, you can't point to an instance of someone getting what they prayed for as though it counts as evidence that God answers prayer. What you've done, by your earlier posts and reinforced by your first several paragraphs of this post, is deny that God should be so defined. That's all well and good, but this means you're also conceding that you have no idea if God actually answers prayer. I say this because if you define God in such a way that praying to him for things doesn't necessarily correlate with getting them, nothing you could observe in the world could ever count as evidence for or against the notion that God is answering prayers.

I mean, just try it—try to articulate something that could happen such that, if it did, you'd have to concede that God doesn't answer prayer. I'm guessing you can't do it, because you've defined God in such a way as to ensure that the notion he answers prayer is compatible with absolutely any possible state of affairs. Basically, you've made your hypothesis unfalsifiable and thus untestable.

Here's where confirmation bias comes back in, because you're still trying to have it both ways. When your hypothesis is such that nothing can rule it out, nothing can support it either. So you've got a bit of a Hobson's choice here—either you have to allow that statistical studies are relevant and concede that God doesn't appear to answer prayers, or you have to take the position that prayer isn't susceptible to statistical study and concede that we mortals simply can't ever know, one way or the other, whether God is answering prayers. In either case, you don't get to claim that God is answering prayers.

And more fundamentally, the difference between a physicist's analysis and a musician's analysis is that one deals with the objective while the other deals with the subjective. "Does God answer prayer?" is an objective question with an objective answer—it's not like in music where there's no real right answer to "Does this sound good?" The musician's approach is the one that doesn't make sense in this context. At the end of the day, there can be no fact of the matter as to whether Stairway to Heaven is a better song than Call Me Maybe. There is absolutely a fact of the matter as to whether there's really a god up there listening to prayers and occasionally giving people things they ask for. Whether we can ever figure out who's right and who's wrong is another issue, but undoubtedly someone is right and someone is wrong.

But as far as confirmation bias is concerned, then, I have a question to ask you: is it just as illogically biased for a person to conclude that God doesn't answer prayer as it is for a person to conclude that he does? If the survey says that such things are no different than random chance, is it just as biased to conclude negatively as positively?

I don't think it works in parallel like you're suggesting. Think about it—if there really were a god up there answering prayers, we could end up with essentially any kind of world. Imagine a world where people who pray for things always get them—that world makes sense if there's a god who answers prayer, but doesn't make sense if there's not. Now imagine a world where people who pray for things don't always get them, but still get them more often than people who didn't pray—again, that world makes sense only if there's a god who answers prayer, at least some of the time. Now imagine a world where people who pray for things get them at the same rate as people who don't pray at all—that world makes sense either with or without there being a god. Last but not least, imagine a world where people who pray for things get them at a significantly lower rate than people who don't pray at all—again, that world only makes sense if there's a god, although in such an instance it would appear he doesn't like being asked for things.
So what we've got is many potential states of affairs compatible with the notion that the universe is ruled by a prayer-answering god. But only one of those potential states is compatible with the notion that this universe isn't ruled by a prayer-answering god. And as it turns out, the one and only state of affairs compatible with a god who either doesn't respond to prayer or doesn't exist turns out to be the state of affairs we observe in the real world. This is prima facie evidence that God, whether or not he exists, doesn't respond to prayer.

Let me try to give you an analogy to explain why this works—imagine that I were to pick a random number between 1 and 10 and ask you to guess. Every time you make a guess, I respond by saying one random English word. It can be any word, subject only to the limitation that if your guess was right, my response must be the name of a color. So if you guess 4 and I respond with "Dog," you'll know you were wrong. But look at what happens when you guess 7 and I respond with "Blue." Strictly speaking, this doesn't tell you that you were right—I can say "Blue" whether you're right or wrong. The only thing I can't do is say something that's not a color when you're right. On balance, you'd look at this state of affairs and conclude that your guess of 7 was most likely right. Could it be a coincidence that my random English word just so happened to be a color even though you were wrong and it could have been anything? Totally. But neither of us would bet on it, because we'd be doing a poor job of playing the odds that way.

This is basically how it works with prayer studies. Whether prayer is or isn't doing anything, what we in fact observe is consistent with what we'd expect to observe if prayer had no actual effect. It's not confirmation bias for me to claim this as evidence that prayer has no actual effect, because if the operative state of affairs had turned out to be otherwise, my hypothesis that prayer has no effect would have been ruled out. This is the most fundamental way hypotheses are tested—make a prediction from them and then see if the real world matches your prediction. In the case of the hypothesis that prayer doesn't do anything, this works beautifully—we find exactly what we expected.

It's confirmation bias to draw the opposite conclusion from the same evidence, however, because the hypothesis "God answers prayers" doesn't predict a world where praying doesn't appear to do anything. The "prayer works" hypothesis may be consistent with the world we observe, but does not predict such a world. By contrast, my "prayer does nothing" hypothesis actually predicts the world we observe.

For this reason, the world we observe is evidence for the "prayer does nothing" hypothesis but it would be confirmation bias to claim it as evidence for the hypothesis that "God answers prayer."

Thanks for great writing.

First of all, scientific studies are inadequate for the subject matter at hand, viz. prayer. They are not the proper approach to understanding either the subject or its effects. It's like trying to prove by science the way I choose ice cream flavors. Sometimes it's my mood, sometimes the temperature, the day, the people I am with, the pictures on the restaurant wall, my preferences, the season, what I had most recently, or what everyone else is ordering. A scientist can't just figure out what my favorite flavors are and make a prediction. With so many factors, and many of them in my subconscious or environmentally motivated, a scientist's prediction is worthless. It's not a subject matter that is predictable.

Secondly, answers to prayer are not always scientifically observable, again making the subject matter irrelevant to scientific study. Many prayers are answered by normal means, since God uses normal means to make things happen. The Israelites walk around Jericho 7 times and blow trumpets. Just then, JUST THEN, a small earthquake happens and part of the wall collapses, enough for them to charge in and conquer the city. Is that an answer to prayer? Jericho is built on a fault-line. Earthquakes happen ALL the time. Does the event make it an answer to prayer? Does the timing? Or was it fortuitous happenstance? You think science can answer this question?

Prayer is also answered at times with unexpected results—not the way the pray-er had in mind. In 2 Chron. 32, Sennacherib, king of Assyria, is laying siege to Jerusalem. King Hezekiah prays for deliverance. In 2 Chr. 32.21, the Assyrian army is annihilated and they retreat. An Egyptian legend says that they were infested by a plague of rodents. Was it an answer to prayer? Science and surveys can't answer that question.

And then 2 Chr. 32.21 says that Sennacherib returned home, went into his temple and was murdered by his sons (confirmed in Babylonian and Assyrian inscriptions). This happened 20 years later, and is still considered part of God's answers to their prayers. Science can have no part in this study. It's outside of its purview.

Another example of prayer being answered with unexpected results. In 2 Sam. 14-18, Absalom is mounting a coup against his father, David. David's wisest (and almost flawless) counsellor, Ahithophel, deserts to Absalom's camp. David prayed (2 Sam. 15.31): "Turn Ahithophel's counsel into foolishness." As it turns out, Ahithophel gives GREAT advice to Absalom (17.1-3), but Absalom REJECTS his counsel, goes with the lousy plan, and they both end up dead. Answer to prayer? Sure (2 Sam. 17.14), but not at all what David had expected. How could science possible evaluate this stuff? It can't.

You asked for a definition of God so that we'd expect people who pray for things to get them at a rate better than random chance would predict. Your mind is set in mathematical analysis of experimental data, prediction rates, and controlled studies with control groups, placebo effects, and confirmation bias. You're applying the wrong parameters in the wrong arenas. Let me say it this way: A musician and a physicist studying Beethoven’s 3rd Symphony will bring radically different observations to the table, and it's completely illegitimate for either of them to apply their analytical perspectives in the other's domain. Your rate calculation logic and control groups studies are the physicist view; my observations and evidence are the musician's frame of reference. Both have their legitimacies.

But as far as confirmation bias is concerned, then, I have a question to ask you: is it just as illogically biased for a person to conclude that God doesn't answer prayer as it is for a person to conclude that he does? If the survey says that such things are no different than random chance, is it just as biased to conclude negatively as positively?

Every Christian claims that God exists and that he answers prayer. How do you know God answers prayer? Your answer can't be an example of some time that a person prayed for something and got it, because that's confirmation bias—if people who pray get what they want at the same rate as people who don't, then for any particular instance of a person getting what they prayed for, you can't conclude that the prayer had anything to do with it.

My philosophical point comes full circle when we start thinking about what it would take to establish that God answers prayer.

First, you'd need God to be defined in such a way that we'd expect people who pray for things to get them at a rate better than random chance would predict. I'm assuming your theological position is that this isn't how prayer works, but the end result of you taking that position is that you also can't know if God answers prayer because your hypothesis makes no predictions. Bear with me on this one for a moment.

Second, you'd need to actually observe that people who pray for things get them at a greater rate than random chance would predict. We definitely don't see this. Again, I'm assuming your theological position is that this isn't how prayer works. But again, the end result of you taking that position is that you also can't know if God answers prayer. Again, bear with me on this one for a moment.

Here's my ultimate point: In order for you to avoid the conclusion that God doesn't exist, you need to take one of two positions: Either God ignores prayers (by which I mean you praying for something doesn't impact whether you'll get it, not that he's not listening to you), or he intentionally answers some of them and not others at a rate he specifically calculated to make it look like he's ignoring prayers. You can't take the position that God answers prayers at a rate other than what we'd expect if he weren't there, because the numbers are in on that one.

But here's the thing—there's absolutely no way to tell the difference in practice between "Praying doesn't impact what happens" and "God answers prayers at a rate calculated to make it look like praying has no impact on what happens." Absolutely anything you could claim as support for one of these positions would be perfectly consistent with the other one. So where does that leave us? One of two places—either God doesn't exist, or he does but we have no idea whether he answers prayers.

If you take a look at one paragraph I wrote above: "Again, I'm not saying that Yahweh must answer all prayers or else he can't answer any prayers. All I'm saying is that answered prayer can't be evidence that Yahweh exists unless unanswered prayer is evidence that Yahweh doesn't exist. You can't have it both ways, where answered prayers prove Yahweh and unanswered prayers are Yahweh being mysterious. That's confirmation bias."

You can re-word for this situation as follows: "I'm not saying that Yahweh must answer all prayers or else he can't answer any prayers. All I'm saying is that any particular instance of a person getting what they prayed for can't be evidence that Yahweh answers prayers unless every instance of someone not getting what they prayed for is evidence that Yahweh doesn't answer prayers. You can't have it both ways, where getting what you pray for proves Yahweh answers prayer and not getting what you prayed for is Yahweh telling you "no" for a reason. That's confirmation bias."