Science communication is not easy, for a lot of reasons. For one thing, reasonable people disagree about what science communicators ought to try to achieve. Should communicators just try to keep people up-to-date on the latest cool things happening in the world of science… or should they try to foster a critical way of thinking about the world? For another thing, it isn’t clear how you would go about science communication if you tried — since, as any grade school teacher could tell you, it is hard to figure out how to get your audience to care. And for another another thing, if the aim is to foster a scientific mindset, then it’s not clear that mass media will be of any use whatsoever. (Presumably, one does not learn chemistry by repeated viewings of Gil Grissom working ponderously over test-tubes.)
These are all important and interesting topics, well deserving of thoughtful and passionate dialogue.
Enter Chris Mooney. Mooney is an activist for communicating science. He is the author of The Republican War on Science, and is the co-author of the controversial book Unscientific America (with Sheril Kirshenbaum). Mooney holds a degree from Yale, a fellowship with the Templeton Foundation, and is a member of the board of the American Geophysical Union. He blogs at the Intersection. Mooney/Kirshenbaum’s ultimate legacy appears to be that they succeeded in starting a passionate conversation about the subjects listed above.
Which brings us to the topic of the present post. In addition to being in the science communication business, Mooney and Kirshenbaum are both critics of atheist activism. Mooney and Kirshenbaum have argued that activist atheism is detrimental peripheral to science communication, and that activist atheists are often uncivil. Their critical remarks have created a tumultuous debate in both online and national print publications. Not incidentally, Coyne, Dawkins, and many others have publicly argued that there is an intimate connection between science and atheism. (Full disclosure: although it shouldn’t make any difference to this post, on this issue — as on most things — I’m in the “Jason Rosenhouse camp“.)
On first blush, it seems as though there are two major issues here: civility, and the role of the naturalist worldview in science. But a little over a year ago, I had the opportunity to speak with Mooney about the role of passion and conflict might have in getting people to think about science. And from that conversation, I learned that Mooney acknowledges a third sticking point.
BN: I was glad to see that you didn’t focus on the deficit model in explaining scientific illiteracy — that’s really good. [Edit 2010: Roughly, the "deficit model" is the idea that science communicators should presume that citizens that are not scientifically literate are responsible for their own illiteracy.] And the alternative is to look at what people do know. So for example the mechanic has a body of knowledge that I can only dream of — I just don’t know how a car works. We ask ourselves how people have all this impressive statistical knowledge about baseball and things (without knowing about science), and the reason is: baseball is useful in some way. People are embedded in a social group and they know that this knowledge will be useful to talk about.
This can also help us understand how misinformation works. For example, the George Will episode. People will say “Atta boy” and pat him on the back for acting like an idiot.
CM: I think you’re right. These things have utility, is what you’re saying.
BN: Exactly. And this leads me to the atheism thing. So you’ve gotten into a bit of trouble with some folks online, because atheism has utility for them. And I’ve found that I’ve learned quite a bit on these atheist forums.
I’d like to hear your thoughts on this. Because you’ve been interpreted as saying to folks like Jerry Coyne: “Don’t make atheistic arguments, because you’re putting atheism in the same truck as science, and people are not going to take science seriously because they’re religious.”
But atheism is a way of getting people interested in science. So Dawkins writes “The God Delusion” and he presents this panoply of interesting bits of information leading up to an argument.
CM: I understand exactly what you’re saying. People say all of these different kinds of things serves a purpose for them — I think that’s absolutely true. And I really like how you framed it, because I haven’t put it in quite that way, but it’s totally right, and it’s one of the reasons why it’s so hard to talk about bad information.
But that doesn’t change my particular view on atheism to point that out. I guess what I’m saying is that a lot of people in what we call the New Atheist movement have formed a community around a set of information, and it has utility for them, in your terms. There’s no doubt about that. You see them doing it so much, so fired up about it.
My argument is that almost in direct proportion to how it’s useful to them, it’s not useful for something else. And that can happen — a community can form around a shared body of information and another community can think it’s awful. That would totally work in your model. And my point is that even as they’re agreeing, scratching each other on the back, creating a dialogue that’s mainly amongst themselves, if you look at how that affects the broader dialogue in the country, it’s a different dynamic entirely.
So I think what I’m saying is: be aware that the way you talk about atheism works for you, and yet it also isn’t working in a different world. I think both those things can be true.
BN: A counter-argument is that you have religious folks who want to defend their views. The Ray Comforts of the world. And to the extent that they want to defend their views in any interesting way, they have to engage with the explicit arguments that are put forward by the atheist community. So that way it becomes something like a dialogue, so that at least it appears as though there’s something defensible going on [on Ray Comfort's end].
So I have this underlying feeling that conflict isn’t necessarily a bad thing. We can’t ever put ourselves in a place where we say, “Oh, no conflict, that’s no good”. And that seems to be what you’re doing — you call yourselves “accommodationists”, or at least that’s the label that’s been put on you. Conflict, to the extent that we want to have a debate, is okay. It’s just a certain kind of unproductive dialogue that sometimes goes on.
CM: Yeah. I think there’s all different kinds of conflicts. And there’s many things you can spend your time debating. We all pick and choose. My point on the general conflict between science and religion in the United States is that I don’t think it’s an incredibly fruitful one, and I don’t think it does the public understanding of science a lot of good to be hitched to the religion-bashing way. I think there are many ways to talk about science in religion in American society that would work better, and I think there’s a lot of evidence to support that, in terms of the way people react.
I’m sure that some people are getting engaged because of New Atheism — I’m sure some people are learning, some people are thinking about science — but I think it’s also clear that a lot of people are not getting engaged or are being negatively polarized. So it’s a difference of goals, in part, that explains the debate I think.
I think it is fair to say that, by far, Mooney and Kirshenbaum sparked the most outrage with their comments over civility. But the ensuing drama has drawn attention away from some of the most interesting questions. How does Mooney think people ought to communicate science? What does “science communication” involve, for him?
One thing is pretty clear. Mooney wants to offer strategic advice about communicating science. Both in person and in his written works, he aims to communicate the art of publicity to scientists, under the auspices of teaching them the art of communicating science to the public. This work is predicated upon the assumption that everyone has the same priorities, in the minimal sense that at least that everyone is on board with the “science communication” project.
But the most important point that I’m going to emphasize here is that his stance is self-consciously political. At least to some extent, there is a “difference in goals” between Mooney and the activist atheists — by which, I think, he means a difference in priorities. Mooney does not think that speaking out against religion is a priority, and that it is on the whole detrimental to science education; while others think it is a priority, and that it supports science education in some respect.
What’s interesting that the one thing that Mooney and the rest agree on is this: that activism over atheism really does have some utility in communicating science. It gives us something to talk about.
You might think instead that his stance is less “self-consciously political” than self-consciously “cultural”.
To paraphrase badly from Rorty’s “Contingency, Irony, and Solidarity” – perhaps Mooney has developed a talent for speaking differently instead of worrying about being right or arguing well. Might he be choosing ‘Narrative that Matters’ as a means towards morale change and progress over ‘narrative that gets dismissed’?
My own experience, which I deliberately kept to a minimum in this post, is that Mooney and Kirshenbaum place little to no accent on the importance of conversation as a means of persuasion. They are in a kind of “mass media”, West Wing publicity mode. Hence they focus on the message, and ask communicators to frame the message to fit with their audience’s pre-existing beliefs. They are afraid that the science narrative will be dismissed if they are too confrontational.
But actually, that’s not a salient contrast. Both sides think that they’re trying to progress towards narratives that matter and away from narratives that get dismissed.
A better contrast between the sides is that they have a clear difference of goals. After all, Mooney recognizes that the new atheist arguments have utility — he understands that these arguments really do convince people. It’s just that, from his point of view, they just don’t have enough utility (or do not serve the purposes that he prefers).
So here’s an interesting slant on “West Wing” modes of engagement. Flip Atheism and you have Religion.
Considering the “science communication” message through that lens, is it Religion or is it Science that’s really the medium?
In America – of the two; which one can can go viral? And, one wonders – have Mooney and Kirshenbaum considered the implications of that?
There is a difference between arguments and reasons which convince some people and what convinces the mass of people.
While a porcentage of the population can be convinced by good arguments and good reasons,
the mass of people are generally convinced by skillful propaganda.
Whether they don’t want to or cannot (for lack of intelligence or lack of rationality) listen to good reasons, the fact is that
they don’t.
Even if we succeed in teaching that mass about “science”, science for them will just be another dogma, which they assimilate without thought just as if they become atheists, they will become unthinking atheists: anyone who has met dogmatic Marxist-Leninist atheists will understand how unthinking atheists can be.
What I am saying is as old in political theory as Plato’s Republic.
Missing from your rundown is the reason why Mooney thinks good science communication is so important. It’s not so people can know about the latest cool stuff, but so they can make informed decisions about public issues. He’s constantly talking about (for example) climate change, vaccination, the oil spill, stem cell research, and the like. These are life and death matters, so science education has to proceed quickly and in a targeted way. The goal is not to create an ideally rational citizenry, but an electorate that makes informed decisions about public policy issues.
Also: I don’t think Mooney objects to all atheist activism, but worries about top science communicators who also go after religion, both constantly and caustically. An example of something he objects to–PZ Myers calling Francis Collins a clown, even though he’s a friend of evolutionary biology and a supporter of stem cell research.
What I will never understand: why the new atheist crowd can’t just disagree with Mooney instead of despising him.
Amos/Dana, I think you both have the same topic. But, assuming that I’m reading her correctly, Dana proposes a counterpoint to Amos’s use of the Republic.
The idea is that by its very nature science can’t go viral. But atheism — like religion — can. Hence, as amos agrees, you can have “unthinking atheists”, compelled by propaganda.
The question is, of the two, which of the positions is more conducive to getting people to take science policy seriously? I mean, if we insist on looking at the situation as miserly Platonic Republicans, then we’ll have to take it as a given that you’re always going to have cultural dopes. So then the question is, which dopes are friendlier towards science policy? Which group is more likely to impede scientific recommendations in government?
Ben:
While you can’t have unthinking working scientists just as you can’t have unthinking philosophers,
you sure can have unthinking pro-science people.
I’m one of them. I haven’t opened a science books since secondary school almost 50 years ago, but I believe what the scientists say about global warming, about evolution, etc.
I believe doctors or at least I believe the Mayo Clinic website.
That’s Mooney’s goal, as far as I can see, to foment a mass of unthinking pro-science people.
Jean, I didn’t mention advocacy for science policy explicitly because it’s not a point of contention. But I did allude to it in the interview. (sp. see the comment on George Will being an idiot.)
I can’t stop the discussion from being derailed into talk about Chris’s high crimes and misdemeanors, since that’s been done elsewhere. Still, I’d much rather have a conversation that focuses on the issues I pointed out in the first paragraph.
Amos, my point is that, at their worst, Mooney’s critics would settle for that too. The difference is that they try to aim for a higher goal as well — e.g., they themselves actually do science education.
Ben, I don’t think a reader can get an accurate picture of Mooney without the two points I mentioned, so far from derailing the discussion, I’m just trying to keep it on track.
Jean, no, I don’t think so. The post is accurate without having to emphasize those points. To my knowledge, there are no lies (by omission or commission), misrepresentations, errors, or mistakes in what I posted above.
On the other hand, if by “accurate” you meant “complete”, then I would agree. But I don’t aim for completeness.
I think the 2 points are essential. You make it seem as if science communicators are either trying to keep people up to date on cool things or trying to teach critical thinking too, but that omits the goal I think is most important to Mooney- creating an informed electorate to deal with problems like climate change as soon as possible, i.e. before it’s too late. You also make it seem as if Mooney is generally opposed to atheist activism. It really does create a more accurate portrait to have examples of what he objects to.
You’re certainly right that public policy is a central issue for everyone, Mooney included. But, that’s not any kind of interesting contrast that vexes science communicators, so it’s not part of the post.
On atheist activism. Well, I’m struck by the fact that he believes there’s a difference of goals. And that’s what I heard come out of the discussion. Atheist activism gets definite low ranking on his list of priorities — there’s no debate about that. It’s not important to him.
But you’re right to insist that this doesn’t logically entail that he is necessarily opposed to atheist activism (and he’s often said he isn’t). So on re-reading, I should have chosen my wording more carefully in places. I’ll do a quick edit for clarity, with thanks for your input.
Okay – maybe I missed the point but at least the ricochet fell closer.
Science is dull. No airtime for that. Check.
Saying that God is a delusion is bound to get some “Airtime”, so by saying “I’m a scientist and other scientists shouldn’t say things like that”, that gets airtime too. Pretty easy formula – point/counterpoint. Media Manipulation 101. Check.
Big Bang Theory is getting good Airtime… perhaps the Sci-Comm-Activist community should get some climatologists and ecologists inserted into the script… wait! Leonard and Sheldon and gang are physicists… so that’d just be ludicrous. Wait! It’s a Sit-Com, Point/counterpoint, Airtime…
it’s cultural, not political. The “Cool” Science of “Big Bangs” and such can’t even compete for the eyeballs unless it’s wrapped in something else and snuck in under Mr and Ms Averages’ radar. How can a meta-concepts like “Science Policy Awareness” or – worse, “Science Literacy Policy”, even dream of living top of mind? Yawn.
Science is boring and difficult and the Public wants to leave that to the Scientists. And – here’s the point – I think Mooney knows that too. His apparent position is constructed as a means towards an invitation to visit Capital Hill where he can have his day, week or however long he can get himself in front of the policy makers. He’s not “self-consciously political”.
He’s crassly political.
I’m not sure why Mooney ever got involved in this “new atheism” thing. I’d guess most people aren’t too bothered in what faith a scientist does or doesn’t hold, the problem’s more that they’re saying things people don’t want to hear – “Hey, your lifestyle’s ruining the world!” “Hey, there’s a big infection!” – and, in fairness, that there’s been a lot of publicised or not so publicised skullduggery in the communication stakes – from the WHO’s connections with Big Pharma to the ClimateGate shebang – that confirms such instincts.
“You’re certainly right that public policy is a central issue for everyone, Mooney included. But, that’s not any kind of interesting contrast that vexes science communicators, so it’s not part of the post.”
I think there really is a contrast here, and it’s a big part of the reason why there’s a disagreement between Mooney and the folks on the other side. Mooney is all about problems like climate change, where science communication is “time sensitive”. We need to get people to believe the science as soon as possible, because otherwise we’re heading for disaster. Mooney’s books and a huge percentage of his blog is devoted to this sort of science, especially climate science, where it’s morally important to change minds quickly, “or else.” I think that focus explains why being “strategic” about communication makes sense to him. If you’re in a hurry, and it’s morally important to succeed, you get strategic.
On the other side, there’s acceptance of climate change, of course, but there’s nothing like the same focus on it. I don’t think there’s any sense at Coyne’s blog that we have limited time to succeed at science communication. Because of that difference of perception, it makes perfect sense that Coyne is more interested in more generally “fostering a critical way of thinking about the world,” as you put it.
In fact, the way you’ve put the two options, it’s a foregone conclusion that the more inclusive one is preferable. Why would anyone aim just to “keep people up to date on the latest cool things happening in the world of science” if they could also “foster a critical way of thinking about the world”? If those were the only possible goals, the choice would be obvious. But no–there’s a third possibility. Your goal might be to save the world!
Now, all that said, of course there can still be a debate. What should our goals really be? Should inculcating climate science (and other science with moral import) be #1? Is global warming really that pressing a problem?
If you agree with Mooney on the goal, then you’ll also (of course) have to think about strategy. What’s going to work best if you want to create a scientifically informed electorate–in the US, now, quickly, and when so much is at stake? Once the problem is set up that way, you can still have a debate, of course, about whether science education should get mixed up with atheist activism. But Mooney’s views at least become much more understandable.
What’s wrong with being “crassly political”, if your political goals are good?
It seems to me that Mooney’s political goals, dealing with global warming, are good.
Dana, yes, that seems like an effective strategy to me. Pop, bang! Conflict! Science!
For example, the people on Fox News are absolutely delighted to have Dawkins on O’Reilly. If I were to guess, not just because it raises their profile. They feel no threat in bringing him on — they suspect that their audience isn’t going to be convinced by anything he says, and they get the chance to look as though they’re providing balance. And yet, Dawkins gets to reach out to people that he ordinarily couldn’t.
Jean, in order for there to be a meaningful contrast between unhurried science communication and short-term policy reform, you’d have to think that they’re mutually exclusive. But I don’t think either Mooney or his detractors would agree that they’re only involved in one and not the other. For instance, when Mooney was criticized for failing to address the role of public education in science communication, he said, (paraphrase), “I wrote that book already” — meaning, it’s still a vital point for him. Moreover, my experience is that he is very skilled at communicating when he’s in a classroom environment. And you see Coyne post with some regularity on the subject of climate change, despite the fact that he is not a climate scientist. http://whyevolutionistrue.wordpress.com/?s=climate+change&searchsubmit=Find+%C2%BB There’s no shortage of earnestness on either side.
And to be totally honest… if there were any real contrast going on here between science communication and short-term policy reform, I don’t think that it would make for a very flattering picture of Mooney and Kirshenbaum. It would give the sense that they really are just proposing a BAND-AID solution to science communication. But actually, I think they understand, at times, that genuine (seemingly unhurried) science education can be a means of achieving short-term policy reform.
Amos, good intentions aren’t enough to make for good politics. e.g., I don’t doubt that Bush had something like good intentions in toppling Hussein, but that doesn’t mean the invasion was good, or a force for good. For one thing, your tactics matter, too.
Ben:
I am aware that good intentions don’t necessarily make good politics.
Frankly, if there is a group of people, who in spite of their good intentions, alienate possible supporters instead of winning them over, it is the New
Atheists. Almost everyone comes away from an encounter with them with a bad feeling.
As a matter of fact, your reply to me smacks of New Atheism: it’s superior and condescending, as if after a life in politics, which I have written about previously, I would not be aware that “tactics matter”. It would be similar if knowing that you are a graduate student in philosophy, I were to patiently explain to you that
Wittgenstein was born in Austria.
That being case, I suspect that Mooney would be much more successful than the New Atheists in convincing people about the dangers of global warming.
Ben, You yourself are trying to get at a difference in priorities between Mooney and his opponents. That’s part of your thesis–that there really is a difference. But do you really think you’re painting the right picture, when you portray Mooney as after communicating science, just for the sake of communicating science? Or even “keeping people up-to-date on the latest cool things happening in the world of science” as opposed to teaching critical thinking?
I think that reading just doesn’t make sense of the data. Mooney’s #1 topic at his blog and in his books is climate change. He’s relentless about it. In fact, people at CfI forums complain that he does too many PoI shows about climate change. He also talks about the oil spill, vaccination, GM foods, and the like. I don’t doubt he’s keen on all science communication, but his blogging and writing make it really quite clear that he’s worried about making the world a better place.
I think it’s way way too dismissive to call this a concern with “short-term policy reform.” For one thing, it’s going to take a mammoth effort and a lot of time to get people up to speed on science so that they make decisions about a whole range of critical issues on a reasonable basis. Far from being unflattering to Mooney and Kirshenbaum, I think it puts them in a positive light to see that they are worried about morally important issues, not just about teaching people cool stuff. It would be pretty weird for Mooney to be so adamant that atheism is getting in the way, if his goal were “cool stuff communication,” not saving the planet, saving lives etc.
I don’t agree that those goals are just as important or just as much in the forefront, for folks like Myers and Coyne. Myers is constantly talking about truth. Why should we talk openly about atheism? On his view: because it’s true. At his recent CfI debate with Mooney, that’s what he said–it’s about truth, truth, truth. At the very least, there is a difference in emphasis here, and I really don’t think it’s the difference between one side (Mooney) just wanting to communicate science for the sake of science, and the other side being concerned with critical thinking as well.
p.s. Sorry for the length.
Amos, I mean neither condescension nor offense. But as a matter of fact, your argument — as it has been stated — is what it is, flaws and all. It should be rebutted with no more or no less than what’s needed to refute it.
So in your example, if I made an argument to the effect that Wittgenstein was born in Poland, then you would not be condescending to point out that this is false. You would be stating a plain vanilla fact. And it wouldn’t matter if I were the King of Philosophy, crown and all — I had better substantiate my claims, or else weaken them, or retract them. And since I hold myself to these standards, and I hold veteran philosophers to these same standards, and hold you to the very same standards, I think that’s an indication that I am showing you the same degree of respect that I have towards myself and towards veteran philosophers.
As it should be.
Jean, the first paragraph is an introduction, where I attempt to clear the ground to give people an idea of what issues you have to grapple with in science communication. This paragraph precedes any discussion of any persons in particular. Some of them apply to Mooney vs. his detractors, while others don’t. Of course, I was most interested in foreshadowing the rest of the post with salient contrasts, but that doesn’t mean that they all fit perfectly.
So it’s not correct to suggest that I portrayed Mooney as communicating science for its own sake as opposed to being for the sake of policy reform. In fact, I didn’t say much at all about the matter of science policy reform, as you initially (and correctly) noted. But I can’t be guilty of both of these things at once — you’ll have to choose one or the other.
What I said was this: “in order for there to be a meaningful contrast between unhurried science communication and short-term policy reform, you’d have to think that they’re mutually exclusive. But I don’t think either Mooney or his detractors would agree that they’re only involved in one and not the other.” (Incidentally, if you don’t like the phrase “short-term policy reform”, I’m open to changing my use of phrase to something else.)
That having been said, I *do* think that you can set up this contrast in the abstract. So, for instance, Roger Pielke begins “The Honest Broker” by distinguishing between four kinds of science communicators: the “pure scientist” (who just shares bland information to the public), the “science arbiter” (who tries to show the public the scientifically credible policy options), the “issue advocate” (who has a very particular policy option in mind), and the “honest broker” (who tries to provide a range of scientifically feasible policy options based on the values of the public). And if I have understood you correctly, you seem to want to distinguish the pure scientist from the issues advocate.
But are there any interesting personalities you might fit nicely into Pielke’s categories? I don’t think so — they all have a passion for issues advocacy in common. Yes, Myers says “truth truth truth”, and then he goes on to advocate positions. Same with Coyne, etc. The difference is, Coyne etc. think that atheism *is* an issue worth advocating, while Mooney doesn’t.
Ben:
If you want to be precise, I never mentioned “intentions”: I spoke of “goals”, which are very different than intentions, so your criticism of me was misplaced.
Although you’re correct that they’re distinct locutions, that makes no difference to the point, since in this context my use of the word ‘intention’ was meant as synonymous for ‘goal’. Rephrase the example and I think you’ll see that it still applies: “I don’t doubt that Bush had something like the right goals in toppling Hussein, but that doesn’t mean the invasion was good, or a force for good. For one thing, your tactics matter, too.”
Ben:
They’re different.
For example, I don’t doubt that Bush had good intentions in invading Iraq, but I don’t agree with his goals.
Most people have good intentions: Osama Bin Ladin probably has good intentions.
Ben, I just think it helps a person assess Mooney if they do know exactly what his goals are. You talk about two goals in your first paragraph, but (as you say above) neither is Mooney’s. You also talk about another goal that isn’t his–spreading atheism. But to really understand Mooney, it surely helps to see what he does identify as the most important goal of science communication. None of the above, but rather quickly producing the sort of science literacy that’s needed in an electorate that’s going to deal reasonably with the pressing issues of the day. That goal comes across clearly in the first pages of his book (click on page 4 and read 4-5, starting with the Pluto paragraph)–
http://tinyurl.com/2gyjmf4
Is that equally everyone else’s goal too? I don’t really think so–not equally. There are differences in priorities and emphasis. As to whether “unhurried science communication” is incompatible with quickly creating this kind of literacy… maybe not. If “unhurried science communication” includes dismantling the religion of your audience, it may be incompatible with getting the desired literacy quickly. I think it’s crucial to this debate what the goal really is, and how quickly it needs to be achieved.
Amos, I gladly accept your corrective.
So to be clear, I reject the idea that good goals make for good politics. As I said earlier, you can choose ineffective or improper means of accomplishing your goals. For example, Bush (to say the least).
Also, what I didn’t mention, but could have, is that there’s a second point in which good goals can make for bad politics. You might have all the right goals, but prioritize them in a totally unreasonable way. So for instance, if you were the President, you might sit and read to schoolchildren instead of attending to an invasion of your country. Everyone might agree that the goal of “do not startle children” is a good one, but most sane people also think it’s somewhat more important to attend to matters of immediate and dire national security.
Jean, as I said, the first paragraph is prefatory material, so it’s best to read it that way.
Also, activist atheism is certainly not one of Mooney’s priorities, as he explicitly says time and again. That’s why it is mentioned repeatedly here: because he disavows being an activist atheist (even though he himself happens to be an atheist).
We don’t agree about the attitude that activist atheists have towards science communication. I think some of them are quite earnest. Myers, especially so. Myers is positioned to make a quick impact, since he’s funny, explicitly political, and passionate. And people respond to that in spades (as Mooney himself elliptically notes).
In fact, if I were to make a comparison, then I would say that Mooney is the one who is trying to play a longer game, because he’s trying to frame things for a national audience.
Ben:
According to Webster’s dictionary,
“goal” suggests something attained only by prolonged effort, while
“intention” implies little more than what one has in mind to do or bring about. So I don’t know if we can say that Bush’s goal was not to frighten the children.
How about Bush’s “objective”?
I agree that one can have a good goal and choose an ineffective means for reaching it. However, generally, in politics, one first seeks out candidates or activists who share one’s goal or goals. After that, in the political process, one can criticize their tactical ability or lack of it.
What leads you to believe that Mooney lacks tactical ability?
From what I can see, he is quite a skillful operator. In fact, much of the criticism of him in the GNU atheist blogs revolves around the fact that he is too astute, too cunning, too pragmatic, too much of a salesman, not “idealistic” or “pure” enough.
Any movement needs both inspirational figures (not Mooney) and political operators (Mooney).
Ben, I’ve read your post several times now, and I’m really not harping on the first paragraph. I’m just saying that nowhere do you seem to identify what Mooney sees as the most important goal of good science communication. I don’t believe the goal is obvious or agreed upon by all parties, and I don’t think you can characterize Mooney’s goal merely negatively–that it’s not spreading atheism. I think knowing what the positive goal is is crucial for anyone who’s going to fairly evaluate his views about science communication.
Obviously, just understanding the guy doesn’t mean being convinced. Sure, Myers is earnest. I’m not sure why you think I’d dispute that. Yes, he convinces some people of some things quickly. The dispute is about whether he convinces the right people of the right things quickly.
We both know this is a huge debate, and I do respect the arguments on the anti-Mooney side. I just don’t respect the way some people don’t seem to want to understand the guy. (I’m not saying you!)
Anyhow…
I’ll have to shoot myself if I spend New Year’s Eve talking about these things. So…I’ll sign off now and wish you a happy new year.
Happy New Year, Ben and Jean.
Amos, I agree with your final sentence.
So here’s my critique at the level of strategy. I think that the most effective way of convincing a population into some point of view would be to use a good cop/bad cop dynamic. However, if the good cop is always beating on the bad cop, and vice-versa, the dynamic is going to be ineffective.
But of course it’s not all just a matter of strategy, it’s also a matter of priorities. So there’s that.
For both of these reasons, there can be quite a lot that is wrong with being crassly political, even if your political goals are good.
Jean, we definitely don’t agree about what’s important to all parties! To use my example, Myers’s successes as a science communicator speak to his talents as someone who is trying to push for policy reform. That’s why I brought it up. Everyone aims to convince the right people of the right things quickly — but there is certainly a disagreement over who “the right people” are. Myers et al. focus on the people who will listen; Mooney et al. focus on the people who won’t.
Also, putting aside one infelicitous phrase that I corrected, I don’t agree that the post was inaccurate… though I will readily admit that it’s incomplete.
And happy new year to you too!
Ben:
I agree with you about the good cop/bad cop strategy.
Wasn’t that the game played by Martin Luther King and Malcolm X during the 60′s in the U.S.? The Reverend King said to the president: if you don’t negotiate with me, you’ll have to deal with that guy over there with the gun (Malcolm).
However, the good cop/bad cop strategy isn’t going to work in this case. I’m not sure why.
Too many prima donnas, too many divos/divas, too many oversized egos, perhaps?
The problem is, Ben, that you’re the only team player in this game.
(That’s a compliment, by the way.)
One salient disanalogy between this situation and the good cop/bad cop scenario is that there’s a difference in goals. e.g., for all practical purposes, Mooney and Myers have different audiences in mind. This is different from our favorite sleazy cop shows because you never see the good cop and bad cop interrogating different suspects.
Anyway. I think that the more common ground people have, and the more insight that they have into what the dynamics are, the easier it will be to find opportunities to cooperate.
Ben, You keep saying (in the post, and in comments) that there’s a difference in priorities, a difference in goals. I was trying to spell out the goals on each side, but you seem not to be biting. So what’s your view of the goals on each side?
Jean,
For one thing, I think it’s fair to say that they have different idealized audiences in mind. Mooney constantly talks about an idealized audience of overworked, diffident, frustrated believers who are alienated from the system. Myers has an idealized audience of people who are willing and primed to engage, both detractors and supporters. So in that sense, they have different priorities.
Still, at the end of the day, they both want to engage people in talk about science policy. Both Mooney and Myers are “issues advocates”, to use Pielke’s term. But unlike Mooney, Myers is also a “science arbiter” (e.g., when he blogs about peer-reviewed research). Though of course, that’s not to say anything bad against Mooney, since he intends for science arbitration to be left to Kirshenbaum.
By the way, is my use of Pielke’s terms helpful in getting at the ideas you’ve been highlighting? Or is it just making the discussion murkier?
I really see an even more ultimate difference of values here (one not quite statable in Pielke’s terms)–and maybe that’s why I find the whole debate interesting. This relates to my own interests in “ultimate goods” and which are essential to “the good life.” I think Myers values true belief just for itself. I don’t think Mooney does quite as much. The way he sees things, we need true beliefs so we can avoid death, disease, species extinction, environmental deterioration, etc. I get that impression especially from listening to Mooney and Myers debate at CfI and then again at PoI. Every time Mooney talks to someone “on the other side” it’s pretty clear he doesn’t value sheer true belief as much as they do, but rather cares more about “non-epistemic” goods–life, health, species preservation, etc. I don’t think this way of contrasting the two sides is tendentious. It’s not a way to make one side win. It just exposes an interesting disagreement. It makes me think of a certain book called “Why Truth Matters.” Why does it matter? It’s an extremely interesting question, and I kinda like to think that underneath it all that’s what this fracas is really all about.
We’re going to be misled if we just judge their argument on the basis of one or two debates. You have to draw upon the whole corpus of what they’ve argued (at their respective blogs) in order to get a clear idea of where they stand.
The only evidence that would convince me that Myers et al. only value the “truth in itself” is if could be shown that they were mute and dispassionate about science policy. But I think it comes out pretty clear that they think the facts are interesting in themselves, but also useful with respect to those non-epistemic goods you mention (life, health, etc.) For example, I think one of the best arguments for science that Myers has ever made, was the story of how he avoided a heart attack by ignoring wishful thinking and acting on early warning signs.
So it seems to me that there’s no real contrast here, because if there’s a difference between valuing truth-for-itself and truth-for-utility, it’s only a difference of degree.
That having been said, if there actually were a contrast, then the implication would be that Mooney doesn’t care about the truth in itself, and only cares about the social role of truth. In other words: he would only care about the truth you can sell to people. In that case, he would only be advocating on behalf of science that works: engineering, health science, meteorology, etc. But as a matter of fact, that’s just not the tune he’s singing. He’s defending science on the whole, including its nebulous R&D aspect. So he has to care about the truth for its own sake, to some extent, or else we can’t make sense of why he cares about science as a whole system.
I think you misunderstood what I meant when I said “Myers values truth just for itself.” I wasn’t saying he exclusively values truth, and values nothing else. I meant he values truth intrinsically. Then I said Mooney doesn’t “quite as much,” which doesn’t mean “not at all.”
Yes, it’s obvious Myers values lots of things besides truth. I read Pharyngula often enough to see that. Obviously he cares about life and health and equality and all that good stuff. But at the recent CfI and PoI debates with Mooney, he repeatedly offers “it’s true” as a complete reason to talk about atheism. He’s downright eloquent in his “praise of truth,” I would say.
By contrast, Mooney wants to know what the costs are going to be of talking about atheism. It doesn’t follow he doesn’t attach any intrinsic value at all to truth. Of course not–that would be going way too far. He’s just more interested (than Myers) in looking at consequences of speaking about this or that truth at this time, in this context, in this way…
I don’t mean to suggest a black and white contrast. It’s a question of how much weight each gives to sheer true belief, compared to other things. That kind of different can be important, even if it’s not stark.
I see. The thing is, it’s a hard call to make, because technical philosophical terms imported from value theory and meta-ethics (like intrinsic vs. instrumental value) are difficult to apply to situations that are open to interpretation, like this one.
It’s hard to use the term, “intrinsic value”, because the concept itself sometimes seems like it’s a philosophical snipe. To take one example, Dewey has somewhat reasonably argued that intrinsic value is valuable relative to a context. Also, I would imagine that a Quinean holist might be skeptical of applications of the concept. The Quinean might argue that if a thing is “intrinsically” valuable, then that might only mean that it is relatively central in our web of beliefs about values — which is to say, that we believe that the intrinsic value is instrumentally valuable to a lot of things in a lot of different contexts.
Because I have unresolved commitments on these fronts at the moment, I’m unwilling to apply the concepts to situations where the evidence is ambiguous. Still, of course, you can interpret a contrast, so long as it’s a contrast of differing degrees. Until I have a complete meta-ethics available to me that resolves the Deweyean and Quinean worries that I briefly sketched above, the interpretation of values in the present case is just going to be speculation.
Somehow the intensity of the hostility that Mooney provokes in certain GNU blogs does not seem to be entirely provoked by differing views on the role of truth.
Why is there such hostility towards Mooney? They even mock his smile, which does seem to be out of place in a discussion over why truth matters.
Yes, but I wasn’t trying to explain the hostility. If that was the question, I’d have a very different story to tell. But I think Ben was signaling he didn’t want to “go there”!
It’s hard to make sense of the debate between Mooney and the GNU atheists without taking into consideration the amount of animosity involved, on the part of some GNU atheists.
This is not just a gentlemanly, abstract discussion about truth, such as that between Richard Rorty and Bernard Williams.
These people are out to “get” Mooney.
Ben, To respond to your last comment. My story about Myers vs. Mooney just presupposes we can talk about what different people value as ultimate ends, and how they weight those various things, etc. etc. I think describing people’s values in these terms is just doing ordinary psychology, not metaethics. I don’t see anything really controversial about this sort of psychological “compare and contrast.”
Amos:
“It’s hard to make sense of the debate between Mooney and the GNU atheists without taking into consideration the amount of animosity involved, on the part of some GNU atheists.”
Then again, there are accommodationists who get considerably less flak than Mooney. How come? Are those other accommodationists better communicators?
Mooney himself strongly confronts both gnu atheism and specific gnu atheists. That’s OK, but is it then reasonable to whine about the confrontational attitude of others? Could it be that gnu atheists despise Mooney because of his hypocrisy, not because of his accommodationism per se?
Matti:
I don’t know exactly why the GNU atheists despise Mooney.
However, in a debate of ideas,
the ideas should be debated, not whether the persons holding them are hypocrites or not. In a debate of ideas, the fact that a person is a hypocrite is irrelevant, is a ad hominem argument and has nothing to do with whether the argument of that person is true or not.
If people want to discuss Mooney’s character, fine, but his character has nothing to do with whether his arguments are worth considering or not.
Guys
This isn’t my thread, so I’m a little reluctant to intervene (though, frankly, not that reluctant), but just to make it clear:
Attacks on Chris Mooney’s character will not be tolerated here.
So if that’s what turns you on, go elsewhere.
(I should say that my intervention here is no reflection on the original post – which was absolutely fine – nor on Ben’s exemplary conduct on this thread.)
Finally, just in case you’re tempted, this is not the place to discuss my moderation of this blog. I’m sure you can find other places on the blogosphere where you can discuss my failings to your heart’s content, so knock yourselves out.
Matti K.,
I dropped the “despise or disagree?” issue way up because Ben wanted to stay away from it…but we’re back there now. So a comment on hypocrisy–
An unfortunate thing about the title of that recent CfI debate in LA–”Confrontation or Accommodation?” or some such–is that it made it seem as if one side is always confrontational, never conciliatory; and the other side is always conciliatory never confrontational. But surely not.
I certainly don’t think Mooney’s views about communication entail one should never be confrontational. His views say we should be strategic about what we confront, particularly when there’s some important goal we’re trying to achieve. That means not confronting evangelicals about Jesus (or whatever), if we’re trying to get them to support teaching evolution. It doesn’t necessarily mean one atheist shouldn’t confront another about methods of communication.
I’ve tried to relate to the idea that atheists shouldn’t criticize atheists, because they’re all in one despised boat together, but I can’t really see it. In fact, criticism can make the whole boatload more attractive to the population as a whole. If there weren’t any open criticism of new atheism, the public might draw the conclusion that all atheists share the attitudes of new atheists, and that clearly is not the case. That perception wouldn’t be helpful to atheists, as a group, because I think lots of people do find new atheism too combative, and not just because Chris Mooney told them to.
Of course, when Mooney says that new atheism is harmful for science education he is talking only on terms of what happens in the US. And even then he is long on claims and short on evidence. I am surprised Stangroom, who once co-authored a book entitled “Why Truth Matters” might have had more to say about that, but clearly truth only matters sometimes.
Why does only what (not) happening in the US matter ?
Kazez:
“If there weren’t any open criticism of new atheism, the public might draw the conclusion that all atheists share the attitudes of new atheists, and that clearly is not the case.”
I agree 100%. In the real world, there are never unanimous monoliths, and one shoud be very careful with terms like “we” and “they”. Therefore I wondered why you used the plural third person when describing Mooney’s assumed motives (see below).
“That means not confronting evangelicals about Jesus (or whatever), if _we’re_ trying to get them to support teaching evolution.”
Even if every atheist would refrain from such confrontation, getting a wide support among evangelists for teaching of evolution would be very inprobable. Thus it is quite futile to demand some kind of team spirit in this matter.
Why not just settle for the normal free society where it is normal for individuals to confront ideologies that he/she does not approve? Accommodationists confront incompatibilists, theists confront atheists. And vice versa, of course.
Exactly right, Matt. Well done.
Is there any rigorous study on the effectiveness of different modes of communication. The “don’t be confrontational” itself should not be a dogma. After all, *anyone* could rule anything whatsoever off limits for debate simply by saying “don’t confront me about that” and I think that’s a mistake, if truth is the goal. If social cohesion is, long term, I think is too, but that’s another story.
Keith
There’s a vast amount of research on this sort of thing (though it doesn’t tend to get cited in these “debates”).
It’s a complicated picture. Neither Chris Mooney’s position, nor the standard new atheist position, is straightforwardly supported.
There’s a good summary of the research in Elliot Aronson’s fabulous book The Social Animal. See the chapter, “Mass Communication, Propaganda, and Persuasion”.
@Keith:
“Is there any rigorous study on the effectiveness of different modes of communication?”
While I’m not aware of any studies that address the role of communication in the specific science-religion context being debated here, there is an entire body of observational and experimental evidence from the field of cultural cognition showing that humans are more receptive to new or conflicting ideas, including scientific information, if that information is presented within their existing cultural identity rather than in a way that conflicts with it. There is an entirely different psychological study by Gal and Rucker (2010) that found that humans are less likely to cling more strongly to incorrect beliefs if the person introducing doubt about those beliefs exhibits open-mindedness. Then as Jeremy said, there are studies (some from the advertising and marketing fields) that suggest the opposite.
I have never actually heard anyone flatly argue against being confrontational. The message I get instead is that perhaps there is more than one way to communicate the same truthful fact, and we should be considering how those different ways may be interpreted by the receiver before we choose one to use. To some I suppose that argument could lead to “don’t be confrontational,” but I don’t think it necessarily has to follow.
There is also the old trope that the above message somehow constitutes telling atheists they shouldn’t be allowed to express their identity, but I don’t think that follows, either.
But it really is complicated; and it intersects with other social and psychological dynamics (particularly cognitive dissonance). So, for example, the greater the discrepancy between the message of the “communicator” and the views of the audience, the greater the cognitive dissonance (since people tend to want to believe they’re rational, right, etc). The question then is how does this fall out: it could be that it makes it more likely that people will change their views or more likely that they will derogate the communicator.
The answer seems to be that that it doesn’t fall out in a straightforward way. So, again for example, a crucial variable here is the credibility of the communicator. The more credible they are, the more likely you’ll get a big shift in people’s opinions. But, of course, it’s not clear how this maps onto the whole atheism debate – for all sorts of reasons.
Plus there are other complications to do with how entrenched the beliefs are, etc.
It really is complicated.
@Jeremy:
I can agree with the view that things are incredibly complicated. The difficult reality is most likely that accommodationism and confrontationalism (forgive me if those aren’t appropriate terms for the two “sides”) are not consistently right-wrong positions in every situation. Perhaps there are really no true “sides,” and the complications of the matter make the correct strategy vary situation to situation.
I just think that the debate has evolved to where it’s polarized to the point where that kind of situational consideration isn’t even a possibility. It seems as though one side must be right and the other side detrimental to science in order for either side to consider anything as progress, and I don’t think that’s the reality of the problem. I suppose the true issue, then, is how to move towards taking the complications seriously into account?
Hammill
Actually, it’s even more complicated than the question of whether accommodationism and confrontationalism are factually supported.
There is also a whole ethical dimension to this debate that (again) just tends to get lost; and it’s an ethical dimension that (partly) depends on how one interprets further empirical data. So, for example, there’s some (a lot of?) data that suggests that religious people are happier on average than non-religious people (Jean Kazez has referenced this on her blog, I think). So if you’re committed to the project of eradicating religion (and clearly that isn’t the position of all the new atheists), or even if you’re not, you’re just a vocal atheist, then there is a strong argument that you ought to take this into account (which is not to say that the data is necessarily reliable or that it means that the eradication of religion would be morally problematic even if it were reliable – but the data is there and it’s certainly arguable that it ought to be factored into a moral calculus about what one is doing when one seeks to affect how other people think about religion).
Indeed, if it turns out that confrontationalism is more effective at undermining religion than accommodationism, then it is possible that for precisely this reason one might favour the accommodationist strategy (the thought here being that this might be the optimal way of achieving particular ends – for example, ensuring that evolution isn’t undermined in public education, while preserving the “good stuff” that is associated with religion).
There are a lot of other issues, complications, nuances, etc., that ought to occur to people as soon as they begin to think about this stuff, but which don’t because the whole debate became poisoned very soon after Chris Mooney’s book was published. I have no time for any of it. (But I appreciated your contribution here, so wanted to say something. Especially since Ben seems to have disappeared!)
I made the mistake of reading B&W, and I need to set the record straight.
(1) Ophelia Benson is not banned at my blog, and never has been.
(2) Anyone can see that my comments above are consistently about priorities–one of the topics of Ben’s post–apart from exactly one sentence.
(3) That one sentence about disagreeing vs. despising led me to no further discussion of that topic (though people at B&W seem to find it very interesting!), after Ben objected. Lots of people despise Mooney, so even that wasn’t about Ophelia.
(4) My reference to “Why Truth Matters” was entirely positive, and can’t be considered taboo at a blog owned by one of its authors.
I would say “unbelievable” but at this point I’ll believe anything. See Hamill’s comment above. All true.
People may rightly condemn the ad hominem attacks against Mooney above as being far beyond the boundaries of fair critique. But people may also rightly condemn Mooney and Kirshenbaum for painting Jerry Coyne’s article in the New Republic as “uncivil”.
To explain why that is, we have to make a distinction that doesn’t get made very often: between tactics and strategy. Roughly, “tactics” are ways of resolving a particular conflict; a tactic can be easily phrased in the form of a command (e.g., “fight”, “flight”, “negotiate”, “die”, etc.) Meanwhile, “strategy” involves reflection on the use of tactics in all the conflicts that are ahead, and some idea of what tactics are appropriate to what situations.
I think any positions that conflate tactics with strategy are wrongheaded. Such positions are remarkable for having a one-size-fits-all policy: e.g., they recommend that everybody must be a “diplomat”, or everybody must be a “soldier”. Luckily, contrary to common belief, very few people have a one-size-fits-all policy. Many discussions are either predicated upon an implicit conception of what an effective strategy looks like, or they outright discuss that kind of strategy.
How can you tell if people are making a distinction between tactics and strategy? One surefire way of figuring out if a person has a considered strategy is by looking to see if they support a diversity of tactics; if they do, then that indicates that they have some consideration of the role that different tactics have in the grand scheme of things. And sure enough, you’ll find that both PZ Myers and Phil Plait do support a diversity of tactics, even though they have developed more dramatic brand names. Also, to use more home-grown examples, I think I make that distinction when I proposed the “good cop/bad cop strategy”, and I think Jean and Matti also recognize that distinction with their interesting counter-proposal.
How can you tell if people are conflating tactics with strategy? One surefire way of figuring it out is if they make no practical distinction between tactlessness and criticism. For, the way I see it, the whole idea of having a strategy is to have a coordinated plan of action — a grand scheme, so to speak — and in principle that this grand scheme is open to rational review by like-minded people. But if an action X is deemed tactless, and the basis for saying it is tactless is mystifying or inarticulate, then that implies that the grand scheme may not in principle be open to review. Hence, by all appearances, it will seem as if there is no grand scheme apart from “the authorities say so”. Hence, this would not be a strategy, but only a tactical command. Hence, by all appearances, anyone who fails in their words and deeds to make a distinction between tactlessness and criticism, is only operating at the level of tactics.
And in my judgment, Mooney and Kirshenbaum have conflated strategy with tactics. For one thing, their response to Jerry Coyne’s article in The New Republic was mystifying, precisely because there didn’t seem to be much of a distinction between the civil and the reasonable. Also, it’s just a matter of fact that their discussions in their latest monograph have not made any appeal to the wealth of utterly fascinating social psychological data that Hammill is curious about, and which Jeremy rightly cites above. Instead, they refer to “common sense”; and since I do not believe that anyone who works outside of the Pew Centre are entitled to that phrase, I find their invocation of it to be unpersuasive at best.
(Also, for the purposes of full disclosure: I’ve been banned from the Intersection, along with many others, so the reader can take my verdict with as many grains of salt as are needed to fit their tastes.)
But, Ben, forgetting about Mooney for a moment, that doesn’t follow.
It’s entirely possible that you do have some grand scheme x, and that you also condemn some action y as being tactless for no good reason. Because people are fallible. They don’t necessarily respond in a way that is consistent in terms of their own framework.
It only implies that the grand scheme is not open to review if this sort of thing happens over and over again. Even then, a grand scheme might in principle be open to review, it’s just that the person who is articulating it isn’t inclined to review it themselves (but so what? – people are fallible).
“…it’s just a matter of fact that their discussions in their latest monograph have not made any appeal to the wealth of utterly fascinating social psychological data that Hammill is curious about, and which Jeremy rightly cites above.”
To be fair, if the monograph you’re referring to is Unscientific America, a good amount of the data I referenced was not produced until after the publication of that book. The latter study that I referenced, for example, was not published until 2010.
However, I do see your point. There were indeed data that existed at the time of publication that were not included but could have been remarkably fascinating to delve into. I think they still could be!
Coincidentally, Mooney tweeted about cognitive dissonance today, so maybe he is looking at some of the social psychology stuff…
Jeremy, if a person’s verdicts are confused even by the light of their own private schemes, then their conduct will just seem like wild muddle to an outsider. That’s the sense in which their conduct will be inaccessible.
It seems to me that one precondition for communicating something meaningful is that I should have at least some private idea about whether I’m coming or going, and that I act in a way that is consistent with that.
Suppose a man says, “Don’t eat the sandwiches!” while also stuffing sandwiches in my mouth. If I’m in that position, I’d have to ask myself: what could this madman possibly want? Does he not know the meaning of the words he’s speaking? Has he lost control of his arms? Is he just trying to confuse me? Etc.
Hammill, yeah I had UA in mind. And, true, some studies are new, I meant the stuff that was available at the time.
Sorry, I don’t think you’re right at all here.
1. You wouldn’t necessarily think he was a madman: you might well think he’s not able to practice what he preaches (which is not particularly unusual);
2. Even if it were true that he was confused, it’s entirely possible to assess his actions in terms of his own *declared* strategy. There is no problem here. You may doubt he really means it. But so what? Even if Mooney doesn’t really mean what he says, even if he doesn’t practice what he preaches, I can still assess the logic of “accommodationism” (i.e., his stated grand strategy), whether it is empirically warranted, etc.
3. And even if Mooney (for example), doesn’t actually mean anything, even if he’s so confused that it makes no sense to think he’s communicating, it is still entirely possible to assess, discuss, examine the ideas that have now become known as “accommodationism”, and which had their genesis in his non-communication.
And so on, and so forth. There’s a lot more I could say here!
This whole strategy/tactics thing is… well it’s dodgy. You say:
Well yeah, but it doesn’t follow that a strategy can’t have strict tactical constraints (e.g., not being aggressive towards religious people).
So, for example, if I play squash against somebody who is more skilled than I am, then I try to wear them down physically over the five games in order to win. This strategy has strict tactical constraints. You don’t go for quick kills. You don’t hit the ball too hard. You play to a length. You make your opponent run diagonals. ETC.
Sure, there is room for some variation within those tactics, but actually surprisingly little. You simply cannot play aggressively for any significant part of the game if you’re using this strategy. If you do, then, de facto, it’s a different strategy.
So it just isn’t true that a tactical constraint (e.g., no aggression) can’t be part of a grand strategy. It can be.
(Note also that if I vary from my declared strategy while playing squash, nobody will think that I’m mad, nobody will simply assume that actually I didn’t mean it when I said I’d play an attritional game, they’ll just think that in the heat of the game I’ve lost sight of my strategy. We’re all fallible, right.)
Jeremy,
You’re surely right about (1) to an extent. Confident judgments of hypocricy are possible if there are on-balance grounds for asserting a person is being hypocritical. But my so-called “in principle” point has to do with situations where we have almost no information to go on. For example, if the Sandwich Man is a complete stranger, and I wake up in a cell with him, and the drama begins… well, I doubt I would know how to explain the whole speech act.
Still: (2) it’s true that you can assess Sandwich Man in terms of how you understand the locutions being used. And it’s true that you can say: these words, as I understand them, don’t fit with this person’s deeds. That’s totally legitimate — no question. I hesitated in this case because I’m not interested in accusing Mooney of hypocricy. It comes close enough to a personal slight that I don’t think it will appeal to many hearts and minds. So I leave that to others, who are more bold than I.
(3) True; some (like Josh Rosenau) have been fairly articulate accommodationists. So the conversation doesn’t have to die just because we hit a dry spell with Mooney/Kirshenbaum. But I have tried to commend them for helping to kickstart the conversation, not necessarily for keeping it alive.
On strategy/tactics. You’re quite right that a strategy can have tactical constraints — in fact, it should have some constraints. A strategy without any tactics is like American analytic epistemology: it sounds impressive, even though it isn’t. But I still have to know, what’s the strategy? Plait has a vague strategy with a tactical constraint (e.g., “Be passionate and reasonable, but don’t be a dick”); I have a tentative strategy that has tactical constraints (good cop/bad cop, but above all don’t be an unreasonable asshole); Jean and Matti have a strategy and probably have some tactical constraints that will come out in the fullness of time.
My point about tactics and strategy is, rather, that if your only strategy is phrased as a tactic, and nobody knows what to do in a variety of interesting and difficult situations, then it’s probably not much of a strategy at all. And that’s where I am when it comes to Mooney and Kirshenbaum. When it comes to matters of civility and critique, their principles are vague, they apply those principles in an unreasonable and inconsistent way, and they’re unresponsive to criticisms.
Of course, that’s not to suggest that we know nothing about their priorities and goals. Actually, there are a lot of areas where I do, and can, interpret them charitably; I tried to defend them against a few above. But when it comes to civility and reason, this is one area where I remain mystified. At best, I’m not sure there’s really any strategy there. And at worst, if there is a strategy, then I suspect that it’s a strategy that is founded in “common sense”, not science.
But, of course, it’s possible to make that “accusation” without using a word such as “hypocrisy”. So, for example, one might say something to the effect that, like everybody, Mooney isn’t always able to live up to his own standards, which is far less inflammatory (but also accurate).
(I’m not making any substantive claim here, by the way. I don’t know Mooney at all (I’ve never exchanged a single word with him); and I haven’t followed his stuff with anything like forensic attention.)
Sure, but that’s a different issue. Your original point was that absence of strategy can be detected in the conflation of strategy with tactics. My point is that this doesn’t *necessarily* follow. The fact that there is no strategy in the case of Mooney (if that’s true), doesn’t negate my point (I won’t insult you by explaining why!).
Hmmm. Even that is… well I guess in most situations it’s a fair point, but… not in all situations.
So, for example, if I know somebody has a weak backhand, my strategy might be to pressure it until it breaks down. That isn’t really a tactic, since within the context of a particular rally, or game, I might choose different methods in order to bring this about, but it’s also very stark, simple, etc.
But I’m not sure we disagree substantively here, so…
Well, I’m pretty certain it’s not founded in science. But, as you’ll accept, that doesn’t necessarily mean it won’t turn out that actually the science does support some version of accommodationism. (Not that this is necessarily the end of the matter, for reasons I talked about above.)
There are other points to make here, but I think I’m probably done for now.
On most of these matters, we seem to agree in most relevant respects.
Still, I may have been too bold in my remarks on strategy and necessity, and if they’re fallacious then I wouldn’t want them to distract from the more important practical issues.
Re: empirical evidence. Mooney’s last couple of Point of Inquiries have been quite relevant to the question how to communicate about science/religion. First there was Brendan Nyhan–the guy Mooney was citing in the CfI confrontation/accommodation debate (the one who found that “corrections” tend to reinforce wrong views instead of changing them). Most recently he talked to Barry Kosmin, who has studied the “nones” and found some very surprising things about them.
As to Mooney’s criticism of Coyne, and how “mystifying” it is/isn’t…. I probably never would have made that criticism, if I’d come to Coyne’s NR article on my own, but I don’t find Mooney’s different reaction mystifying. What he would have wanted, probably, is a review that emphasizes the positive, and downplays the negative. Is it really mystifying to want that?
Here’s the analogy that comes to my mind. “Animal Liberation” author Peter Singer wrote a glowing review on the back of Matthew Scully’s book “Dominion”, despite the fact that Scully is a conservative (he’s the speech writer for Sarah Palin and George Bush II), and Singer’s a liberal; Scully makes Catholic natural law arguments on behalf of animals and Singer makes utilitarian arguments; Scully appeals to religion and Singer is an atheist. Singer deliberately overlooks all those differences even though he’s on record as thinking some of them generally do hinder good thinking about animals. It’s quite obvious he’s so kind to Scully because his book is fantastic in one respect–he’s a great writer about cruelty to animals. Plus, Scully’s whole ideology makes him the ideal person to reach the conservative, religious audience that tends to be least sympathetic to the animal cause.
I think Mooney would have Coyne treat Collins exactly as Singer treats Scully, and for analogous reasons. Maybe that’s the wrong thing to want from Coyne, but why is it “mystifying” that Mooney should want it? I really don’t see anything very puzzling about that.
Mooney blogged about the “nones” stuff today.
Actually, the recent secularisation “debate” that occurred on the blogosphere was another occasion where people’s ignorance of previous (sociological) research was quite stunning.
Most of the protagonists just seemed absolutely oblivious of the fact that sociologists have been talking about these sorts of issues in an incredibly nuanced way for about fifty years now (actually more than that if one wants to count Weber’s stuff on desacrilisation).
The work of people such as Martin, Wilson, Bellah, Herberg, Berger, Luckmann, etc, etc, is all directly relevant, yet I suspect that most of the people pontificating on these matters have never even heard of these people, let alone read their research.
Jean, because there’s a distinction between grace and civility. Singer’s behavior was graceful; Coyne’s behavior was civil. It’s not just a verbal slip, since the former’s supererogatory, and the latter’s due diligence.
Jeremy, IIRC, I would think that anyone who has read “Why Truth Matters” must have at least a passing acquaintance with the names and profiles of Berger and Luckmann. (And H.M. Collins, etc.)
Kazez:
“Maybe that’s the wrong thing to want from Coyne, but why is it “mystifying” that Mooney should want it? I really don’t see anything very puzzling about that.”
There is nothing puzzling about Mooney’s dislike of Gnu atheists speaking out publicly. What is puzzling is that a journalist in a free society does not see the futility of complaining about the _publication_ of other articles. Generally, journalists express their dislike of articles by writing their own articles with counterarguments.
Let’s say some other journalist would advice Mooney to tone down his rhetoric about AGW in order to boost some short-run common good, say the general trust in the economy. Do you think Mooney would regard such advice reasonable?
Matti – Addressing somebody by their surname alone when talking to them directly is bordering on the uncivil. If you want to continue to post here, you won’t do it again.
IMO, the biggest difference between Mooneyesque accommodationism and New Atheism is not one of goals, or even priorities, but of overall strategy.
Mooney is a pessimist about major social change with respect to science and religion. He’s an atheist but doesn’t seriously entertain the idea that atheism could become mainstream and have beneficial effects by undermining the authority of religion. He therefore is more worried about backlash against atheism and science than about the detrimental effects of religion. Half a loaf is all we’re going to get, and only if we’re lucky and very careful.
The “New” Atheists are more optimistic. They look at the history of social change in the U.S. and in more secular countries and think that the U.S. could be more like Britain (or better, Sweden)—much less overwhelmingly religious, and perhaps even mostly secular.
Jeremy is right that there’s data out there relevant to these issues, but I personally don’t believe that Mooney is really interested in what the relevant science really says—or at least, he’s not interested in telling other people about the data that don’t support his thesis.
I think Mooney’s pretty unscientific in how he talks about these things, cherry picking his data and refusing to address the main counterarguments and the most relevant data that might undermine his theses.
Mooney paints a dismal picture of how people form, maintain, and entrench their beliefs based on simple small-scale psychological research about individuals in laboratory settings. He draws certain plausible lessons from those things, but ignores larger-scale and longer-term data that clearly indicate things are not nearly that simple, and often go the other way.
He also ignores arguments that undermine his arguments, such as Overton’s arguments about shifting the “Overton window” of the “acceptable” range of public opinions.
Mooney’s preferred strategies are often unsuccessful in political practice.
Conversely, many of the most important changes in public opinion in the last few decades undermine Mooney’s simplistic theses.
IMO, Mooney constantly makes an “intuition pump” argument that amounts to little more than pumping up the commonsense idea that “you catch more flies with honey than with vinegar,” and supporting it with cherry-picked data, while ignoring more relevant sociological data.
The most dramatic changes in public opinion in the last 50 years seem to refute Mooney:
1. The rise of the religious right, using tactics that were far from accommodating, and often extremely confrontational. If Mooney is right, how do you explain Rush Limbaugh’s and Glen Beck’s incredible success, and the fact that “liberal” is now a dirty word? Lee Atwater and Karl Rove were not successful in reshaping US political discourse by using moderate, accommodating ideas presented in honeyed tones.
2. The success of the gay rights movement, despite strong and specifically moral opposition grounded in religion. If Mooney were simply right, the gay rights activism of the last 30 years couldn’t have shifted public support of gay rights from a very small minority to an actual majority. The gay community was told for decades by Democratic strategists that their time had not come, but they gave up on knowing their place, came out, and gained substantial acceptance
(Astonishingly, most of that happened in a single generation, and despite the greatly increased clout of the religious right.)
3. Recent data on secularization in Britain shows that many sociologists and political commentators were wrong to thing that the secularization trend of the 70s, 80′s and 90′s was a short-term blip. Many assumed that as the secularized people aged and raised families, they would become more religious. Some did, but most didn’t. That appears to be true of most of Western Europe, making the U.S. an outlier in terms of how it has failed to secularize—until recently, that is. Recent US data suggest that the US may finally be secularizing as Western Europe did in the latter 20th century. (The current youth generation is both more secular and more accepting of irreligion than any previous generation; if the US follows the European pattern, this will be durable change, not a blip, and the trend will continue.)
4. Sweden shows that fairly rapid secularization can go very far indeed, and how it can be done in at least some places. Just a few generations ago, Sweden was overwhelmingly Christian; now it’s predominantly atheistic. This didn’t happen by exclusively accommodationist, incremental shifts. The most influential figure was probably Hidenius, who was far from accommodationist—criticizing religion in much the same way that the New Atheists do. If Mooney is right about the New Atheists, how do we explain Hidenius and the shifts in Swedish opinion both away from religion and toward science.
Mooney has stonewalled for years about Overton Window strategy vs. an accommodationist “line” with extreme fears of backlash. Mooney appears to be polically naive, at least as far as his public explanations of strategy go.
(I doubt he’s actually that naive, but he apparently doesn’t want to even acknowledge the existence of alternative strategies. He prefers to cast the New Atheists as failing to think strategically, rather than having good reasons for differing with him on strategy.)
The point has been raised dozens of times by other bloggers, and scores f times on Mooney’s own blog, but so far as I know, he’s never actually addressed it.
That’s one reason why a lot of us eventually stopped taking Mooney seriously, except as an annoyance. He never takes the main arguments seriously, and frequently misrepresents them. (E.g., claiming to be “mystified” as to how anybody could disagree with his simple and obvious arguments, when he knows very well that there are counterarguments he is avoiding at all cost.)
@Paul:
I agree that the Overton Window is something to be considered in line with data from social-psychological and other studies, but I’m confused after reading your several points regarding what refutes accommodationism. Perhaps my understanding of what constitutes accommodationism is incorrect. For example, if increasing trends in secularization refute accommodationism per your points in 3 and 4, does that imply that accommodationism is inherently against secularization? Or are you stating that “New Atheism” is the cause of secularization, and thus accommodationism is flawed because it assumes that “New Atheism” could not lead to increased secularization? I would assume the latter is what you’re arguing, but is general secularization even the goal of interest in this discussion, or is it more narrowly defined to science literacy? Do all accommodationists argue that “New Atheism” is harmful? And are there comprehensive data that explicitly show that aggressive atheism is the catalyst for secularization in your cases over other factors? It seems we’ve established the paucity of data supporting either stance in this thread, so those data would be interesting to see.
Regarding Limbaugh/Beck: I believe the most important question to ask is not are they simply popular or successful, but have they been successful at swaying those in the middle or elsewhere to their cause? There should be a distinction between being popular within a pre-established audience and being popular by increasing the size of one’s audience, i.e. is “liberal” a more universally dirty word now, or is it just more intensely dirty within those who have always leaned in that direction? I suppose one could deem Limbaugh/Beck a success in some respects (book sales, speaking engagements, etc.), but if looking at the 2006 midterms (pre-Beck but with Limbaugh) and at the 2008 election (Beck and Limbaugh), their position suffered severe political setbacks. There has not been a consistent rise of the Right during their tenure. In fact, Jon Stewart’s Beck-esque rally attracted far beyond the number of attendees of Beck’s, and Stewart takes a much more “accommodationist” political tack, so what, if anything, does that state?
I still agree with Jeremy that the issue is much more complicated than it seems on the surface, if not based on anything more than the conflicting data that exist. However, you are correct that these issues aren’t being rigorously addressed, I don’t think by either side, and they should be for the debate to move forward.
@Paul (@Hammil)
Yeah, there isn’t anything in those 4 points that refutes accommodationism. Nor, for that matter, is there anything that supports it.
You can’t just flag up certain – apparent – social trends, and then argue to cause and effect, or draw straightforward conclusions about the nature of social change (if that’s what you’re doing – it’s not clear).
It’s also curious that people who think that it is plausible to argue that vocal atheism might be the cause or catalyst for large scale social change (i.e., large changes in beliefs and behaviour) seem to believe they’ve solved the issue of whether social being determines consciousness or consciousness determines social being.
This is quite a feat since it was being endlessly debated when I was doing sociology proper.
I say again – this stuff is much more complicated than most of the people involved in this debate seem to realise.
Jeremy & Hammill,
I entirely agree that this stuff is much more complicated than many people involved seem to realize. My main point is that that applies to Mooney in spades.
I’m not trying, here and now anyway, to make a strong case for “New” atheism or against accommodationism very generally.
I’m specifically criticizing what I call Mooney-style accommodationism, which includes a major component of New Atheist-bashing, and does it based on very simplistic arguments.
While I do myself interpret the phenomena I cited as undermining very strong forms of accommodationism (like Mooney’s), I think they at least suggest that Mooney’s analyses, his “evidence,” and especially his framing are grossly inadequate for supporting his position.
Mooney presents himself as an expert on communication and on the politics of science and religion. He also criticizes New Atheists as both philosophically naive and strategically naive or irresponsible. He condescends like crazy, and stonewalls about the actual bones of contention.
If he understands Overton Windows or their importance in New Atheist strategic rationales at all, he doesn’t let on. For literally years and years, he has stonewalled on this crucial point and argued against a caricature of the New Atheists, feigning mystification as to how those nasty bullheaded irresponsible uncivil meanies could fail to agree with his oh-so-reasonable positions and strategies.
In fairness, I think some New Atheists (notably PZ Myers) have said some pithy and simplistic things that support that caricature, but if you actually read their more detailed explanations, they’re not as naively idealistic (or irresponsible) as Mooney makes them out to be. They have actual reasons for disagreeing with Mooney on strategy, and IMHO their arguments are actually better.
A lot of the argumentation on both sides is simplistic, incomplete, and unbalanced, but IMHO Mooney is the worst of the lot, and his condescension about both philosophy and strategy is insufferably annoying.
Mooney often “frames” things in the worst sense, in a several ways. (I think framing is important, myself, but I think Mooney typifies its dark side, where it degenerates into deceptive spin.)
Mooney makes unsubstantiated or grossly simplistic assertions that appeal to ignorant onlookers’ naive schemas, e.g., that you catch more flies with honey, that the US is a very religious country, that people find important meaning in religion, and if you can’t beat them, join them.
Explicitly or implicitly, he frames the out atheist cause as hopeless and/or almost purely counterproductive, generally ignoring serious arguments that in the long run you can beat them, that the US doesn’t have to be a very religious country, and that weakening the mindshare of religion may in fact be important to the overall acceptance and valuing of science.
It may be that those arguments are ultimately largely wrong, our outweighed by more refined and correct versions Mooney’s arguments, but we’ll certainly never find that out by listening to Chris Mooney.
As I said in the earlier comment, the New Atheists have good reasons for being more optimistic about these things than Mooney, and even for thinking that their strategy is consistent with Mooney’s goals in the long run, while his own short-run tactics are likely self-defeating in the long run.
He presents himself as the wise strategist, and the New Atheists as naive and irresponsible, when in fact they’re playing a longer strategic game, with more similar goals and better rationales than he admits.
As for the evidence I cited, I think that the pattern of secularization in Western Europe suggests that the US could follow a similar trajectory, however belatedly, and that it may already be happening. Maybe the time is right for the New Atheism.
I fully acknowledge that that could be entirely wrong, and I certainly don’t claim that the New Atheism is responsible for the recent US trend—I think it’s mostly not. (I think that the internet and other features of pop culture have more to do with it, and that it started before the “New” atheism became New.)
I can interpret those data as supporting either accommodationism or atheist activism, or both, but I can’t see how they’re consistent with Mooney’s conveniently pessimistic framing of the US as a very religious country, where people find meaning in religion, and implicitly as one that must always be that way such that we can’t beat them and should instead join them.
He carefully neglects to mention the interesting facts that (1) other similarly religious countries have secularized a lot in a single lifetime, and significantly in a single generation, (2) people in those countries generally do not report that their lives are meaningless—in the long term, they do find other ways to make their lives meaningful, (3) that may have useful causal relationships to polical phenomena, and (4) it’s arguably happening right here already.
I agree that these issues are complicated—more complicated than I’m making them here—but Mooney, the alleged expert whose advice we should follow, doesn’t even scratch the surface. He skates right past the real issues, systematically, for years on end. He apparently can’t substantiate his claims, or even demonstrate that he understands the basic issues.
Perhaps a really good analysis of the most relevant sociological data would ultimately support Mooney’s position. Perhaps the US is different from Europe or Japan in such crucial ways that secularization like theirs can’t happen here, or that an exclusively accommodationist strategy would be more effective, or that such secularization wouldn’t ultimately benefit science.
I don’t currently think so, myself, but I’m no expert and I’d be very interested in serious arguments about it. That’s exactly what I expected from Mooney years ago, but he has systematically avoided providing.
Instead, Mooney keeps beating the same two major straw men—that the New Atheists are philosophically naive in certain crucially relevant ways that they’re actually not, and that the New Atheists don’t have a reasoned long-term strategy or evidence for its likely effectiveness.
You never hear Mooney discussing Overton Windows and explaining why his “moderate” strategy and suppressing “extreme” positions is actually better than a diversity of positions and tactics tactics as Overton would recommend. (That’s especially odd given things Mooney does say about the fragmentation of mainstream media and self-selecting of messages; that has a lot to do with Overton Windows.) He won’t even mention Overton, despite being asked the relevant questions frequently for literally years on end. (His stonewalling about it has been pointed out with increasing frequency for a year or two as well.)
You also never hear Mooney admitting that the basic philosophical positions and arguments of the New Atheists are very common among professional philosophers, and his own views are not. Instead, he quote-mines a few convenient philosophers such as Rob Pennock or Barbara Forrest for simplistic broad statements, and makes it sound like the New Atheists are naively and illegitimately crossing well-known and generally agreed boundaries between science, philosophy, and theology.
(For example, he misleadingly quotes Forrest about methodological naturalism, to the effect that science can’t study the supernatural, but he ignores Forrest’s own paper that argues exactly the reverse—that the continuing success of naturalistic explanation and failure to verify supernatural claims does in fact scientifically support philosophical naturalism, just as most of the New Atheists argue.)
Either Mooney is way over his head both philosophically and strategically, or he refuses to divulge his actual sophisticated reasoning to his audience. I suspect a bit of both—his reasoning is somewhat simplistic, but he crafts an even more simplistic message for public consumption, talking past his critics to his naive target market.
Given that he’s been doing this for literally years when at least some of his critics have been arguing quite seriously and earnestly about real and important issues, I think it’s past time for Mooney to put up or shut up. He should seriously discuss Overton windows, and why so many philosophers are basically in agreement with the New Atheists, or be recognized as a simplistic, deceptive propagandist with nothing substantive to offer.
By the way, since I accused Chris of quote-mining, I should point out that I don’t claim that was intentional.
What I take as a quote mine of Pennock may have been an innocent mistake, which Pennock set him up for. When I looked up the quote, it seemed in its immediate context to be as broad a statement as Mooney made it out to be, and I accused Pennock of being wrong. Only later did I go back and see what I interpreted as an important caveat several pages earlier.
What I take to be Chris’s quote-mining of Forrest may have been innocent as well, originally.
I do think it’s significant that Mooney has never acknowledged that Forrest doesn’t agree with the point he was quoting her to support, after bashing the philosophically uppity New Atheists on the basis of her superior philosophical expertise.
Paul W. Have you listened to Mooney’s latest PoI about the Nones in the US, with Barry Kosmin? That interview makes it pretty clear that Mooney is interested in the empirical evidence about how and why religion in the US is decreasing. Kosmin gives no credit to vocal atheism, and all the credit to a surprising assortment of other factors. That research wasn’t covered in Mooney’s 2009 book, as far as I know, but you can’t complain now that Mooney isn’t interested in empirical research about the impact of vocal atheism (or the lack thereof).
Jean,
I’ve been meaning to listen to that podcast for a few days now. I’m very curious about Kosmin’s interpretations.
(I hope they’re not like Elaine Ecklund’s interpretations of her data, which I consider utterly bogus—her own data undermine her simplistic theses, and seem to support quite different conclusions.)
I may get to that podcast today; thanks for reminding me.
As for Mooney’s interest in empirical research, I’m skeptical that he’s very interested in evidence that undermine’s his theses.
His favorable treatment of Ecklund (at least initially) really put me off. I felt he was approving of her invalid spinning of her own data, either out of ignorance or just because he wasn’t really concerned with the truth. Either way, that’s a bad thing, and he didn’t seem interested in several clear and pointed critiques of Ecklund (e.g., by Jason Rosenhouse and other bloggers, or by commenters on Mooney’s own blog, like me).
Several things like that make me distrust Mooney and any data he cites. (That doesn’t mean I’m not interested—just that I’m not at all trusting of what he reports.) He does not seem willing and able to interpret data validly, or to acknowledge mistakes when they’re pointed out.
Paul
I don’t in the least bit accept your characterization of the debate – I think it’s blindingly obvious that the reaction to Mooney’s ideas and work has been wildly disproportionate (laughably so, actually) – but four points:
a) I don’t think anybody here has argued that Mooney has made his case (or even that they are inclined towards accommodationism);
b) The secularization debate isn’t just about what the data is telling us about religion today (it’s also about what we might reasonably think the data from the past tells us about religion then); also the secularization debate isn’t just about data – it’s also a nuanced conceptual debate;
c) The vocal atheism vs. accommodationism debate (or however one wants to characterize these positions) is also (partly) an ethical issue (simply saying “It’s about truth” over and over again, doesn’t make it so – and it *is* philosophically naive to think that it is possible to dissolve an ethical debate simply by asserting a particular value preference);
d) It is sociologically naive to think that social change is just a matter of Overton Windows; or indeed, “strategies”, or “tactics”, etc. Or to put this another way, I’ve seen absolutely nothing from the new atheists – or indeed from Mooney’s side – to suggest that the explanation for any supposed secularization lies fundamentally – and I’m entirely aware that the word “fundamentally” disguises a multitude of complexities in this context – in the promulgation of ideas (whether via vocal atheist strategies or accommodationist strategies). In other words, it’s entirely possible – contra Mooney, but also contra some of the more foolish new atheist bloggers (e.g., Coyne) – that this whole debate is largely irrelevant (at least in the medium term).
I was recently digging around (unsuccesfully, so far) to find more about Barry Kosmin’s research because yes, there’s a concern that Mooney might pick the interviewees that help him support his own views, and then ask them the questions that give him what he wants. So I’d like to see the raw data, and find out what other people have to say about it. Yes, I did think Eklund was a bit of a spinmeister, and her data didn’t show quite what she wanted it to.
I think Paul is spot on, here. Paul admits that he would enjoy a serious discussion of empirical evidence, and that matters of the Overton Window are only part of the picture. I suspect that Paul would be delighted to hear about the conceptual issues, too — as we all would.
Part of the problem is that there aren’t any systematic, accessible books out there that go through the problems and possibilities of science communication in a concrete, scholarly, engaging way. Perhaps that should be Jeremy’s next monograph!
Well no, because I know virtually nothing about it (plus I’m hoping never to write another book!). But, as far as I can remember, I haven’t ever written anything substantive about the New Atheist/Mooney debate. Really, I’m much more interested in what it tells us about the psychology of group behaviour.
The conceptual issues surrounding the secularization debate will be covered in most introductory sociology of religion texts. Certainly it’s stuff that used to be taught at high school level in the UK (probably still is, but I’ve not looked for many a year).
The classic texts are by people such as Bryan Wilson, David Martin, Will Herberg, Robert Bellah, Peter Berger, Rodney Stark, etc, (obviously these guys deal in a mixture of empirical and conceptual analysis).
Sure. But I doubt everyone knows everything about the subjects. And I was thinking more in terms of science communication (or: knowledge transfer, knowledge mobilization, knowledge management). There are a few texts out there that focus on more provincial concerns, but what’s needed is someone with a background in social science who is willing and able to lay out the central problematics of the field, and some salient, reliable findings. And then write it in an accessible way.
(Though Piekle’s “Honest Broker” is probably a good start.)
Maybe you should write the book? If you fancy it – particularly after you’re done with your PhD – let me know.
You know I’m the series editor – with James Garvey – for Continuum’s contemporary social issues series? I’m not sure such a book would quite fit, but we’d certainly consider a proposal.
I didn’t know that. That would be fun! But we’ll see. My dissertation is probably not going to be in that direction, which means it would be a totally different project.
Waterloo is in the process of hiring a chair in science communication for our philosophy department. That would make this kind of research initiative easier.
Jean,
I listened to the Kosmin episode of PoI. More on that in a bit.
If you haven’t found Kosmin et al.’s report specifically about Nones, you can find it here:
http://www.americanreligionsurvey-aris.org/american_nones_the_profile_of_the_no_religion_population.html
You can find a link to the basic ARIS report in the sidebar.
I’ve had a look at it and found no big surprises so far. As I expected, some of what I took to be spin by Mooney and by Kosmin himself doesn’t seem to be borne out by the data presented.
Jean,
About the PoI episode… it wasn’t any worse than I expected, but I cringed a bit at a few things Mooney and Kosmin stressed but IMHO failed to explain the (in)significance of. Not that most of those things were literally untrue, or uninteresting, but it seemed to me that they expected certain inferences to be drawn that are not true.
In particular, they stressed that most nones are not atheists, and only a small minority identify as atheist.
To me it’s more significant that about half the nones actually are atheists in the broad sense of not believing in god, and even more don’t believe in a personal god—only about quarter do. (I wish it was clearer what “Deist” meant, though.) The large majority of Nones are not unchurched religious people, but actually irreligious people.
That’s entirely consistent with (but nothing like proof of) my current belief that this is an opportune time for atheist activism. I think a lot of people are atheists, but don’t call themselves that because they’ve bought anti-atheist propaganda and thing that “atheists” are dogmatically certain, have no good arguments against religious belief, and are a disproportionately obnoxious buttheads.
One big groaner for me was when Kosmin said that a lot of Nones dislike prominent vocal atheists. I suspect that’s true, for small values of “a lot,” but I’d really like to see data.
Rather than data, we got a bogus anecdote.
Kosmin said that those people said things about not liking the “atheist warlords,” which I think pretty obviously means Dawkins and his ilk. He even said that was “their words, not mine.”
Yeah, right. I would bet serious money that no more than a very few respondents said anything that vivid, maybe only one, but Kosmin pulled a Mooney and presented that exceptional anecdote as though it were a typical response, and he was just reporting it dispassionately.
If they actually got a statistically significant number of responses like that, I’ll bet it would have been in the formal report, not just the mass-market spin. I’m almost certain that Kosmin was editorializing way beyond the data, disclaiming doing so, and spinning things quite misleadingly. Ugh.
I’ll admit that some significant percentage of Nones may think explicit, vocal atheists tend to be annoying—I’m quite sure some do. I just don’t think that is evidence one way or another about Mooney’s agenda or the actual advisability of vocal atheism.
Maybe we especially need to reach the reachable people—irreligious people who don’t currently identify as atheists, but are less likely to respond very negatively than religious people, because they don’t have a big emotional stake in religion.
I could be wrong, of course, but I don’t think Mooney is doing much to clarify the issues.
Wait a sec–the statistic is that 2% of Americans are atheists, 15% are nones. For purposes of understanding the influence of “new atheism” it’s clearly a bad idea to lump together atheists and agnostics, and give the statistic of about half the nones being atheists.
What I found interesting and believable (after initial surprise) is that many people became nones because of (1) the scandal in the Catholic church, (2) overly powerful evangelicals during the Bush years, (3) anti-clerical attitudes, (4) general opposition to “organized” religion.
I thought it was awfully smart to distinguish “belief” from “belonging.” Being a none is an issue about belonging, not belief. That makes a lot of sense.
I was also struck by the fact that Kosmin said most of the nones are not terribly philosophical, and see the two extremes in the same light-they don’t like the evangelical extreme or the “new atheist” extreme.
As to the anecdote–well, yes, I’d like to see the data. It’s OK to report that someone referred to “atheist warlords” (it’s is an amusing phrase–why not quote it?), but you’d want to see how many people said stuff like that.
I certainly found it very interesting that the nones are 60/40 M/F, and also that the Freedom from Religion foundation is 79% male. So the further you get to the “new atheist” end the more male you get.
When I have time I’m going to follow your link. I certainly do want to know how robust all Kosmin’s data is–but especially wonder how closely his interpretive remarks match the data.
Quite so.
One reason why I’m peeved with Mooney is that I and a number of other people have raised these kinds of issues with him—including attempting to understand actual empirical data from ARIS, The World Values Survey, Ecklund’s actual data, etc.—and he has studiously ignored the issues we’ve raised, for years.
Mooney is obviously not interested in seriously arguing his position with serious people, using serious evidence. He is not even a little bit interested in clarifying difficult issues. That is just not what he does, ever.
Instead he repeats the same simplistic claptrap we’ve refuted many times, for an audience that doesn’t know any better.
Maybe it shouldn’t be surprising that he does this, given what he says about how answering opposing arguments generally entrenches the beliefs that you refute, even if you do it well. (Which I think is sometimes true, but often not.)
For scientists interested in data, or philosophers interested in serious argument, it’s can be extremely frustrating. Mooney’s strategy is not to play the game, and to pump the same old intuitions with the same simplistic arguments.
That’s good framing for certain purposes, and sometimes even necessary—e.g., in writing an op ed or a TV spot, or if given two minutes on Nightline, you generally don’t have enough space or time to address criticisms of your view. If you do, you won’t have time to make your own positive arguments. You often have to talk past your critics, or you’ll just get derailed by a Gish Gallop, and raise the salience of criticisms without successfully rebutting them.
Unfortunately, Mooney argues that same way in meta-discussions about politics and framing, even in books and on his blog, where he has enough space to actually make his case, and for an audience that’s interested in the details. He just won’t do it, ever. He’s always in simplistic propaganda mode.
Maybe that’s good strategy with the best of ultimate motives, and he’s ultimately right, and he’ll save the world doing things that way.
It’s still frustrating as hell for anybody who’s seriously interested in the things he discusses, and thinks some of what he says is wrong. You can’t just trust him, and you can’t get a straight answer to a straight question about his evidence and reasoning.
He’s lost a lot of former supporters and sympathizers that way.
Jean,
That seems like a bit of an understatement. (Perhaps you meant it that way.)
One message Ecklund has tried hard to push is that learning and doing science does not erode religious belief. She tries to attribute the low religiosity of scientists to their being disproportionately from irreligious homes.
Her own data show clearly that while irreligious household of origin is positively correlated with scientists’ later irreligiosity, that’s nowhere near strong enough to explain the huge correlation between science and irreligiosity.
Her data even show that most nonreligious scientists come from specifically Christian households, not irreligious households as she strongly suggests.
If she’s going to guess that such correlation implies causation by household of origin, the clear lesson is exactly the reverse of what she concludes: science does in fact erode religiosity, dramatically, in a single generation. About half of religiously-raised scientist become irreligious, and IIRC it’s pretty clear many others become less religious.
(I’d guess that if she gathered and reported more detailed data, it’d show that the overwhelming majority of religiously-raised scientists either lose their religion entirely or become significantly less intensely and/or orthodoxly religious—that science usually erodes religiosity substantially even when it doesn’t entirely destroy it.)
That’s consistent with a variety of other data, E.g., from the General Social Survey and Larson and Witham’s surveys of National Academy members.
Across the board, at all levels from high school dropouts to NAS and Royal Society members, learning and achievement in science is correlated with irreligiosity. It’s one of the strongest correlations in the social sciences, even stronger than the oft-noted correlations between science and being white and male and from the middle class.
That’s the elephant in the room, but it’s the opposite of what she wants to say, so she ignores it and says what she wants. Wow.
“Mooney is obviously not interested in seriously arguing his position with serious people, using serious evidence. He is not even a little bit interested in clarifying difficult issues. That is just not what he does, ever.”
I frankly just find that a weird assertion. At the CfI debate with Myers (etc) he presented more empirical evidence than Myers did. Since he started doing PoI he’s interviewed many scientists who study religion, reasoning, and communication–about 5 or 6. On the other side, I certainly don’t see any more seriousness about the science of communication.
Jean,
I didn’t like PZ’s talk at the CFI thing, either, and immediately criticized him for it.
Like Jeremy, I think the simplistic “it’s about truth” line is bad. Nobody on either side really thinks you can speak the whole truth, all the time, to everybody, in every situation.
(Not even PZ, who admits he’d be a lousy politician, and who can support politicians who he acknowledges can’t politically afford to say the things he does.)
I think PZ is usually better than that. He often does make mostly strategic, consequentialist arguments for a different strategy than Mooney’s. (Overton and similar arguments.)
As for Mooney’s citing of evidence, I think it’s generally so selective and superficially spun as to be worse than nothing. He never deals with the elephant in the room, which the NA’s (including PZ) have brought up over and over again for years.
For example, Mooney likes certain convenient little psych results that show that in certain conflict situations, discussion of the conflict tends to reinforce people’s existing views, even if those views are refuted. From that he apparently draws the convenient conclusion that being confrontational is generally bad. (Without necessarily saying so explicitly and generally—that would make it less plausible—but making it the “obvious” conclusion when sapplied to the “New” Atheists in particular.)
Mooney systematically ignores similar experimental data that seem to support the opposite view—including a bunch of very basic and very well-known social psychology results from the 1960′s and 1970′s about bracketing, conformity, and groupthink. They seem to support Overton reasoning: if you don’t voice the “extreme” views, the group tends to converge on a new center position, midway between the views that are voiced. The center thus shifts away from the people who self-censor their (perceived) “extreme” views.
IMHO, both kinds of things are obviously frequently true, but the details matter, and the whole subject is fraught with complexity, uncertainty, and risk.
(I take that to be part of what Jeremy said earlier about how maybe none of this actually matters—on average, there may be roughly equal and opposite reactions to anything any of the particular players do, and it’s hard to even tell when particular players’ strategy matters, much less which strategy is actually preferable.)
Scarily enough, Mooney and the NA’s could both have losing strategies; there might not even be a winning strategy in the particular case. Maybe 50 years from now, we’ll still have more creationists than evolutionists in the US, no matter what we choose to do.
Mooney doesn’t acknowledge this, but the NA’s generally do, and fairly often at least do so implicitly. (E.g., saying that accommodationism has been tried, and evidently hasn’t worked very well, judging by the steady popularity of creationism over the last several decades, and that they think it’s time to try something else.)
The individual psychology of belief fixation is complicated, and the social psychology is far more complicated. If things were as simple and one-sided as Mooney makes them out to be, politics would be simple, and that’s just false. There are a lot of two-edged swords flying around, for basic, deep reasons.
It has always been thus, and the answers are not obvious, and that’s one reason why movement politics always has this kind of conflict between more radical elements and more centrist elements.
(I find it very interesting how Mooney generally doesn’t discuss general principles of movement politics, with historical examples. Given how he presents himself as an expert on political rhetoric who can tell us how things work, how can he not? My own guess is that he doesn’t like any subject that’s too complicated for him to put a very simple one-sided spin on.)
IMHO neither side has been nearly as even-handed about this basic problem as I’d like, but Mooney is by far the worst.
I don’t think that either of the above kinds of “little” “scientific” data addresses the big picture very well—given the complexity and messiness of the subject, it’s too easy to support pretty much any strategy you want. To see how the rubber actually meets the road in large-scale opinion-making, you need large-scale evidence.
The New Atheists do try to supply that, to an extent, over and over again. (Although I quite agree with Jeremy (as I understand him) that neither side has actually made a compelling case. Their examples—like mine above—are more interesting and thought-provoking than conclusive.)
Like the New Atheists, I think the history of political movements and opinion shifts strongly suggest that Overton was substantially right, and that the kind of political triangulation Mooney advocates is frequently a losing strategy in the long run. (Note the qualifiers!)
While that point has not been proven and shown to apply in the particular cases, it merits a whole lot of serious discussion, but Mooney has always studiously ignored it.
I’m not saying that he can’t refute the NA’s Overton arguments, and show that centrist triangulation is likely a better strategy at the bottom line, in this case. I’d very much like to see him try, as I expected him to do years ago. That would be very interesting.
(Maybe he has some deep insights into the gay rights movement and the rise of the religious right, and the history of Western European secularization, etc., such that they actually support his view against the NAs’, and he can explain it all for us reasonably well.)
But Mooney prefers to act as though there are no serious counterarguments, and his opponents are philosophically and politically naive, unreasonable jerks, apparently with bad motives, because their behavior is otherwise “mystifying.”
Seriously, I think that by careful omission and lots of spin, Mooney has engaged in what amounts to a systematic smear campaign against the New Atheists, for years on end.
By straw-manning their strategic and philosophical arguments, he makes them out to be stupid and pigheaded and mean, and not just mistaken. (His obsession with superficial “civility” strongly reinforces this, especially given the way he cherry-picks and distorts anecdotal evidence. The Tom Johnson affair was just the dramatic tip of a fairl big iceberg.)
However “politely” he phrases it—he doesn’t use naughty words, of course—that’s just horrendously, substantively uncivil. IMHO, Mooney is the least civil major figure in the debate, on either side, which makes his pontificating about superficial “civility” extremely galling. (IMHO, he’s “civil” in the way James Carville is “civil.”)
The substantive ball has been in Mooney’s court for literally years, and never moves. (He won’t actually address either the basic philosophical or the basic strategic points the NA’s make.)
The evidence he adduces—e.g., that some irreligious people don’t like vocal “atheists,” or that some scientists are in fact religious—is entirely consistent with the New Atheists’ position and arguments, as far as they/we can tell.
He often states the dead obvious as if the NA’s were so incredibly stupid or pigheaded as to simply not see the dead obvious, despite the fact that they’ve always generally acknowledged his basic points and explained why they’re not convinced about the big picture, long run bottom line.
(We know we’re an unpopular minority, duh. That doesn’t mean that Mooney’s right that we should just lay low and leave it that way. The large majority had negative opinions of gays a few decades ago—including a lot of liberals—and laying low never worked.)
That’s where I think Mooney consistently falls down. Rather than showing that his model of the situation is the only one that fits the data, or that the NA’s model does not fit, he presents his preferred model and some data that superficially seem consistent with it, and declares victory. Then he makes his critics out to be obnoxious dopes who just don’t get it, or just don’t care.
That’s just obnoxious. It may be good political framing, but it’s lousy substantive argumentation.
–
I’m curious… Are there any particular data that you think Mooney has presented which actually particularly favor his position over the typical NA position? What has he ever said that’s actually relevant to whether and when centrist triangulation is preferable to Overton strategies?
(I’m also curious whether you think he’s ever made a good case about science/religion “compatibility,” in the sense of “compatibility” that the NA’s are obviously concerned with. So far as I know, he’s never addressed the actual point squarely, and has frequently misrepresented it.)
–
Despite PZ’s performance at the CFI thing, I think it’s pretty clear on the whole what the NA’s are generally saying about honesty and telling the whole truth.
They’re not generally saying that telling the whole awkward truth is always the best policy. (Not even PZ, in his usual, somewhat more nuanced writings.)
They’re partly saying what Jeremy said above—that Mooney has not made his case—and that in lieu of a compelling argument for holding back, laying low, etc., they will default to vocal truth-telling. (As they see it, of course.)
They’re saying that all other things being equal, honesty is the best policy in the long run, for a variety of mostly practical reasons, and Mooney hasn’t come close to making a good case for overriding that strong default.
–
I’m oversimplifying a bit unfairly above. There is a strain of New Atheist whining, especially about Mooney, that amounts to having truth-telling as an irreducible, absolute moralistic trump card. Jeremy is right to find that simplistic and unfair, and prefer a more nuanced philosophical discussion. My point is that the more strategic, consequentialist discussion of the value of whole-truth truth telling is there, in various by various Atheists writings, and Mooney just ignores it. He just doesn’t engage the better argument. (Which IMO is typical of him whenever he discusses the New Atheists.)
Given that, it’s not surprising to me that some New Atheists have gotten sick of trying to make the more nuanced argument, and just hit back as simplistically as Mooney has hit them, but that doesn’t make it a good thing to do. (I have to confess I’ve gotten frustrated and done that a little bit myself, although I do try to regularly make it clear that it’s not really that black and white, and the central controversy is strategic.)
–
Some New Atheists are also saying—or at least I am—that when Mooney raises the spectre of selective truth-telling and PR spin, it’s hard to accept what he says at face value. He doesn’t seem to default as strongly as some others to telling the full truth, even when asked, so that people can check his reasoning and come to their own conclusions.
We can’t of course say with any certainty what’s really happening in Mooney’s head and by what standards he’s “honest” or not, but we do have to take his observable behavior into account. We’ve tried to get some straight answers to what we consider crucial questions from him, and in general we just can’t.
That kind of concern is inevitable in very “political” discussions, and Mooney is a very political animal; it is natural to wonder how Mooney really sees things, and whether what he says to us about spinning things is mostly uninformative and misleading meta-spin aimed at us.
Irrespective of whether you get simplistically moralistic about it, it’s a practical problem that many of us have with Mooney.
And that’s what I really didn’t like about PZ’s talk at the CFI debate. I got that same kind of feeling.
I’m pretty sure PZ doesn’t really believe that truth is all that matters, full stop, and I thought he was indulging in some cheap holier-than-thou grandstanding. I give him the benefit of the doubt, in that one instance, for being mostly simplistic, and temporarily unusually carried away with his own rhetoric, but it was a bit disturbing.
One reason that I usually trust PZ more than I trust Mooney is that I have a stronger sense that he’s a straight shooter, for better or worse, and usually a bit less simplistic. That was an unfortunate exception.
@Paul:
I’ve laid out of this thread for some time and perhaps it’s been declared dead, but I’ve finally read through your latest response and your earlier one to me from days ago.
First, I’m noticing that there is an odd fixation with Chris Mooney in arguments against “accommodationism.” Not odd on your part in this thread since he was the topic of the OP and you limit your discussion to “Mooney-style accommodationism” in your much earlier response to me, but odd in the sense that dissenting arguments against accommodationism, in general, ultimately seem to boil down to an argument against Chris Mooney, i.e. not “accommodationism is flawed because X” but “I cannot personally stand behind accommodationism because Chris Mooney did/is/acts/might be Y.” In some cases, X in the first statement simply becomes the latter statement about character trait Y. Of course not all arguments do this, but it seems to be an abnormal proportion. IMO accommodationism has been argued by far more than Mooney, but the arguments of others mostly go unaddressed or, if they are addressed, eventually migrate to Mooney in some form, per above. Even in sight of the personal distaste many have for Mooney, however justified, I just can’t get behind it as an effective critique of the accommodationism argument. This opinion holds for almost any argument or individual, as I detail below; arguments from character are usually based on uncertain assumptions that by definition can never be thoroughly verified empirically and thus fail to directly address the core position.
As a generalized observation, your above comment is – in my reading – a long attempt to deconstruct accommodationism a la Mooney, predominantly via deconstructing Mooney’s character by calling his motives or tactics into question. I just find these arguments-by-character unconvincing because they rely on basic assumptions about character which are unknown and, ultimately, tread close to ad hominem, although I don’t believe yours makes that big of a leap (you seem sincere in your concerns and keep them somewhat decoupled from your concerns about the actual arguments). However, you even testify to the aforementioned uncertainty when you state “We can’t of course say with any certainty what’s really happening in Mooney’s head and by what standards he’s “honest” or not.” In light of that statement, let me pose this question: if there can be no certainty to the assumptions used to build an argument against character, why should that argument be considered rigorous, or the admittedly-unknowable character even considered a relevant component?
In light of the above, I can agree with your stance that Mooney et al. have yet to offer convincing datasets to support their arguments. As I’ve said well upthread, IMO there are little data to support either stance in this debate, accommodationist and New Atheist, and what datasets do exist offer conflicting evidence. Given this, I see the appropriate evidence-based stance to be one that considers both accommodationist and confrontationalist approaches until one side can produce convincing and rigorous data, although it is only realistic that individuals can and will have more polarized opinions. Or, perhaps more realistically, there is room for both approaches in a way that varies in a situational nature. I suppose time and data will tell.
Paul, I get a bit lost in your comment… It seems like you’re responding to Mooney as if he were an opponent of all of new atheism, or even all confrontational communication. At least in UA, the message is much more narrow. In a nutshell, it’s “Don’t tell Americans they have to choose between science and religion, because if you do, most will choose religion. And look–lots of smart scientists choose both.”
Relative to that message, I don’t see the relevance of Overton window stuff. That’s a defense of extreme messages–like “religion poisons everything” or “religion is a scam.” It’s not a defense of giving people potentially alienating choices.
As to Mooney’s evidence…I think in UA he was writing more or less in the mode of an op ed writer. Just as in the last few days you see people writing about the bad tone of politics, without bringing in empirical studies, he was giving a largely armchair analysis of what happens when science communication gets intertwined with religion-bashing. If armchair analysis is taboo, a lot of people are going to be out of business.
More recently, I gather from his PoI guest choices that he’s interested in empirical evidence about all the relevant issues. The Ecklund, Nyhan, Kosmin, and Specter interviews are all relevant here. Maybe at some point he’ll put it all together and make a more empirically substantive argument.
My guess is that even if he does so, his fiercest critics aren’t going to be satisfied. I actually think he knows that, and avoids engaging endlessly for that reason.
You asked what I think about Mooney’s position on compatibility. I expressed doubts way back when, here…
http://kazez.blogspot.com/2009/07/atheism-loud-and-quiet-part-2.html
On the other hand, does it really matter? I’m not sure you can be a good philosopher and have both religion and science crammed into your head, but I think you can be an exemplary citizen. So it’s not clear to me that there are “public reasons” to worry about people gaining science literacy without losing religion.
I think Paul’s comment was very helpful stuff, and definitely worth reading. And it’s more or less appropriate to this thread, since the thread is about Mooney. But as a rule I’d like to see the conversation shift away from Mooney and more onto the more general issues that Paul raised (sp. the Overton window).
Unfortunately, unlike Jean, I don’t see the difference between “a defense of extreme messages” and “giving people potentially alienating choices”. Presumably, what makes extreme messages extreme is that they’re controversial, and hence force a choice that risks alienating them. Claims like “religion poisons everything” was what UA set out to criticize, since it implicitly contains an ultimatum: keep your religion, or be healthy.
Anyway. Hammill is onto something with his remarks. Once you put aside the blood feud between certain internet personalities, the line between quiet-accommodationism and atheist-activism is pretty thin. I think that the line would fade entirely into irrelevance if a) folks came around to the idea of adopting a diversity of tactics (discussed above); and b) it was acknowledged on all hands that institutions (e.g., the NCSE) have a responsibility to represent scientists’s world-views fairly.
Hence, if a substantial number of scientists say that science is incompatible with some (or all) religious doctrines, then perhaps the NCSE should make a note of that. And Jerry Coyne’s major issue is that the NCSE misrepresents scientists in that way. Because accommodationism is a misrepresentation of a solid minority of scientists, we have public reasons to care about the matter (to use Jean’s phrase).
@Ben:
I think there are two issues. The first, as you note, is the issue of accommodationism contra atheist-activism. I agree that the line dividing the two is, in reality, not much.
However, in my eyes there’s a second issue of ‘accommodationism’ (quotation marks because I can’t think of a better term) contra those representing noncompatibility in science communication, which is in my mind distinct to a degree from the first issue. There’s very much a gulf between those two sides. Just look at the terms used to describe positions: unfair representation, misrepresentation, intellectual dishonesty. I think the terms “lies” and “deliberate distortion” have been thrown around in the past. I realize that much of that language stems from the “blood feud” you mention above, but the division seems great and very real. IMO the blood feud is getting in the way of actual progress.
While I can’t speak to the use of all those terms, I do think there’s fair reason to think that the NCSE is misrepresenting (or simply failing to represent) the beliefs of American scientists.
To the extent that the language of the NCSE is a sticking point, then the divide is quite substantial.
(I don’t mean to imply, by the way, any aloofness from the drama when I use the phrase “blood feud”. I take a very definite side. Still, that’s not going to be an interesting or productive conversation, so let’s not have it!)
Hamill,
I think you’re misreading what I’m saying about Mooney, though I may not have been sufficiently clear.
Other people brought up the issue of many NA’s contempt for Mooney—e.g., Jean explicitly asking why so many people despise him rather than just disagreeing with him.
I’m giving an answer to that.
It’s not an ad hominem. I’m not saying Mooney’s (allegedly) dishonest, therefore wrong. I’m also not saying that because Mooney’s (allegedly) dishonest, all accommodationists are dishonest.
I’m explaining two crucial points on which the NA’s sincerely disagree with Mooney—with frequently-presented good reasons IMHO—and find him unaccountably unresponsive.
They think that (1) science and religion are philosophically incompatible and factually incompatible in light of scientific knowledge. They also think (2) that incompatibility is worth playing up, not playing down, because they take Overton arguments more seriously than Mooney’s political triangulation arguments.
Mooney doesn’t play the debate game, and proceeds to frame his opponents as naive, ignorant philosophically unqualified, ideologically marginal uncivil extremists, which is just false.
They’re mainly popularizing quite mainstream philosophical and scientific ideas, and they’re perceived as far less civil than they actually are, because the things they say are unpopular.
If anybody’s philosophical views are marginal among philosophy experts, it’s Mooney’s. Yet he’s the one who pulls rank with arguments from authority and ad hominem attacks. (E.g., on “scientists” who make “philosophical” statements above their pay grade.)
I think it should be quite understandable why people who disagree with him are annoyed with him.
He’s a journalist constantly trying to pull philosophical rank on scientists, when scientists are making mainstream philosophical arguments that he doesn’t even try to actually refute.
Even Mooney’s friends don’t actually agree with him when he says that the NA’s are wrong, and makes it sound like the philosophical experts are on his side against those uppity ignorant NA scientists. They’re not.
As I mentioned, he likes to quote Barbara Forrest to make it sound like she authoritatively says that science can’t undermine religion broadly, when in fact she explicitly says quite the opposite in her philosophical publications. He also likes to quote people like Phil Kitcher (the eminent philosopher of science) on the possible utility or inevitability of religion, but fails to mention that they too agree with the New Atheists that in fact, science undermines religion quote broadly, by undermining supernaturalism.
See Jean’s post that she linked to in a comment above, for example. Even Jean doesn’t agree with Mooney on science/religion compatibility. She says the same kind of thing that Mooney lambastes Jerry Coyne for saying.
Likewise, even Jeremy (above) doesn’t think Mooney has made his case on the strategy point.
Yet Mooney persists in slamming the NA’s, as a group and frequently singling out prominent individuals, for being uncooperative with his strategy, making them out to be willfully irresponsibly unconcerned with the consequences of their saying what they actually think.
If Mooney had made his case, he might have a point. He hasn’t, and won’t, so that makes him obnoxious for acting as though he has and using his undefended assumptions to justify what are essentially ad hominem attacks on New Atheists.
IMHO, Mooney systematically refuses to address the NA’s actual arguments, and proceeds to vilify and demean them on the basis of contentious points that he has not established, and hasn’t even seriously tried to address the objections to.
I think that should be obvious to anyone who’s followed his blog and New Atheist blogs over the last few years. It’s has been a really tiresome pattern for years.
Even Jean seems to recognize this to some extent. Perhaps she’s right that Mooney doesn’t engage with his critics because he thinks he can never convince them—that no matter how good his arguments, they’ll never get on board.
Even if you think it’s true that the NA’s are simply unreasonable and unreachable, it should be obvious that they don’t think so, and that they are understandably frustrated by Mooney treating them that way.
Who wouldn’t be? Even if Jean’s right, isn’t that a kind of ad hominem argument, in itself—the “you’re just not worth arguing with” gambit?
Isn’t that exactly the kind of thing that normally pisses people off?
Perhaps the New Atheists shouldn’t take it personally and let it actually piss them off, but is that reasonable reasonable to expect?
(I’ve been to enough philosophy colloquia and APA meetings to know that even professional philosophers tend to get rather pointedly negative about such things.)
Until Mooney makes a serious effort to actually defend his two main theses, don’t be surprised if people flip him off for criticizing them for disagreeing and behaving accordingly.
As for accommodationism more generally, I think that while Mooney is the biggest offender, other so-called “accommmodationists” are often guilty of the same style of argument.
For example, like Mooney, Genie Scott is fond of a bait-and-switch on the “compatibility” issue. Rather than addressing the sense of incompatibility the sense the NA’s plainly mean, she switches to the sense of compatibility that they’ve always granted—everybody knows some scientists do good science, even outstanding science despite being religion.
(Even PZ Myers has always and frequently granted that point, e.g., praising Francis Collins and Ken Miller and as scientists and popularizers of, while disagreeing about whether their theology is really compatible with science generally. IMHO, he’s right on both points.)
I think it’s noteworthy that most New Atheists still like Genie Scott but don’t like Chris Mooney.
One reason for the vitriol against Mooney specifically is that he doesn’t just say things they disagree with. He’s perceived as going way out his way to attack New Atheists generally, and to single out specific New Atheists, when they’re right and he’s not only wrong (in their view) but hasn’t even tried to make an actual case.
He’s the most prominent and consistent self-appointed anti-New Atheist hit man for accomodationism. It’s only natural that he’s a lightning rod for resentment.
For example, we were led to believe that Chapter 8 of Mooney’s book would make the arguments and give the actual evidence for what he’d been saying against the New Atheists for years.
It didn’t. It was mainly a rehash of the same things he’d said on his blog and elsewhere, which we thought had been well refuted, in what Jean rightly describes (above) as “op ed” mode.
After that, Mooney behaved as if he had made his case, and didn’t want to waste time revisiting subjects he’d already thoroughly addressed.
Gee.
I think New Atheists were right to find that frustrating. It made it pretty clear that Mooney never would actually make his case for either his main philosophical criticism of New Atheists, or his main strategic reason for casting them as obviously counterproductive and therefore willfully irresponsible.
His behavior since then has done nothing to dispel that impression.
Paul, I actually appreciated your attempt to answer the question “why despise, not disagree?” I asked the question sincerely, not to be incendiary, and I think what you said was interesting (and also not incendiary). I did read it all, felt enlightened by some points, etc. I can just about see why many people would find Mooney unsatisfying (not enough direct arguing, etc.) That’s not to say that I think there’s any defense for the way he’s been bombarded and personally vilified. One little substantive point–I think Mooney’s philosophical ally is Rob Pennock, a guy with a good reputation. So it’s not true that all the philosophical ballast is on the other side.
Jean,
I’m not saying that the philosohical ballast is all on one side—I’m just saying that
(a) Mooney’s wrong to make arguments from authority and imply that the New Atheists are going against a philosophical consensus. (b) There isn’t a philosophical consensus, quite, but if anything, it’s Mooney that’s going against the preponderance of expert opinion, and he’s the one who doesn’t know and accept what the qualified experts know.
I’m sure that there are some generally respected philosophers whose opinions on the subject do resemble Mooney’s, and maybe Rob Pennock is one.
I’m not sure how closely Pennock’s views actually do correspond to Mooney’s, though.
As I mentioned before, I looked up Mooney’s favorite pull quote from Pennock’s Tower of Babel, and it seemed in its immediate context to mean exactly what Mooney presented it to mean.
I quite like Tower of Babel overall, and generally respect Pennock, but I strongly disagreed with that statement.
Later I worried about it—I was afraid I was arguing with a straw Pennock—and reread the whole chapter, and I am just not sure. A few pages earlier he says some stuff that makes it sound like he may only be making a feasibility argument, rather than demarking a basic scope boundary.
That is, there are too many supernaturalist theories, and no scientific way to find the “good” (promising) ones, so we give up on supernatural theories in science, because would otherwise waste all of our time going down an infinity of rabbit holes and get no useful science done.
If he only means that as a practical argument, I agree with him, as far as it goes, and he too disagrees with Mooney.
If he means that as a strong argument that the supernatural is off-limits to science, in principle, he’s agreeing with Mooney and I think he’s wrong.
(The fact that we can’t identify particularly good supernaturalist hypotheses to study doesn’t mean we shouldn’t provisionally conclude that they’re all bad and the basic idea is probably wrong. That’s what we normally do with failed paradigms.)
Mooney and some other accommodationists like to make methodological naturalism out to be some kind of in-principle boundary that protects the supernatural from scientific scrutiny. Like most philosophers of science, I think that’s wrong. Methodological naturalism is just a heuristic scientific guess based on the chronic empirical failure of supernaturalist hypotheses, and the continuing success of materialist science based on that heuristic strongly suggests that supernaturalism is just false.
I don’t know what Pennock actually says about all that, and I don’t think it really matters for our purposes—certainly no prominent New Atheist says that there is unanimity about the incompatibility of science and religion among respected professional philosophy.
It is Mooney, not the New Atheists, who appeals to some kind of authoritative philosophical consensus. The NA’s don’t generally make arguments from authority, or want to, but if they did resort to that, they could do much better at it than Mooney. That’s what makes Mooney’s dismissive condescension about expertise so ironic and annoying.
By the way, I do respect Rob Pennock a lot, as a philosopher, and quite like him personally to boot. I talked to him a few times and (IIRC) had him over for dinner, many years ago. He’s a good guy all around, and if we actually do disagree on this particular point, that’s okay by me. I just don’t want Mooney making a minority view out to be a consensus, and using it to dismiss the actual arguments.
Rob Pennock has a new article about this stuff in Synthese and I’ve been meaning to read it for a while. I know him too–or knew him, way back when when were both at UT. Yes, I don’t really care for the line about methodological naturalism not leading to philosophical naturalism, but I should read him more before making up my mind.
Jean,
I’d be interested in your take on Rob’s Synthese article.
I just had a look at your blog and saw the discussion of the modal ontological argument, which I need to think about.
That makes me wonder if you were still at UT when Rob Koons got tenure, ate his brilliant metamathematical brain, and started spewing apologetics—including a spiffy new modal logic version of the Cosmological Argument.
He claimed that with modern modal logical techniques, you could make a clear and compelling version of the Argument. Whee! (I know some of his colleagues there who were, like me, perhaps necessarily unimpressed.)
Yes, I recall the apologetics, and still remember some very weird things he said at that time. But wait…I’m getting very curious who you are. Still at UT…used to be at UT…c’mon, give me a clue.
Clues? You want a clue? Too fun.
OK, I’ll give you two.
1. My Erdos number is 3, but
2. I’m not actually very mathematically gifted.
(You didn’t specify that they had to be useful clues.)
I have a theory about your true identity, thought I think there’s only about a 20% chance I’m correct. If my theory is true, than your pseudonym is completely misleading. True or false?
p.s. My theory meshes with your two clues, as far as I can tell.
It’s not an ad hominem. I’m not saying Mooney’s (allegedly) dishonest, therefore wrong. I’m also not saying that because Mooney’s (allegedly) dishonest, all accommodationists are dishonest.
That’s fair. I apologize if I somehow implied that I thought you were making that extrapolation. I wasn’t. I have seen it before (at least IMO) but not really from anyone in this thread. I personally tend to fall on the noncompatibility side of the aisle but I don’t always align with the noncompatibilists on associated issues. I suppose that lends credence to the idea upthread about how complex this issue is.
As for your comment about civility:
They’re mainly popularizing quite mainstream philosophical and scientific ideas, and they’re perceived as far less civil than they actually are, because the things they say are unpopular.
I disagree to some extent. I don’t think it’s a stretch to say that there’s a stunning amount of incivility on both sides of the aisle, at least in intervals. You reference some of the incivility yourself, as well as what is reflected back (per your “lightning rod” analogy). With that said, the incivility seems to be mostly incivility about people and not about the core ideas. And, as you correctly mention, civility is really in the eye of the beholder.
As you and Jean have been discussing above, I think the true interesting parts of the debate lie at the philosophical conflicts and camps therein. I would add that there’s a fascinating political dimension to the debate, even though politics is often seen as a dirty word. Part of me almost wishes that the entire debate were taking place in a refereed journal rather than the blogosphere, due to the inherent incivility that the internet brings. Perhaps that could corral the personal disputes a bit and focus a laser beam on the core issues. I suppose there are positives even for the current forum, however.
Going back and addressing a few points I’ve left hanging a long time…
Jeremy:
I agree, except that my understanding is that the data isn’t there, and that the data actually may support the idea that in the long run, cross-culturally irreligious people are happier than religious ones.
As I understand it, people like Rodney Stark (of the Baylor surveys) like to make a simplistic argument that religiosity is is correlated with happiness (or self-reported life satisfaction or whatever), based on US data, but a finer-grained analysis doesn’t bear that out in the simple way he means, and cross-cultural data seem to show that it’s probably false in general, and maybe even the opposite of the truth.
(My impression is that Stark’s work has largely been discredited and it’s increasingly obvious he’s in denial about the fact that secularization is happening in the US, that the Nones are NOT mostly religious Christians who are simply unchurched, etc. All the other surveys seem to show that secularization is really happening in the US, the Nones mostly don’t believe in a personal God, etc. As I understand it, when Kosmin stresses that most of the Nones are not religious, and that most of the ones who are “religious” in some sense are nowhere near orthodox, but “Deists,” he’s basically saying Rodney Stark and the Baylor surveys are just wrong, without naming names. Most of Stark’s other favorite ideas seem to have been discredited as well—e.g., the idea that religiosity is “social capital,” which doesn’t hold up cross-culturally. Less religious societies seem to work better by a variety of measures, and attributing so much good stuff to religion seems to be a naive attribution error.)
As I understand it, the best data on the religion/happiness issue show that while more religious people do tend to report being a little happier, happiness is not correlated with religiosity per se as much as religious participation—and in fact, that people who are less orthodox and less certain in their beliefs tend to be happier than more orthodox and certain people, when you factor in religious participation.
The leading explanation for that (from what I’ve read) is that people who are more social are happier, so going to church and church-sponsored activities mostly for socializing may be good for you—in a religiously oriented culture like the US—but if you take the religion itself too seriously, that’s likely bad for you.
That’s consistent with data from much less religious countries, like Sweden. In less religious countries, less religious people aren’t noticeably less happy than religious people. They find other things to organize their social lives around—hobbies, political movements, etc.—and end up about as social and about as happy.
(Of course, one issue with the sociality theory is which direction(s) the causality runs. Are more social people therefore happier, or are happier people therefore more social? I suspect a substantial effect both ways, but don’t know what the science shows.)
Like the data showing that gays tend to be less happy on average than straights (or used to be), the correlation can easily be explained—if you’re a member of a maligned minority, who doesn’t fit in to most social contexts in your society because you’re different, it’s really not surprising if you’re less social and/or less happy. Of course you’re more likely to be alienated and lonely. That’s why a lot of gays have traditionally migrated to places like New York and San Francisco and Houston (especially neighborhoods like the Village or the Castro, and Montrose where they’re not powerless and can “be themselves” with less cost and risk.) It’s a lot more satisfying to be accepted than to have to hide who you are or be rejected and reviled.
One of the things in the Kosmin ARIS report on Nones that was of special interest to me was the discussion of African American Nones. As I already knew, there’s a whole lot of Christian orthodoxy and observantness in the US black community; a whole lot of black social and political life is organized around and through churches. What I didn’t know is that the black community also has a substantial minority of Nones—it’s not simply more religious across the board, but polarized. A reasonable percentage of black people do reject the highly religious aspect of black culture, and become not just less religious or less observant, but irreligious. I guess my wife isn’t as much of an outlier as I thought, and it makes me wonder how that works. Does the rampant religiosity and orthodoxy just generating a reaction against religion in general, or what?
Similarly, but more generally, it’s not clear to what extent the increase in Nones in the US as a whole is due to a reaction against increasing religiosity by other people, or increasing visibility and power of the religious right. If so, we may NOT be seeing the kind of long-term secularization trend that has happened in other industrialized democracies over the last, but at least partly a short-term reaction. My guess is that’s not mostly true, and what we’re seeing is the kind of enduring and likely self-amplifying secularization thing that happened elsewhere, but I don’t know, and I don’t know how to know.
If I get Kosmin right, there are reasons to think this is following the European pattern, and the changes are likely to endure and feed back. In particular, rates of religious attendance are actually much lower than people like Stark report from simple self-reporting surveys. Americans tend to overreport church attendance by a factor of about two. (Interestingly, people in other industrialized democracies don’t seem to; their self-reports are much more accurate, maybe because there’s less stigma attached to non-churchgoing.) If you use better survey techniques, or just a rough estimate from approximately known aggregate church attendance rates, it’s clear a whole lot of people aren’t really going to church, or not nearly as much as they say—only about one in four is actually in church on any given Sunday. In particular, even fewer men are in church; a lot are leaving it to Mom to church the kids.
That’s interesting, the story goes, because church attendance is believed to be a pretty good predictor of how religious the next generation will be, and whether the fathers go to church with the family is also a predictor. If mom and dad don’t both take the family to church regularly, the kids are especially likely to become irreligious. (Maybe because Dad sets a “bad” example by clearly valuing the one true American religion—football—over Christianity.)
I’m no expert, and don’t know if that’s all true, but it sounds pretty good to me.
As for why it’s really happening, I don’t know that, either. As I said before, I don’t think most of it is because of the New Atheism. I suspect it has a lot more to do with the rise of mostly cable television, the internet, and other factors I haven’t thought of or thought enough about.
I wouldn’t be surprised if cable television is a significant factor, because there’s a lot more “liberal” or “controversial” programming out there that might offend conservative religious sensibilities, but they can’t pressure the cable networks to remove it the way they used to pressure the Big Three networks’ Standards departments. Comedy Central and WB and Showtime are going to target less conservative demographics, and tell the conservatives not to watch it if they don’t like it. That, in turn, makes it less feasible to pressure the Big Three, who have to compete for market share. They can say that if you don’t want your kids to see it, program your v-chip. And a whole lot of people don’t want to do that, or don’t get around to it, so their kids end up watching Queer Eye and the L word and the Daily Show, absorbing values that make orthodox “morality” seem quaint, even hokey.
More recently, the internet has made it easy for a lot of people, especially younger people, to access stuff that couldn’t even get on cable TV. (Or to bypass their parents’ V-chip settings and watch TV shows they can’t watch on TV.)
That leads to the kind of fragmentation and polarization of media that Chris Mooney bemoans—it is polarizing, often leading to people choosing media that amplify their existing prejudices, but is also liberating. It allows people to move out of the narrow mainstream into alternative ideas and identities in a way they couldn’t before, or not as easily.
How important is that, really? Beats me, but it sure sounds like it could be important.
Another basic issue in all this is why we didn’t follow the usual first-world pattern of secularization decades ago.
One common explanation (with statistical arguments behind it) is that religion thrives in countries where many people are poor and especially insecure. (Poverty per se is less important than inequality, and knowing that you’re not keeping up with the Joneses.) More religious countries generally have more inequality and poorer social safety nets. People whose fortunes and status are less threatened by poverty, inequality, and risk of misfortune tend to find less need for religion. (Or so the story goes.)
If the causality actually runs that way, it may not be surprising that the US has secularized less, more slowly, and later than other industrialized democracies. It’s had greater income inequality and/or poorer social safety nets than almost any other industrialized country, for many decades.
If that’s all there is to it, that may work against the New Atheist’s arguments, and favor a more accommodationist approach—maybe we should work for social justice, and get less religiosity “for free,” as a side benefit. I’m skeptical, because I think the causality runs both ways. I think that lower religiosity makes it easier to see social justice issues as social justice issues, rather than as matters of personal salvation and “charity” being a personal virtue, but not a societal obligation. Voluntary alms for the poor just don’t generate as much social equality and security as progressive taxation and universal health care.
In the US at least, I think there’s a nasty connection between the religious, antigovernment, and antitax/supply-side voodoo economics phenomena. (Republican strategists pretty clearly think there is.) I could be wrong about the significance of that, but I’m skeptical that we can get a lot of social justice without weakening the hold that orthodox Christianity has on our culture. (Not just fundamentalism, but the kind of basic orthodoxy most Americans have.) Maybe US Christianity can evolve to be less concerned with orthodoxy and intolerance and more concerned with social justice, but I’m not holding my breath, and I think that the most likely way to make that happen would be by Overton strategies. I could of course be wrong, but that’s how I currently lean.
One factor that has been used to explain the US’s slow secularization after World War II is the Cold War, and America’s leading position on the anticommunist side. More than in other countries, patriotism was emphasized and equated with anticommunism, and anticommunism was equated with anti-atheism. After the cold war, with communism less of a visceral threat, rhetoric about those horrible godless communists lost currency and vividness, and that may account for the more rapid trend secularization of the last couple of decades.
I don’t know how that really works, though.
Another factor I’ve been wondering about that may contribute to our slow secularization, and may have contributed to the cold war polarization, or just had an additive effect in the same direction, is our style of representative democracy.
Most other first-world countries have parlaimentary democracies, not the kind of representative democracy we have, with an entrenched two-party system. I suspect that has some ramifications that are relevant to why secularization progressed sooner/faster elsewhere than here in the US.
In the US, you can go pretty far in marginalizing small minorities among your own supporters, so long as you’re not as strikingly anti-them as the other major party. They have nowhere to go. For example, blacks might vote for the somewhat less-racist party, even if both parties are significantly racist, or gays might vote for the less somewhat heteronormative party, even though both parties overtly dis and marginalize them, and mostly keep them invisible.
It’s rather different in a parlaimentary democracy, I’d guess. Blacks or gays might vote for a minority party that actually supports them, or refuses to dis them, and get actual representation in parlaiment—if they’re a bigger part of the support for a smaller party that at least gets somebody in office, and maybe have some leverage in forming coalition governments, they have some assurance that somebody will regularly speak up for them, not shut them out of the political discourse, and maybe even make substantive demands of coalition partners on their behalf. My impression is that it results in something a bit more like an open marketplace of ideas, with a little less inequality of immediate political clout.
I could easily be wrong about that, but it seems to me that if that’s sorta app right, Overton strategies may work significantly better under parlaimentary democracies than two-party systems—ideas that would otherwise be “unheard of” become familiar, and get argued for, have to be taken seriously by some larger number of people, and may eventually become common, rather than always being relegated to a lunatic fringe that can mostly be hidden and ignored like a crazy uncle in the attic.
I don’t know nearly enough about parliamentary politics to have anything like firm opinion on this, but I thought it might be relevant somehow, and thought I’d toss it out there. It seems possibly relevant to the polarization/broadening of political discourse that you get from more fragmented media. With more diverse media, you can’t squelch minority views and values as effectively as if you have a dominant “mainstream” media where “relevance” is mostly about two-party political discourse. (E.g., center-left vs center-right.) If so, diversification of media channels may have some effects somewhat similar to parlaimentary democracy; that, in turn, could amplify Overton effects so that things in the US will follow something somewhat more like European pattern.
(I’d like to think so, but something entirely different might happen. Maybe diverse media and a two-party electoral system are a disastrous combination.)
re Rodney Stark and the Baylor religion surveys, here’s a pretty damning press release, with a link to a full paper, arguing that Stark’s work is unreliable at best.
http://www.centerforinquiry.net/news/new_report_raises_legitimacy_questions_of_much-touted_baylor_university_rel/
I’d be interested in any opinions on it, especially people with more sociological expertise than me. (E.g., Jeremy)
It seems pretty consistent with my impressions from reading about several other studies by different groups, though I certainly don’t recall all the details or who said what.
Dear God, Paul, that’s a long post!
If you’re inclined to write this stuff down in a systematic way (i.e., so that it’s self-contained), then I’d be happy to publish it here as a guest post. I feel bad you’re putting what must be a lot of effort into this, but it’s all popping up at the bottom of a thread (not that it’s my fault or anything).
(I’ve only just seen your later stuff because I happened to be looking at the folder my blog notifications come into when the notification for this post arrived.)
By the way, what do you make of this story (given the whole secularisation thing)?
http://www.dailymail.co.uk/news/article-1340944/How-believe-guardian-angel-looking-us.html
(Okay, I’m not being entirely serious. But it does remind me of a poll a little while ago that showed that more people in the UK believed in heaven than God, which I thought was quite funny.)
It’s the religious participation, not the religiosity…
I think this is like saying what people really enjoy at baseball games is the camaraderie, the hotdogs, the music, being outside, being away from work, the team spirit…not watching the game. Even if that were true, the question would be “where else can you get all that stuff, in one package, except at a baseball game?”
Likewise, even if it’s true that people enjoy the 50 elements of religious participation, not the religiosity, they’re not available outside of churches…at least for now.
(Actually, I find it hard to believe the religiosity itself does nothing whatever for people, as it makes a lot of sense that it would be consoling to believe in heaven, the power of prayer, that everything happens for a reason, and the like.)
Well, anecdotally, we know that it does something for some people. I know people whose children have died, who only (seemingly and self-professedly) manage to keep going because they think they’re going to see their child again in heaven.
Of course, there are issues here to do with counterfactuals and the other sorts of things that religion does to people.
(The fact that it’s participation that’s important has been taken as a sign that secularisation, even in the States, occurred a long time ago. So the idea is that religious attendance isn’t motivated by religiosity.)
Jean:
I’m not sure that’s the question, or the most pertinent question, and I’m not sure what the analogy is really supposed to be or to show.
Surely, if you’re looking for exactly the same combination of features that baseball offers, you don’t have a lot of alternatives.
It is interesting, however, that cross-culturally, people find a lot of the same things, in relatively similar combinations, elsewhere. (E.g., soccer and bratwurst and German beer, or cricket and British sausages and beer, or instead of baseball and hot dogs and American beer. And what about a multiplayer videogaming LAN party, delivered pizza, and sixpacks? Now there’s some camaraderie and team spirit.) The fact that some people are currently very attached to a particular familiar combination of things doesn’t mean that there aren’t other equally good combinations of things, and perhaps better ones, that at least some people might come to prefer. (Especially young people who aren’t as sentimentally attached to the “old ways,” such as many young people in the US who grew up playing soccer and multiplayer video games as well as baseball.)
I think we need to be sure not to overglorify the benefits of church and church-related social activity. Most people don’t go for it—3 out of 4 Americans won’t attend services this week, including 2 out of 3 religious Americans. The fact that more actively participatory people seem a bit more satisfied, statistically, may not mean what the promoters of religion would have you believe. I suspect it’s mostly a matter self-selection—more social people tend to be happier with their lives, irrespective of religiosity, but religious ones are likely to be social in specifically church-related contexts. Most would do probably be comparably social and satisfied if they weren’t religious, too. It’s also not clear that the causality runs from being social to being satisfied—it may go mostly the other way, e.g., with less satisfied people tending to be less social.
Even to the extent that specifically religious social opportunities increases socializing, and to the extent that that causes happiness, there are downsides. Many people don’t especially like church or church-related socializing, which is why most don’t usually go, and a lot of the ones who do go don’t really want to—they do it because they think they’re supposed to, or think it’s better for them than it actually is, or because they’re pressured to. (E.g., parents going to church “for the kids,” people going to church to avoid disappointing their spouses, kids going because their parents make them, or to be “good kids” etc.)
I think it’s generally a good thing if we weaken the social expectation that church attendance is expected, and weaken the cultural message that it’s A Good Thing to do and Good For You. (Which seems to be happening over the last few decades. Yay.) It’s certainly not nearly as good for most individuals, or for society, as promoters of religion would have us believe, and the message should be “Don’t want to go to church? That’s perfectly okay. Don’t go.” It’s fine if people who really get a kick out of church do go, but they should stop pestering and pressuring and boring the rest of us, and people who don’t go shouldn’t be stigmatized.
If religion really were anywhere near as good for people and for society as it’s cracked up to be, it would show up in the statistics for more and less religious countries. It doesn’t.
It would also show up in various more detailed statistics about this country. Religion promoters make a lot of claims about religion that would have observable consequences if they were true, and they don’t, so they’re apparently not true. Raising kids to be abstinence results in more out-of-wedlock pregnancies, abortions, and STDs. Churches that promote stable marriage and family values have just as much divorce among their members, etc. Praying for people’s recovery from heart surgery doesn’t help them—if they know about it, it worsens their outcomes a bit. People who believe in God and Hell are no less likely to end up in prison or on Death Row—they’re a bit more likely to. The list goes on and on and on.
Given how hard it is to study such things and interpret the statistics, I don’t think we can make any firm claims that religion is bad for individuals or societies, or that it’s neutral. It’s possible it could actually be a good thing, though it certainly doesn’t appear that way to me. I do think we can infer it’s nowhere near as good as it’s cracked up to be, and irreligion is nowhere near as bad.
I don’t have any ethical qualms about criticizing religion or promoting irreligion, in terms of its typical net effects on individuals—near as I can tell, it’s’ a wash at worst, and I can comfortably default to calling them like I see them. I could be a bit wrong, but that doesn’t mean I should play it safe and shut up—I could be wrong in either direction, so it’s at least as likely that shutting up would be a mistake.
For me, the big ethical concern is long-term, big picture effects on societies, and I think that promoting irreligion is very probably a good thing in that regard—less religious societies do seem to work better, and a lot of American political problems are rooted in religion.
For me, the remaining ethical concerns are indirect—what is the right time and the right strategy for criticizing religion and promoting irreligion?
So do I, but I find it very easy to believe it doesn’t do nearly as much as it’s supposed to, or for as many people, and the benefits are outweighed by costs, to those people and to others.
My reasons for thinking this are not just big-picture statistical things. I can believe the statistics partly because they’re consistent with what I know of individual psychology.
Jeremy:
I think such anecdotal evidence is usually quite wrong. This is exactly the kind of judgment that people are just astonishingly bad at, and subject to enormous confirmation bias. Scientifically, we’d expect a ton of anecdotes like this whether it was true or not that religion had that effect, so even huge numbers of such anecdotes are not actually evidence of anything. They’re only to be expected, either way.
This is a well-known and huge bias that psychologists have to know about. If you are ill or depressed, and do anything about it that you think is likely to help, you are very likely to get better and to attribute your getting better to whatever it is that you did.
(I’m not talking about a placebo effect here, just people’s basic inability to make accurate causal attributions about this sort of thing.)
For example, suppose you’re get sick, and after a few days, you finally go to the doctor and she gives you a pill or a shot. Within a day or two or three, you will likely feel much better, and think hey, that worked great!
The problem, of course, is that with or without treatment, you’d probably feel better pretty soon. When people feel really bad, they usually feel less bad soon—the most probable change from a low point is upward.
The same problem plagues clinical psychology, and there is a very strong tendency for any psychotherapy to appear to work very well on average. The only way for a therapy not to have a lot of apparent successes, and quite good on average, is for it to be systematically quite harmful.
In light of that, it’s entirely unsurprising that people think religion helps them get through hard times, like the death of a child. Of course they do. At the time, it seems like they’ll never get through it, but it turns out they do, so they attribute that to whatever they did that they believe can help—praying, talking to their pastor, etc.
The truth is that balancing on one foot several times a day would probably help about as much, and if you really believed it could help a lot, and did it for that reason, it would seem that it clearly did help a lot.
The truth of such things is just not something people have introspective access to, even though they think they do, and their self-reports must be discounted almost entirely. Religion “obviously works” for believers mostly for the same reasons that all sorts of crank therapies and quack medicines “obviously work.”
There’s no substitute for actual comparative, statistical data, and the data do not bear out religion’s claims—it’s clear that the benefits of religion are at best grossly exaggerated, and it’s perfectly understandable how that happens. That’s only to be expected, due to normal cognitive biases at the individual level.
Hmmm… what happened to all my paragraphs? I don’t know why it’s breaking the lines there but not putting spaces between the paras. (In my browser anyhow.)
Sorry about that.
[Edit: No problem; fixed. -- Ben (Jan 21/2011)]
Paul – I do wonder whether you’ve actually ever met a mother who has lost a child?
In my experience (my brother was murdered), it really isn’t a matter of “getting through it” (as you put it), though I guess that might be how some mothers experience it.
Moreover, there is a very specific comfort to the idea that you’re going to see your child again that has nothing to do with the sort of thing you’re talking about (because, of course, people who are bereaved also do all the other standing on one foot sort of stuff you’re talking about – and it doesn’t – always – work as well).
(I should say that my mother isn’t religious. I know other people who have lost children.)
Your general point is right, of course, without evidence this is anecdotal (hence my remark about counterfactuals).
Jeremy,
I have certainly known people who’ve lost loved ones, including children, and I’ve lost some loved ones myself.
I don’t doubt that for some people the expectation of seeing the loved one again in the afterlife is a consolation, even a substantial one.
I do suspect that the consolation it actually offers is often exaggerated.
When people are grieving in such situations, they’re not mostly grieving about effects years in the future—they’re grieving about the now, and the short-term, and the foreseeable future. The distant future is not nearly as emotionally real to them.
They’re emotionally attached to someone, who’s suddenly not there, and it’s like an amputation, with a phantom limb.
Believing that you’ll see the person again eventually is like believing your amputated leg will grow back, but only eventually—say, after a couple of decades. It’s mostly consoling in an abstract way, which is somewhat helpful, but it doesn’t address the main issue: you want your leg back NOW, or VERY SOON, or at least SOON, because you were counting on that leg. But you know that for as far as you can see and emotionally “feel” into the future, emotionally, the leg isn’t going to be there. Not today, not tomorrow, not next week, not next month, not next year, and not next decade. You’re going to go without for a very, very long time.
I have no doubt that people do clutch at that not-really-very-consoling straw and wring all the consolation out of it that they can, and maybe for some people it’s especially effective.
I think most people will just clutch at different straws and wring approximately the same amount of consolation out of them, though admittedly somewhat less on average because they’ll have a smaller selection of straws at hand.
If the death was sudden, they might focus on the fact that at least it was quick, and the victim didn’t suffer. Or if it was not sudden, and was painful, they’ll focus on the fact that at least the victim’s suffering is over, and they’re not in pain.
I also know from a couple of friends who do end-of-life care (one’s a social worker at hospices) that religion sometimes makes things substantially worse—and not rarely, either.
My impression is that it’s grossly underreported—almost never discussed at all—how often a tragedy causes people a painful religious crisis, and they have to suffer through that in addition to suffering through the loss itself. Some people get terribly angry and mad at God. Others stop believing in God, and feel lied to, betrayed, and left twisting in the wind. They expected solace, and they’re getting just the opposite at the worst possible time.
That’s bad timing even for people who do get past their religious crisis and stay religious, or who decide they do believe in God after all and can make their peace with Him somehow. A the crucial time, they’re confused and frustrated and horribly disappointed in religion itself, on top of the pain of losing a loved one.
It is unclear to me whether religion is a net benefit, on average, in such situations. Maybe it is, but I’m not convinced, and I’m pretty sure that as usual, religion is systematically overrated. That’s ubiquitous in our culture.
I haven’t followed this dialogue in detail.
However, I have lost a child, and while I’m an atheist and while I don’t believe in an afterlife in general, I do believe that my son still lives, in some sense.
If you had asked me about the possibility of an afterlife before my son’s death, 9 year ago, I would have laughed and mocked the whole idea.
I haven’t turned to organized religion and never will, if only because it bores me to sit still during religious services.
I am not advancing the afterlife as a philosophical or metaphysical position. I’m only trying to explain what I sense is a common reaction of people who have lost children. It’s not like losing a parent or a grandparent.
I can fully understand how someone with more tolerance than myself for psalms and sermons could turn to religion under the cirumstances.
Jeremy,
Another fairly common circumstance in which religion amplifies tragic losses is when religious people think that the lost loved one is going to Hell, or worry that it’s likely.
This is not rare, either. A substantial percentage of Americans really believe in Hell and that it’s an easy place to go; many of them do suspect that various people they know and love are likely, or even probably going there, if they don’t Get Religion and shape up. If they die without demonstrating necessary righteousness and religiosity first, that’s bad.
Probably the single most common version of this is when somebody commits suicide. A lot of orthodox Christians think that suicides automatically and immediately go to Hell and suffer tremendous torment. They don’t go to purgatory, but straight to Hell, and once you hear that the victim died by suicide, you know that the torture is happening RIGHT NOW.
Young people are especially likely to die by their own hand, if they die at all—it’s one of the top causes of early death. They are also likely to have living parents who still feel parental towards them and responsible for them, so the combination is not at all rare.
Yes, it may be consoling to think your child’s in Heaven, but I’d guess that the emotional impact of thinking your child is suddenly and irretrievably in Hell is much bigger. Especially if you think it may be your fault because you didn’t raise them right. If you even suspect that they’re in Hell, and that you may have contributed by failing somehow, that’s gotta be very disturbing.
@Amos – Yes, to all of that.
@Paul – I’ll respond later. But, just briefly, you may know that there’s some classic sociological stuff that looks at how coroners define suicide (it tends to be cited as a counterbalance to Durkheim’s claim that suicide is a social fact that can be explained by other social facts), and one of the points that comes up is that coroners can come under enormous pressure from Catholic families not to define their child’s death as suicide (for the sorts of reasons you’re talking about).
But this argument isn’t about the balance of the effects of religion. It’s about whether religion can be a consolation. The fact that it often isn’t a consolation, doesn’t mean that it isn’t sometimes a consolation. And one of the ways it just is a consolation is that it offers people a way to think that death isn’t the end. (The counterfactual argument here doesn’t alter this point. It just means – if it’s true – that the consolation could be found elsewhere.)
(And, incidentally, I’ve looked at a lot of the data on fear of death, and while it’s true that some people do fear death because of what they think is going to happen to them, it’s also true that people fear death because of the fact that it represents the extinction of the self. Though, I hasten to add, the data doesn’t particularly seem to show that religious people fear death less than non-religious people – there is some evidence that the moderately religious fear death most of all – which actually I think might say something interesting about the ontology of religious belief – which is another huge complication in all this stuff we’ve been talking about).
Jeremy:
I have no doubt that religion can be a consolation—does anyone actually doubt it?—but I didn’t think that was what this thread was about.
This thread has largely been about strategy and ethics in “New” Atheism vs. accomodationism, and for that, what effects we can realistically expect matter a whole lot—as you and Jean stressed early on.
It seems to me the particular issue of whether religion can be a consolation is off-topic. Nobody on either side is saying it can’t, and as near as I can tell, nobody on either side actually thinks whether it major merely can be a consolation is of major strategic or ethical importance.
What crucially matters is mostly the likely effects, in either direction, and which way the overall balance goes.
Paul, My point was that the bundle of ingredients people get from going to church are not available, as a bundle, except at church (synagogue, etc)…at least right now, in the US. You’re underestimating what all is in the bundle. It includes–
(1) social interaction, (2) talking about life, death, values, (3) ways of dealing with life-cycle events like birth, death, and marriage, (4) peace and quiet, meditation, escape from everyday concerns, (5) connection to people distant in space and time, (6) synchronized singing, movement, and the emotional elevation this produces (see Jonathan Haidt), (7) avenue for”good works”–building houses, giving blood, etc., (8) beauty–music, art, architecture, etc.
People want all 8 things (and more) but also want them in one bundle. You don’t get the same effect by (1) by making friends, (2) taking a philosophy class, (3) giving blood, etc. So I don’t think the data that church-attendance makes people happy can be dismissed, on grounds that they could get the same boost from pursuing all these things in other ways.
I actually belong to a religious organization, and know many other people who do, so have a pretty up-close-and-personal feel for what the benefits are. But aside from that, I find the data compelling, especially the data that says there’s a “dose-response” effect–more church attendance makes for greater happiness. I also find it compelling that people rank prayer-meditation as one of their top activities, in random sampling studies. Here’s the data–
http://1.bp.blogspot.com/_ig4Mz3Ac464/TOWOGVH5pKI/AAAAAAAACB4/bNz5WBeI1Zk/s1600/Picture+13.png
“Yes, it may be consoling to think your child’s in Heaven, but I’d guess that the emotional impact of thinking your child is suddenly and irretrievably in Hell is much bigger”
Right, but statistics show that almost everyone things they and their children are going to heaven. To modify something Sartre once said, “hell is for other people.”
Footnote–that table is from a 2004 study by Daniel Kahneman et al. The category is prayer-worship-meditation.
I don’t know. But the point arose as a counterpoint to your argument that the benefits of religion are to do with its practice.
It isn’t always about the practice (whatever the story of the balance of the thing is), and that’s worth pointing out.
Paul
This business about balance isn’t straightforward because of a thought about the moral significance of misery.
So you’ll know there’s this idea that there is a problem in thinking that say a tiny increase in happiness for a million people would justify inflicting severe pain on one other person (even if the balance of happiness to misery, however that might be calculated, came out massively in favour of happiness).
I think there’s the same worry here with the consolations of religion. You have to be careful about taking that away even if it seems as if it is worth it in terms of a general increase in happiness (not that it’ll necessarily come out that way).
Of course, I am fully aware there are huge complications here. Not least, you’ll reasonably point out that religion can make people miserable, etc.
But this leads on to another complication. This isn’t necessarily a zero-sum game. It’s undeniable that many horrors are perpetrated in the name of religion; and that religion frequently makes people miserable. But it doesn’t follow that one could not champion more benign forms of religion under the umbrella of accommodationism. (So, for example, I think Karen Armstrong, for all her sins, would probably sign up for this sort of project.)
And, of course, it’s true that there are huge problems and complications here too. But it’s not an idea that can just be dismissed.
The other thing worth saying is that I’m not sure that the more angry of the atheists really do understand the consolations of religion, albeit they might pay lip-service to them (don’t they tend to bandy around that word faitheist as a sort of absurd rhetorical device in an attempt to dismiss these sorts of arguments?)
There are two thoughts here:
1) It’s one thing to say one understands something. It’s another to experience it in the way that Amos describes;
2) I think people tend to be seduced by a blithe sort of counterfactual thinking; you know, so there’s this assumption that the consolations of religion can easily replaced by other sorts of consolations (I posted about this a long time ago: http://blog.talkingphilosophy.com/?p=6)
(Right, that’s it for now. Just got in from a run. Need to shower!)
Right about the blithe counterfactuals. There’s a difference between X being in principle replaceable with Y, and its being true that if a certain group of people didn’t have X, they would have Y.
Jean, could you tell me which Kahneman et al. 2004 paper that table is from? It doesn’t seem to be from “Toward National Well-Being Accounts,” from American Economic Review, as I initially thought. (That has a similar table, but I can’t tell how to interpret the table you linked to.)
Here’s the article–
http://faculty.smu.edu/jkazez/KH.pdf
By the way, sorry to (seem to) ignore so much that’s in your comments. I am absorbing much more than I have time to respond to.
From a fairly quick look at that paper, after looking at the other in some detail, I have some of the same reservations.
Kahneman et al.’s retrospective diary approach seems more likely to be usefully accurate for most things than some other methodologies—especially the typical simplistic survey questions, which are known to be pretty useless—but not nearly as reliable as the real-time beeper sampling methodologies (like Csikszentmihalyi and some others use).
Kahenman et al. acknowledge that beeper studies are the gold standard, and theirs is meant as a much cheaper rough-but-useful approximation.
I think the remaining errors are likely to be especially high for questions about religious practice. We know from other studies that Americans grossly overestimate their frequency of church attendance, etc.—by about factor of two. That’s a whole heck of a lot.
I would guess, in lieu of any real data, that they also grossly overreport the duration and affective quality of specific events of religious participation, and that would significantly affect studies like Kahneman et al.’s.
The potential is certainly there. For example, if people recall going to church, they may forget the tedium of most of the experience, and remember only the high points, especially if that fits better with their schema of the experience and their self-image, etc. I’d be suprised if they didn’t to some significant extent, and more so about religius practice than say, work or commuting or shoppig, although I suppose it could go either way.
Kahneman’s data in those two papers are not meant to be representative of Americans in general, on a random sample of days—they’re more an example of a new methodology for a specific demographic.
The data are only for employed women in Texas, on one reference day which is a work day. (I think that turns out to be interesting in detecting likely bias.)
If you look at the line in the table for “pray/worship/meditate,” there seems to me something funny going on.
Note that only 23 percent of the sampled Texas working women report any such activity on the reference day. That’s unsuprising to me—it’s a work day, after all. Even if most of them do something religious or meditative fairly often, I wouldn’t expect most to on a random work day.
Now look at the mean hours/day. It says .4 hours, i.e., 24 minutes.
24 minutes?
I’m rather skeptical that even a quarter of typical working Americans actually prays or worships or meditates for 24 minutes on a typical work day, but I suppose it could be true of Texas women. (Texans and women both being more than averagely likely to participate in such things.)
But that’s not what “mean” means in that column. The mean is across all respondents, including mostly zero values for people who didn’t report doing that at all on that day.
And there are more than 3 times as many women who didn’t report such an event as women who did.
That means that the 23 percent who account for an average of 24 minutes must each have pray-worship-meditated for an average of about 100 minutes each!
E.g., for every religious/meditative woman one who “only” prayed, etc. for an hour that workday, there’s another who did it for two and half hours, or something like that.
On a work day. WTF?
My reaction to that is not a chance. Just no way. That can’t be right, and must be grossly inflated somehow.
I can’t prove that, and I suppose that it’s just possible that a substantial percentage of employed Texas women pray or whatever for over two hours on a work day, but until I see some really, really good evidence, I have to think it’s probably false.
It’s a much better guess that the methodology doesn’t work well for that kind of question, for the same kind of reason regular survey reports grossly overreport church attendance. (It also raises questions as to just how well the methodology really works at all.)
Given such skepticism about the validity of the methodology for such questions, it’s also reasonable to guess that the reported affective states are overestimated, as well—maybe by a factor of about two, like church attendance.
If that’s true, and if the affective scale is approximately linear, as the methodology seems to assume—such that reporting means is even meaningful—then the “positive” rating of 4.35 might reflect a “true” rating that’s much lower, around 2.17.
If that’s true, then religious activity would still have a fairly positive affect rating—like everything else listed—but would nonetheless be the lowest thing on the chart, by far not one of the highest.
(That reveals that the “positive” ratings of the sixteen activities are all similarly high—the ratio between the highest to the lowest is less than 2:3. Are Texas working women on major mood-stabilizing drugs, such that sex is less than 50 percent more fun than commuting?)
But maybe I’m wrong in doing that linearly. (I’m not sure I’ve read the summation issues right.) Or maybe the overreporting is nowhere near a factor of two.
Suppose that we instead just guess that the positivity of religious practice is overreported in a fairly small way—by, say, one point on a 7 point scale.
That would only lower its positive rating from 4.35 to 3.35—which still puts it at the very bottom of the list, just below “working” and “commuting.” It’s no longer way off the chart, but it’s still the very least positive experience of the sixteen, by a very small margin.
Now suppose that we only guess that its positive rating is retrospectively overreported slightly, by an average of just one half of one point on the 7 point scale. It’d go from 4.35 to 3.85, not to the bottom but still below two thirds of the other items on the list, right between “computer/email/internet” and “taking care of my children.”
I could believe that, and I don’t think you can make a good argument it’s implausible.
If we interpret the slightly-adjusted data straightforwardly as an indicator of how people might use their time better, these women wouldn’t be any worse off spending less time on religious practice and more time with their kids, or at the computer—and they’d be distinctly better off spending less time in religious practice and more time watching TV.
I could easily believe that, and I don’t think it’s at all implausible—certainly not from the crude data presented.
Even if we don’t adjust the data at all, and just interpret it straightforwardly, even the women who chose to participated in religious practice on the reference day wouldn’t be much worse off with less religion and more TV. (A 4.35 rating vs. a 4.19—close to a washm, or at least arguably in the noise of the crude methodology.)
And—again, with no data adjustment at all—they’d be better off giving up religion and using the time just “relaxing” or “socializing.” (And they’d be a lot better off having more “intimate relations” instead.)
Of course, that’s very simplistic. Maybe they can’t give up religious practice and just relax, or hang out with friends instead. (And maybe there’s nobody around they’d want to have more “intimate relations” with. Rats.)
Religion is clearly not better than sex. It’s not even better than hanging out with friends, or just chilling out. It’s not much better than watching TV.
Of course, there are huge problems with counterfactuals here. Can you trade one thing for another? Does the static utility of something imply marginal utility—would more be better, or are you getting the optimal amount, and more wouldn’t be much more gratifying? Does the narrow utility of whether you enjoy it reflect its overall utility, in terms of how it affects your experience at other times?
With religious practice in particular, and with Texas women in particular, it wouldn’t be shocking if a lot of them like religious practice because it’s hard for somebody else to make demands on that time. E.g. it may be easier for Mommy to relax at church than to just relax at home, because at home somebody will make demands on her time that it’s difficult to refuse.
Still, the most striking thing for me about the Kahneman data is how similar the ratings are for so many activities. Only sex (“intimate relations”) stands out from its neighbors in the rankings.
IMHO, it’s pretty obvious that
(1) Kahneman et al’s methodology is still pretty vulnerable to even fairly small reporting biases—Kahneman et al. really should present a sensitivity analysis if they’re advocating this methodology as a reasonable substitute for ESM, and
(2) given known reporting biases about religious practice, in particular, we can’t draw any real statistically valid conclusions at all—the factor-of-two self-reporting biases known from other studies are larger than the differences across all activities in this one—except that
(3) the utility of religious practice is not so obviously, dramatically high that it shows up clearly and unambiguously in a crude, limited study like this. Religion may in fact be a good thing, or it may not—we can’t really tell from this kind of data, especially not by straightforward interpretation of closely-spaced rankings—but it’s not such a fabulously good thing that we can tell it’s good at all, either.
I’d like to say not like sex, which does stand out. Sex is clearly better than religion! That would at least be funny.
Unfortunately, sex is another thing that you wouldn’t be the least bit surprised to get systematically inaccurate self-reports about, and about which marginal utility may diverge wildly from static utility measures. Dang.
The thing is, I read the very same thing everywhere–the happiness boosting impact of religions seems to show up on all sorts of studies, using all sorts of methodologies. This methodology looks at what people were doing in the last 24 hours, as I recall, so the potential for under/overreporting is pretty low. In any event, you get the same result from real time beeper-sampling, I believe (now I need to dig that up–I think I read it in Csik.). I believe it’s Richard Layard who calls the the happiness boosting power of religion one of the most robust findings in all of psychology!
My favorite study along these lines is discussed by Carol Graham, in “Happiness Around the World.” She talks about a study that showed that in countries with higher church attendance, everyone ELSE is on average happier. In countries with more atheism, everyone ELSE is on average less happy. Wow! (I don’t have the citation–but the book is fascinating as a whole.)
But getting back to this study–I don’t have time to scrutinize it at the moment, but yes, it’s a study of 1000 (?) women. It’s been cited a lot to make the point that spending time with children is a relatively low-rated activity. Daniel Kahneman has a stellar reputation, so I’m not inclined to think there are stupid errors in the study. But you raise good points…so, I will look again some time.
Yes, sex is better than religion–according to the Texas women. I’m happy to let the results land where they may. Why not?
Jean:
My impression is that the large majority of such results are from the kind of survey that we should discount entirely, as likely measuring biases more than facts, especially if there’s better evidence available.
No, it actually isn’t. It’s still quite high. (Did you read Kahneman et al.’s classic 1987 book Judgment Under Uncertainty: Heuristics and Biases? There’s other work by Shimon Ullman and others around that time, that makes me think retrospective reports are pretty useless when it comes to this sort of thing.)
Many of the big basic biases that cognitive psychologists demonstrated in the 1970′s and 1980′s are exhibited in retrospective self-reports at the end of a brief experiment that the subject just participated in.
People really aren’t good at correctly remembering what happens over the course of an hour, e.g., while watching a TV show or listening to a lecture. You get a lot of primacy, recency, and salience effects, among other biases.
(They tend to remember (1) the first part, and (2) the last part, which they just saw, plus (3) a few non-representative, especially “memorable” bits in the middle. The stuff in between is much harder to remember and gets a whole lot less weight—retrospectively, people see a few brief peaks, with the broad valleys that actually accounted for most of the time mostly telescoped away.
People are not even good at remembering what happens over the course of a few minutes. Many classic experiments demonstrating primacy and recency effects only take a minute or so.
Even worse, people particularly suck at attributing their affective states to their actual causes, even in real time—it’s very easy to make people think they’re feeling something for a different reason than they are actually feeling it, and to make incorrect factual assessments based on that.
(For example, if you have hetero guys walk across a narrow bridge across a gorge, and encounter a modestly attractive woman either on the near side or the far side, they’ll rate her as substantially more attractive if she’s on the far side. They don’t realize how much lingering generalized physiological arousal they have going after they get off the scary bridge, misinterpret the generalize arousal as sexual, and think that the woman aroused them, and therefore must be pretty sexy. They don’t know what about her is such a turn-on—though they may latch onto something confabulate a story—but they do think it’s her turning them on.)
The basic lesson from a lot of results like that, demonstrating various biases, is that people just don’t have very good introspective access to this sort of thing. Unconsciously, they are guessing based on a variety of indirect clues, using a variety of unconscious assumptions, much as they’d guess about somebody else’s throught processes based on their observable behavior.
That’s why in beeper studies, the task of the beepee is generally only to immediately report how they feel, at that particular moment, without thinking about why, or whether it’s appropriate.
You don’t ask them what they’re feeling that way about, because if you do, it’ll bias their responses—even immediate responses—based on what they consciously think they feel, or would or should feel in those circumstances. They’ll adjust things automatically and unconsciously, and won’t know they’ve done it—worse, they’ll typically be sure they haven’t.
The last thing you want is for them to think about where they are and what they’re doing, because that will immediately change their thoughts and feelings, and they won’t really remember what they were actually feeling a few seconds before, when the beeper went off.
I forget who said it, but as I learned it in grad school, the basic rationale for beeper studies has two parts:
1. Don’t ask them later; they won’t remember.
2. Don’t ask them now; they don’t know.
If you don’t realize these things, and take self-reports at anything like face value, you’ll trick yourself—you’re not mostly measuring the things that people are actually experiencing, but what they tend to naively guess about those things.
Kahneman et al.’s methodology only partially addresses the first point—and arguably, not terribly well, given various strong biases demonstrated in short-term experiments—but beeper studies are just not practical a lot of the time. And it does nothing to address the second point, which I’d expect to be especially problematic when it comes to assessing religion.
I’m curious about that. I’d be interested in what the best data actually do look like.
One reason that people don’t do beeper studies more often is that it’s hard to get compliance—e.g., people who will stop what they’re doing and write things down in a diary, in church.
One of the most robust findings in all of psychology is that people tend think their own children are above average. Should we then conclude that the large majority of children are above average?
Another of the most robust findings in the social sciences is that people tend to think that their own cultures are superior, and that the central, distinctive tenets of their own religions are true, and that the comparable distinctive tenets of others’ are false.
The robustness of a finding may not reflect ground truth, but pervasive systematic biases.
That’s what I’d tend to expect of anything about religion, because religions evolved to take the credit for good stuff, avoid any responsibility for bad stuff, and make themselves seem indispensible.
(We see it all the time. Jocks thank Jesus when they win, but don’t blame him when they lose. If a horrible plane crash happens, it’s just a tragedy, but if one person survives, it’s a miracle.)
Religion seems like exactly the kind of thing you’d expect grossly biased reporting about, at every level, without anybody intending to deceive.
People are constantly told that religion is good and good for them, and have essentially zero introspective access to what they’re trying to report on. That’s a recipe for finding trends that reflect their biases more than the actual facts. Given such a bias, you will find consistent patterns in the random noise, irrespective of what else is actually going on.
Because there’s reason to question what the data actually measure, and what the implications are.
Supposing that we do accept the data at face value, as a simple indicator of value, though.
Are you happy with the fact that the data show that socializing with friends or just chilling out in private are better than praying/worshipping/meditating?
Should we now tell people that religion’s better than working or commuting, or most “chores,” but not as good a use of their free time as just hanging out, alone or with friends?
Keep in mind that if we’re talking about the straight utility of religion’s direct effects on individuals, we’re not mostly putting it in competition with work and chores and obligations, but with preferred recreational activities of all kinds. That’s some pretty stiff competition, and religion’s not looking very good.
(We need a followup study comparing religious practice to various optional, recreational activities people commonly engage in, rather than to work and commuting and shopping and child care. If these data are anything to go by, I’d expect religion to end up at or near the bottom of the list.)
I’d be happy to conclude that religion generally loses to tailgating or fly fishing or sewing bees or something that you particularly like, but I don’t honestly think this study tells us anything quite that strong, either way. (More’s the pity.)
Paul, It makes sense to worry that people are biased in the direction of saying they enjoy prayer/worship/meditation, because they’ve got a theory that says they’re supposed to enjoy them. However, I’m still not inclined to dismiss the data, for several reasons.
(1) This study reveals that people don’t enormously enjoy the time they spend in child care activities, yet they have a strong bias to say that they do. We’re supposed to enjoy being mommies and daddies! That makes me think this method can reveal how people really feel about their activities, not just what they think they’re supposed to feel.
(2) Other studies show that there’s a “dose response” effect–more church attendance correlates with more happiness. If people are just biased in favor of saying they enjoy going to church, it seems like you wouldn’t get this pattern of greater overall happiness with more frequent church attendance.
(3) Spending time in a synagogue as I do, I’m “up close and personal” with people during worship, prayer, etc. They sure do look joyous to me, and when I talk to people about this, I just get confirmation that they enjoy the activity.
(4) If you’re going to say OK–people do enjoy the activity, but that’s because they’re doing something socially approved, etc. etc. I’m going back to the childcare example. Childcare is socially approved, but studies show people don’t enjoy it as much as they “should.”
**
I’ll see if I can dig up any beeper studies of how people feel during worship/meditation/prayer. Yes, that’s the gold standard.
Jean:
Thinking about what I wrote yesterday, it occurs to me that most of the experiments of the kinds I mentioned are measuring people’s assessments of one-shot events, e.g., was that a good movie, or a good sermon, etc.
People may do significantly better on repeated events, and I think they might. E.g., if you think you like going to church, but are actually bored, you may notice that next time you go to church—”hmmm… usually I like church, but this is pretty boring.” After enough instances like that, you may bring your schemas more in line with reality.
As I understand it, though, there are still systematic distortions. For example, in gambling against house odds, your occasional modest wins are likely to be more memorable and emotion-laden than the very frequent, boring small losses, so that you weight the former too much and the latter too little, and may feel like you’re winning overall when in fact you’re losing. (Or something roughly like that.) Without a reality check, like counting your money, you may never notice you fairly consistently lose. (And if you’re compulsive, that conscious knowledge may not matter as much as the emotional sense that you’ve been winning.)
BTW, I’m not saying that the distortions would necessarily favor religious activities over others, such as watching TV or sex. The effects could ultimately go the other way, at least relatively speaking.
(E.g., overweighting the exciting bits of a movie, or the peaks in sex, as opposed to what actually goes on the bulk of the time.)
It could be that going to church is the kind of fairly repetitive thing, with relatively few memorable strong peaks, that loses in the recall-the-good-bits competition.
The same could easily be true of prayer or meditation. If they’re mostly about being relaxed and soothed, with no slick intros, exciting peaks, or dramatic conclusions (to exploit your primacy, availability, and recency biases), people could grossly
underestimate how much they enjoy them. If so, social encouragement to do it may compensate for individual failures of recall and evaluation. That wouldn’t be terribly surprising, either.
Just a clarification here: I’m not just worried about the obvious ways in which religion bills itself as the kind of thing you’re “supposed to” like. There’s a slew of biases that don’t depend on anything like that—just odd artifacts of how memory works, etc.
For example, if you have volunteer subjects do something rather tedious for rather a long time, and surprise them by paying them something for their time, how much you pay them affects how tedious they recall the prior task as being—the more you pay them, the more tedious they think it must have been.
Apparently, they don’t actually remember how tedious it was, and are unconsciously using the same sorts of clues other people would use to guess how tedious it was. If you pay them more, they infer that it was a more onerous task. If you pay them very little, they infer that it wasn’t, and was actually mildly interesting and pleasant.
(This is sometimes interpreted—mostly wrongly I think—in terms of crude cognitive dissonance, and people suppressing the actual memory of the task to make it fit the schema imposed by the reward. I don’t think it’s mostly that—I think it’s just that they can easily slot it into the pay-for-services schema and don’t waste much cognitive effort recalling anything at all, or unconsciously guess that their memory is faulty. That’s not a bad guess—external clues are often better indicators of what happened than memory, so the heuristic makes good sense.)
If the actual church or prayer experience is typically unmemorable, recall of those experiences may be especially prone to this kind of unconscious guessing. If you can get people to go at all, they may remember it as more pleasant than it is, because after all, they went “voluntarily” and probably even donated money. They may not remember why they themselves actually went, and unconsciously guess that if they went voluntarily, it was probably pretty good.
–
Unfortunately, this Kahneman et al. paper doesn’t include data for Sundays—the one day that most churchgoers would likely be in church.
That makes it a clearly weird sample, and presumably most of the religious observance in the sample is the kind of thing even most churchgoers don’t do, or don’t do much of, on work days.
One weird thing about private prayer is that if it’s particularly boring, most people will probably stop. Any tedium or other unpleasantness may be mostly self-limiting. In that case, the actual experience could actually be better than it’s remembered (at least relative to many other activities) because there wouldn’t be a lot of tedium or frustration that goes unremembered.
I’d expect church attendance and private prayer to be subject to very different biases in recall, and maybe with net effects in opposite directions. A really good study of religion’s direct effects on affect would have to separate those things out.
I don’t find this terribly reassuring. The numbers for child care could be too low, or about right largely due to luck, or even too high. If the data are as noisy as I think, and child care is the very mixed bag I think it is, none of those things would be surprising. It doesn’t make a very good cross-check on the methodology, without itself being validated.
It also doesn’t make a good cross-check for the more basic reasons I explained above—the various known biases could have very different effects on recollections of child care than on religious observance.
For example, if you’re bored or frustrated in church, you likely won’t do anything about it. Observable behavior is socially suppressed, so there won’t be key dramatic events that lead to particular moments that you’re likely to recall.
When interacting with children, boredom and frustration are likely to lead to squabbles and spats, with more observable and memorable drama. Other not-pleasant things are more likely to be memorable, too, e.g., Johnny barfing up his cereal, and you having to clean it up while worrying about whether you should take him to the doctor.
The really damning thing about this particular result in this particular paper is that I just don’t think the duration numbers for worship/prayer/meditation are believable.
IMO, there is no way that the roughly one quarter of women who do it at all are doing it for an average of an hour and forty minutes on a workday, (Unless perhaps the numbers don’t mean at all what they superficially seem to mean.)
Until that’s satisfactorily explained, I think the only reasonable thing to do is to dismiss the results as incomprehensible, and very likely very wrong. Somebody is grossly overestimating their time spent in weekday religious observance, and we shouldn’t trust their recall of its affective quality any further than we can throw it. Or maybe Kahneman et al. just misplaced a decimal point somewhere.
I’m guessing the time estimate is off by at least a factor of two, and probably more, so it’d be utterly unsurprising if the affect estimates were off by a point or so out of seven, rendering them completely useless for our purposes.
If those duration numbers are really not way, way off, they’re the most interesting result in the study, and cry out for investigation.
I can believe that there’s a correlation, but I don’t think that anybody’s actually shown it’s a “dose response” correlation.
I suspect that
(1) the causality doesn’t mostly go from “dose” (time at church) to “response” (happiness). I suspect the causality at least partly goes the other way, and is mostly due to a third factor. (E.g., people prone to throwing themselves into social activity are happier, irrespective of whether that social activity is religion-related.)
(2) This is largely an artifact of comparing apples and oranges, due to breaking things into a convenient but biased set of categories.
As I understand it, the Kahneman et al. paper we’ve been discussing is fairly typical in its breakdowns of kinds of activity. The crude breakdowns include obligatory or partly-obligatory activities like working, commuting, shopping and child care.
It’s not too surprising that religious observance, which is entirely optional, might be a bit more fun than things that are often chores that you have to do. (Even if, like childcare, they may sometimes be enjoyable for their own sake too.)
The interesting comparisons are between entirely optional, voluntary activities—e.g., religion vs. socializing, relaxing, and/or watching TV—but those are not broken down nearly enough to show the likely “dose response” effects that religion shows.
For example, I know some people who are very into sports, who get together with other people to watch games every weekend, and also watch several games a week on weeknights.
Likewise, I know some people who are very into roleplaying games, or board games, who get together with gaming groups two or three times a week to play.
I would surprised if those people didn’t show a similar (apparent) “dose-response” pattern, if the categories were suitably refined. They know what they like to do, and they do it a lot, and I do think it makes them happier. (Would that my own preferences were so clear and easily satisfied.)
I suspect that what most of those people have in common is mainly a matter of personality. They’re the kind of people who can find something they like, deem it “good enough” to invest a lot of time in doing, and not get bored with doing it a lot, for years on end. I suspect that most of them would find something else to like and do a lot of, if they weren’t doing what they’re doing.
For these kinds of purposes, we shouldn’t be comparing religion—a particular kind of optional recreational activity—with “socializing” or “relaxing,” which everybody does, or even with “hobbies,” but with particular enthusiasms and fandoms—sports fans, roleplaying gamers, low riders, bikers, ham radio operators, model plane builders, quilters, pub crawlers, etc., etc.
We don’t really need to break it down that far, but we need to distinguish between people who have a particular favored form of recreation and do that a whole lot, and people who don’t.
(We also need to break religious observance down a bit, at least distinguishing between public and private observance, etc.)
If we did that, I suspect we’d realize that religion isn’t really so special after all, in the way it seems from the usual misleading analyses. The seemingly special “dose response” effect would be largely an artifact of singling out a particular kind of optional activity and identifying people who do that a lot, but not doing the same for other particular optional activities.
This is a classic statistical trap. If you treat some salient category specially—even just making it a category—and you do not treat other things very comparably, you will most likely find statistical evidence that makes it seem special, because you miss similar patterns that would pop right out with a more appropriate breakdown.
This is a very well-known problem in psychology, especially trait psychology. How many dimensions of personality are there? As many as you go looking for. How many dimensions of intelligence? Again, as many as you go looking for.
You can make almost anything look “special”—usually quite inadvertently—if you don’t compare it to other things at the same granularity, in appropriate dimensions. The real trick is in knowing the relevant dimensions ahead of time. A whole lot depends on your theoretical model of the domain, and whether it’s close to accurate.
(This isn’t just a problem for the social sciences. It’s much more common than most people realize in “harder” sciences. A lot of my own research has been about detecting methodological problems of that sort, and re-doing stuff that had seemingly been “done to death,” but systematically done a bit wrong—and getting strikingly different results.)
I’ve spent a fair bit of time in churches, quaker meetings, and reform synagogues, and I agree that people by and large seem happy to be there and friendly.
I don’t think they’re obviously happier or friendlier than people in a lot of similarly focused secular optional social contexts. They don’t seem all that joyous, though of course they do a lot of social smiling, etc., and that’s known to feed back and actually effect mood.
I don’t really know—I don’t know if anybody really knows—but I think people are prone to attributing too much to specifically religious activities, and not enough to base rates about voluntary and/or social activities.
That’s the usual artifact of attaching a salient label, as Tversky and Kahneman showed in the 1970′s—given a few simple bits of data, people tend to ignore relevant base rates and attribute way too much to the salient specific category. It would be surprising if “religion” was not very subject to that effect.
See my earlier remarks as to why you’d expect various biases to affect those things rather differently.
Cool. Keep us posted.
Paul, About duration… There are a lot of blacks and hispanics in this sample (and they’re all Texas women). Here in Dallas, I sometimes see people who carry bibles around with them, especially black students/staff, and sit and read periodically. That might explain how it is that 1/4 of the sample spent 20 minutes in prayer/worship/meditation. So that’s not church attendance or private prayer. It’s reading your favorite bits of the bible! Dunno–maybe this is very enjoyable. I’m prepared to find out otherwise, but hey–it would be a little weird how many people do this (voluntarily!) if they didn’t enjoy it.
The results here about childcare mesh with other data as well–lots of studies show the same thing, including (as I recall) beeper studies.
A lot of your doubts are really doubts about research in psychology more generally, and would call into doubt all the findings in Kahneman’s article, wouldn’t they? I would think so.
As to “dose-response”–I do think a bunch of studies show that. The more church attendance, the higher people score on self-reports of happiness. That’s (mild) corroboration for saying people do enjoy the time they spend in church, and aren’t merely misreporting enjoyment. It would be surprising if doing X more often made people overall happier, if they didn’t enjoy doing X. (But not impossible–exercise might increase overall happiness without being enjoyable at the time.)
As to religion being special, or optimally enjoyable, or some such–I wasn’t saying anything that strong. It could very well be that various leisure activities are just as enjoyable, but they’re not interchangeable. So if someone loses faith and can’t enjoy going to church any more, they can’t just fall back on monster truck rallies, which may be just as fun for the people who go to them. (I don’t rule that out!) So “deconverting” can have costs for people. I think that was the point that got us into this entire discussion.
Maybe we are done. If so–it’s been interesting. If not…not.
How about meditation for something which is not enjoyable at the moment, but increases overall happiness?
Jean,
I can easily imagine that 23 percent of that sample spends an average of 24 minutes a day reading the Bible or whatever, but as I said before, that’s not what Kahneman et al.’s results say—the mean time is across all respondents, i.e., with zero values for anybody not in the 23 percent who reported doing that at all.
To get a mean of 24 minutes from 23 percent of the sample, the mean among that 23 percent is over 4 times higher—about 104 minutes, or about 1 3/4 HOURS a day, on workdays!
They’re apparently spending almost 9 hours of free time during the work week on worship/prayer/meditation. It’s almost like a quarter-time job.
That’s really a whole heck of a lot. Even for women, even in Texas. Even for bible-carrying black women in Texas. I could imagine that if they were retired or “housewives,” but not if they’re employed an average of 6.9 hours a day and spending another hour and a half commuting.
I don’t think it’s true, or if it is, this is a very weird sample with some very interesting things going on, and we should be very careful about generalizing from it.
If true, it would raise some serious questions—what are the costs of spending over 8 hours of precious free time during the work week on Bible reading or whatever? How much time are they spending on religion on the weekends? (A few hours on Saturday, and all day Sunday, so that all told, it’s like a half-time job?) What are the women who do that not doing, such that they have so amazingly much time for it? (You’d think that’d have to cut into times for higher-rated activities, like “socializing” and “relaxing.”)
I have to guess that the figure is somehow grossly inflated, casting grave doubt on the affective quality reports for that time as well.
Either that, or it’s mostly a matter of double-dipping, somehow. If so, it probably doesn’t mean what it seems to, and it would be good to know what it does mean.
“They’re apparently spending almost 9 hours of free time during the work week on worship/prayer/meditation. It’s almost like a quarter-time job.”
That’s the wrong inference. These women don’t necessarily spend 1.75 hours on w/m/p every day. It could be that a different 23% spend that much time on w/m/p each day.
Now that I think about it more, I would guess they’re spending that time in a bible study group (very popular in Texas). That makes better sense, as far as the amount of time goes. Bible study means you go to someone’s house, drink coffee, chat, talk about the bible–it’s pretty time consuming.
Actually, I don’t really know what’s going on, but I’m not inclined to dismiss results from such a highly respected guy in such a highly respected journal. By the way, this was picked up by lots of newspapers, not because of the religion thing, but because of the childcare results.
You are right, of course, that I was being simplistic about how to divide the reported stuff among the surveyed women.
No matter how you slice it, though, something very weird is going on. Consider the other extreme. If it’s a mostly different 23 percent every work day, that means that the overwhelming majority are doing this sort of thing at least one day a week—putting in an hour or three of whatever it is.
I don’t think that’s true, even among bible-toting black women in Texas, and it’s certainly not anything like representative of what Americans do. (3/4 of whom aren’t even in church on a given Sunday.)
The really weird thing is the distribution of the times, with over three quarters of the women reporting no religious observance and the rest reporting a whole lot on average.
I just don’t find the distribution plausible, unless the data don’t mean what they seem to. You can’t have that many zeroes and that many hour-or-twos without
(a) a very homogenous sample with very strongly bimodal individual behavior, or
(b)a very bimodal sample of people with very homogeneous individual behavior, or
(c) a mix of both, where the resulting modes happen to be weirdly similar and reinforce each other.
Neither extreme is plausible, and I don’t think anything in between is either. It’s a freakish distribution that (a) cries out for explanation and (b) pretty clearly isn’t typical of the population at large. Until it’s explained, the data are useless for our purposes.
Those are the kind of bizarre numbers that usually turn out to be wrong or somehow very misleading, even if they come from somebody with a Nobel Prize. They just don’t make any apparent sense, and if they do somehow make sense, it’s not clear how we should make sense of them.
–
I think you’re overly impressed with Kahneman’s stature. He’s a very smart guy who’s done a lot of good work—and I’ve admired him for decades—but that doesn’t fix the limitations of his methodology, or put any bounds on the remaining errors, and I’m sure he wouldn’t claim otherwise.
It also doesn’t seem to me that Kahneman et al. are staking their reputations on any particular numbers in their paper. They’re presenting it as a new metholology, applied to a particular example population, not a well-validated methodology applied to a representative population.
Some other people (e.g., Layard) are overinterpreting and over-weighting their data.
(That suggests to me that there’s not a lot of better data for them to fixate on, but I haven’t done my homework.)
Nobody who understands the area would expect the data from his methodology to be able to reliably make fine judgments, e.g., between an activity that’s a 3.4 and one that’s a 3.9 on a 7-point scale. The phenomena in question are just way too weird and difficult to measure.
Nobody’s going to criticize Kahneman et al. a lot for that, because everybody knows that you have to take such things with a pillar of salt, and look at lots of other data of various kinds to help guide your interpretations.
They’re going to be fairly charitable about it, and not call the methodology useless, partly because it’s not. It’s clearly useful for some relatively crude but important things—like noticing striking hedonic treadmill effects within each of multiple domains of experience—and only offers suggestive hints about others, like more precise comparisons between different kinds of experience.
I was just looking at another Kahneman et al. paper where they discuss this sort of thing, and I found it interesting that their first example of difficulty in measuring utility was an experiment where people put their hands in painfully cold water, and the water temperature is gradually raised. The subjects adjust a knob frequently to reflect how unpleasant the experience is at that moment, and later they also try to recall the overall unpleasantness of the whole experience.
Actually integrating the moment-by-moment reports gives systematically and strikingly different results from the retrospective accounts. People’s recollections are mostly sensitive to the extremes and the final value, and almost entirely insensitive to the durations of somewhat less extreme discomforts—e.g., how slowly the water temperature is raised near the beginning vs. the end, which has a big effect on how long they’re in fairly serious discomfort.
One interesting thing about that example is that the whole time-varying discomfort experience only lasts a minute and a half, and people have simply terrible recall of most of the immediately preceding couple of minutes. That is normal, and it’s therefore quite normal for people to misjudge experiences with moderate pleasantnesses or unpleasantnesses of different durations.
Another example they give is of reducing the recalled discomfort of a colonoscopy, by tacking on an extra minute of low discomfort at the end. Without significantly changing the total unpleasantness of the experience over time, and in fact making it take longer, you can substantially reduce the recalled unpleasantness, and make people more willing to submit to another colonoscopy in the future. Again, this effect takes place pretty much immediately—certainly on a much shorter timescale than the daily retrospective method can cope with.
It’s clear that Kahneman et al.’s recollection diary method does nothing whatsoever to address that very common kind of bias, which takes effect on a much shorter timescale. They conspicuously don’t claim it can address that ubiquitous problem, and refer to ESM (beeper sampling) as the “gold standard.”
I think it’s pretty clear that problem could introduce multiple biases in their reported numbers for worship/prayer/meditation.
If people are spending a weirdly long period of time in that “activity” (whatever it actually is) that may means there’s a whole lot of unremembered pleasantness or unpleasantness in the middle, and relatively few memorable peaks. There’s no way that the subjects can be expected to accurately recall what it was really like for most of the time they actually spent, and retrospective reports may suffer from a different amount of bias than you’d get with shorter-duration events.
That’s bad enough, but if the subjects are engaged in a non-representative kind of “worship/prayer/meditation,” we have no guide for comparing that to what other, more typical people do. If your speculation that they’re going to bible study group for several hours is correct, that may be a very poor indicator of what happens with most religious people, whose religious observance takes a different form, e.g., praying or meditating alone, sitting in church without frequent interaction with others, etc.
The sad fact is that “hedonic psychology” is in its infancy, and we just don’t have many well-established important results yet.
That’s the kind of situation in which you should be skeptical of simple and interesting new results, even from good researchers in good journals. Many are going to turn out to be more or less wrong.
A large fraction of published results in psychology are largely wrong, especially simple, exciting new ones. Replication and theoretical/methodological refinement are tremendously important.
To a significant extent, yes. Psychology is a really difficult subject, and there’s a lot of poorly-evidenced stuff out there, because really good evidence is often hard to come by. There’s a lot of stuff published in refereed journals that should be considered preliminary and suggestive, not anywhere near definitive.
A big fraction of the results in psych journals are more or less wrong, and don’t get replicated. Later studies typically show smaller effects, or no effect, and eventually people iteratively come up with better data and better theories.
Part of that is due to biases in what gets submitted or published. Negative results are mostly filtered out in various ways, and positive results get through. The net effect is to initially exaggerate many effects. Once a claim is out there, and tentatively accepted based on positive results, negative results are publishable—it’s suddenly interesting to show that the claim is wrong—and you get a more realistic balance eventually. (And often a more sophisticated picture.)
Constructive replication is just tremendously important in psychology, because seemingly unimportant variations in methodology often reveal that earlier results depended on things nobody had any idea of the importance of. (Or that some people guessed the importance of, but nobody else cared.)
It shouldn’t be surprising if the recent spate of results in “hedonic psychology” about happiness go the same route—many will turn out to be wrong, or overstated, or subtly mis-stated. It’s a very common pattern with exciting, novel results in an “infant” field.
I’m not dissing psychology by saying that, just acknowledging how the field works, mostly because it has to work that way. Doing good psychology is hard.
My understanding is that there’s a significant correlation, but that it’s actually relatively small.
People talk about it a lot anyway because there are so disappointingly few known long-term correlations between happiness and anything else. E.g., in identical vs. fraternal twin studies the effect of genetics is several times stronger, but even that doesn’t account for a majority of the variance.
It’s also frequently noted that the correlation may not actually be a “dose-response” correlation, where religion is the dose (i.e., the cause) and happiness is the response (i.e., the effect).
I personally—not unlike a number of experts—don’t think it is mostly a dose-response phenomenon. I also think that to the extent that the causality really does run that way, it may be a dangerous thing. I think there are other interesting factors in play, and nobody’s really sorted them out yet.
If I recall correctly, one of the stronger correlations in all of this is that so-called “intrinsic religiosity” is better correlated with happiness than so-called “extrinsic religiosity.” The difference is largely one of how you take your religion—is it true and deeply compelling, or is it largely instrumental. (This comes at partly from twin studies, IIRC.)
Unfortunately intrinsic religiosity is fairly well correlated with so-called “right-wing authoritarianism” as well.
In other words, the kind of religion that really makes people happiest may be the kind that the right-wing fundamentalists are selling, for deep reasons related to scary psychological traits.
To the extent that those things are matters of fixed predispositions, that bodes ill for liberal religion. To the extent that we promote religion generally, we may be considerably more effective in promoting nasty authoritarian, intolerant religion, and inadvertently make things worse.
To the extent that it’s not a matter of fixed dispositions, and people can like either, it’s still rather scary—even to the extent that people can initially buy into very authoritarian or relatively nonauthoritarian religion, the ones who do buy into the former may be happier, and more committed, and in the long run that too may promote nasty religion over nice religion.
Yikes.
I don’t think it’s very good corroboration at all, given that I think that
(a) people are pretty bad at both kinds of self-reports, and
(b) they’re likely to be biased, in their unconscious guessing, in same direction by the same assumptions. It could easily be a consistent bias with no grounding in reality.
It’s not that I think people are typically grossly wrong about whether they’re happy or unhappy, in generally or in church. It’s that I think the rather small-but-significant differences that result in these kinds of correlations are likely to be results of comparably-sized biases.
Maybe I should stress that—I think people usually can more or less detect whether one thing makes them a lot happier than another, especially in the long run, despite a lot of weird biases. (E.g., I think most chronic depressives eventually figure out that they’re really not as happy, on average, as most other people.)
Unfortunately, most of the differences and resulting correlations we’re talking about are small enough that we shouldn’t be surprised if what we’re actually measuring is mostly biases.
I’m not sure I’d find that too surprising. Consider Amos’s example of meditation, from his latest comment.
Meditation could be like exercise—something people do because they think it’s good for them, even if they don’t enjoy it much, and would rather be doing something else.
I think church is like that for many people—they do it somewhat more often than they otherwise would, because they think it’s good for them (and/or for society.) In any given instance, they might prefer to be doing some other voluntary activity.
I don’t think that’s a radical idea—I think it’s the way that many of us were brought up. People do think religion is good for them, whether they especially enjoy it or not.
It’s supposed to affect your behavior in positive ways, and color your experience outside of church in positive ways, and thus “pay for itself” even if you have to sacrifice something more fun to do it—which people often do, and I always thought everybody knew that.
I don’t get the sense that people are usually expected to value going to church just because it’s so darned fun. It may be somewhat pleasant, like a lot of things that aren’t work or commuting, but going to church isn’t mainly supposed to be about having fun in church.
Likewise, I never got the impression that praying was supposed to be particularly fun. Maybe not unpleasant, maybe mildly pleasant but not something you do for its sheer pleasantness, either. It’s something you do because you owe it to God, or whatever, and it’s supposed to build character or something, such that it affects how you live your life, with long-term, bigger-picture rewards.
That is related to the whole “intrinsic religiosity” thing. As I understand it, if you’re doing religion because it’s fun, or pleasantly social, you’re doing it wrong (in IR terms). The higher correlation between IR and overall happiness allegedly has something to do with not using religion instrumentally that way.
I don’t know if that’s true, especially because I wonder about bias in the self-reports of overall happiness by that kind of person. (I could easily imagine that they’d be especially prone to reporting that they’re especially happy, maybe because they think people like them are right with God, and that others must be less happy. Or maybe not. I just don’t know.)
If it is true, sending people to church on the assumption that it will help them and make them happy may be largely self-defeating. To the extent that they do it for that reason, it may be less likely to work—they won’t be very committed or “live” their religion in a way that affects their happiness more generally, and they’d be better off going for whatever they personally find most enjoyable in the moment.
Hello Paul,
No doubt, as you say, that the meditation “feels better” for most people because they think that it is good for them.
That’s also most probably true about attending church.
However, are even our most basic pleasures context free?
A good meal tastes better if we eat it among friends; a wine tastes better if it comes in a fancy bottle and is served by an elegant waiter.
The sensation of the orgasm seems more “satisfactory”, more “complete” if it occurs in a loving relationship than in masturbation.
That is, self-report is always or almost always dependent on how the self which reports frames the experience or the evaluative context in which the self which reports places the experience.
Sorry: I don’t have the technical vocabulary that you do, but I’m sure that you get the idea.
Paul, I will read your whole comment later, but in response to the first part–
The bible study theory does make sense of the fact that such a small fraction (23%) is spending so much time on p/w/m on a work day. Speculation: the next day, a different 23% will do the same because they’re bible study happens to be scheduled on a different night. At least that might account for a lot of the 23%…
Another speculation–what if you asked about a broader category: spending time in a like-minded group talking about major issues of life and death? That way, going to an Amnesty International meeting would be included, as would getting together with your Tea Party gang. My guess: you will get similar results. People enjoy that kind of thing a lot: what counts is like-mindedness and talking about important life and death topics.
So does that undermine my position? Not at all, because I think people are very different. If you take away bible study from those who choose it, they’re not going to revert to the AI meeting or the Tea Party meeting. Likewise, if my AI meeting is canceled, I’m not going to go to the bible study or the Tea Party meeting.
Moral of the story–if I slip atheist serum in the coffee of my bible study friends, they will simply lose one of their favorite activities, and not necessarily make up for it with an alternative.
Yeah, I ignored all your points (I did skim)…I’ll come back and read more carefully later.
I came across a Gallup report from October last year: ‘Religious Americans Enjoy Higher Wellbeing’. that suggests Americans who are the most religious also have the highest levels of wellbeing. I thought the same might be of some small interest to some of you. Here’s the link:
http://www.gallup.com/poll/144080/Religious-Americans-Enjoy-Higher-Wellbeing.aspx
There’s also an older European study which suggests that active ‘participation in faith-based organizations is not largely a function of a society’s level of trust in them’. More ‘off topic’ perhaps but I thought it interesting anyway.
http://www.gallup.com/poll/13117/religion-europe-trust-filling-pews.aspx
@Paul:
“In other words, the kind of religion that really makes people happiest may be the kind that the right-wing fundamentalists are selling, for deep reasons related to scary psychological traits.
To the extent that those things are matters of fixed predispositions, that bodes ill for liberal religion. To the extent that we promote religion generally, we may be considerably more effective in promoting nasty authoritarian, intolerant religion, and inadvertently make things worse.”
I have to disagree here, although if these statements are backed by large amounts of evidence, I’ll stand corrected. Your statement may be true for, say, hardline evangelicalism of the type entrenched in the U.S. (but again, I’m unaware of much evidence in support of it), but we’re talking about a generalized “religion” here, no? What about Buddhism, Baha’i, Chinese Traditionalism, Sikhism, Hinduism, etc. etc.? We’re talking at least one third of the world’s population, and (with some variation, of course – this is several billion people!) hardly all religions reflect or even exhibit “right-wing fundamentalism.”
I suppose what I’m saying is that I see arguments against “religion” really most often take the form of arguments against fundamentalist, evangelical Christianity of the highest order. That would be fine, but I don’t think that it’s fair to argue against a generalized “religion” via an argument about a narrowly defined belief system in a sea of wildly diverse belief systems. It might be an effective argument against Christianity, but not simply “religion.”
Perhaps I’m being too picky here, however.
Paul: If I recall correctly … “intrinsic religiosity” is better correlated with happiness than so-called “extrinsic religiosity.”
Paul, you do, of course, recall correctly. Interestingly, psychologist Gordon Allport, who is associated with the extrinsic/intrinsic distinction, also found (in his studies back in the 1950s) that those scoring high as ‘extrinsically’ religious were also more likely to exhibit prejudice. Probably a blip.
“intrinsic religiosity is fairly well correlated with so-called ‘right-wing authoritarianism’ … To the extent that we promote religion generally, we may be considerably more effective in promoting nasty authoritarian, intolerant religion”
Thomas Bouchard’s studies strongly suggested that ‘extrinsic religiosity’ is closely associated with environmental factors and ‘intrinsic religiosity’ is more closely associated with inheritance. Further, I gather he also found that tendencies towards fundamentalism are also more likely to be inherited. It seems obvious we can effectively promote religious (and ideological) ‘brands’, behaviours of extrinsic religiosity and flavours of intrinsic religiosity. But I believe nastiness and intolerance are better attributed to people than religions as such and suspect these attributes are, to a significant degree, inherited. Extrinsics use whatever the local religion is as a means to justify and promote their existing prejudices. And perhaps fundamentalists would be fundamentalist somethings whatever their local enviroment, whether it is one of Islam or Catholicism, Marxism or Nationalism or (in these times of virtual localities) conspiracy theory ‘x’.
Likewise, I never got the impression that praying was supposed to be particularly fun. Maybe not unpleasant, maybe mildly pleasant but not something you do for its sheer pleasantness, either. It’s something you do because you owe it to God, or whatever, and it’s supposed to build character or something, such that it affects how you live your life, with long-term, bigger-picture rewards.
That is related to the whole “intrinsic religiosity” thing. As I understand it, if you’re doing religion because it’s fun, or pleasantly social, you’re doing it wrong (in IR terms). The higher correlation between IR and overall happiness allegedly has something to do with not using religion instrumentally that way.
I don’t believe this is how real people think about these things, at least not the real people I talk to and come into contact with. I myself go to services at a reform Temple and know my own reasons. I talk to other people who go about their reasons. I also talk to Christian friends. Plus, I can see with my own eyes–the mood during services is joyous. I also know the rabbis personally and can compare their demeanor to other peoples’. There is a lot of joy there, and that observations fits perfectly with the psychology I read and trust. I really like Jonathan Haidt on the subject of religious experience–he’s an atheist who “gets” it.
Paul, I’m struck by the fact that this is all a bit messy, and is ghettoized at the bottom of my post from last year. Did you want me to synthesize your comments here into a new post?
Ben, When Jeremy suggested letting Paul do a post a long time ago, it made perfect sense. But he declined, and we’ve had a dialogue here. He’s spent a lot of time on it, but I have too. If you now synthesized his comments into a post and left mine behind, that would obliterate the “dialectic” that’s developed, and obviously be unfair to me.
Yes, my immediate thought was that any synthesis would be unfair on the people who have responded to Paul (primarily Jean).
I don’t think Paul has anything like made his case, FWIW.
Hamill,
I think there are several things going on in this thread, some about specifically American religion, which a lot of the data is for, and some not.
One of the themes has been possible US secularization paths, and how probable further secularization is in the US. For that, comparisons with other first-world countries, mostly with a Christian background, seem pretty relevant.
–
A different issue is what’s true of religion in general vs. Christianity vs. fundamentalist Christianity.
I do think that some things I say apply to religion fairly broadly, though sometimes with much less confidence for lack of familiarity with relevant data.
I would say, however, that it sounds like you may underestimate the extent of fundamentalism, or, more generally, conservative authoritarianism, in non-US and non-Christian religion.
Eastern religions are often not as friendly and benign as they seem from the US.
For example, there are Hindu fundamentalists, and they’re a significant political problem in India. (If you think about the traditional caste system, you shouldn’t be at all surprised that Hinduism can be pretty nasty at least some times in some places.)
Likewise, consider the Tibetan theocracy, with the Dalai Lama as the head of government as well as religion, before the Communist takeover. It wasn’t all good for everybody, with IIRC 20 percent of the population making up the religious/governing class and mostly being supported by the other 80 percent who were mostly subsistence farmers. It was a pretty backward authoritarian place, with Buddhism turned into an authoritarian power structure not entirely unlike the Roman Catholic church of old, and its leader basically deified. (Not to mention pretty much supernaturalizing Buddhism and deifying the Buddha, irrespective of what he actually intended, with an elaborate metaphysics and ontology of supernatural beings and powers.)
I think a lot of Americans have a tendency to romanticize Eastern religion, and be unaware of the ways religion is often similar around the world.
I don’t have a lot of specific scientific data or a tight argument, but I am skeptical that religion in general is a good thing, or that the nasty aspects of American religion are a peculiar outlier.
I don’t have the confidence that a lot of accommodationists seem to that religion can generally reform itself and progress into something benign, and stay that way.
I have a stronger inclination to believe that all religions exhibit a tendency toward authoritarianism if not scriptural fundamentalism, which must constantly be opposed, with varying and unpredictable degrees of success, by moderates and reformers.
That opposing tendency is natural, too, as are resulting schisms and conflicts; I’m just saying that the tendency of many people toward toward religious conservative authoritarianism is very widespread, not just a matter of specifically American or Christian conservative politics or religion.
(If you’re interested in that sort of thing, you might want to check out Martin Marty et al.’s Fundamentalism Project, studying fundamentalisms and similar authoritarian religious movements in all major religions, around the world. It’s not just a US or Christian or even solely Western phenomenon.)
–
This is an interesting issue that I think gets to the heart of a lot of the Gnu vs. accommodationist conflict.
The accommodationists often say two crucial things:
1. The Gnus criticize all religion as though it were biblical fundamentalism or literalism, and
2. The Gnus should stop dragging science into the fight against religion.
One problem is that #1 is just false, but easy to make seem true by quote mining or just broad generalization.
The Gnus do criticize specifically scriptural inerrantism sometimes, because inerrantism, etc. is popular and needs to be criticized.
That is very far from all the Gnus talk about, and
1) many of their frequent criticisms apply to all popular kinds of Christianity, and
2) their most important arguments apply the vast majority of popular religion around the world—including Hinduism and at least most popular forms of Buddhism—because the Gnus undermine the supernaturalism that almost all popular religion assumes, usually explicitly but sometimes implicitly.
(Most unpopular theologically very “liberal” religion assumes something supernatural, too, though it is often veiled. Karen Armstrong’s seemingly science-compatible apophatic religion is an example, and is fundamentally incompatible with science if you actually understand it.)
The key to the Gnu position is scientific materialism, not atheism per se, and the evident falsity in light of science, of almost all religion. (As most people understand the term “religion,” or as the vast majority of religion practice it.)
But of course that’s exactly what the accommodationists most want the Gnus to shut up about—the very strong connection they see between science and atheism. (Rightly, IMO.)
The Gnus frequently argue against supernaturalism fairly generally—most especially the ideas of
(1) dualistic human souls,
(2) disembodied souls like many religious ideas about immortal human souls or nonembodied gods
(3) gods, i.e., powerful and often disembodied minds with special powers over matter
(4) divine sources of knowledge, especially moral knowledge, via overt revelation or by supernatural mystical insight.
Their big argument against these sorts of things is that modern science seems to indicate that they are not true—if such things existed as most religious people think they do, in any form similar to what religious people think they do, science could probably find evidence of them, and since it doesn’t, they very probably don’t exist.
(I’m running together science and basic rationality a bit, here, but I think that’s basically the right thing to do. There are no boundaries between them such that science stops being applicable and “philosophy” takes over, and science normally does use basic philosophical reasoning to rule hypotheses out, rather than testing every hypothesis explicitly and directly. The compatibilist accommodationist like to make it sound like there’s some boundary the Gnus are illegitimately crossing between science and philosophy, but IMO they’re not and the compatibilists are resorting to really, really bad metaphysics and philosophy of science.)
We now have clearly better explanations of the things that souls and gods were invented to explain, and reason to believe the basic hypotheses were false.
We also have pretty good scientific explanations of why people would tend to believe in such things, whether they were true or not. That too helps make atheism—and particularly scientific materialism—more credible than supernaturalism.
(I’m not saying those particular explanations are very well-evidenced or proven—just that it’s not scientifically implausible that people would systematically believe untrue things of the sorts that religions come up with. That means that the popularity of such beliefs is not a good indicator that they’re true; we have no need of that hypothesis.)
Or so the Gnu story goes, anyhow.
–
These basic Gnu positions and arguments do apply to religion quite broadly, although the applications are often not made explicit because the Gnus are usually arguing in a social context where Christianity or western Theism is the dominant religion.
Consider Buddhism, for example, which many westerners think is very different from Western religion—and it can be—and that it can be or even generally is compatible with Science.
As I understand it, most Buddhists believe in dualistic souls of a sort, and most specific forms of Buddhism assume that.
(Even if they paradoxically believe that the soul is nothing and not a thing, and everything, all at once, and/or that everybody’s souls are really the same soul and that’s also the same thing as reality, or something like that.)
As I understand it, this is partly because early on, most forms of Buddhism sucked in Hindu metaphysics from the surrounding culture. Karma and reincarnation got in too, even though the Buddha disclaimed (IIRC) any knowledge of an afterlife.
Even the minimal forms of Buddhism that didn’t suck in a lot of Hindu metaphysics typically assume the existence of (a) a life force, which is also a dualistic, recognizably supernatural thing, and/or (b) a dualistic human soul.
As I understand it, most actual Buddhists of even fairly austere sorts do believe that the human mind has essentially magical powers—by meditation and whatnot, you can bypass the merely rational mind and get more direct “knowledge” unavailable to a merely rational mind, by what amounts to supernatural ESP of a sort.
I do understand that a comparatively small number of Buddhists don’t believe that, and think that Buddhist practice is largely a collection of mind control techniques—for beneficially controlling your own mind, not for brainwashing others—and that that doesn’t require any magic or dualistic souls. I don’t think that’s common at all, partly because materialism is not common, and Buddhism adapts to whatever’s around.
Buddhism is unusually “compatible” with other religions in that minimal forms make few strong, clear, unambiguous claims, e.g., as to whether there’s a dualistic soul—but that also means that Buddhists generally do incorporate some of the ambient basic assumptions about dualistic souls that are so common around the world, not least in the East.
I think the issue of essentially supernatural ESP is important in understanding a lot of mystical religion, both Eastern and Western.
It is very common in mysticism around the world to think that the merely material brain can only do so much with perception and cognition, and that the mind has something else—what Armstrong calls the Intellectus—that transcends merely material perception and thought, and directly apprehends Deep Truths.
E.g., it’s widely believed—by Eastern and Western mystics alike—that mystics can meditate, achieve a cosmic spiritual state, and discover deep facts about the ultimate nature of reality, by something like unmediated mental perception, or even introspection down past the level of normal matter as scientists understand it. On any scientific account of the mind, that’s just not remotely plausible. Perception and introspection are both error-prone information processing phenomena, and there’s good reason to think that such perception doesn’t exist and that if it did, it couldn’t be anything like reliable.
It’s much more plausible that when mystics around the world “find out” these things, they’re just wrong, or no more than randomly likely to be right, and that the noted similarities in mystical traditions have a lot more to do with mystics having similar brains in similar modes than that they are directly apprehending a deeper level of reality.
–
Given that the Gnus criticize very basic ideas like supernaturalism, including dualistic souls and deep spiritual insight, I think it’s quite unfair to say that they only criticize Western religion or orthodox Christianity, or even that their arguments really only apply to fundamentalists and literalists.
That’s nonsense. Apparently the critics of the Gnus don’t actually read what the Gnus write. The gnus make a bunch of different arguments, many of which address issues of very broad scope.
Acceptance of scientific materialism rules out the overwhelming majority of religion all around the world, including most theologically liberal religion in all major traditions.
Sometimes the Gnus make that explicit, and sometimes they don’t, but that’s very clearly their general position, even when they tailor their particular arguments for audiences in a predominantly Christian culture.
–
It seems to me that Mooney and his ilk—the especially anti-Gnu accommodationists—want to make it sound like the Gnus don’t make general arguments, because they don’t want people to know about the arguments the Gnus are actually making against religion in general.
Mooney apparently doesn’t want you to know that the Gnus frequently do make arguments broadly applicable to supernaturalism and thus religion very broadly, because two very awkward issues would immediately come up:
1) whether those arguments are in fact correct—does the success of materialist science undermine the plausibility of supernaturalism and thus most religion?—and
2) whether science could in fact conflict religion with religion in that very broad and direct way at all, even in principle, which they claim or strongly imply it can’t.
Even addressing #1 would be very awkward in terms of #2, so strategically, perhaps it’s best for compatibilist accomodationists not to go there.
–
For Gnus, that’s just infuriating, because it makes it sound like they’re not even making the arguments that are in fact their central themes, and always have been; the specifics about inerrantism, fundamentalism, Christianity, etc., are just important special cases worth addressing explicitly.
The Problem of Evil, for example, applies much more broadly than to only to literalist fundamentalism—it undermines all basically orthodox Western Theism.
It is a quite direct attack on the central tenets of even minimally orthodox Theism, of the sort that the vast majority of Americans do believe. Not just fundamentalists. Not even close.
It’s also an argument which most professional philosophers think is a pretty darned good argument, to which the Theists still don’t have a good counter.
(That’s why theodicy is such an amazing zoo—even theologians generally can’t agree on what’s an acceptable counterargument, i.e., a counterargument that isn’t obviously wrong.)
Yet the compatibilists accommodationists often ignore it, and other very good, basic, and general arguments that appear in all “New Atheist” books, and proceed to claim that the Gnus are really only arguing against “literalistic” religion.
Have a look at, say, Dawkins’s The God Delusion. I think it’s a bit philosophically sloppy, myself, but it’s far from being applicable only to literalism or fundamentalistm.
Then have a look at Mackies’s The Miracle of Theism—a book famous in philosophical circles for dealing with theist’s positions and arguments.
Notice that many of the points and arguments are the same. Dawkins is surely less thorough—no surprise in a pop bestseller—but he hits most of the same imporant points, in mostly of the same basic ways, as serious professional philosophers discussing these things.
That’s actually not at all bad for bestselling book popularizing key philsophical points that most professional philosophers do agree with. And however pop and sloppy it still is, it’s still it’s vastly better than most of the books arguing the other way.
E.g., Armstrong’s, Haught’s, or any of the orthodox folks theologians’ anti-Gnu books—or chapter 8 of Mooney and Kirshenbaum’s book, which is a philosophically shallow hatchet job that carefully skirts around the actual bones of contention.
The idea that folks like Chris Mooney would characterize folks like Dawkins as philosophically shallow and naive, and off-target about religion is just laughable.
Mooney specializes in being philosophical shallow and off-target, and especially in unfairly making others appear so, when they are not.
Just once, I’d like to see an anti-Gnu acknowledge that the Gnus do make the arguments they actually do make—e.g., about the Problem of Evil and the scientifically apparent nonexistence of dualistic souls—and that they apply do very broadly. (E.g., to Theism fairly broadly and religion fairly broadly, respectively.)
How many times do how many Gnus have to say those utterly basic things before the anti-Gnus will even admit they’ve been said?
Given what the Gnus’ central themes are, and what they actually frequently say about science and religion, the anti-Gnu compatibilist line amounts to:
1. The Gnus never say anything applicable to religion broadly, and
2. they should stop saying it so loudly and so often!
(Is it really any wonder people get utterly sick of Mooney et al. criticizing them that way, for literally years on end?)
Jean:
I didn’t actually decline, though I didn’t know how to write a good coherent piece right away and do justice to the subject(s)—especially given the tangles here. I did continue the dialogue and have appreciated the dialectic.
Jeremy:
You keep saying things like that, but I’m not quite sure what you mean by them.
I don’t know what you think I’m supposed to be making a case for.
I’m not mostly claiming to make a compelling case “for” anything—I agree with you that this stuff is really complicated and interpreting the evidence is pretty iffy. A lot depends on your conceptual model of the situation, and it’s pretty arguable.
I’m more making a case against things on that basis—i.e., why Gnu atheists don’t buy the accommodationist strategy, don’t think they’re just ignorant ornery assholes for failing to do so, why they feel misrepresented by Mooney et al., etc.
I’m not asking you to agree with the Gnus. I’m explaining why they don’t agree with the anti-Gnus, and that it’s not just a matter of Gnus not knowing and not caring about philosophy of science, or about political strategy, or about non-fundamentalist religion. They do have their reasons, and they’re not simply bad reasons, even if they’re not sufficiently convincing and compelling to the other side, or to you.
If you can see why they see things differently—even if you don’t actually agree with them—and if you can see why they are not terribly happy with the way they think they’re misrepresented by the anti-Gnu accommodationists, I’d think I’d been fairly successful.
It seems to me that you stress how the Gnus haven’t made their case, or that I personally have not made my case, but don’t find it as important to say that the accommodationists haven’t made their case either.
Where do you think the burden of proof lies, and why?
IMHO, the anti-Gnus frequently tell the Gnus to shut up*—that what saying is false and harmful, or at least obviously harmful, and they therefore shouldn’t be saying it—-and they do that without actually addressing the Gnus’ issues of whether it’s actually false, and whether it’s actually harmful in the big picture and the long run.
(*A note about “shut up”: Mooney likes to say he’s not telling anyone to shut up. Bunk. He’s not commanding them to, of course—how could he?—but he’s clearly telling them to, if only in the sense you’d tell a friend to shut up to keep him from blurting something that you realize would end up getting him in trouble for saying it, without stopping to explain to his satisfaction why it’s in his best interest. The point is that the Gnus stress why they think they’re right, and Mooney mostly ignores that and insists that the Gnus are harmful and should stop saying what they’re saying, without addressing the real reasons they think they’re right, both factually and strategically.)
Do you actually disagree when I say that the anti-Gnus (notably Mooney) have never made their case about actual science/religion compatibility, yet continue to say the Gnus are wrong and should shut up (in the sense above) about incompatibility?
Do you actually disagree when I say that the anti-Gnus have still never addressed Overton issues, even after years of criticizing Gnus’ actions that are based largely on Overton-style reasoning that the anti-Gnus won’t even acknowledge to the extent of saying Overton’s wrong or inapplicable?
I understand that you don’t think Overton strategies are the be-all and end-all. I don’t either.
On the other hand, I don’t think anybody’s made the case that Overton strategies are entirely useless or even clearly inferior, either. Do you?
Do you actually disagree when I say that the anti-Gnus frequently misrepresent the Gnus as only making criticisms that apply to “literalism” or “fundamentalism,” and fail to acknowledge that their central themes and arguments are in fact much broader in scope, do have a lot of currency among philosophical experts, and do address the large majority of religion (in the US at the very least)?
—
From various things you have said, it sounds like you actually agree with me about a lot of this—notably that
1. science and religion—at least common supernaturalist majority religion—have some deep and broad conflicts in pretty much the way the “New Atheists” claim.
2. Mooney, e.g., is factually wrong to say that the Gnus are so wrong about that,
3. Mooney has not made his strategic case that his political triangulation is better than the Gnus’ Overton strategies, in the long run.
(I’m not asking whether you agree with his strategic bottom line, but whether he’s made a convincing case for it against people who take Overton seriously.)
@Paul – Well, actually, I just meant that I don’t think you’ve shown there are compelling reasons to think that the happiness/religion research is not revealing something real about the way in which religious practice feeds into people’s sense of well-being!
J, if I’m not mistaken, I think Paul is critiquing the idea that religious people are happier. That’s not entirely equivalent to a discussion of wellbeing — you can have objective theories of human welfare (e.g., based on economic indicators). And it’s not equivalent to a discussion of peoples’ “sense” of their own wellbeing — after all, one of his primary points is that people can be totally mistaken about themselves.
Sure, Ben, I wasn’t choosing my words particularly carefully there. The point I was making was that my remark was only directed towards the last bit of this whole discussion (i.e., from the post that prompted me to ask whether Paul wanted to write a guest post).
I wasn’t suggesting, generally, that he hadn’t made his case.
@Paul:
“If you think about the traditional caste system, you shouldn’t be at all surprised that Hinduism can be pretty nasty at least some times in some places”
But that’s the point, no? Only at some times? In some places? Even if violent, extremist fundamentalism is the norm (and I think there is a good argument that it is generally not, at least at present), that isn’t an effective argument against “religion.” It is simply an effective argument against extremist, violent fundamentalism. Even extremist, violent fundamentalist forms of different religions have their differences! After reading your response, I don’t think this is what you were really trying to argue, but it had a tendency, at least to me, to drift in that direction.
“The Gnus criticize all religion as though it were biblical fundamentalism or literalism…One problem is that #1 is just false, but easy to make seem true by quote mining or just broad generalization.”
I don’t believe that gnus criticize religion as though it was all fundamentlism. I agree that such an accusation is false. I do, however, think that gnus have a tendency to errantly use criticism of fundamentalism as ammunition against very nonfundamentalist (and even antifundamentalist) forms of religion. There may not seem like there’s a difference between the former statement and the latter, but I believe there is, and that the distinction should be important.
“(Gnus’) most important arguments apply the vast majority of popular religion around the world—including Hinduism and at least most popular forms of Buddhism—because the Gnus undermine the supernaturalism that almost all popular religion assumes, usually explicitly but sometimes implicitly.”
I agree, but I believe that Jean has laid out a good case for the potential errancy of such arguments as effective critiques of religion above (and elsewhere).
“But of course that’s exactly what the accommodationists most want the Gnus to shut up about—the very strong connection they see between science and atheism.”
This has been a long thread, but I believe somewhere very far and away upthread I stated that I’ve rarely heard an accommodationist tell gnus explicitly to “shut up.” IMO it’s more that honest critiques of communication strategy are getting taken to the “shut up” conclusion erroneously, when there are in fact several alternative conclusions available.
“Just once, I’d like to see an anti-Gnu acknowledge that the Gnus do make the arguments they actually do make—e.g., about the Problem of Evil and the scientifically apparent nonexistence of dualistic souls—and that they apply do very broadly”
I’m will very readily acknowledge that gnus make broadly applicable arugments (as I’ve said upthread, perhaps they occasionally too broad!). However, I think that there is much, much more going on with “gnus” than simple opposition to supernatural claims. If the gnu “position” were simply that supernatural claims are bogus and not supported by evidence, there would hardly be any need for labels, like “gnu,” to define subsets of the skeptic community. Instead, I believe that much of the uproar among factions and source of the labels lies not in the message but how it is being communicated, which in a roundabout way gets back to the topic of the OP. My personal characterization of the gnu position – perhaps oversimplified for sake of argument – would instead be that supernatural claims are bogus/unsupported by evidence, and because of the strength of truth in this statement, one needn’t concern themselves with how the statement is being communicated. The latter portion in italics seems to be where much of the disagreement and uproar festers from. “Accommodationists” (if that is a correct label. I have always considered it as having to do with philosophical compatibility but it appears to be used for much more) want gnus to consider the varying, nuanced methods of communication and carefully choose an approach that minimizes collateral damage, while gnus want to be free to state the truth however forcefully and bluntly they wish, because it is the truth. I think there are arguments in favor of both positions, but the difference is far from a simple position about materialism and the validity of supernatural claims.
” The anti-Gnu compatibilist line amounts to: they should stop saying it so loudly and so often!”
But I think that’s an oversimplification of the counter-gnu position (I won’t use “anti-gnu” because IMO that language implies a good-evil dichotomy that I don’t see as wholly present in the debate, at least not consistently). There are much deeper issues at play in terms of communication than “stop being so loud” and “stop talking so frequently.” The issue seems much more nuanced and complex than this, perhaps by several orders of magnitude; however, I believe it often erroneously gets simplified down to the statement above, and the debate suffers because of it.
Paul:
As Ben says, you’ve made a fairly good case that people can be totally mistaken about themselves, but what you haven’t shown is that religious people tend to make more mistakes about themselves than non-religious peoples.
Amos, I don’t know if Paul needs to do that. Part of what makes the social psychology of bias so interesting is that it often poses a challenge for just about all of us — including experts who ought to know better. It’s just incredibly easy to engage in (counter-)motivated inferences, to see patterns where there are none, to confuse the familiar with the ordinary, to assume that our own experiences are the benchmark of normality, and so on. In other words, we’ve all got buggy hardware — it takes a lot of effort and training to compensate. So it seems to me that it’s not quite right to say that Paul has to prove that the religious suffer from a special handicap. Rather, those speaking on behalf of the religious would have to prove that they are specially gifted at avoiding human errors.
On first blush, I’m sure that must sound like terrible news for a great many scientific studies, which fail to control for these biases. But I don’t think either Paul or I want to dismiss evidence that falls short of the mark. You just have to be careful about how you go about interpreting the data, and coming up with tentative explanations. (That having been said — and I’m only speaking for myself now — I absolutely do dismiss anecdotal evidence.)
Ben:
When you say that you absolutely do dismiss anecdotal evidence, I assume that you are just referring to happiness or similar studies (sadness studies, for example) based on such evidence.
Actually, Paul was giving reasons to think that religious creates special biases. In the study we’ve been discussing, people have to rate their activities in the previous day. If you believe it was admirable of you to go to bible study, that might bias your assessment of how much you enjoyed it. That’s not the only bias we have, of course, but it’s a bias not present when you’re assessing how much you enjoyed being with your friends the previous day. I think he was in fact saying that some of the evidence for religion increasing happiness reflect this sort of biased remembering (and other religion-specific biases).
But actually, there is research that gets around that issue. In fact, there are at least three types of research that shows a positive impact for religion.
(1) There’s research on how much people enjoy religious activities at the time (that’s what we’ve been focusing on).
(2) There’s research that studies whether religious people are happier overall, not just at the time they’re worshiping, praying, etc. The measure here varies–it’s sometimes “satisfaction with life,” sometimes “happiness,” sometimes various health measures are used. Here there’s going to be much less of a problem about biased remembering of what church was like (because that’s irrelevant).
(3) There’s also research about whether the number of religious people in a society has an affect on the happiness of everyone else. Amazingly enough, the answer is yes–see Carol Graham’s book “Happiness Around the World.” Here the bias in question can’t play a role, because the happier people aren’t religious.
As to anecdotal evidence: we’ve been talking about the science of happiness, but of course we also have our sense of what adds to happiness (or doesn’t), which influences our decisions in everyday life. I don’t think there’s anyone who doesn’t draw on “anecdotal evidence.” In fact, our anecdotal evidence is one of the things that colors how we react to research. I very much doubt that Paul reacts equally skeptically to all psychological studies. How one reacts, when you demand more evidence, is colored by your own experience. Your own experience is obviously composed of lots and lots of “anecdotal evidence.” We don’t go around doing controlled scientific study to get our own personal “folk theory” about people. And everyone does have a folk theory about other people.
Whoops, I messed up italics. Just “was” (first sentence) was supposed to be in italics.
I guess what puzzles me about this discussion is that normally, what people report about their experiences is taken to be “true”, even though we know that it is not scientifically precise.
If I say I was happy during my vacation, others accept that, even though my passport was robbed, I got sick, the beach was overcrowded, sight-seeing bored me, and I spent hours in traffic jams on very hot days.
That is, everyone knows that when X says that he is happy attending church or football games or rock concerts, X is relating a narrative about his experience, not calculating the quantity and)or intensity of pleasure vs. pain.
I need to clarify my position in a couple of important ways.
I am not claiming that the direct effects of religion individuals don’t tend to make them somewhat happier on average. I wouldn’t be the least bit surprised if they do, a little. (And a bit more than a little for some people, and not at all for some others.) I don’t claim to be able to quantify it, or to really defend even approximate quantities.
What I am saying is
(1) That’s only part of the story of whether religion in general makes people in general happier, on the whole, and IMHO not the biggest part. Indirect effects of religion are huge. Religious people may have “positive illusions” that make them somewhat happier, and that may be good for them, but they also vote, indoctrinate their kids, etc., which may be bad for others, and IMO a net loss for society on the whole. (Or if I’m wrong, a net win that’s also likely more important than the direct effects on individuals.)
(2) I do think that the extent to which religion’s direct effects make individuals happier is often overrated—people can’t really know how happy it makes them, with any precision, and tend guess too high for any or all of several reasons. That really could go either way, I must admit—it’s also quite possible they tend to guess too low when recalling how much they enjoy religious practice, at least relative to other things.
For example, watching TV may seem retrospectively like more fun than it actually is in real time, and be more addictive than actually fun.
In behaviorist terms, TV gives you a “variable ratio reinforcement schedule,” which people (and other animals) are notoriously vulnerable to.
By contrast, meditation and (some common forms of prayer) may make little memorable impression at all, even if it’s consistently fairly pleasant, but not dramatic; its rewards would then tend to get retrospectively telescoped down to insignificance. It might be more pleasant on the whole than watching TV, but you might never remember that, or not in an emotionally motivating way.
In behaviorist terms, meditation gives you something like a “fixed interval reinforcement schedule,” which is notoriously not very reinforcing for the amount of reward actually obtained.
The difference in how much different basic reward schedules reinforce behavior is one of the most famous and robust results in all of psychology, and it IMHO it means that all of these kinds of retrospective studies are clearly methologically weak. If they don’t take into account the reward schedules—how intensely peaked the rewards are, and how those peaks are distributed—you can generally expect substantial errors when comparing activities with different basic reward schedules.
They are ignoring one of the strongest, most basic, and best-known regularities in psychology. IMHO, that puts the burden of proof squarely on the people promoting the methodology. How could basic reinforcement schedules not matter when comparing many varied activities with clearly different temporal distributions of rewards?
That typically is a significant effect, so why not here?
(3) Most of the psychological studies of these things just aren’t very convincing, for reasons that the researchers themselves often point out but then subsequently either fail to address, or which get lost in the popular reporting. The effects they show may be real, or may not, and their magnitudes may be off significantly. Even small biases may change the relative orderings of the things people have to choose between.
(4) One reason I’ve talked so much about this is just that the Kahneman et al. paper that Jean linked to, in particular, doesn’t impress me nearly as much as it does her. Unfortunately, I think that’s led us down a path where we’re overemphasizing the issue of how much we should trust a straightforward interpretation of that particular paper, and maybe losing sight of somewhat bigger and more general issues.
I think that particular paper has some apparent problems that make it bad for our purposes. It’s got methodological weaknesses and an clearly non-representative sample with regard to some of the very issues we’re talking about.
If you want a good example of the methodological controversy about these subjects, google up articles about the Danes and their happiness or life satisfaction.
The Danes notoriously have the highest average self-reported life satisfaction in Europe, by a significant margin. In a variety of studies, they’re a substantial notch above the Swedes, who lead the rest of the pack by a small margin, and the rest are also fairly closely-spaced.
(If I was confident of such data, and simple interpretations of them, I could confidently say they show that less religious first-world countries are clearly happier than more religious ones. Sadly, I have serious reservations about results I find appealing, as well as those I don’t.)
A lot of researchers just don’t believe that the Danes are that much happier than other Europeans. They point out how weak the methodology is, and attribute the consistent and striking results for the Danes to any or all of a variety of biases.
For example, Danes may have a smaller range of expectations about happiness, and thus rate somewhat less intense happiness than others do at the top of a point scale. (Given their very egalitarian culture, one wouldn’t find that very surprising.) Once they do that, their rankings of less extreme happiness are naturally biased upwards, relative to others’, by well-known and predictable bracketing effects.
If that’s true, the higher life ratings Danes give may not reflect higher actual happiness, or even higher perceived happiness—or even higher recalled happiness at all.
They may only reflect somewhat more modest expectations about extremes of happiness it’s plausible to achieve.
I’m pretty sure nobody knows if that’s what’s going on.
Given that it’s so easy to question such a seemingly clear result, which shows up in a variety of studies—the Danes really do score significantly higher, whatever it actually means—you really ought to wonder what other kind of errors could likely get in there, and how you would ever know.
It’s quite plausible that the Danes are actually utterly typical of Northern Europeans, and not ahead of the Swedes at all. But if that’s true, and comparable biases affect the other countries’ results—even to a much lesser degree—there’s a much bigger problem. Any of the numbers could be substantially off in either direction, and unless the Danes are the only people subject to some significant, relevant bias, some probably would be, changing many of the relative orderings—and we’d never know.
The only reason people latch onto the Danes and question those results, in particular, is that the results are so striking that many scientists simply find them implausible. (Many Danes, too.)
The implications of that are huge. If the methodology is so untrustworthy that experts must question the most obvious and striking results because they’re so obvious and striking, what should they do with the modest and seemingly more plausible results?
Obviously, they shouldn’t trust those either—if you can’t trust a consistently big difference to be real, you certainly shouldn’t trust a smaller or less consistent difference to be real. The methodology is clearly woefully imprecise, as far as anyone can tell.
Suppose that the cause of the Dane’s high ratings is just that they have a somewhat narrower range of expectations, relative to people of other countries.
The same kind of problem could happen when measuring different groups within a country, such as the US.
In fact, if those different groups do have somewhat different expectations, on average, you would expect that they would give somewhat different answers about their current happinessautomatically.
More even-keeled people, with less extreme ideas about what constitutes great happiness and great unhappiness, would automatically tend to give more extreme ratings of their actual happiness. If their answers were significantly above the midpoint of the scale—as most people’s clearly are—they would then tend to amplify that by their somewhat lower high-end expectations, and give significantly higher scores.
Consider the implications of that.
Small but seemingly significant differences in means are likely if groups have differences in their typical range of expectations, which could in turn depend on any number of things.
People with a broader range of emotional experiences quite plausibly would give lower ratings for the very same degree of perceived or recalled happiness. People whose friends and family have more variance in their happiness might do the same, if they bracket their expectations relative to what they observe happening to others rather than themselves.
Likewise, more socially conscious people might bracket their experience relative to whatever broader groups of people that they tend to compare their lot to.
Similarly, people who tend to think about concrete situations might bracket their happiness relative to rich and beautiful vs. poor and ugly people, as they imagine them, while more psychologically-oriented people might bracket their happiness relative to depressive people vs. happy people, irrespective of their concrete situations.
Now consider the very common religious practices of being concerned with the less fortunate, and counting your blessings.
What are the plausible effects of those?
One plausible effect is that it actually makes people happier—they calibrate their concrete expectations about income, health, etc. relative to the less fortunate, and are pleased to be what they then consider concretely fortunate, and it actually makes them happier with what they concretely do have.
But does it really? (I personally think that effect is overrated.)
Another very plausible mechanism is that thinking about the less fortunate and counting your blessings makes you calibrate your expected happiness differently—lowering your expectations about how happy people are likely to be, and making you more satisfied that you’re as happy as you are. Without making you happier—or even making you think you’re happier—it makes you rate your happiness higher, just by shifting the scale you use to write your perceived or recalled happiness level down.
Either seems plausible, and able to account for higher self-reported “happiness” for religious people, but they are entirely different things. One involves making people happier, or think they’re happier, and the other does nothing but affect which box they they think they should check, for the amount of happiness they’re reporting.
Either interpretation is consistent with the given data, so the data don’t give us any indication at all which is actually going on, if either.
That is a well-known problem in the field—there is no known way to scale the results appropriately relative to people’s differing ranges of expectations. Worse, there’s generally no way to tell what kind of bracketing the subjects are doing—is it relative to an imagined range of concrete situations, or to an imagined range of happiness per se—and either way, whose situations and happiness are they judging their own relative to?
The fact that is problem is not mentioned, or not stressed, in many papers doesn’t mean that it’s small, or that it has been solved. It’s neither small nor solved.
Unfortunately, even beeper studies are subject to these problems of bracketing—even in the moment, we just don’t know how people are judging their felt happiness relative to a range of expectations, and choosing a box to mark. Avoiding the problem of inaccurate recall doesn’t solve the bracketing problem, and raises more questions—do people bracket the same way when recalling as when judging their instantaneous happiness? It’s quite plausible that they do not—and IMO rather implausible that they do—so that the bracketing bias varies in an unknown way on recall, on top of the other inaccuracies of recall.
(At a guess I’d think beeper studies are more accurate in that regard, too, but I’m not sure.)
Amos:
I don’t think I really need to show that. What I’m mainly saying is that it’s quite plausible that religious people tend to be some differently mistaken about religion—maybe even less mistaken—or that individuals may be differently mistaken about religion than about other things, such that the numbers could be wrong in either direction. (As I argued above, they could be less mistaken about religious practices than about, e.g., watching TV.)
With respect to the Kahneman et al. data specifically, I would be delighted to take the data at face value and simply conclude that people enjoy religious practice more than they enjoy watching TV, but less than either socializing or just staying home and relaxing. I could then recommend that people give up religion in favor of socializing or just chilling out.
But of course I don’t trust the data, or the simple interpretations of them, and even if I did, it’s not simply a matter of which activities are simply “better”—a whole lot depends on what is or isn’t interchangeable, what people would or wouldn’t actually interchange, and what bigger-picture effects they all have. (None of which is very clear, and I realize Jean has raised issues I haven’t addressed yet.)
Funny you should mention that. One of the last couple of articles I read about hedonic psychology starts off with that—recollections of vacations—as a motivating example of why you should not take such accounts as true, because they are systematically unrealistic. People tend to remember the distinctive and interesting things about vacations, which are typically fun, or seem in retrospect like they would have been fun, and forget the other stuff that actually makes up the bulk of the experience.
(E.g., you might remember seeing some awesome piece of art, but forget how tedious the long wait to see it was, and that you didn’t really enjoy it as much as you should have because you were distracted by needing to pee.)
I’m not sure what you’re saying here. I think people do often discount others’ recollections of such things substantially—e.g., a friend’s raving about how fabulous Paris was doesn’t generally get me to book a flight there—but they also often don’t discount such things as as much as they should.
It does seem that we agree that retrospective reports of happiness are just not very reliable, which is the main thing.
If what you’re saying is that everybody knows that retrospective reports of affect are very unreliable, at least in terms of total pleasure derived, and that it’s odd that people make an exception for “scientific” data that clearly don’t, I couldn’t agree more. It is odd.
There seems to be a certain methodological difficulty involved in researching how much people enjoy some religious activities at the time. You cannot really expect a Buddhist in meditation or a Christian or Moslem in ‘selfless’ prayer to report on how much they are ‘enjoying’ the experience at the time. They cannot simultaneously be having the experience and reporting on it. As with other non-religious activities (say having sex) all, it seems, you could do is ask them afterwards. There is perhaps a bias problem related to religious activities – somebody might be reluctant to admit it wasn’t much fun – but that seems true for other activities, such as the aforementioned case of sexual congress. In some respects though, and I speak as an entirely unreligious person, asking how much they ‘enjoyed’ their praying seems somehow off the mark. It doesn’t seem quite the same thing as, say, watching a movie or going on a night out. It may not be so much a ‘pleasurable’ experience as such but one that leaves the individual feeling more at peace, happier, than they were before. Indeed, it seems there may be religious activities that are not pleasant at the time. Consider say confession for a Catholic, I can imagine that, like a secular therapy session, an individual may not enjoy talking about their moral failings at the time of doing so, but may feel unburdened, happier if you like, after the event. Prayers of petition, for a dying relative say, may also not be ‘pleasant’ to engage in but may leave the individual in a a more peaceful state. And perhaps, for persons involved in deep scriptural study – like somebody struggling with Hegel – it may not be so much a case of being ‘pleasant’ at the time, but what there may be is a satisfaction ‘happiness’ at what has been achieved during the session of work. All that said, again as an avowed atheist, watching say the Southern Black Baptists in enthusiastic singing worship – they certainly look to me like they are enjoying the experience immensely.
James
Jeremy:
Oh.
Well.
Oops.
BTW, does my last comment about reinforcement schedules and the expectations/bracketing problem (e.g., with Danes) make it clearer why I am seriously concerned about odd and unpredictable biases in these methodologies?
Different cultures place differing values on being happy.
I’m Jewish and I live in Chile. Jews, especially those from Yiddish-speaking backgrounds (oy vey), tend to consider sorrow as accepted daily currency, while Chileans, although less than other Latin Americans, value seeming to be or appearing to be happy or upbeat.
It’s more complicated than that, and it’s basically unconscious: I’m not claiming that Latin Americans “pretend” to be happy.
However, given that the United States is a multicultural society, do happiness studies consider cultural differences as a relevant factor?
Jean, I’ll let Paul speak for himself, since he hardly needs my defence.
Thanks for the references. About Carol Graham’s citation — although I can’t tell for sure, she seems to be referencing a working paper by Clark and Lelkes (2009): http://www.pse.ens.fr/document/wp200901.pdf. I look forward to reading it when it passes through peer review.
Amos, for the purposes of that comment, when I said “anecdotal evidence”, I don’t mean studies that draw on testimony. I mean just saying offhand, “in my personal experience, generalization (x) is the case”. When it comes to certain hot topics, that’s pretty much the most problematic sort of statement you can make. Offhanded generalizations shouldn’t be considered to be sufficient grounds for the rational persuasiveness of x as a datum, unless it is our only basis for opinion.
Anecdotes are a first resort, since in some benign sense all experience is anecdotal. Anecdotes are also a last resort, since sometimes you won’t have any other sources of evidence at hand. But when you have other options, anecdotal evidence is just a distraction.
Ben:
I understand what you’re saying, but doesn’t that stance make it difficult for those with a radically different vision, for those who are outside the established order of intellectual discourse, to be heard?
After all, between a pile of sociological studies and, let’s say, one outsider, we would necessarily always accept the version of the sociologists.
Yet experience shows us that often formal disciplines neglect the viewpoint of oppressed groups or outsiders, for example, of women or gay people or blacks until fairly recently.
Complex mechanisms often exclude outsiders from access to formal disciplines, from outright discrimination to lack of resources to “old boy” or “old girl” networks, to just not “fitting in,” etc.
I don’t want to romanticize the outsider stance, but I sense that there are creative people who are outside the formal intellectual disciplines and they tend to give
anecdotal evidence.
In fact, this post is merely anecdotal evidence, and you can challenge me, asking me what studies, what papers back my reflections.
Amos, I don’t think I’m excluding anyone who has a viable standpoint. True, the point of my remarks above is that I want to exclude people who have radically deluded world-views; i.e., people who believe in things that are completely against the evidence. But there’s no reason to think that people who have a non-standard epistemic standpoint are necessarily deluded. As history has shown us, sometimes a non-standard standpoint can explain the evidence better than all competitors, and other times they offer yet another reading of ambiguous data.
Maybe it will help to see what it would take for me to be wrong. If I’m wrong in my remarks on anecdotes, then it would mean that we should seriously reconsider strong scientific evidence every time that we encounter somebody’s idiosyncratic generalization of an episodic memory. But that would be utterly batty.
Ben,
If Carol Graham is relying the study you linked to, that could be a problem.
That study uses data from the European Social Survey, which includes data from some former Communist (Soviet Bloc) countries.
From a quick skim, it looks to me like they ignore the enormous significance of this and just throw Eastern and Western European data together. (I could be wrong.)
I need to read more closely, but I’d suspect that most of the “majority atheist” countries/areas that have especially low happiness are former Soviet Bloc countries, and that the greater happiness of somewhat less atheistic countries just means that Western European atheists are happier than Eastern European atheists. No surprise there.
If that’s what’s going on, it would be ridiculous to generalize that atheists are unhappier among atheists than among non-atheists.
Paul, Yes, that’s it. I’ll read it too. There’s lots of juicy stuff in there.
Jean,
Table 5 lists the 26 countries they included, and I think it’s pretty damning if you really look at the numbers.
The average life satisfaction for all of them is 6.5 on on a 11-point Likert scale.
If you order them on that number, you get this list:
8.5 Denmark
8.0 Switzerland
7.9 Finland
7.8 Sweden
7.7 Norway
7.6 Austria
7.6 Netherlands
7.5 Belgium
7.5 Ireland
7.3 Spain
7.1 United Kingdom
6.9 Germany
6.8 Italy
6.8 Slovenia
6.6 Israel
6.5 Czech Republic
6.4 France
6.3 Greece
6.2 Poland
6.1 Turkey
6.0 Estonia
5.8 Slovakia
5.6 Portugal
5.5 Hungary
5.1 Russia
4.3 Ukraine
There are 11 countries that score at least half a point higher than average, and every single one of them is a Western European country. The top 5 are Denmark, Switzerland, Finland, Sweden, and Norway.
There are 6 countries that score a half a point or more lower than average; half of them are former member states of the USSR (Russia, Ukraine, Estonia), and two of the other three are former East Bloc countries (Hungary and Slovakia).
No surprise there. It’s exactly what I’d expect.
If they’re ignoring the huge differences between the post-Soviet East and the West, lumping them together, and looking at correlations between atheism or religion and happiness, as they seem to be, the paper is garbage.
How can this paper really be “garbage”? They read this paper at three conferences. They thank the top guy in the field (Ed Diener) for his comments. Carol Graham cites their results in an excellent, meticulous book. Both authors have great credentials (if you look at their webpages. On the other hand, the paper is not published yet, from what I can see. The right attitude seems to be to find it interesting, but not treat the findings as fact….except all the survey data. That’s going to be hard to challenge.
I don’t see how including soviet and eastern bloc countries can be the whole reason for the pattern. One of the findings is that “The religious are more satisfied with their lives in pooled multi-country European data, even after controlling for age, income, education, labour market status, health condition, marital status and country.” And country!
Does the spillover effect also hold up, regardless of country? They seem to be saying so. They’ve tested this across regions within countries, not just across different countries. It seems like that would take care of your objection. It might be necessary to see more data than is in the tables and figures to be sure.
One thing I found interesting is that the Danish are as religious as they are–64% go to church, 44% pray, 60% belong to a denomination. This goes against the idea that the Danish are the poster-kids for happy atheism.
You know, Jean, I’m getting pretty tired of your arguments from authority.
I don’t care what anybody else thinks or says. I can read papers like this and do basic math. I’m pretty sure I’m good at it. I have refereed comparable papers by Nobel-level scientists, and made the same kinds of methodological criticisms, and been thanked for them. I once overturned 30 years of research, including scores of studies, in a certain narrow area in my field. I did that by pointing out that people had been perpetuating invalid methodology pioneered by a Nobel-level scientist, an indubitable genius who made a mistake way back then, and other people followed him off the cliff for decades. He wrote a letter saying I was right, which helped me get tenure in a top department in my field.
I know what I’m doing, and I really, really don’t like arguments from authority. I’ve proven authorities wrong a number of times—I’ve made something of a career of it—and I am utterly unimpressed.
So back to the actual data…
The table I mentioned was Table 6, not Table 5. (Sorry.) It’s the most useful data in the paper.
The data after that aggregate East and West together, and don’t account for the effects of communism or its collapse, so most the correlations they’re talking about are not at all what they seem to be.
If you aggregate data that way, and interpret correlation as causation, you are implicitly blaming atheism for Communism and its collapse.
You can make an argument that that’s appropriate, if you want, but without a really good argument to that effect, the interpretation they offer is nonsense.
For our purposes, their conclusions are worse than useless, unless you think that secularization in the US is likely to follow the Eastern European pattern—including domination by a totalitarian dictatorship, and eventual collapse.
If you want to argue for that, by all means go for it.
I think it’s far more likely that secularization in the US will follow rather more like the Western European pattern. Extrapolating from post-Soviet Eastern European data to the US is just not a reasonable way to understand the US at present, or to predict its future.
So back to Table 6.
Consider just the East Bloc countries, which shouldn’t be directly compared to the West, for reasons that should be dead obvious by now.
There are only three East Bloc countries that have average happiness ratings or slightly higher, for Europe. (But obviously significantly higher than the East Bloc average.)
One of them is the Czech Republic, which is the least religiously observant, by a substantial margin. But another is Poland, which is the most religiously observant by a substantial margin. The third is Slovenia, which is only moderately observant.
At the other end of the scale, for the East Bloc, we have the decidedly unhappy Russia and the utterly miserable Ukraine. Russia has moderately low observance and Ukraine has moderately high, by East Bloc standards.
That should tell you a whole lot, right there. There are simply too much variance and not enough data points to draw conclusions about whether religiosity is even interestingly statistically related to happiness in the East Bloc, much less draw much more subtle inferences about whether having atheistic neighbors makes atheists unhappy.
It’s quite clear that East/West effects on happiness are much more dramatic than religiosity’s effects, at least in the East—and the East is clearly where the numbers about atheist majorities are coming from.
All these data show is that it kinda sucks to live in the ruins of the Soviet empire, even decades after it collapsed.
Nobody is surprised by that. Everybody who was paying attention learned that long ago, and it wasn’t the least bit surprising even then.
The rest of the data analysis is straining at gnats while ignoring the elephant in the room, while it’s stomping on the gnats. It’s utterly clueless garbage, and in no way valid support for Social Capital theory and all that.
As to my “arguments from authority” and your being an authority yourself…
No, I’m making no arguments from authority. I’m just relying on credentials in the way it’s rational for people to rely on credentials. (Social epistemology is all about that…) I am a philosopher, not a psychologist. I read a lot of positive psychology because I’ve taught some and written a bit in that area, but I’m not an expert. Therefore, when I read a technical article, it is eminently sensible of me to pay attention to credentials. It’s definitely not an “argument from authority,” in the sense that I’m saying “X has such and such credentials, therefore X is right.” The reasoning is rather that I have to assign an article some level of being “worth taking seriously” based on who wrote it, who cites it, where it was read, where it was/wasn’t published. On all those measures, this article is worth taking at least somewhat seriously.
Now, obviously all those considerations aren’t going to be outweighed, for me, by your judgment, for the simple reason that you’re anonymous and I don’t know what your credentials are. I asked you who you were way up, and you didn’t tell me. It’s clear from the way you write that you have lots of psychology background, and went to graduate school, etc. etc. All along, I’ve assumed that, but no more. Now you say you’re actually a tenured professor in a top department. Truth be told, I had not made that assumption, for various reasons. Now that I know that, it changes things just a tad…but honesty, just a tad, because you do remain anonymous. I’m sorry, but anonymity has both its advantages and its disadvantages.
I’ll read the rest of your comment when I have time. One of my problems is that I’m no expert on statistics, so I’m quite happy to read what you think. I’m just a bit leary of such a total dismissal as “utterly clueless garbage.” If you’re going to judge it that way, I think you actually owe us an “error theory”–an explanation how it is that such stupid stuff is being read at major conferences and taken seriously by people like Ed Diener and Carol Graham. Care to speculate?
Paul
That sort of thing is right on the edges of what is okay at Talking Philosophy.
Be nice!
I am not sure that Paul owes anyone an error theory unless and until the paper cited above actually passes through the rigors of peer review. In the context of a debate over what science tells us about human happiness, there’s not much social epistemology to discuss until we get our hands on actual science.
Ben, The paper has been read at three conferences, cited and commented on by the leading people in the field. If you look at the lead author’s website you can see that he’s refereed hundreds of articles for top journals and publishers. Even if the paper hasn’t been published in a peer reviewed journal, it would be odd if it were “utterly clueless garbage.” At least in the philosophy world the equivalent paper (3 conferences, cited by top people, etc) would be very unlikely to be utterly clueless garbage. People read and take seriously unpublished philosophy papers all the time.
Jean:
I’m sorry, but that sounds a lot like an argument from authority to me.
I’m not asking you to be swayed by my judgment. I’m not asking you to care about my credentials, which I only brought up because credentials seem important to you. After all, touting the credentials of the people I’m disagreeing with—in response to an actual argument I’m making—amounts to an ad hominem against me, by invidious comparison. (However unintentionally.) I thought I’d point out that I’m not a clueless noob, although that shouldn’t matter, if my arguments are right. I think they are.
I’d really rather those credentials were entirely irrelevant to you. If I was a random undergrad making these particular arguments, they’d be just as good and should be just as persuasive.
I’m only asking you to take the arguments seriously, on their own merits, irrespective of who presents them.
–
The arguments I’m making are not complicated, and should not be difficult to follow, even without any sophisticated understanding of statistics.
(Especially the crucial East/West thing. It’s not rocket science.)
So just think about it a bit… If religion per se had no effect on happiness, but the whole Soviet empire/collapse thing had happened, such that the East was more atheistic and also less happy, what would you expect to see?
You’d expect the people in the East to be less happy, and you’d have to expect more unhappiness on average in more atheistic countries, just because there are more atheistic countries in the relatively unhappy East.
Again assuming religion was unrelated to happiness, you’d also expect relatively happy or unhappy countries to exhibit a variety of religiosities, and relatively religious or irreligious countries to exhibit a variety of happinesses.
That is exactly what we do see—within the East, we see relatively happy countries of remarkably variable religiosity—including the very most religious and the very least religious countries—and relatively unhappy countries of moderate and variable religiosity, as well.
Likewise, among relatively religious or relatively atheistic countries, we see a striking mix of happiness values.
For example, the two countries with the lowest church attendance are the Czech Republic and Russia, which have the second-highest and second-lowest happiness values. The two countries with the highest church attendance are the Ukraine and Poland, which have the very lowest and third-highest happiness values, respectively.
If we go by prayer instead, it’s roughly similar and not flattering to religion. The least prayerful countries are the Czech Republic and Estonia, with happiness ratings near the top and the middle, and the most prayerful countries are the Ukraine and Poland at the very bottom and toward the top.
–
OK… after writing the above, I decided to just type the East Bloc data from Table 6 into a spreadsheet and regress it.
It turns out—no surprise to me—that using data from that paper, within the former East Bloc, countries’ life satisfaction ratings are negatively correlated with both church attendance and prayer.
If we assume that a correlation implies causation of happiness by religiosity, as the paper does for Europe as a whole, it turns out religion causes negative happiness in the East, by either measure that they show.
Interesting, no?
I typed in the data from Table 6 for Western European(not East Bloc) countries, and it turns out that in the West (as well as in the East) countries life satisfaction is negatively correlated with both church attendance and prayer.
In either the East or the West, it appears that if the correlation implies causation, religion is causing negative happiness.
Yes, I get the argument about the soviet and eastern bloc countries and it sounds good. I also noticed what you’re saying about those different countries with high religiosity & low well-being and vice versa…and was puzzled. But here’s the difference between you and me. For me it’s like I’m watching a tennis match. The article authors hit the ball to you, you hit it back to them…well, now I want to hear what they would say back! I don’t know what they would say, because I don’t have the expertise to figure it out. That’s where their credentials come in. Because of them, I figure they would surely have something to say, and I want to know what it is before I make up my mind.
Now I’m curious what happens if you take all the east block and soviet countries out of the data. Do you then lose all the spillover effect? Or do you see it still? What if you look at any one country and compare regions (which they’ve identified). Do you see their spillover effect? This is the kind of thing I’m not clear about, and feel like I should both read the paper again and hear their defense before making up my mind. I’d love to see this discussed by them somewhere.
Jean, to be clear, we’re not talking about what people within a field say to each other about their work and what they take seriously. Rather, we’re talking about the conditions under which a person outside of a discipline can justifiably defer to the persons within it.
If an outsider is going to justifiably defer to an insider in the context of a conversation about science, then they should defer to the best the field has to offer. In this case, I’ve suggested, papers in peer review are the best that science has to offer.
The way I see it, the only other options are to:
a) abandon the pretence of science (i.e., by reaching Clark et al.’s conclusions on more philosophical grounds, or by relaxing the standards of science down from peer review),
b) abandon any attempt at rational persuasiveness (i.e., by admitting that Clark et al. have a weak argument, but your personal hunch is that there might be something to it),
c) abandon deference altogether (i.e., by tackling Paul’s arguments head-on).
To some degree, you seem to be doing both (a) and (b). And that’s fine. It’s just an extremely modest position.
“At least in the philosophy world the equivalent paper (3 conferences, cited by top people, etc) would be very unlikely to be utterly clueless garbage. People read and take seriously unpublished philosophy papers all the time.”
Sure. But I can think of at least one paper in professional analytic philosophy that has passed through peer review, was produced by one of the top names and in one of the top journals, and yet is complete and utter compost. (Actually, this happens all the time in science, too — the difference is that the scientists retract their mistakes.) I’ve also heard accounts of the opinions of scholarly reviewers that suggest that these reviewers trip over their pants every morning. Presumably, if that’s the state of peer review, then the state of conference review is not going to be any better. But I’d still advise the deferential dilettente to defer to blind peer review over any other measure.
Paul W., it looks like what you are doing with Table 6 is plotting rows of the “Life Satisfaction” column against the corresponding rows in either the “% churchgoer” and “% prayer” columns, possibly getting very zig-zaggy line plots, and then calculating a straight-line fit to the points on those plots. Basically, that’s just naively plotting a measure of happiness against a measure of religiosity without controlling for any other factors that affect happiness. By contrast, the paper’s authors treat “Life Satisfaction” as a function of several different variables, namely individual religion, regional religion, income, labor markets, etc., and determine correlations based on that multivariable model.
Furthermore, the fact that the model already accounts for such things as income, markets, health, and so on, means that several of the effects of former Soviet communism are already taken into account.
I just ran the numbers for the whole table, and the result reminds me I need to back WAY up and clarify something A LOT.
—
I can’t test their hypothesis separately for East and West, using the data they present in their paper—which is basically what you’re asking for, I think.
Their hypothesis is not that religion is positively correlated with happiness at the national level.
Which is a good thing, because that’s obviously false; religiosity at a national level is generally positively correlated with bad things—poverty, inequality, social dysfunction, dissatisfaction, and unhappiness. More religious nations are generally more screwed up.
The usual response to that is not to say that religion causes those things, but to assume it doesn’t, even a little bit, and that those things instead cause religion, or that some third factor causes both—religion obviously can’t cause poverty, inequality and misery, can it?
That is generally not a statistical argument, but one of intuitive plausibility.
However, when people find a positive correlation between religion and something good, it’s quite common to assume that religion is quite important and does cause that thing, often all by itself—any degree of correlation is assumed to reflect causality to that extent—and not to invert the relationship and say that good things cause religion, as bad things do.
In other words, heads religion wins, tails it doesn’t, and maybe even wins. (Because it’s often implied that if negative things make you turn to religion, it must be benefit you to do so.)
–
Now we get back to the numbers I just ran, for the whole table. They too show a negative correlation between happiness and religiosity by either measure, without needing to separate out East and West.
I was a little surprised by that, because I thought that the numbers I was familiar with, for mostly Western European countries, would get seriously messed up by throwing in the relatively unhappy and relatively atheistic post-Communist East.
They don’t get as messed up as I thought, though, which is interesting. The negative correlation between religiosity and satisfaction still shows up, even if you do lump East and West together.
But that in itself isn’t necessarily inconsistent with what those people are saying—they are not claiming that religion at the national level isn’t correlated with bad things. They just assume the causality doesn’t run that way.
The usual way this kind of thing runs is that people note the correlation between religion and bad things at the national level, but discount it entirely—religion isn’t causing bad things.
Then they notice a correlation within national populations that religious people are on average a little bit happer, and don’t discount it at all—religion is causing good things.
If we were even-handed and had no preconceptions about whether religion’s effects were good or bad at either level, and just assumed it was causal, with correlation equaling causation, we’d come to a clear conclusion:
1. Religion makes individuals happier than other individuals in their countries, but
2. its indirect effects tend to make everybody in the country less happy, and that effect is large on the whole.
In other words, religion people gain an advantage at the expense of the general good.
I’m not saying that’s true. I’m only saying that it’s consistent with the general trends—and that those trends therefore offer no evidence that religion is good on the whole, unless we interpret some correlations as conveniently non-causal and others as conveniently causal.
Those conveniently pro-religion assumptions may be true, but they’re not supported by the statistics—they are a matter of a particular asymmetric causal interpretation of the merely correlational statistics, which in themselves say no such thing.
That’s what scabs my ass about most discussions of the supposed statistical evidence of the benefits of religion—the statistics just don’t say that, the particular interpretation does.
That interpretation might be right, or mostly right, but there are other causal interpretations of the very same data that make religion look pretty darned bad, and no simple statistical way to choose between them.
Maybe religion really is good, but maybe what we have is a massive case of confirmation bias–counting the hits and ignoring the misses.
We’re assuming our conclusion—that religion causes good things it’s correlated with, and doesn’t cause bad things its correlated with.
—
That brings me to the paper in question.
WTF?
I really need to look at the paper much more carefully, but if I understand what they’re basically saying, I don’t see how it’s possible in light of the above kind of interpretation—that the mostly bad national effects are not caused by religion and the mostly good intra-national effects are.
They seem to be claiming not only that religiosity makes religious people happier, but that their religiosity makes irreligious people happier too.
If that were true, you’d think it would show up in the national statistics—national average happiness levels would be boosted by religious people, if religion tends to boost irreligious fellow citizens’ happiness, too.
But it clearly doesn’t, at least in terms of basic correlations—religion is correlated with lower happiness at that level.
Something is wrong here, but I’m not sure quite what yet.
I suspect that part of what’s going on may be that that their conclusion depends on exactly what measures they choose—e.g., what counts as an atheist and what counts as a deleterious percentage of them—so that the countries that qualify come disproportionately from the unhappy East.
They also presumably depend on exactly what they do and don’t factor out—e.g., income, and how that interacts with the East/West thing or other poorly understood regularities or randomnesses.
Paul W.: “I can’t test their hypothesis separately for East and West, using the data they present in their paper–which is basically what you’re asking for, I think.”
No, what I’m saying is that it looks like you are just plotting “Life Satisfaction” vs. “% churchgoer” or “Life Satisfaction” vs. “% prayer.” If that’s your approach, then it’s just wrong and isn’t a serious test of the paper’s thesis.
J.J.:
(BTW, that “what you’re asking for” was not directed at you, but at Jean. I hadn’t seen your comment when I wrote that part, and wasn’t missing your point to that extent, at least not at that moment.)
You are quite right that I “naively” scatterplotted (no lines) and (linearly) regressed the data to see the simplest trend.
However, it’s not clear to me that the “controlling” the authors do is any more appropriate, without a very careful explanation of the causal model that makes the controlling clearly fair.
I certainly don’t think that you should count on controlling for income, etc., to correctly cancel the effects of Soviet Communism, which had a whole lot of weird effects with a whole lot of downstream consequences very relevant to the study, and quite poorly understood.
There’s just no way to do that appropriately without a validated model of the various effects, and that we certainly don’t have. Not even close. Even if everything else about their paper is right, it seems to me that that invalidates it right there.
As a basic sanity check, the authors should at least have separated out East and West as I did, to see if their results are an artifact of combining them. If they controlled well for the right effects, they’ll get the same results by doing East and West separately, and seeing what effects stay or go away.
Until they do that, I can’t believe their analyses. East and West are just too different, including being fairly grossly different on most of the interesting axes of their data. One size can’t fit all.
I have another major concern with the controlling you mention for things like income and health.
Suppose you believe, as I do, that irreligiosity isn’t just correlated with higher income, greater health, etc., but partly causal of it, if only very partially.
(E.g., irreligiosity might be mostly a side-effect of more egalitarian and less dysfunctional cultures, but also partly a cause of it, if only to a small extent.)
Now suppose we straightforwardly control for income and health.
Whoops… there go the positive effects of irreligiosity on income and health, and indirectly on happiness.
But wait. Every time you control for something, you’re at least implicitly assuming that it’s not caused by the things you’re comparing.
Unless you know that income and health are not even partly caused by irreligiosity, such controls assume that possible benefits of irreligiosity do not exist, and factor them out.
That’s just not fair. It’s assuming part of the conclusion that religion is good, by assuming that irreligion is not responsible for the stuff you factored out.
I need to look at the paper a lot more closely, but I just don’t see how they can get around that.
I agree that my statistics are grossly simplistic—that’s why I’m not reporting any numbers or suggesting anybody take them too seriously. I don’t think they prove much of anything—mostly that there’s certainly no trend the other way in those very basic numbers. If you want to find a trend favorable to religion, you have work harder.
On the other hand I don’t see how you can justify the assumptions of fancy factorings-out, either.
If you actually understand the paper and these issues better than me, and think the methodology is well-justified, and doesn’t assume its conclusions in the above statistical way, do feel free to explain it.
I will look at it more closely.
Paul W
You do seem to know your stuff about research methods and evaluation. You are absolutely right: you have to discern the relationship between irreligiosity and religiosity and the goods that contribute to happiness – you can’t just factor those goods out and make simple claims about irreligiosity or religiosity ‘causing’ happy societies.
I find it plausible that particular individuals can find happiness in or from religious practices or activities, even the claim that this good cannot always be fully substituted for by non-religious practices and activities seems unobjectionable.
Still superstition, putting revelation over reason, blind acceptance of bad dogma and deference to bad authorities (whether they are in Rome or Moscow, the Bible or Das Kapital) – all of these things hold back science, and therefore health and wealth – if you just LOOK the damage to scientific progress and life-satisfaction caused by religion (and pseudo-religion) is plain obvious.
Religious dogma also prevents self-interested rational behavior – say using contraception when you’ve as many mouths as you can feed. You can see the cost to ‘life-satisfaction’ that Papal edicts have in Africa, however happy particular individuals may be at communion. Prima facie, Religion does NOT cause happy societies – its withering and watering down does.
I’d really like to know if they would get their “spillover” result if they looked at smaller subsets of the data. I agree with Paul W. that breaking it down into east and west would help show whether there’s any real spillover effect. I’d also like to see the results country by country. If living among religious people increases well being for both the religious and the non-religious, then you’d expect to see that when you look at the data from one country at a time. At least you’d expect to see it in many countries, not just when you look at the entire pool of Europeans.
Now, here’s what puzzles me — maybe they did actually compare people in different regions of the same country? I don’t know how to interpret this passage on p. 8–”The country dummies pick up any differences in cultural context (e.g. the way in which subjective questions are answered) and other institutional factors. The religious effect is then identified from differences in religiosity between regions within the same country.”
Maybe someone can explain.
It seems that it might be interesting to compare East and West Germany to see how the effects of living under Communism work in what could be seen as one single culture.
Sorry for the rant my last post descended into.
The American Sociological Review (December 13, 2010) carries an article ‘Religion, Social Networks, and Life Satisfaction’ (Chaeyoon Lim/Robert D. Putnam) that I thought might interest some of you. It is freely available online.
They take “the positive association between religiosity and life satisfaction” as “well documented”. But:
“Our findings suggest that religious people are more satisfied with their lives because they regularly attend religious services and build social networks in their congregations. The effect of within-congregation friendship is contingent, however, on the presence of a strong religious identity. We find little evidence that other private or subjective aspects of religiosity affect life satisfaction independent of attendance and congregational friendship.”
I (as demonstrated) am not qualified to usefully comment on its methodology or validity but did think the conclusions might interest others than myself.
Ah, the link might help:
http://asr.sagepub.com/content/75/6/914.full
Jean,
Their data are for sub-regions of countries—something like 3 to 16 regions of 26 countries.
This does let them compare more and less religions regions within countries, and that’s the main novelty of their results.
Their basic shtik, as I understand it, is to control for major variables like income, health, and education level (see p. 8), and see how the religiosity of a given region—not a whole country—affects happiness of both religious people and irreligious people in that religion.
(Where by “happiness” I actually mean retrospectively self-reported “life satisfaction,” with all the potenential bracketing problems, etc. which they can’t control for.)
A big question is what it really means when they compare regional happinesses after controlling for things like income and education. Income and education are negatively correlated with religiosity.
If the correlation is only due to lack of education and low income causing religiosity, or some other factors causing both, it makes sense to control for those things. On the other hand, if the causation runs the other way, and religiosity causes lower education and lower income—which it clearly does in some kinds of cases—it doesn’t. That’s factoring out benefits of irreligiosity that in turn affect happiness.
Consider the simple time cost of attending church and bible study classes. Where does that time come from? If it comes out of the time that would otherwise be spent watching TV, it may be a pure win for the individual.
But if part of the time might otherwise be invested in, say, studying that will eventually result in a better job, it may be a net lose.
A common assumption is that religion more than pays for itself, economically—e.g., if it takes time and money, that will somehow be more than paid back by improved work ethic or whatever.
I for one don’t buy it. I think religion tends to be costly in several ways, and when I look at the negative correlation between religiosity and education or income, I think the causality goes both ways.
Their statistics seem to simply assume that religiosity is a pure win—that it doesn’t cost anything in terms of, e.g., education or income or health that in turn effects happiness. That’s obviously not a safe assumption.
IMHO, the last paragraph of the whole paper is by far the most important. In it, they acknowledge that they focus on benefits of religion, not costs, and they haven’t paid any attention to “opportunity costs”—e.g., whether whatever time and energy and money people put into religion might have otherwise been spent on something that would increase their happiness more.
They make it sound like it’s probably no big deal, and that those costs are probably not very big—in fact, they can’t “clearly identify” any. (Which is why the paper up to that point apparently assumes they don’t exist at all.)
I think that shows a stunning lack of imagination. I can sure think of costs of religiosity, including well-known financial costs to countries that tend to benefit their more religious regions.
In some Western European countries, for example, some churches are directly subsidized by the government, with a cut of individuals’ taxes going to their nominal religion—whether or not they’re observant or even believing.
Churches naturally spend more of that money where there are more religious people. (E.g., building and maintaining more churches where there are more poeple who go to church.)
Any investment like that is bound to have other regional effects—some of the money paid for religious stuff in a region circulates in the economy of that region, being paid to whoever provides goods and services, being re-spent by them, etc.
It’s also bizarre to assume that the “social capital” of religion is not largely political capital that benefits the religious (and more religious regions) at the expense of irreligious (and less religious regions). Political power is largely a zero-sum game. Whatever organizes people better and gives them more say in government tends to automatically take power away from somebody else, with many downstream effects.
(E.g., imagine a region with a dominant religion whose members tend to unite in support of a particular party, which is also thus regionally dominant, and more effective in national politics. If that religious support benefits that party, and the party benefits that region, where most of its constituents are, then religion benefits the region—including non-religious people who get a bigger share of the national budget, etc.)
I can’t say how large these effects are, or even which way they go. (I can certainly imagine such religiosity undermining political effectiveness, too.)
Still, it’s just implausible to simply assume the the effects are nil.
Paul, What I’m trying to understand is this–could their spillover result all come from lots of facts of this sort (I’m making this up): when regions of country X are compared to regions of country Y, greater religiosity correlates with greater happiness.
OR are they only comparing regions of X to regions of X, and regions of Y to regions of Y? The charts and tables make me think all regions are compared to all regions. The quote –
“The country dummies pick up any differences in cultural context (e.g. the way in which subjective questions are answered) and other institutional factors. The religious effect is then identified from differences in religiosity between regions within the same country.”
–makes me think only regions within a country are compared to each other. If so, wouldn’t that get around the problem you raised about east vs. west?
Re: controlling for other variables. You’ve almost convinced me this is a huge problem, but then it cuts both ways. They control for health, but studies in some countries show a positive correlation between religion and health, and health and happiness. So controlling for it masks the positive impact of religion.
Irreligiosity may correlate positively with wealth worldwide, but within affluent western countries, I think that’s not the case. So in this particular study, controlling for wealth may not mask the positive impact of irreligion.
Opportunity costs–well, maybe, in some countries. There are lots of cultural factors. In some places, church is where you solidify business relationships. In others, you waste time on incantations instead of taking your kid to the doctor.
It seems like you’re going to get your best understanding of religion by doing a lot of within country studies and then separately looking at giant regions like Europe.
Jean, Amos:
Have a look at Figure 2 of the Clark and Lelkes paper.
It shows the distributions of regional religiosities within each country in the data set they used. The vertical length of the rectangle represents the spread of religosities between regions for the top and bottom quartile marks. (I.e., the middle half.)
Most Western European nations show a relatively narrow range of regional religiosities. Most former East Bloc countries show a bigger spread.
(BTW, ES is Spain, not Estonia, which is EE. Ukraine is UA, Germany is DE. Sl is Slovenia, not Slovakia, which is SK. CH is Switzerlannd.)
It shouldn’t be suprising that the biggest variation in regional religosities is for Germany (DE). The more atheistic East Germany was only reunited with the more Christian West Germany two decades ago.
Because of that, you wouldn’t be surprised if Germany turned out to be especially important, one way or another, in the comparisons between regions in given countries. You can expect the Western regions to be very different in terms of happiness and in terms of religion, irrespective of whether religion causes happiness.
If you took Germany out of the data set, I woldn’t be surprised if the averages changed substantially. I can’t tell though, because they don’t break the data out in any way that reveals the importance of particular countries.
Now consider the general themes of the paper—having religious people around is good, even if the majority is of a different religion from you. Apparently religious people are pretty tolerant, and don’t tend to harm religious outsiders—they positively beneift them.
Now think about Eastern Europe.
When I think of Eastern Europe, religious tolerance is somehow not the first thing that comes to mind. When I think about religious tolerance and, say, Poland or Russia, I think of antisemitism.
So let’s look at the statistics for Jews in Poland and Russia…
Funny, there aren’t any. I seem to recall their being millions and millions of European Jews.
Where on earth did all those millions of Jews go?
Oh yeah, they went to the U.S., or to Israel, where they’re the majority, or to anyplace they could get to other than Eastern Europe.
Or they went into ovens.
That matters. Even if religion has changed a lot since then, it matters to the statistics now.
Look at the statistics for Poland, for example. If you want to look at religious tolerance, you have to take into account the fact that Poland is just overwhelmingly Catholic. There aren’t enough Jews left there to tell you whether they’re tolerant. They’re gone.
That points up a very general problem of measuring religious tolerance.
Measures of religious intolerance are pretty much self-limiting—if intolerance is high enough, people leave, are pushed out, or are killed, or countries split, and then there are fewer people to complain within the resulting countries.
If you want to know how tolerant religious people are, you have to take that into account, which these kinds of statistics make no effort whatsoever to do.
Clearly that matters with the Jews—religious tolerance in Europe may appear higher than it actually is, because millions and millions of Jews don’t fill out European Social Survey forms any more.
Now consider more recent events in the East Bloc, such as the conflict in the former Yugoslavia, and the resulting fragmentation of that country into a bunch of countries.
What do the statistics for, say, Serbia and Bosnia tell us about religious tolerance?
Nothing at all, because those people were too busy, too recently expelling and killing their neighbors of different religions, and dealing with the aftermath, to fill out ESS forms.
Even if we had the data, what would they say?
Well, there probably wouldn’t be a lot of Muslims in Serbia complaining about Christian intolerance of Islam, even though Christian Serbs committed genocide against Muslims pretty recently in Bosnia.
I think that makes it a bit clearer how they can throw Eastern Europe into the mix and still come up with results that suggest that Europe is a religiously tolerant place.
What’s largely going on is that where you find a lot of happy people of the non-dominant religion, that’s not mostly a fact about how tolerant that religion is. It’s a fact about that religion in that country being unusually tolerant.
Less tolerant countries with the same dominant religion mostly don’t show up much in those figures, because the people they don’t tolerate have largely left, or even been killed.
That’s really not painting a fair picture of religions’ tolerance.
As usual, heads religion wins, and tails religion doesn’t lose—even when, on reflection, it quite clearly deserves to.
Ha! That’s a great point.
Paul:
If I ever need a private detective, I’ll call you.
Jean,
I’m not ready to comment in much detail on the specific statistics they use—I’ve mostly been too busy trying to understand what the data can and can’t possibly actually show, and what their assumptions are going in.
I need to refresh my memory about multivariate regression issues before I say much definite about the stats and what they might actually reveal, even if they’re technically correct. (As I said before, I’m not actually very mathematically gifted. I usually get by OK teasing out assumptions that reveal that that fancy statistics don’t and can’t prove what they’re purported to prove.)
E.g. “Except where religion is so intolerant that it grossly distorts the statistics in favor of tolerance, it’s moderately tolerant.”
You’re right that it definitely cuts both ways. No matter what you “control” for, or to what extent, you’re going to flatter some theses at the expense of others.
Basically, we’re effectively talking about fudge factors—exactly how and how much you “control for” this or that, based on which quantitative causal assumptions.
That said, I need to pay closer attention to the technical issues and try to figure out in what sense they’re really “controlling for” what, and what that’s likely to actually mean.
Wish me luck.
My impression is that national irreligiosity correlates positively with national average wealth, but maybe only up to a point, but also correlates with low variance of wealth—i.e., more secular countries are more egalitarian, with fewer strikingly poor people, relative to the mean—and that has positive effects all the way up the scale. If I’m not mistaken, the combination of of increased wealth and/or decreased inequality tends to cause greater happiness in more secular countries, all the way up the scale to the very least religious countries.
(In the West, anyhow. God knows what’s really going on in the former East Bloc. It’s a zoo.)
That’s all assuming that you believe the statistics about self-reported happiness or satisfaction, of course, and I have big reservations about things like expectations and bracketing.
I’d be delighted to conclude that irreligion causes reduced inequality, and that in turn increases general happiness, but (a) I’m putting an debatable causal interpretation on correlations, and (b) the correlation between equality and happiness could mostly be a bracketing effect in reporting.
Dang.
I particularly have reservations about differences in expectations between Eastern and Western Europe—you really wouldn’t be surprised if people who lived in the East Bloc under Communism have ingrained generally lower life expectations than others, but then you wouldn’t be surprised if they had unrealistically high expectations of capitalism and democracy, and are typically disappointed at present. Depending on just how those opposed factors affect affect bracketing, they could make people report either higher or lower life satisfaction than Western Europeans. What you shouldn’t assume is that they’ll report the same happiness the same way. They probably won’t, to some unquantifiable extent. It’s a basic stumper.
Right, it’s going to vary. A lot. And really weirdly, which is why I’m skeptical that you can just toss a bunch of data into a regression equation and get anything like a sensible answer.
Right. To even get better idea what the relevant questions might be, about regional religiosity, it’d be a great thing to display all the regional statistics on a map. There’s no substitute for looking at geographic effects geographically.
I’m pretty sure that you’d see interesting, thought-provoking patterns popping right out, which would give you a handle on which assumptions might be especially questionable in which countries, etc.
One thing I’d really like to see is a map of regional contributions to the specific conclusions—e.g., are the regions that most strongly support the generalizations in certain countries, and are the ones that undermine the theses different from West to East, or what?
It’d be especially interesting to see data for within-country comparisons and across-country comparisons displayed visually.
For example, if you’re trying to factor out national-level effects so that you can directly compare a region in one country against a region in another, displaying a map of regional values may give you a big clue as to whether you screwed it up—if the national boundaries pop right out, that may tell you something about unaccounted-for national-level effects.
And of course if you just map fairly basic raw data, you’d notice holes immediately, like the lack of data for Bosnia and Serbia. Then some of the things I said above about intolerance effects distorting the statistics would occur to a lot of people.
This area needs a lot of basic exploratory data analysis before trying to come to interesting conclusions.
I’m not doing anything all that special. It’s much easier to be a competent detective if just you remember not to ignore murders, especially if there are millions of them.
Paul, I wish I understood the statistics better. I sleep with a statistics book next to my bed. Sadly, I mean that literally. The book sits there. I sleep. When the authors talk about “country dummies” what comes to mind is guys in pick-up trucks with “Guns ‘n’ God” bumper stickers. But I’m sure that’s not what they mean.
How boring, we’re getting to the point of agreeing. You’ve pretty much convinced me that there’s a huge amount to worry about here, before accepting what they say about the spillover effect. I would really like to see all the details…now I’m on the lookout for a thorough book about religion and well-being.
One little thing about inequality–Carol Graham’s book is full of marvelously counter-intuitive findings. Like how in the US, inequalities bother nobody but liberal, rich people. Sadly enough, it really might not be true that irreligion causes greater equality causes greater happiness. There’s a great deal in there about low expectations having a wonderfully (?) salutary affect (as you sort of say).
One more thing about how this paper is unpublished, so not peer-reviewed (in response to Ben). It was read at 3 conferences, which presumably means it massed through peer review. At least that’s how philosophy conference papers are chosen. I do wonder, though, why it never made it into a journal. If I get energetic, I may just have to email the author and find out the answer to that and other questions about the paper.
If the paper’s results are true, it’s certainly horrendously bad news for atheists. People shouldn’t live near them, and they shouldn’t even live near themselves. Eeek.
That income equalities in the U.S. bother nobody but rich, liberal people seems weird to me.
If they were to say that they bother nobody but liberal people,
I would accept that, but not all liberals (or people who are left of center) are rich or even upper middle class.
Everyone whom I know in the U.S., perhaps not a representative cross-section, is liberal or left of liberal and none of them are rich and although all of them are middle class, in general they could not even be considered upper middle class.
What’s more, they seem to have lots of friends, of similar socio-economic level, who share their political posture and thus, are bothered by income inequality.
They include a lawyer who works for NGO’s, computer programmers, a librarian, a few secondary school teachers, a journalist, a psychologist, an ex-university professor, a retired private school teacher, etc. Not hardcore poor to be sure, but hardly rich.
Jean — I said “I’d still advise the deferential dilettente to defer to blind peer review over any other measure.” (italics added)
Jean, Ben,
I wouldn’t put too much stock in peer review, either, in the short term. It’s a whole lot better than a lot of things, but still quite fallible.
Studies have been done that show that a substantial fraction of results in peer reviewed scientific journals are basically wrong—and I’m sure the fraction in psychology is not especially low. (We’re not talking a very few percent, but low tens of percent.) The data are faulty, the analyses are making invalid assumptions or whatever, and/or the interpretation is somehow bogus. The effects supposedly demonstrated don’t exist, or are much smaller than claimed, or even demonstrably go the other way when you get the big picture.
It’s really not uncommon for there to be a flurry of psych papers on some subject, and then a decade or so later, a flurry of new papers showing that the old papers were basically garbage.
For example, in the 1960′s, after Kennedy’s assassination, there were various papers that tried to explain “flashbulb memories”—how people remember in detail where they were, who they were with, and what they were doing when they heard about a momentous event. (E.g. Kennedy’s assassination, or the Pearl Harbor attack, or whatever.) There were seemingly speculations as to how and why such an ability would have evolved, etc., etc.
A while after that, there was a flurry of new papers showing that flashbulb memories basically don’t exist, or are at the very least wildly exaggerated as to the amount of detail and accuracy.
Some clever psychologists who were skeptical of the first flurry of papers looked up various kinds of records that would pinpoint where some people actually were, and who they were actually with, and what they were actually doing when Kennedy was assassinated, and tracked those people down and asked them what they remembered.
Many of them remembered vividly where they were, etc.—they seemed to have classic flashbulb memories—and they were generally wrong, e.g., not even in the same part of town as they “remembered” being, or with different people doing soemthing else. Their memories weren’t in fact much more detailed or accurate than other people’s, but they thought they were lots better.
That sort of thing can make you pretty cynical about psychology, but the fiasco was a good thing in the long run. It partly motivated important studies that reveal how memory typically works, and what the effects of recalling events repeatedly are. (With great significance for things like “recovered memory syndrome”—the “recovered memory” phenomenon turned out to be mostly crap, but this time people noticed it sooner because of the similarities to “flashbulb” memories.)
We now know that people’s memories are generally quite fragmentary and schematic, and that when they recall an event, they’re actually unconsciously filling in a whole lot of blanks with what they’re unconsciously guessing probably happened from the few and fallible clues they actually remember. Then they tend to remember the details they unknowingly unconsciously made up last time, and the more detailed memory seem more definite and reliable each time. They forget how uncertain they were the first time, and the amount of detail they seem to remember makes them unconsciously guess that it’s a clear and therefore reliable memory. After enough times, it’s very clear, very detailed, and seems very reliable.
I think the right attitude to take is usually one of skepticism about new and sexy results. Peer review usually works, often in the short run, but fairly often only in the long run—plenty of bad, mistaken papers get published in peer reviewed journals, and even whole faddish slews of them sometimes, with an apparent “expert scientific consensus,” only to get overturned a two or three years later, or even two or three decades later.
In psychology, it’s not rare for big trends to follow that pattern on a timescale of several decades, e.g., Behaviorism, which was a reaction to the pseudoscientific aspects of Freudianism that had become orthodoxy, only to be undermined by cognitive psychology, because hardcore Behaviorism was an overreaction and ran way too far the opposite direction.
oops… “seemingly speculations” should have been “seemingly well thought-out speculations.”
(They were clearly speculations, not just seemingly speculations.)
One might take a look at psychological studies of the “causes” of homosexuality during the last 40 years to see
how things change.
By the way, I also wouldn’t put too much stock in somebody being acknowledged/thanked for reading and commenting on a paper.
If somebody takes the time to read and comment on a paper, it’s just good form to thank them for it.
It doesn’t necessarily mean that they agreed with you, even a little bit, and it often doesn’t mean that you made the changes it would take to convince/please them.
(I’ve certainly been acknowledged in papers I thought were pretty bad, and acknowledged comments I thought were pretty bad. I’ve also read published philosophy and psych papers and books where people thank persistent critics for their comments, clearly without coming to agreement on the obvious bones of contention.)
Great example. Another along the same lines, without the obvious sociocultural baggage, would be autism, or schizophrenia. They turned out to be mainly organic brain problems, after a many decades of people explaining them at entirely the wrong level.
(Or going back to the sociocultural baggage end of things, you could look at the history of “scientific” racism in published journal papers. Yikes.)
Yes, Paul. Hence I said: “I can think of at least one paper in professional analytic philosophy that has passed through peer review, was produced by one of the top names and in one of the top journals, and yet is complete and utter compost… But I’d still advise the deferential dilettente to defer to blind peer review over any other measure.”
Everybody uses circumstantial evidence to decide how seriously to take a paper–what level of “presumption of innocence” to extend to it. My circumstantial evidence was primarily Carol Graham, a uthor of the excellent book “Happiness Around the World.” Since she respects the study and covers it in the book, I’m inclined to doubt it’s garbage (though it could certainly be flawed). For people who don’t know Graham or that book, this consideration isn’t going to “work” as it does for me.
Judging by the papers I’ve read—though I haven’t been thorough—I’m guessing there isn’t really an excellent one.
The problems I see with the Clark and Lelke paper are not particularly surprising—the more closely I look at the “scientific evidence” about the effects of religion, the surer I am that the whole area has serious problems at this point, and is hard to distinguish from a broad pattern of confirmation bias. (There are certainly some very interesting statistics, but they generally don’t have convincing causal interpretations, so far as I’ve looked.)
Bother them consciously, or correlate negatively with self-reported happiness?
My impression is that in the US, there’s some too-common ideology about people deserving “the fruits of their labors” that allows them to excuse inequality, in abstract political conversations, but that the inequality probably really does make people less happy, if self-reported happiness is to be believed, and if we assume the correlation reflects causality.
I could be wrong about that.
(BTW, I don’t have Graham’s book yet, but I did just get ahold of Haidt’s.)
Two points:
(1) if the national-level advantages of irreligion are largely due to bracketing by low expectations, I suspect that the advantages of religion at the personal level are similarly overstated, and my scientific wild-ass guess would be that they’re even more overstated.
I can imagine, for example, that the Dane’s aren’t really extraordinarily happy, and that they just report their felt happiness with a higher number because of lower expectations.
On the other hand, I’d guess that if they do that, Fundamentalists Christians in the US do the same thing, for much the same reason, based on their relative standing among their local peers—and that the local peers have a bigger effect.
I think people judge their own standing relative to people they actually know more than they judge them relative to national averages.
That’s consistent with the basic background about the importance of regional data that Clark and Lelkes present near the beginning of their paper, e.g., that unemployed people are less miserable in areas where unemployment is high.
If bracketing is a big problem, I wouldn’t be surprised if the main way churchgoing affects people’s reported happiness levels is by bracketing—they judge their success/status by the members of their church, who are more likely to be their economic peers.
For example, your average U.S. Pentecostal doesn’t go to college or make much money, and likely didn’t finish high school. The more they go to church, the more they hang with other people who also don’t go to college, likely didn’t finish high school, and don’t make much money—they are thus less likely to feel less low-status and poor.
That could be the main cause of higher reported happiness among more church-attending people.
In the U.S., there are strong correlations between religious denominations and income level, and I’d guess even stronger correlations if go by particular congregations—e.g., the bell curve for Presbyterians is presumably broader than the bell curve for most particular Episcopalian churches in particular local areas.
So basically, if the Swedes are more reporting higher happiness because their economy is more egalitarian than ours, our fundies are probably also reporting higher happiness because their particular congregations don’t have particularly high economic standards.
That could be totally wrong, of course, but I really am not much worried by Clark and Lelke’s data and interpretations about how horrible it is to be an atheist among atheists. I think it’s nonsense.
(2) Besides the above double-edged sword about expectations inflating religions’s numbers too, there’s another reason why I’m skeptical of Clark and Lelke’s paper on that point.
I find it interesting that when they try to measure the alleged regional advantages of religiosity, they use prayer or church attendance, which are generally more meaningful indicators of religiosity in Europe than denominational labels.
In Europe, nominal denominational affiliation is fairly often meaningless—to assume that a Lutheran is an orthodox Christian, or a believing Christian at all, is rather like assuming a “Jew” in the US really believes Moses parted the Red Sea. It might or might not mean what it seems.
The denominational label fairly often just means what ethnic/cultural context somebody was raised in—or just says what church their family still goes to for weddings, out of family tradition.
So what does it mean to define “atheists” as people with explicitly “no religion” as their denomation, and then look for correlations between high levels of “atheism” and general unhappiness within a region?
I have to suspect that they tried their usual measure of religiosity—prayer—and got nothing interesting. And they tried the next best—churchgoing—and again, they got nothing.
So they switched to denominational afffiliation, which has a very interesting property, to get a “significant” result.
I think the interesting property is that people in Eastern Europe who aren’t religious generally don’t claim even a nominal religious affiliation. They’re more prone to call themselves irreligious than to give their family’s traditional church’s denomination’s name.
Many Western European irreligious people do still maintain a nominal affiliation with a denomination, like “cultural Jews” in the U.S.
If I’m right about that, I’m pretty sure their chosen threshold picks out disproportionately Eastern European regions with low levels of religion, and not Western European ones with comparably low levels of belief and observance—with all kinds of post-Communist misery attached.
Maybe I’m a bit paranoid, but I suspect that if they could have gotten the same result using their usual “best” measure, they would have reported that. Instead, they had to substitute a worse measure that happened to get a seemingly interesting effect.
Can you tell that I don’t trust this paper any further than I can throw it? (You might even guess I’ve thrown it a couple of times.)
I could be wrong, but that’s one reason why I want to see a map of the results—I want to know if their choice of measure and threshold picks out predictably unhappy places for unsurprising reasons.
Of course everybody uses circumstantial evidence. No doubt. But then again, people usually don’t try to defend the use of circumstantial evidence when there is better evidence available.
As I said above, it’s fine to believe what you like about the paper — nobody will stop you. But that’s your privately motivated choice. There’s not much of a scientific social epistemology going on with your belief, or (for that matter) with the claims expressed in the solid chunk of books on the New York Times best seller list. Rather, these beliefs and claims are quasi-epistemic. They’re just the sorts of things that people do when they’re up to stuff.
Ben, You seem to be trying to take us back to a point in the conversation, 4 days ago (Jan 28), when the discussion was briefly more fraught and less substantive.
You gave the link to the paper Jan 28, 4:28 pm. You then said you’d read it “when it passes through peer review.” (That’s what I was returning to when I said, last night, that conference papers are typically peer-reviewed.) But OK, you added “blind” at some point, and I suppose not all conference papers are blind peer reviewed.
But anyway… On Jan 28 8:28 pm Paul said that “from a quick skim” he had seen a certain problem. He was rather modest about it “(I could be wrong)”. The next morning, he said IF he understood the paper correctly, then the paper was garbage. Note: he was modest about whether he was understanding it correctly.
My reaction was (Jan 29, 9:33 am) to wonder how, after a “brief skim,” you could determine the paper was garbage? I then listed a lot of stuff about the authors’ credentials.
There was nothing at all wrong with this appeal to credentials. Paul was not actually certain he was reading the paper correctly. I was essentially saying “maybe you’re not reading it correctly, because people with such good credentials tend not to write total garbage.” This is true, and a good rule of thumb. I would say: “Watch out for reading worthy opponents in such a way that they appear to be idiots–that probably means you’re straw-manning them.”
The upshot of my appeal to credentials was not at all “they’re right” but “let’s dig deeper–if it’s flawed, it may not be quite that devastatingly flawed.” Since then, we’ve moved way beyond talking about the authors” credentials. Paul has read the paper more carefully, I’ve read it twice.
Now that we’ve done much more “due diligence,” I’m more ready to think the paper might be deeply flawed. But I confess–I do continue to wonder. I have a hard time squaring that with Carol Graham’s high esteem for it. I don’t think her book makes the kinds of errors that Paul has talked about. It’s meticulous and very highly regarded. So if the paper makes those errors, why didn’t she notice? I think that’s a perfectly good question, and reason to withhold final judgment about the paper.
And no, my reasoning is not somehow “privately motivated”. What I know about Graham would ideally part of our shared evidence, and then we could all make the same inferences. If it’s not, that certainly doesn’t mean I ought to see myself as having private reasons–as if I just liked the authors’ last names or something.
But anyway–why go back to all that? We’ve moved on to an interesting discussion about the substance of the paper.
Paul, That’s an amusing theory about why religion increases happiness. (I’m not being sarcastic…) I have to say, though, that the parking lot of our temple is full of BMWs and Lexi, which always makes us feel relatively poor and destitute. We seethe with resentment on the way home.
“I have to suspect that they tried their usual measure of religiosity—prayer—and got nothing interesting. And they tried the next best—churchgoing—and again, they got nothing.
So they switched to denominational afffiliation, which has a very interesting property, to get a “significant” result.”
See pg. 25. They found the spillover effect using all three measures of religiosity. Table 8 shows spillover effects of denomination on people with religion and with “no religion.” Table 9 shows spillover effects of churchgoing on “not churchgoers” and “churchgoers.” It also shows spillover effects of prayer on “not prayers” and “prayers”.
Hope you will enjoy Haidt, who writes “warm and fuzzy” like all the positive psychology folks. Graham is not warm and fuzzy in the least–in fact, the book is horrifyingly complex and technical…but extremely interesting.
http://www.nytimes.com/2010/12/06/opinion/06douthat.html?_r=1&partner=rssnyt&emc=rss
I remembered reading this column about how religious conservatives are becoming more highly educated and wealthier.
I don’t know whether the studies are true. I just send it to add to the conversation.
I should probably clarify what I said about being pessimistic about an excellent book on this subject, and problems with the field.
There may be a good book, and Graham’s may be it—I don’t know. I doubt there’s an excellent one because I’m pretty sure the subject is simply not well understood yet, for various reasons I’ve spelled out.
I would expect even a very good book to turn out to contain errors, including some significant ones, which we’ll only find out about in the future.
–
Also, with respect to the issue of bracketing and its effects on measures of the (seemingly) higher happiness of more secular societies, vs. the (seemingly) higher happiness of more religious people within a society as in the US…
I didn’t make the argument I started out to… I was multitasking and spaced out.
It’s not just about concrete expectations about things like education and income, but about how you rate actual, perceived, or recalled happiness on a particular scale—what are your expectations in terms of actual happiness, and in particular, what expectations for others’ happiness are they framed relative to?
Nobody on earth has a good way of comparing their own happiness to anybody else’s, given the fallible and nonrepresentative measures of others’ happiness we get. Nobody has a good reality check. Nobody knows, or can know, what the “objective” reference scale is.
That means that asking somebody how happy they are is a projective test—differences may say more about how they tend to mentally frame things than about how they actually feel.
When you answer a question like that, you generally implicitly assume a scale between a certain plausible amount of misery and a certain plausible amounts of satisfaction and joy.
If you think you’re pretty average for your group, and that your group is pretty average for the purposes of the question, you’ll tend to write down something close to what you think that people feel on average.
If, rightly or wrongly, you think you’re less happy than average, you’ll write down a number that reflects that. And if you rightly or wrongly think that you’re atypically happy, you’ll write down a number that reflects that.
—
If, on the other hand, you think that you are average for your group, but that other groups are less happy on average, you will tend to write down a higher number for the same perceived/recalled happiness—just because you assume that people in other groups are typically rather less happy.
The number that you write down may thus depend on how you perceive other groups relative to your own. If you rate them lower, that’ll bias the scale, and you’re rate yourself higher, even if you don’t actually feel any different at all.
Now consider the people who statistically seem to be a bit happier because they’re more religious, and maybe especially if they’re more “intrinsically religious.”
Do they really think even think they’re that happy, or do they just think outsiders are more unhappy?
Consider all the rhetoric in religion, especially popular, conservative Christianity, about how people in other groups don’t know how to live, and suffer the consequences. Jesus suffuses their lives with joy, and non-Christians live benighted lives in darkness. You should spread the Gospel to bring people Jesus’s message, and make them feel better—know they’re loved, know how to live better, more fulfilling lives that will make them happier. That’ll keep them from falling into sin with all the misery it causes, let them focus on what really will make them happy, and help lift them out of the mire of their relatively joyless, relatively godless existence.
By implication, outsiders’ lives now must kinda suck, relatively, for lack of the wonderful things those Christians think accounts for whatever joy they actually have.
(I think this is pretty obviously true, even if it’s not obvious how the rubber meets the road for our purposes. One of the “surprising” basic results of hedonic psychology is that almost nothing matters as much to people’s long-term happiness as they tend to think it does. Another is that the things that do matter generally matter less than they think. Clearly people’s framing of these issues—at least their conscious framing—is quite typically quite wrong. If we know anything at all in hedonic pschology, so far, it’s that most of what people think makes them happy is largely matter of attribution error, not reality. It’s an exercise in projecting assumptions.)
You shouldn’t be surprised to find a bracketing effect there, where groups that are more smug about lifestyle and happiness issues rate themselves higher, just because they think others are worse off.
Bracketing problems may thus affect people more if they have a more black-and-white worldview, where their group or culture is assumed to be superior, and outsiders’ beliefs and cultures are assumed to create misery.
I would therefore expect that if this effect is significant, it makes exceptionalists rate their happiness higher than people with a humbler, more egalitarian, multicultural view of the world.
In the US, there’s certainly a fair amount of that sort of exceptionalism associated with orthodox Christianity. (Often associated with nationalistic American Exceptionalism.)
Right-Wing Authoritarian types, especially, may tend to give themselves higher happiness ratings, not because they feel any different on average, but because they’re pretty convinced the rest of us screw ourselves up by not being like them. In their ideology, America is the best place in the world to live, their religious culture is the best culture to have, and people of other countries and/or other religions are obviously falling short of whatever they’re achieving—they’re wrong, stupid, dysfunctional, and immoral, and they’re of course comparatively miserable because of it.
If I recall correctly that Right Wing Authoritarian types are on average “happier,” that could be mainly why—they’re not really expressing any great happiness, but revealing an assumption that the rest of us live in greater misery.
People with a more realistic view of the world may just give more realistic (implicitly relative) ratings. They guess that their own countries and their own cultures don’t generally make people a whole lot happier, and don’t tend scale their answers as much that way.
–
That could be totally wrong, of course. It could be that when people rate their own happiness, they access an instinctive internal “happy-meter” and are relatively insensitive to such bigger-picture issues. Maybe they tend to use an instinctive happiness scale that’s generally similar across most people. Or they might tend to frame their answers in terms of what has the most visceral emotional weight—how they perceive the people they actually interact with, or are close to, e.g., at home, at work, or in their churches or hanging out in bars, and their perceptions of people they don’t know but identify with, in their own cultural group(s).
People do generally tend to judge themselves more by the standards of people in their immediate vicinity or cultural group rather than in terms of abstract considerations about people thousands of miles away in other countries.
(Just look at the news—Americans tend to care less about 4000 brown people dead somewhere than about one missing teenaged white girl.)
The upshot of all that is that I think we need to look at foundational results in hedonic psychology—do we really even know what people are basically doing when they write down a number reflecting “happiness” or “satisfaction”?—before taking applied hedonic psychology very seriously at all.
Paul:
Wouldn’t it be better to ask Jeremy if you can start your own thread?
You’re saying a lot of things that should be talked about, but in this long, unending thread, only four people are participating, you, me, Jean and Ben. An all-star cast, to be sure, but others might have something to add.
Amos – I already offered Paul that chance. Indeed, if he wants to become one of our bloggers, then that’d be cool.
Jeremy:
Yes, I know that you made that offer above.
I think that Paul should reconsider it.
As you suggested above, he should outline Jean’s contribution to this conversation.
If that is complicated for one reason or another, maybe Jean can summarize her own position.
Jean, my latest comment was a direct reply to one you made yesterday. That’s why it was brought up, and keeps coming up — you keep talking about it!
Ophelia has linked to this conversation a couple of times, I linked to it once. I don’t think we’re really in a closet here.
Graham’s book deals extensively with the issue of expectations and how people compare themselves to other people to arrive at their estimation of their own life satisfaction. It’s one thing to gather a lot of data, but positive psychologists are also interested in explaining it (of course).
Yes, religious people believe they ought to be happier than others, so that might enter into their self-assessments. But Americans think they’re happier than non-Americans, Texans think they’re happier than non-Texans, so it’s hard to tease out the “I’m religious, so I’m happy” part of religious people’s self-assessment. I’d be extremely surprised if that explained the whole difference between their happiness levels and those of non-religious people, because religion is a package of things, not just a unitary thing. Lots of the things in the package are independently known to be happiness boosters–like socializing, optimism (boosted by the thought that God is in charge), helping others, and expressing gratitude. Religious people get a more regular dose of the whole package, because the church does that packaging for them. Non-religious people have to run around gathering all the elements independently, so stand a smaller chance of getting them all. That’s where atheist “fellowships” come in–they are also “packagers” of happiness boosters, except for the God part. It would be lovely beyond words to see if membership in one of those is just as felicific as membership in a church.
Jeremy:
And I’m still considering that, and thinking about what would go in or out of a self-contained piece.
For the moment, I’m really enjoying being able to explore the ideas with a select group, dialectically.
I’ve really been very pleased by this thread.
Wow. I’m honored. That is a very intriguing offer. I’d need to think about what it would entail.
My output tends to be extremely sporadic—I go through fits of posting comments, then not saying much of anything on the internet for long periods–and I also have great difficulty editing things down into self-contained pieces without feeling I’m no longer doing justice to the various deep and interrelated subjects.
That’s a major reason I don’t have my own blog. I’m sure I’m not capable of frequently posting good stuff.
(And I hate flame wars, but sometimes feeling compelled to participate in them. Pithy short posts, however good and well-written, often function as flame bait, and I tend to write at length and defensively—arguably overexplaining, in too much detail, to forestall misinterpretations and such.)
Being able to post occasionally when I have something especially good and coherent and defensible could be very cool, though—especially if I could do it preferentially when I feel I have the time and energy to deal with fallout of more than a very few people actually reading it, beyond the bowels of somebody else’s threads.
Paul – Occasional posts would be absolutely fine here (the advantages of a group blog). You’ll notice that I don’t post very often.
We really don’t tend to get flame wars within threads here (partly because I complain loudly if that looks like it’s going to happen!).
If you think you want to go ahead, then I can give you posting privileges, and you can post as and when (and if) you wish.
I’ve really enjoyed it. If Paul W becomes a blogger, I’ll certainly be reading. And now, let us pray…
Just kidding.
Writing defensively is fine.
Christopher Hitchens has gotten rich from writing defensively and he does it well.
Does Hitchens ever propose anything?
To take a more notable example, Nietzsche basically writes defensively. Check out his rules for warfare in Ecce Homo,
Why I am so wise, chapter 7.
I’m sure everyone has read them, but for those with a rusty memory:
First, I only attack causes that are victorious; I may even wait until they become victorious.
Second, I only attack causes against which I would not find allies, so that I stand alone–so that I compromise myself alone–I have never taken a step publicly that did not compromise me: that is my criterion of doing right.
Third: I never attack persons; I merely avail myself of the person as of a strong magnifying glass which allows one to make visible a genral but creeping and elusive calamity……
Fourth: I only attack things when every personal quarrel is excluded, when any background of bad experience is lacking. On the contrary, attack is in my case a proof of good will…
Jean:
You know, they should have titled it “Let Us Pray to Jesus, or maybe not, even if we don’t pray or go to church at all, so long as we identify culturally as Catholic or Protestant.”
But really, it shouldn’t say anything. If my latest understanding of the issues is correct, what the paper is reporting is likely just an artifact of failing to judge relative income against the right (regional) standard.
I could be wrong about that, or even have gotten something basically backward in my reasoning, but at this point it looks to me like the paper’s got a big non-result that mistakenly appears to be a result due to modeling error.
—
I’ve been looking harder at the statistics, trying to sort out what’s really going on, and it isn’t increasing my happiness or my life satisfaction. I hate this paper.
Looking at Table 7 is interesting, even if we take it at face value, which I don’t.
—
But before I go into that, a couple of preliminaries.
Jean, you are right that they’re looking at regional variation within countries.
What the regression basically does is assume that certain factors are the important ones, and try to find a weighting of those factors that gives results similar to the actual data.
Regression isn’t magic, and involves a whole lot of assumptions that are frequently false to some degree or other—e.g., assuming that various factors are basically independent and simply additive, rather than depending on each other or varying in weird nonlinear ways.
If your regression formula is not appropriate to the domain, you’re putting garbage into the regression algorithm and will get garbage out, often with very high statistical significance that means less than nothing. If your model isn’t right, “statistical significance” just means you’re generating systematically wrong numbers with great repeatability.
One of their terms in the regression formula is relative income—they’re basically factoring income out, and IMHO they’re doing it wrong in at least two ways:
——
(1) They assume that the relevant distribution for determining income quartiles is the national distribution, not the regional one. I think the regional one is more appropriate, or some function of both weighted toward the regional standard.
Consider, say, Kentucky and New York state. The same income (in dollars) will buy you a better material quality of life in Kentucky than in New York. Most things cost more in most places in New York than in most places in Kentucky. But most people in New York get paid more, too.
If we don’t account for that, and just lump New Yorkers and Kentuckians into national income quartiles, something funny will predictably happen—people in Kentucky will look unrealisticaly happy compared to New Yorkers, at the “same” income levels, because the same number of dollars doesn’t mean the same thing in both places.
New Yorkers are not satisfied as satisfied with a given income in dollars as Kentuckians—because they’re not stupid. If they only get Kentucky-level wages and have to pay New York-level prices for everything, it makes them materially poorer, and they do notice that they don’t have the big screen TV, the nice car, the big house, etc. They know they’re not well off.
Now consider the fact that New Yorkers are also, on average, less religious than Kentuckians. Failing to account properly for income will introduce apparent regional dissatisfactions that have to be attributed to something else by the regression algorithm.
The regression algorithm will tend attribute that difference in satisfaction to whatever other factors are different between Kentuckians and New Yorkers, including religiosity.
Big oops.
Regional effects like that aren’t exclusive to the US. It is common in many places for more urban areas to have higher prices and wages, and also for less urban areas to have substantially higher religiosity. (It has always been thus. Religious conservatives have been railing against the sinful cities for thousands of years.)
If you introduce a bias that makes it look like people in less urban areas are easier to satisfy, you will also be introducing a bias that says that more religious people are easier to satisfy.
I wouldn’t be surprised if that completely explains the main results in the paper.
Judging income quartiles relative to national standards makes people in low wage, low cost regions look happier than people in high income, high cost regions whether they’re religious or not.
Of course it does. Mystery solved, I guess. It’s quite predictable that we’d see that kind of regional effect, to some extent, without assuming religion is actually beneficial or harmful in any way. It’s likely just an artifact of improper normalization that happens to be correlated with religion.
—
Just imagine making the same assumptions with, say, gays or Jews. They tend to live in more urban areas—not least because they don’t like living with the people who live in the country—and therefore will seem harder to please if we just judge their incomes by national standards.
So gays and Jews are probably as hard to please as atheists.
Does that mean they’d be better off being straight or Christian? Probably not.
If somebody published a paper based on these kinds of statistics, titled “Let’s not be gay” or “Let’s not be Jews,” what do you think the reaction would be.
I think there’d be a whole hell of a lot of questioning of assumptions.
But if you say that sort of thing about being irreligious, people are much more likely to just accept it as probably true.
That’s confirmation bias. Papers like this can be published dissing
“atheists,” but it would be Not Okay to so casually “demonstrate” that Jews or gays or Catholics were particularly grumpy and hard to please.
A paper saying the same kinds of things about those other groups would be scrutinized much more closely, I’m pretty sure, and would likely never even make it into a conference without a much more convincing discussion of methodology.
—
Finding such an effect therefore proves precisely nothing at all, unless somehow we can quantify these issues and show that the effect is unexpectedly large, or unexpectedly small.
I don’t think we’re even close to being able to do that with any confidence.
——
(2) The other basic way I think they’re mis-handling income is to assume income effects are independent of religion—i.e., that religiosity doesn’t cause you to have a higher or lower income relative to others in your society.
The regression algorithm basically determines how strong the correlation is between income and satisfaction, relative to the correlations between the other factors and satisfaction, and divvies up the credit for happiness accordingly.
Whatever credit is given to income alone—and that’s a whole lot—is not given to, say, irreligiosity, even if the latter may in fact tend to cause the former.
They make that mistake with several variables, e.g., education level, which is also strongly correlated with happiness.
Without knowing to what extent education causes irreligiosity and/or wealth, it’s impossible to know how to correctly model these effects.
Whatever assumptions you put in—by assuming that the factors are independent, and just plugging them directly into the regression formula—will mostly determine the answers you get out.
——
Backing up a bit, there are some other very basic problems with the paper from the get-go.
–
On page 1, they start talking about happiness—that’s obviously what most of us are interested in—but then say they used ESS measures of life satisfaction.
The unwary might think that was because the ESS didn’t actually ask about happiness per se, but did ask about life satisfaction, which is fairly well correlated with happiness, so they used it as a proxy.
Not so. The ESS does measure (self-reported) happiness, and they simply chose not to use the obvious measure.
Big, big red flag there. Really big. Huge.
Almost nobody—not even the authors, I’ll wager—really thinks life satisfaction is more important than happiness. (What good is being unhappily satisfied?)
Eventually, on page 6, we get to this:
That’s as close as they get to justifying not using the obvious measure. No actual justification.
Note that they say that they say “life satisfaction” is sometimes pick up less ephemeral feelings, but they don’t say who believes that, or when or why. They don’t even say they themselves believe it, or ever have believed it. :-/
From what I know, I’d guess that “happiness” is the better measure—if you ask about happiness, people are more likely to just tell you how they feel. If you ask about life satisfaction, you’re framing things differently, and they may consciously or unconsciously tell you how they think they should. (Maybe because with they frame it that way, they do actually feel it that way.) Or they may be more subject to recall biases.
Asking whether somebody is “satisfied” may prompt them to think in terms whether they consider themselves lucky, whether or not that makes them quite as happy as it should.
(Think about asking somebody “how’s it going?” and them answering “Can’t complain.” Not feeling justified in complaining isn’t necessarily the same thing as not feeling like complaining, and even when it is, it isn’t necessarily the same thing you’d feel if asked whether you’re happy.)
It may also be more likely to try to recollect a bigger picture. That may make them less likely to be biased by their current happiness, but more likely to be biased by well-known biases in recall. I’d guess the latter is worse, on average.
–
My guess is that Clark and Lelke are hiding a basically negative result, which happens all too frequently, even inadvertently—people naturally focus on the “interesting” numbers, when the fact that the other numbers aren’t as interesting may tell you something much more important.
(E.g., that the an expected effect either isn’t happening, or is being counteracted by another effect going the other way, or is just too small to reliably detect with that methodology. Even if you can’t put much precision on the answer, it may let you put an upper bound on something. For example, you may show that whether or not religion is actually good for you, it’s clearly not as good for you as many people think.
I suspect that using the obvious measure—”happiness”—they didn’t get the result they wanted, or didn’t get any interesting result either way, so they chose the other measure and reported that instead.
It’s not lying, but it’s not telling us what we’d really need to know, either.
The really righteous thing to do is to report both. All you have to do is click the mouse a couple of times, type in a different label, and click “run,” to run the regression again.
It’s trivially easy, and the obvious thing to do, and yet they chose not to do it, or chose not to tell us what happened when they did.
I’d reject the paper just for that, if they didn’t fix it. They may not be hiding a negative result, but they give careful readers every reason to suspect that they are doing exactly that, and disregard their conclusions.
(Like justice, good methodology must not only be done, but be seen to be done.)
I have similar concerns about their apparently quite selective use of different measures of religiosity.
If you look at the regression results in Table 7, one thing is very striking about the religiosity measures: by the strongest correlations with life satisfaction from a measure of religiosity come from denomination, and most strikingly for Catholics and Protestants—the effect is less than half as large for “other” religious people. For “others,” it’s one of the weakest correlations in the chart.
The correlation of life satisfaction with attending religious services is considerably weaker than the one for denomination, and the correlation with prayer is considerably weaker still—prayer is only about 40 percent as important as denominational identity.
It too one of the smallest correlations in the chart.
Some comparisions:
(1) the effect of just being in the second income quartile rather than the third is more than twice as large, and the effect of being in the top income quartile as opposed to the bottom one is eight times as large.
(2) The effect of being married rather than single is about nine times larger than the prayer correlation. (And weirdly, it’s very negative. What is up with that? Is there a basic modeling error in there? Does it have downstream effects on the numbers about religiosity?)
We’re talking about fairly modest religiosity effects, especially when using the prayer measure, and that has serious implications.
One is that even small biases in sampling, self-reporting, or inappropriate modeling may make nonsense of the results.
That basic methodological problem aside, the numbers about different religiosity measures would be interesting, if they means anything like what they seems to.
It would seem to mean that actual religiosity is not as important as churchgoing, and neither is as important as nominal religious identity.
Even people who don’t much do religion—don’t go to church, ever, and don’t pray, ever—seem to benefit from being nominally a member of a denomination.
IIRC, the former two are generally considered to be better indicators of actual religiosity, and the latter is pretty noisy—it’s not clear what it means, and it is clear that it often means different things in different places.
And yet the correlations with the worst measure are the strongest.
What on Earth does that really mean?
I have several half-baked theories, but I don’t really know. One basic possibility is that it has more to do with labels than actual religiosity, with different implications depending on what those labels really mean in high income and high cost vs. low income and low cost regions.
—
If the regional income normalization error is as important as I think, it may explain the surprising importance of denomination, and in particular why Catholics and Protestants score higher than “other” religions like muslims and Orthodox.
The latter, like gays and Jews, may tend to live in more urban areas, as they do in the US, and likely for the same reasons.
Another mystery solved, maybe.
—
If this kind of interpretation is plausible—and I think it’s at least that—it makes nonsense of the authors’ suggestion that Catholics and Protestants are so all-fired tolerant that people of other religions or no religion are happier living around them.
As I understand it, a major reason that those minorities tend to live in urban areas is that people in less urban areas tend to be more intolerant of them.
Rightly or wrongly, you probably don’t want to be a Jew or a Muslim or an atheist, or even Greek Orthodox Christian in, say, your average small town in Kentucky.
You’d probably rather live in a city, where maybe you can live near more people like you.
–
If that’s right, the authors’ interpretation of their statistics in terms of religious tolerance is exactly backwards.
They’re not measuring effects of religious tolerance, but of religious intolerance!
Irreligious and minority religious people are basically being herded into larger cities, and sometimes into ghettoes, where their happiness simply counts for less in the statistics because of an accounting mistake about income.
That’s the regional analogue to the phenomenon I was talking about before, at the national level, e.g., with respect to Bosnia and Serbia.
Apparently low levels of unhappiness in more religious areas may reflect intolerant religion’s ability to make people go away and not stay there and complain.
At either level, if you have happy minorities, it likely says more about tolerance in that area—places minorities are willing to go to, and stay—than about religious tolerance in general. It may be the flip side of religious intolerance in the places those people already left.
—
And I haven’t even started to deal with the East/West issues.
One thing that’s likely to be screwy, and matter one way or the other, is that the East Bloc countries and (especially) Germany have much considerably variance in regional religiosity.
If I understand the stats correctly, that means that the East will get undue weight in the statistics—the regional phenomena will all be stronger. If they’re based on a accounting mistake about relative income, those mistakes will just be that much bigger and weigh that much more.
BTW, Jean, my statistics say that y’all citified Jews need to stop kvetching and praise Jesus.
Paul, I’m afraid that I’m going to have to part ways with you at this point.
Whether or not Clark and Lelkes are justified in choosing “life satisfaction” instead of a measure of “happiness”, their methodological choice is motivated by conventional wisdom. The reason that 20th century economics switched from measures of happiness to measures of preference (and, at their most radical, revealed preferences) is because comparing pleasures is a fraught business. The utilitarian dream has been that the epistemic problems involved in measuring happiness will be resolved down the road, but it’s hard to do.
In fact, I think that you provided a great many reasons to think that the subjects of (non-ESM) studies will get their own levels of happiness wrong. So long as you’re being asked about your happiness either in the past or the future, you’re going to screw things up. People who are drunk with nostalgia will transform the misery they experienced in high school into a quaint wellspring of fun anecdotes; the immediate dissatisfaction with being moderately frustrated in traffic will seem far worse than the threat of global warming in 10 to 20 years; etc. And even when people are asked about their happiness in the moment, then (except in cases of extreme happiness and extreme misery) they’re probably going to be implicitly using these other ideas as a comparison class.
These sorts of problems are also, in part, why (in The Moral Landscape) Sam Harris decided to abandon “happiness” as a standard of wellbeing, and instead used “welfare” as a primitive concept. It’s also, in part, why people like David O. Brink have opted for objective measures of wellbeing.
Another problem, which is perhaps less noticed (but which you pick up on), is that facts about “happiness” are hard to distinguish from “moral oughts” even in principle. If you go back to Mill, you find that he simply assumes that the two come apart. Hence, you can arrive at conclusions about The Morally Right on the basis of The Good without being circular. But if it turns out that the morally right and the good are intimately intertwined, then that’s an argument in favor of adopting a standard that does not discriminate between them. It may be that what you call “biases” are just a part of the data that has to be considered, given that there is no internal felt compartmentalization between pleasure and the good.
If Clark and Lelkes adopted the measure of life satisfaction, it’s perfectly understandable. And if it’s a rubbish measure (and it might very well be), then it’s at least popular, reasonable rubbish. You can’t come down hard on them for failings that are shared more widely.
But anyway — I completely agree that they should report both measures. Maybe the data is available online. If it’s just a few clicks away, then…
Ben:
Could I ask you an absolutely innocent question?
What do you mean when you say that “facts about happiness are hard to distinguish from ‘moral oughts’”?
Thanks.
Ben,
I don’t think Clark and Lelke’s choice of “life satisfaction” is justifed by the conventional wisdom.
As I understand it, there is pretty clearly not a conventional wisdom on that particular point in the field of hedonic psychology, or in the related area of economics. (And to the extent there’s anything close, I think it leans the other way.)
One of the distinctive things about hedonic psychology and related approaches to economics is that they frequently appeal to something like Benthamite utility, and defend that choice over traditional measures like conscious preference or revealed choice.
The main point of hedonic psychology in economics is that people’s revealed choices and conscious preferences do not correlate well with what actually makes them happy.
Along with that, you get top people like Csiksentmihalyi and Kahneman emphasizing basic emotional things, especially happiness, over more abstract things like “life satisfaction.”
As I understand it, that’s a major motivation for Csiksentmihalyi’s beeper studies, and for Kahneman’s presentation of his retrospective methodology as a cheap approximation, relative to that “gold standard.” (Or at least it’s part of how they spin them.)
If Clark and Lelke are appealing to conventional wisdom in economics more broadly, I’d say they’re in the wrong area.
In the specific area, they have to at least acknowledge it’s not obvious that “life satisfaction” is a better measure than “happiness.”
(If it’s not obvious to Csiksentmihalyi or to Kahneman, it’s just not really obvious.)
I’m not saying it’s not arguably a good choice—there’s not a consensus that it’s the wrong choice—but they need to at least make the argument.
When I say that “happiness” is the “obvious” choice, I don’t mean that it’s obviously the right choice.
What I mainly mean is that it’s the one most people would naively pick, especially if they read the Clark and Lelke’s own introduction.
If the naive choice is wrong, that needs to be explained. And given that some prominent experts in the particular area think the “naive” choice is actually right, one thing is obvious: whether or not “life satisfaction” is the right measure, it’s not the obvious choice and if there’s an obvious choice, it’s “happiness.”
Choosing “happiness” may only be obvious in the naive sense, but choosing “life satisfaction” isn’t obvious in either sense. It’s the kind of thing that would understandably puzzle most non-experts, and would seem wrong to at least some leading experts as well.
Or I could just be wrong—I’m not sure I have a handle on the “received views” in the field. I’ve only read a few papers. Maybe I’m misremembering Csiksentmihalyi or misunderstanding Kahneman a bit, or maybe they’re more mavericky in that regard than I realize.
BTW, Ben, my last comment may make it sound like I think we’re disagreeing more than I actually think we are.
I do realize that we’re pretty much on the same page—i.e., that reporting both numbers would clearly be better.
Sorry for going on and on. :-/
Amos, it’s just the assumption that what makes us happy tends to be what is good, and what is good tends to be what makes us happy. For example, some people will tend to think that their perverse pleasures are not pleasures at all, or they’ll convince themselves that following their moral duties make them happier than they really do. For example, people are commonly suckered in by implicit personality theories, and by the “halo effect”: beautiful people are thought to be moral and good and ugly people thought to be shifty and mean, even though there’s pretty much no reason to think that.
[The slipperiness of subjective measures of satisfaction is] a philosophical problem as much as it is a psychological one. People will assume that those morally good pleasures are also things that make people happy, in some qualitatively special sense of “happiness”. Hence Mill proposed that higher and lower pleasures could only be discerned on the basis of the preferences of competent judges.
For instance, as a kid I read quite a lot of Archie comics, thinking (I guess) that it made me a better person. Then one day I thought to myself: “Why am I buying these stupid things? They have never made me laugh or made me happy.” It actually came as a shock to me that I was compulsively doing something that made absolutely no difference to my happiness. But it wouldn’t surprise me if I said I was more satisfied after I had bought the comics. Stupid brain.
Thanks, Ben.
That goes back to Socrates (or farther back), doesn’t it?
Virtue is happiness or virtue is necessary and sufficient for happiness. Evil people are unhappy even if they don’t realize that they are unhappy. If you’re unhappy, somehow you deserve it: you’ve done something wrong, if not in this life-time, then in a past one.
Speaking only for myself—and only as I’m using the words right now—I’d rather be happily unsatisfied than unhappily satisfied.
If I’m happy with the way things are and what I’ve got, despite being able to see how things could and should be better, that seems dandy. Maybe I can find a way to make them better, and be even happier.
If I am satisfied with the way things are, but nonetheless unhappy, that suggests to me that I’m just screwed. I have neither present happiness nor much expectation of future happiness.
Unfortunately, I think the words mean different things to different people, and at different times.
It’s usually unconsciously and automatically assumed that satisfaction involves happiness, to some extent.
It’s rather like pain, in a funny way. If you ask people if they’d rather be in pain or be unhappy, they likely won’t know what you mean—and rightly so. How much pain? How divorced from consequent unhappiness? How do you arrange that, anyhow, and what would that be like ?
Many people have difficulty imagining, say, severe pain without consequent major unhappiness about that pain—if it’s really severe pain, isn’t it impossible not to suffer a lot too? It often seems obvious that the suffering is part and parcel of it actually being pain, at least if it’s really intense pain.
As it turns out, that’s not true. (If I’m recalling this correctly.)
You can give people certain drugs that make them able to feel pain quite clearly and intensely, but not really mind much. It doesn’t make them unhappy to quite vividly feel intense pain, and they often find that pretty darned interesting, without it being particularly unpleasant.
But most people don’t know that, and if you ask them “Would you rather be in severe pain or be severely unhappy?” it’s just a bizarre question you shouldn’t expect them to know how to answer.
They might say they’d prefer to be unhappy than to be in pain just because the more concrete idea of pain hits them more viscerally, with consequent associations of unhappiness that “get through” to them better than discussing “unhappiness” directly.
Until they’ve experienced serious pain without much unhappiness, or at least clearly understood certain thought experiments to clarify the idea, I wouldn’t trust their answers at all. Anything could happen when you ask the question.
A similar problem arises with asking random people about “satisfaction”—do they really even understand the idea of satisfaction without the usual consequent happiness? (I’m not even quite sure what that means to me, and I’ve thought about it a bit.)
That’s one reason I prefer the measure of “happiness,” and especially “how happy are you at this moment” in beeper studies—rightly or wrongly, I have the impression that I understand the concept, and that other people are fairly likely to understand it similarly, so that I have some idea what they’re telling me.
I get the impression it’s not an uncommon view in the field—asking about life “satisfaction” makes the question more ambiguous, hence more likely to result in different kinds of oddly different answers from different people, or from the same people at different times, for unknown reasons that may mess up your results so that they don’t mean anything in particular.
If that’s right, then whether or not you think happiness is the highest good, it’s arguably something we’re somewhat more likely to be able to meaningfully measure, hence more useful most of the time.
I know I’ve seen some exploration of those ideas, with some data that were supposed to be relevant, but I forget where—and whether they particularly supported my position, and whether I bought the argument. Oops.
Good heavens, and there I thought we had all settled down to pray!
Paul, I think your income (and herding into cities) point might be devastating, but I’d really like to hear what the authors would say in response. I emailed the lead author a few days ago and didn’t link to this discussion, but asked if there had been online discussions of the paper or written responses. If I hear back, I’ll let y’all know.
I don’t know that this sort of a result is so unkosher, when other maligned groups are in question. There was a recent study showing that people tend to gain weight just as a result of having fat friends and neighbors. Well that’s miserable. Moral of the story: stay away from fat people! (If you put both results together, the moral of the story is: stay away from fat atheists, which is even more miserable.)
Re: satisfaction vs. happiness. I actually think there are good reasons to focus on satisfaction sometimes. That relates to many issues about well-being. If you are in graduate school busting your butt and eating beans for every meal, you might have high satisfaction, low moment to moment happiness. It may be satisfaction that you yourself really care about. Same with people caring for a young infant. No sleep, lots of stress–you’re satisfied more than you’re happy. So–it’s not just chicanery to focus on satisfaction (for some purposes).
On the other hand, moment to moment happiness is interesting too. If you’re deciding whether to do something new, like (say) having a child, you might really want to know BOTH how satisfied parents are and what it’s like to be them, moment to moment.
I just had a potentially important thought about the “relative income” problem I see with the Clark and Lelkes paper.
Are the other papers that purport to show individual benefits of religion making the same mistake?
If so, and if that’s as serious a mistake as I think it is, they’re all garbage.
I can imagine that’s true. In lieu of any better information, it may be standard procedure to “factor out” relative income using national data, not regional data—or better, even more local data, for the cost of living in the particular towns people live in.
If income is being “factored out,” the means of doing so should at least account for the actual local cost of living.
If not, any tendency of religious people to live in lower-cost, lower-wage areas such as smaller towns will bias the results in favor of religion.
(Somebody check my reasoning on that, please.)
I’m curious what studies people are most relying on that purport to show that religion makes individuals happier within a given country, such as the US.
(Jean, if you’d look at Graham’s book for that—what’s her main evidence for that basic idea?—I’d be interested in which particular papers she weights heavily, or if that’s not clear, which few she cites the most.)
If everybody’s “factoring out” income that same way, it’s pretty understandable that Clark and Lelke would do it that way too, without really thinking about it.
It may be the kind of thing that somebody initially did explicitly do because they couldn’t think of anything better to do with the limited data available, and acknowledged the dubious assumptions behind it.
(E.g., “we don’t have the local data we’d really like, so we’re doing the best we can with national data. Future research should gather better data, with local cost of living, and address the possible errors.”)
When people do sort of thing, it’s not rare for it to snowball—other papers do the same thing, for the same reason but explicitly “following” so-and-so’s earlier paper, as though that justified it, and less explicitly or emphatically warning about the potential problems themselves.
The caveats sometimes do get lost along the way, and it becomes the standard methodology that nobody questions at all after a while—it’s just how everybody does it, and often how you have to do it—those are the most precise data available—so how could it be so wrong? How could people have made such a basic mistake?
The inability of people to imagine such things—and see how their predecessors have effectively been led down the garden path—is one of the things that leads to faddish things like the “flashbulb memories” mess.
Jean, I agree that both happiness and satisfaction are interesting, for rather different purposes.
Given just one or the other, though, I’d usually go for happiness.
E.g., if you’re busting your butt and “satisfied” you’re doing as well as can be expected, but stressed the hell out—been there, done that, on the tenure track—you’d usually like to know if you’re right to think that it will also, eventually, actually make you happy. (Been somewhat disappointed on that front, myself, like a lot of people.)
I think that for many people, probably most, satisfaction is important relative to happiness in the short run at least partly because it’s guessed or hoped that satisfaction in the short run is based on correct prediction of long-term happiness.
(E.g., if you knew the things you’re doing would not make you happier in the long run, as you expecdt them to, you’d be less satisfied now with how things are going.)
It seems to me that most people value satisfaction at least partly on that kind of assumption of future happiness—if not for oneself, then for others.
E.g., I might bust my butt to give my kids advantages, and be satisfied doing that despite not being especially happy at the time, but only because I’m guessing that those advantages will ultimately make my kids happier, and that difference will likely outweigh my short-term happiness deficit.
The idea that it’s really just a treadmill with permanently deferred happiness—nobody actually getting to the happier state, and everybody mistakenly thinking they’re making themselves or others happier in the future—is rather horrifying IMHO. Nightmarish, in fact. (If it’s true, there is no God!)
Or I might be more calculating than most, and projecting.
Paul, I guess I’d need to see your sources. My impression was collected downstream from Jon Elster’s remarks on the subject. IIRC, according to him, cardinal utility (which is a kind of preference-ranking that is augmented by judgments of pleasure-intensity) was largely abandoned in favor of ordinal utility.
Though, to be sure, the preference-rankings of economists and decision theorists are extremely boring. We want to know interesting things, like whether or not people all over the globe are gratefully and graciously making themselves miserable. You won’t find out the answers to that if they just list off their settled judgments. But if our attempts to measure interesting things will rely upon methods that are not internally valid, then it’s understandable that social scientists would settle for more modest goals using a method that is (by and large) internally valid. And, granted, while there’s a lot of debate you might have about whether or not a pattern of choices are indications of preferences, at least the debate isn’t an intractable one.
You’re absolutely right that I was concentrating on economics (and its effects on demography), not hedonic psychology. And Elster’s paper was written in the late 80′s, which was right before Kahneman started his stuff on hedonic psychology. Still, the difficulties in measurement that initially caused people a century ago to shift over to mere preference-rankings have not disappeared. (Beeper studies aside.)
It’s interesting that we seem to see the same problems, yet put accent on different solutions. So you say:
That’s probably right — I don’t trust people to have the kind of introspective skills that will let them distinguish happiness and satisfaction. But that’s the fault of happiness, not satisfaction. Happiness is a flighty little pixie that resists analysis; satisfaction is a lumbering ox, easy to find. That’s because you can always define satisfaction in a non-ambiguous way. That is, for instance, you can define satisfaction in negative terms — as the lack of misery amongst persons with desires. So on the one hand, if people are anhedonic or they are miserable, then they are unsatisfied; but on the other hand, if they are in a complicated mood (happily frustrated), obviously happy (ecstatic), or in a muted state (blase and unambitious), they’re satisfied. If you make that sort of thing explicit, then it seems to me that worries about the effects of Danish expectations of their own levels of happiness will seem besides the point.
Amos, that’s a good point. Socrates was pretty naive in many ways, and this was probably one of them.
You can even extend the discussion to Kant’s moral psychology, in some weird and wonderful ways. But I suppose that would have to be another post.
Doesn’t satisfaction often produce happiness?
If I’m satisfied about doing a good job or working for a good cause, I tend to become happier.
That’s my experience at least.
Amos:
Yes. One of the weirdest and scariest issues, which is central to all of this—both scientifically and philosophically—is whether happiness is generally a side-effect of unrealistic expectations.
If people think that what they’re doing will make them happier in the long run, but it won’t, they may be nonetheless happier in the short run, because they think so—and on average that may be where most of their happiness actually comes from: the short-run happiness, over and over, from being consistently mistaken about how things will concretely turn out in the long run, or how much happiness they’ll actually derive from those outcomes.
If that’s true, telling people the truth about the most important things in their lives may be quite bad for them. Unfortunately, letting the believe falsehoods about the most important things may also be bad for them and others in the long run, in different ways.
The idea that that those are both serious possibilities, and maybe both true to a substantial extent, is one of the things that makes me believe in evolutionary psychology, in a general way, and disbelieve in God. It makes sense evolutionarily, but it would require an astonishingly perverse God.
For me at least, satisfaction has little to do with expectations.
When I speak of satisfaction, I refer to feeling “good” about myself, because I live up to the standards that I think that I should live up to, because I do what seems “right” or “correct”, because I achieve what I think is “good” to achieve. That has nothing to do with expectations of the future.
Feeling satisfied about myself, having “scored a goal” in an online debate, having solved a domestic problem, having done a good deed (giving my seat in the metro to someone older than myself)
makes me “happier”: I feel “better about my “self” and that makes me “happy”.
By the way, “happy” for me means a good meal in good company with good conversation. That is my basic measure of what it means to be happy.
Eating the meal (if I were alone) would be “pleasure”, but eating the meal in the context of good company and a good conversation makes me “happy”.
If I were younger, being in love and having sex would be my basic measure of happiness.
Amos:
I would guess that for most people, feeling good about doing something isn’t entirely independent of expectations of it having an effect—or at least being the kind of thing that plausibly has such an effect, even if you can’t tell whether your indidual actions actually do.
Take the Kantian “do what you would will that everyone do.”
What you would will everyone to do generally depends on what effect you think it would have for everyone to do that, even if you don’t judge individual actions by whether they really do have that effect, as opposed to being in accord with the general rule.
So, for example, you might pick up your litter because you think people should, because if every one did the effects would be good.
But if you realized that litter would actually be beneficial if everyone left it on the ground, you might change your mind about the rule. (E.g., if you thought litter ending up in landfills caused bigger problems than litter scattered around.)
I don’t think many really see the rightness of acts as being independent of their consequences, in general, although they may be more concerned with general effects of rules than with direct effects of individual acts.
But then I’m not an ethicist. :-/
Anyway, whether I’m getting all that quite right or not, I think that even for people who aren’t as Utilitarian-leaning as me, consequences do usually matter substantially in some sense, at some level of abstraction.
For example, I think people value taking good care of your kids especially highly because they think that on average, taking care of kids makes the kids and/or other people happier on average, at the time and in the long run.
If it turned out that child-rearing practices had no major or systematic effect on how kids turned out, or even on whether they’re happy at the time, or made others happy or unhappy, that fact about general effects of general actions would be a big deal to most people—why invest as much time and energy and money in taking care of kids, if it doesn’t make them happier, or make them turn out better? Why not invest all that in something that does make something better? Wouldn’t we feel better about helping people we could help? (Wouldn’t it be irresponsible to help people we can’t actually help, when we could be helping people we actually can help?)
Whether you’re more worried about the likely consequences of your particular individual actions, or the general value of particular kinds of acts, it generally matters somehow.
Paul, I tried to find an answer to your question about Graham, and didn’t really find it. Digging around in the “religion & happiness” literature I have around the house, there is very little about the way people control for wealth. Most of the literature seems not to, actually. The relationship between wealth and happiness is the main topic of Graham’s book, and it’s really intricate. So I think if you wanted to deal with all three things coherently (wealth, religion, happiness), that would be …. complicated.
Here’s a nice New Yorker article about Graham’s book–
http://www.newyorker.com/arts/critics/books/2010/03/22/100322crbo_books_kolbert
Gotta run. I’m taking a trip tomorrow, so will sadly have to miss the next chapter of this discussion.
Amos, the relationship between satisfaction and happiness is quite interesting. Although this is not to disagree with you, it’s worth noting that I don’t think that satisfaction is necessarily tied to goal-satisfaction. A goal can be satisfied without a person being satisfied by the goal — you might decide that the goal is not really worth doing halfway through the action, but still accomplish the task through sheer strength of will.
The relationship between satisfaction and happiness isn’t very clear to me. Satisfaction is easy to get a grip on, because it is tied to behavior, cognition, and affect. It’s behavioral, because satisfaction is the outcome of a healthy relationship between a person and the satisfaction of their goals (as you indicate); it’s cognitive, because satisfaction is always in relation to something (e.g., you might be satisfied by your lot in life, satisfied by your plans and projects, to the current state of affairs, etc.); and satisfaction involves affect, but only in the sense of being associated with desire. By contrast, happiness is just affect — it is non-cognitive, because you can be happy for no reason. And happiness does not often emerge as the end-result of our attempts at goal-satisfaction — that’s the paradox of hedonism in a nutshell.
To make matters worse, it seems that many utilitarian writers have implicitly conflated the concepts, or at least they have had an entirely different lexicon. Hence, Sidgwick argues that happiness does not have a “feeling-tone” (not uniquely characterized by affect), and Mill claims that the things that are desired for their own sake become a “part of happiness” (happiness is cognitive).
But I think they’re weird. Satisfaction has a great many identifiable features, while happiness is mere affect. So it’s easier to talk about satisfaction than it is to talk about happiness. For instance, speaking for myself, I find it enormously difficult answering the question, “How are you?” If I said, “Happy”, I would be lying, which I guess means that I have a pretty high standard for what counts as happiness. But if I said, “copacetic”, I would be entirely right. Perhaps I’m just a uniquely insensitive person, but I doubt it.
Paul:
I’m much less future oriented and much less consequentialist than you
imagine and much more a creature of habit.
My goals are set first by habit and second by a sense of myself, that is, the goals reinforce my sense of identity.
I feel satisfied if my sense of identity is reinforced, if I feel good about myself. In general, I am my own witness, although the “I” which witnesses has internalized the hypothetical viewpoint of others, of relatives, of friends, of signficant others. If my actions satisfy, those imaginary witnesses, I feel good about myself and if I feel good about myself, I feel happy.
I assume that the viewpoint, the recognition of those imaginary witnesses, is a recognition of ethical worth. I have no other standard than that of my imaginary witnesses.
By the way, I’m speaking of my psychology, as I perceive it, not of a consistent philosophical position. I don’t have a consistent philosophical position about ethics.
I don’t have any long-term projects nor much sense of a long-term future, except that I will get older and die.
I imagine that my way of being isn’t completely idiosyncratic and that many people see the world as I do.
I came across another paper on the intertubes, “Religiosity and Life Satisfaction Across Nations” by Adam Okulicz-Kozaryn, a postdoc at Harvard.
Using data for 79 varied nations around the world from the World Values Survey, he found that religion is negatively correlated with life satisfaction, but positively (and more strongly) correlated within more religious countries.
Okulicz-Kozaryn criticizes the methodology of Clark and Lelkes (2009) and other similar papers, saying they’re doing the regression wrong. They’re throwing variables at different levels (individual and national and/or regional) into a plain single-level regression equation, and that leads to failure to control things properly and introduces spurious effects.
He uses multilevel regression, in which you have a two-level regression equation that directly models the hierarchical structure of the data.
He also calls for more “disaggregated” (i.e., finer-grained) data, pointing out that people don’t just live in countries or regions, but in cities, towns, etc. Without fine-grained data, even appropriate multilevel modeling just doesn’t have the right data to work with.
To the extent I understand it, I take all that to be consistent with my concerns about biases introduced by failing to take into account things like local costs of living, making anybody who lives in high-cost areas look harder to please, so that you flatter religions common in lower-cost areas at the expense of religions in high-costs areas such as big cities.
I don’t really know what to make of A. O.-K.’s data though, because there are still tons of methodological issues as we’ve been discussion. His results are for countries around the world, which is even worse in some ways than lumping Eastern and Western Europe together, in terms of getting a handle on what’s basically going on.
One weird result he stresses—but may misinterpret IMHO—is that people who rate the importance of religion or God at the maximum level also tend to give the most extreme answer to the life satisfaction question.
Weirdly, they’re likely to say that they are maximally satisfied (10 out of 10 or maximally unsatisfied (1 out of 10).
A. O.-K. seems to take that at face value, to mean that very religious people are likely to be very happy or very unhappy.
I would not be at all surprised if something very different is going on, to some considerable extent:
(1) Some very pro-religion people are trying to skew the poll by giving the answer that will affect the results the most, e.g., that they’re extremely religious and extremely happy.
I find that very plausible. Look at the book ratings on Amazon some time: lots of people generally give books either 1 of 5 or 5 of 5 stars, to maximize the effect of their votes, and to shift the average in the direction they think it should go. (Presumably they don’t really think that all those books are either completely bad or totally awesome.) I suspect most don’t think of this as lying so much as trying to correct for other people’s errors as much as possible. (Shades of Overton Windows.) I think religion books are especially subject to this, and they’re certainly not an exception—it’s quite clear that many people are voting on the POV, not the quality of the book, and maximizing their vote.
(2) Some religious people are interpreting the question differently. E.g., if asked if they’re satisfied with their lives right now, they may emphatically say NO to signify that they fall short of what they owe God, or to signify that they’re unhappy with their situation because everybody else isn’t a zealot, too, or whatever.
There could be other things going on, too, of course, e.g. maybe some anti-religion people trying to skew the poll by claiming to be extremely religious and extremely unhappy, or any number of other weird things.
–
Unfortunately, what that says to me is don’t believe it—I strongly suspect that a significant number of people are answering the question in some weird way, to some extent and we don’t know how and why, so we don’t know what to make of it.
If there are similar but milder weirdnesses with people who don’t give the absolute most extreme answers, those effects could easily shift averages by something like a point on a 10-point scale, and that’s usually the kind of difference in means we’re looking at when comparing religious and irreligious people.
Worse, the results of these things are often reported interms of percentages of people who are very happy or very unhappy—that leaves out all the moderate answers in the middle, and emphasizes the extremes, maximizing the effect of people giving spuriously maximal answers.
Yikes.
Quite interesting as always, Paul. For anyone interested in the paper Paul is referring to, click here. (Incidentally, I’m nerdily delighted by the fact that we’re talking about life satisfaction written by someone with the initials A.O.-K.)
I doubt you’ll find a magic bullet theory to explain the religion-satisfaction connection. There are no precise laws in social science, and very few imprecise ones. Still, it’s an immensely productive first step to spell out the social psychological mechanisms that exist between certain contexts, reports of satisfaction, and religious practices.
Following the train of thought that I’d given above: It’s not clear that measures of life satisfaction tell us anything about happiness as a psychological state. That’s quite right. Hence, I’ve argued that we should essentially just give up on the happiness question. In the absence of further information, satisfaction is something like what Campbell et al. (cited by A.O-K.) think it is: the lack of misery. So if we’re going to go around positing mechanisms, let’s start by talking about the data as if it had to do only with misery and the alleviation of it.
Also, I have to admit being dissatisfied with where we’ve left the religion-inequality connection. From what I gather, there’s very robust reason to believe that religion in religious countries helps people cope with income inequality. As you argued earlier, it’s not at all clear whether the one can be disentangled from the other (e.g., whether inequality causes religion, or vice-versa). But that’s a very frustrating stalemate. Instead, surely we might say, one function of religion is that it gives people a way of reconciling themselves with two kinds of invidious comparisons: a) income inequality, b) inequality in solidarity (i.e., increasing the sense of belonging). That seems to be consistent with the data in a non-trivial way: it makes sense that liberals would be more concerned with inequality than others, it makes sense of the fact that religious people are most satisfied when they’re in religious countries, and it makes sense of the data in a way that makes no reference to our favorite snipe (happiness).
(Or maybe not. Tell me why!)
Ben,
I’m not sure I’ve correctly absorbed all you’ve said about happiness vs. satisfaction, but I still value happiness over satisfaction, to the degree they’re separable, and in principle I think they are to a considerable extent.
My perhaps mistaken understanding is that happiness—in an important sense of the word—is a basic affective thing, which has to do with our constant need to choose, prioritize, and plan.
(We must essentially prioritize things on a linear scale of like/dislike, approach/avoid, etc. Even if it’s a forced fit due to forced choices, I think it’s something we innately do, or approximate.)
It’s pretty plausible to me that there’s something resembling an innate happy-meter that (1) matters enormously and (2) is more directly accessible to introspection than satisfaction, which is a more complex, higher-level thing that’s far more susceptible to various framing and confabulation effects.
(I could be wrong. I haven’t kept up with the literature or thought about it enough.)
Unfortunately, both happiness and satisfaction are susceptible to bracketing effects in reporting, which may be large, and to recall biases if you don’t do them with beeper studies.
The fact that momentary happiness is “fleeting” is related to what’s good about it, IMHO. I believe in something approximately like Benthamite utility—what I most care about is basically the summation of happiness/unhappiness over time, however fleeting the particular contributing feelings are.
For me, “satisfaction” is mostly important insofar as it either correctly records actual happiness, or reliably creates it. Unless I know that “satisfaction” is very strongly correlated with happiness—and that the correlation holds in the particular cases under discussion, such as Church vs. TV vs. just socializing—it’s quite a poor substitute for what I really care about.
That’s terribly unclear to me. I think what we mostly have is correlative statistics that can be plausibly interpreted that way, and can be plausibly interpreted very differently—and we’ve got to be careful not to kid ourselves that the statistics can distinguish between those.
That’s what I see Clark and Lelkes (2009) as doing—confirming their own biases with their biased interpretation of selected correlations, and explaining away correlations that seem to go the other way.
To the degree that we can simply trust the correlative statistics to reflect causality of happiness by religion, it’s absolutely clear that religion is not nearly what it’s cracked up to be. At a national level, it’s inversely correlated with happiness, and at an intra-national level, it’s fairly weakly correlated with happiness if at all.
The net effect of religion seems to be roughly neutral or mildly negative, and any other intepretation of the statistics requires choosing which correlations to regard as causal and which ones to regard as going the other way, or as epiphenomenal.
That is the elephant in the room.
It certainly is. What’s even more frustrating to me is when people ignore the fact of that statistical stalemate, and use the statistics to confirm their causal biases, blithely talking as if it’s not stalemate at all.
I don’t think that we can say any of this “surely,” or even close.
One basic problem in this area is the hedonic treadmill. We know that people are typically wrong in their attributions of what makes them happy, and we should be quite skeptical when the statistics seem to confirm our biases about anything, including religion.
Our default guess should be that in the long run, religion’s direct effects likely does not affect people’s personal happiness much, either. The fact that so many people believe it does shouldn’t make us less skeptical, because we know people simply suck at such judgments and that everybody clearly including some of the the scientists involved is prone to confirmation bias.
Given the apparent ubiquity of counterintuitive hedonic treadmill effects, if we do find something that seems to have a positive effect, we should suspect that it’s an error of measurement, analysis, or interpretation.
Given the hedonic treadmill, we know that data about happiness are very noisy—what we’d naively expect to be major trends are in fact generally adapted out, and disappear.
What we end up measuring is generally second- or third-order effects, because the first-order effects mostly get cancelled out.
That means that whatever we do manage to measure is very likely not to be what it naively appears to be—we’re measuring fairly subtle patterns in very complex things that we clearly don’t understand well.
In signal detection theory terms, that’s a recipe for (1) finding patterns in noise, and (2) misinterpreting the real patterns that you do find.
Finding and understanding patterns in the behavior of adaptive systems is especially tricky, and given the hedonic treadmill, we know that’s exactly what we’re faced with. We know this is a very hard kind of problem, where the ground systematically tends to shift under our feet.
Given that that the first-order effects are known to typically be adaptively canceled out, we know we’re mainly measuring second- or third-order effects that may mostly be results of odd failures of adaptation in particular kinds of circumstances, for reasons we just don’t know, and hence can’t generalize from.
I’ve done research on much, much simpler adaptive systems, and it’s very hard to do meaningful statistics about them—people typically make major modeling errors, very often by assuming things are independent which cannot really be independent, due to feedbacks.
We know we’re dealing with adaptive feedbacks—if there’s one striking result in hedonic psychology, that has to be it—and that independence assumptions are more than suspect; they’re typically going to be wrong, to some unknown extent.
We’re actually dealing with several poorly-understood and interacting adaptive systems at several levels, which is really, really weird.
I’ll try to make that more concrete, below, but keep in mind that I’m only talking about known unknowns—things I’ve already thought of but can’t usefully model. There are presumably many unknown unknowns—adaptive interactions nobody has thought of.
–
Consider religion’s alleged ability to make people more satisfied with what they’ve got.
What are the predictable feedback effects of that?
One predictable but unquantifiable effect is to make people less dissatisfied and hence more complacent—to accept things as they are.
Marx was certainly right to some extent that religion is the opiate of the masses, and that religious institutions coevolve with other institutions—at least sometimes—to defend whatever power they have, and often act to preserve the status quo.
To assume that the short-term satisfaction-increasing effect of religion does not reduce people’s drive to change things in the longer term is to say that Marx was entirely wrong.
I don’t think he was. I’m not a Marxist, and don’t think he was entirely right, but I’m also pretty sure that he wasn’t entirely wrong, either.
If we take the psychology of happiness and motivation seriously, as this kind of work typically purports to do, we have to acknowledge that issue.
To the extent that religion makes people more satisfied, it makes them somewhat less motivated to change things. It may have other countervailing effects—e.g., giving them a way to network, organize, and be more politically effective—but the big question is how big those effects are, and under what circumstances.
It might be that under most circumstances, religion does more good than harm, by generally lessening people’s dissatisfactions but making them more likely to act positively anyhow.
That remains to be shown.
It is utterly bizarre for Clark and Lelkes to simply assume that there are no such costs because they haven’t thought of any. In an economics paper, of all things.
If you think of “religion” and “economics,” Karl Marx and “opiate of the masses” should be among the first things that come to mind—and if your analysis assumes that he was entirely wrong, you should say so very explicitly.
Most economists don’t assume Karl Marx was entirely wrong about that—they think it’s pretty obvious he was onto something, even if they think that positive effects of religion outweigh the negative ones, as Marx didn’t.
Anybody who thinks that religion doesn’t evolve to defend itself at the expense of others’ interests and its overt goals at least sometimes isn’t paying even passing attention to the Catholic child-rape scandals. It is absolutely evident that the highest priority of the Catholic church, in at least those cases, is to preserve its own image, status and power—quite explictly prioritizing things that way, in Canon Law.
Marx obviously wasn’t entirely wrong, which raises the question of quantifying these obvious kinds of adaptive feedback—satisfaction leading to complacency at the individual level, and power leading to conservatism at the institutional level.
To simply assume things are independent and factor them out is to beg perhaps the very biggest questions in the field of economics and religion—e.g., what is the net effect of religion on individual motivation, and what is the net effect of economics on religion at the institutional level?
To blithely make independence assumptions and run a regression as though it meant something is just pseudoscientific question-begging.
I think it’s all up in the air. For example, it may be that in countries with a single strongly dominant religion, religion coevolves with power structures to defend the status quo most of all, and do relatively little to actually change things for the better. On the other hand, it could be that in such countries, where one denomination has a lot of clout, religion has more power to change things for the better, and often does. I would be surprised if both things didn’t happen, sometimes, and often with an odd mix of issues about which they take stances, with both positive and negative effects.
Consider the Catholic Church, for example. On the one hand, it exerts some pressure for social justice, and decries simplistic laissez faire capitalism. On the other hand, it invests much of its clout in opposing abortion rights, contraception, gay rights, etc., causing millions and millions of deaths and creating millions of orphans, as well as increasing the incidence of teen pregnancy, STD’s, and bigotry about such things.
If you also take into account the Catholic church’s frequent condoning of authoriarian regimes—often because they’re pro-Catholic and/or anti-Communist—the picture gets even weirder.
Without even addressing the mixed effects and estimating the net effects of Catholicism—the single most important denomination in most of these studies—it’s utterly ridiculous to assume that religion doesn’t have major costs both at the national level, by its effects on policy and even form of government, and at the individual level, by fostering satisfaction and hence complacency.
It’s like assuming that the cow is spherical. If we know anything about this domain, it’s that these things are obviously not independent.
Garbage in, garbage out.
Paul:
Marx’s affirmation that “religion is the opiate of the masses” does not work for the Catholic Church in Latin America regarding income inequality.
In many Latin American countries, where the vast majority of the population is nominally Catholic, you find a leftwing Catholic Church and a rightwing one, the leftwing version often associated with the Theology of the Liberation in its current watered-down version. However, the Theology of the Liberation, even watered down, is far to the left of Barack Obama when it comes to income inequality.
For example, in Chile, where I live, Eugenio Pizarro, a leftwing priest, was the Communist Party presidential candidate in 1993 and got, as I recall, about 4% of the vote.
(There were two other leftwing candidates that year, which divided the total).
The current leftwing president of Paraguay, Fernando Lugo, is a bishop or ex-bishop. (I think that he was suspended upon being elected).
Amos, I’m certainly not claiming that “religion is the opiate of the masses” is the whole story, or even close—just that there’s apparently some truth to it, and that’s enough to make nonsense of certain independence assumptions.
Positive claims about religion have the same kind of effect—to the extent that they’re true, they make nonsense of assumptions that we can ignore feedback effects, and take statistics based on independence assumptions at all seriously.
There are ways I think it’s pretty clear that Catholicism has negative effects, in many places at many times, and other ways that it has positive effects, likewise in many places at many times.
To ignore those effects, or to assume they just cancel out, is to beg many questions.
Consider the basic question of whether specifically Catholic religiosity is caused by poverty and inequality, or is caused by it.
In at least some respects it’s pretty clearly both. There is no way they’re independent.
Poor people have fewer choices of what to do with their time, and going to church and church-related activities is relatively attractive. The Catholic church clearly markets itself fairly effectively to poor and uneducated people. Poverty is pretty clearly among the likely causes of Catholic religiosity.
On the other hand, it’s also quite clear that the Catholic church promotes overpopulation with its anti-birth control stance, thus creating more poor people. Catholic religiosity is thus pretty clearly among the causes of poverty and inequality, as well.
Those things have different effects in different places and at different times—e.g., both typically being larger in poorer countries with lower standards of education—but it’s certainly not safe to assume that the effects are zero, or that they exactly cancel each other out, such that they can be ignored.
The authors of the papers we’ve been discussing generally don’t make any attempt to take into account obviously economically important effects such as promoting high birth rates in countries with low incomes and high inequality.
How much of the alleged benefit of Catholicism for poor and disadvantaged people is actually only enabled by the Church’s tendency to promote the creation of poor and disadvantaged people?
IMHO, that effect is pretty clearly not ignorable. It may be counteracted by equal or larger positive effects, such as the Church’s attempts at protecting poor people’s interests, but if so, it’s up to the people ignoring such costs to show their work.
It’s ridiculous to assume that national income averages and distributions are not affected by religion—they quite obviously are, although the net effects are variable and difficult to determine.
Again, the idea that economists can simply ignore such well-known (but poorly understood and poorly quantified) effects is simply bizarre.
This is largely an exercise in looking where the light is brightest, and assuming nothing else matters.
“Catholic religiosity is among the causes of poverty and inequality”.
In Chile there is a higher percentage of nominal Catholicism among the rich and the upper middle class than among the poor, since Protestant evangelic sects have captured a sizeable segment of the low income market.
So it seems unlikely that there is a relationship between poverty and being Catholic in Chile.
Wealthier Catholics (in Chile) tend to affiliate themselves with
sectors of the Catholic Church that are rightwing in terms of the issue of economic inequality.
I would say (although I have no statistics) that high birthrates in Chile are common among both the wealthier and among the very poor, but not among the poor, who now use birth control.
I think that if you take as your model of “religion” U.S. Protestantism, you going to miss out on some aspects of how “religion” may function in other cultures.
Now, there is an historical question about whether cultures where Catholicism is predominant are less properous than those where the Protestant religion is predominant. That seems true, with certain exceptions, for example, France.
However, although Catholic cultures may tend to be less properous than Protestant (or non-religious) cultures, within Chile, a Catholic culture, Catholicism is not related to poverty nor is Protestantism with prosperity.
Amos, you seem to be missing my main point, which is that these things (e.g., inequality and religion) interact, with the causality going multiple directions.
I’m not saying that I know which way the net effect goes, or that it’s simple, or that it isn’t quite different in different circumstances.
I’m saying the opposite of those things.
I’m saying it’s complicated, and nobody knows what the net effects are, and how they vary depending on various factors.
The point being that you can’t blithely make independence assumptions—whichever ways the causality goes, and whatever the net direction and magnitude of the effects, these things are clearly not actually independent.
When I say that Catholic religiosity is among the causes of poverty and inequality, I’m not saying that that’s the bottom line—maybe it’s also among the causes of wealth and equality, in different ways, and maybe at the bottom line that counts for more.
Either way, pretending that religiosity and income are independent—and “factoring out” the effects of income to reveal the effects of religion—is just not valid.
Paul,
I
Putting aside my earlier reservations — I think I’m prepared to agree with you that it’s peculiar that anyone would take for granted that satisfaction measures are superior to happiness measures when they’re doing hedonic psychology. I think there are strong, more or less adequate reasons why they might have done so, but they definitely still should have said what those reasons are. (That’s how I’ve gotten into this axiological tangent.)
Like you, I definitely agree that there’s an in-principle distinction between happiness and satisfaction. And like you, I think that there’s some empirical connection between these two concepts — for instance, the hedonic treadmill. But that empirical connection between happiness and satisfaction is not good enough to support robust inferences. In other words: I just think that testimonies (when the surveys are done properly!) are a more reliable indicator of satisfaction than they are of happiness, so we might as well put a priority on measuring satisfaction.
Of course, I’m as disappointed at this result as the next guy. I have a great many sympathies for utilitarianism, and the classic utilitarian can’t really find any traction without reliable happiness measures. Maybe in the future, when we better understand the psychological mechanisms that connect happiness to satisfaction, we’ll be able to base our explanations and ethical decisions on reliable happiness measures. But I don’t think we understand that yet; despite his genius, I don’t think Mill got the story quite right. Since we can’t have a useful ethical theory without it, we had better make do with the things we actually do understand.
I think one thing that has to be cleared up is what connection satisfaction and happiness actually have. On the one hand, I don’t think satisfaction is trivially expressed by testimony, but on the other hand I think that in a substantial minority of cases, satisfaction is consistent with merely putting on a performance of getting by. So, for instance, sometimes we delude ourselves into thinking we are satisfied with some state of affairs, and in the process of deluding ourselves we actually become more or less contented, and hence satisfied. (At the level of lifestyle, this is sometimes referred to as habituated preference formation.)
II
Briefly putting aside my reservations about happiness-talk. I’m not sure the hedonic treadmill counts in favor of your skepticism. My understanding of the hedonic treadmill is that it shows that people adapt their expectations to their circumstances, and hence (ostensibly) maintain a more or less reliable level of happiness. Presumably, the mechanisms that constitute the hedonic treadmill will include a) sour grapes and b) ratcheting up ambitions in light of new opportunities. And since you’re using the hedonic treadmill for skeptical reasons — to show that people are deluded about what makes them happy — then presumably you have something like sour grapes in mind, because people tend to think that sour grapes is a kind of self-delusion. But actually, sour grapes is only arguably delusional — it honestly depends on your interpretation of the connection between reasons and desires. e.g., Who’s to say that sour grapes isn’t perfectly rational hyperbolic discounting?
These interpretive questions cause me enough brain cramps that I’m inclined to throw my hands up in the air and to say — “Oh, screw it! Let’s just ask people about satisfaction!” And you’ll notice my remarks about the religion-inequality connection in the previous post were concerned only with satisfaction (and coping), not happiness per se.
III
On the plausibility of interpretations of statistics. Surely, you would agree that some interpretations are more plausible than others. But I think that some interpretations are more objectively plausible than others. I’ll try to make my ambiguous, back-of-the-envelope suggestion a bit clearer so that it has a chance of being defensible.
I think we agree on a lot of points when it comes to your concrete explanations. To pull out the highlights: I agree that high satisfaction is consistent with complacency, and complacency can be oppressive and odious. That’s why I think Marx’s opiate theory helps to explain why the people who aggrieve inequality are more secular! I also agree that Marx’s “opiate” theory is largely right — because it allows people to cope by making sense of inequality, and by giving a community of support.
But as you say, it’s true that the opiate theory would seem to be wrong in cases where religion acts as a device for agitation. Fair enough. I think my language was insufficiently precise to get my claim across: when I say that religions let people cope with income stratification, I leave it as a live possibility that dissent is a method of coping. In other words, dissent is an unusual (and sometimes effective) kind of therapy. So my suggestion is not in itself an opiate theory, even though it is consistent with an opiate theory.
Paul:
It’s all very complicated, as you say. In fact, the most enlightening thing we’ve all learned in this long conversation is how complicated it all is.
One thing to take into account is that “Catholic religiosity” in Chile is very different than “Catholic religiosity” in, say, Ireland. What’s more, “Catholic religiosity” in a wealthy Chilean family is different than that in a low income one, the Chilean class structure being very rigid.
The theology and the rites may be the same, but given the cultural differences, people live their religiosity in very different ways.
By the way, I’d like to point out yet another convenient asymmetry in how Clark and Lelke interpret their data.
On pp. 12-13 they try to explain why Catholics are happier when the majority is Catholic, but Protestants are happy when the majority is either Catholic or Protestant. (According to theirs statistics and interpretation.)
Their first hypothesis is that it’s because Protestants are more religiously tolerant—they don’t mind being around lots of Catholics.
They don’t offer the opposite hypothesis: that majority Catholics are more tolerant and beneficial, and treat Protestant minorities better than Protestant majorities treat Catholic minorities.
But when it comes to “atheists,” that’s exactly the sort of hypothesis they do offer—if atheists are happier with a religious majority, it must be because the religious majority treats them well.
In other words, if Protestants are happy in the minority, it’s because they themselves are tolerant. But if atheists are happy in the minority, it must be because the religious majority is tolerant.
Funny, that.
(Bear in mind, of course, that I think their basic result is likely not real at all, but an artifact of failing to control for local cost of living and its differential effects on minorities who tend to live in high-cost areas. Still, it’s quite interesting how diametrically opposed explanations occur to them when talking about happy Protestant minorities vs. talking about happy atheist minorities. Apparently, Protestants are obviously not the bad guys, but atheists obviously are, in statistically identical circumstances.)
Ben, here’s an example that may make my skepticism in light of the hedonic treadmill more understandable.
It’s a very well-known fact that when testing IQ’s or aptitudes of schoolchildren, higher ability is quite strongly correlated with shoe size.
A little familiarity with the basics of the domain makes it clear why this is so: shoe size is strongly correlated with age, and age is strongly correlated with test scores.
Older kids do better on the tests, and older kids tend to have bigger feet, but neither causes the other.
Now it gets interesting.
If you compensate for age in the obvious easy way, by looking for correlations within class year cohorts, it’s still true that test scores correlate positively with shoe size, although to a smaller degree. Still, the correlation is quite statistically significant.
The problem there is that doing it by year isn’t right. Kids in the same class are not all of the same age. Older kids within the class still have higher scores on average, and bigger shoes too.
Now what happens if you control for actual age, to the day, by actual birth date?
Depending on exactly how you attempt to do that, you may still still get a statistically significant correlation between age and shoe size, one way or the other.
Why? Because kids of a given age age don’t all start school at the same age. One born just before a school year starts will usually go to school almost a year sooner than one born a few days later, just after a school year starts.
The ones who go to school at a younger age are at relative disadvantage educationally, because their brains have not matured as much, and they typically having a somewhat smaller shoe size too, because their feet haven’t matured as much either.
Oh, and by the way, Leo children are demonstrably and significantly stupider than Virgos, on average, to a very high statistical standard; we have plenty of data that prove that. (Most Leos are born just before the school year starts, and most Virgos are born just after.)
I hope this illustrates what I mean about first-order vs. second-order and third-order effects.
If you don’t compensate very precisely for large, first-order effects, you don’t know if the results you get reflect real second-order effects, or just weird artifacts of failing to control for first-order effects with sufficient precision and accuracy—you may just be measuring artifacts of the experimental methodology.
The bottom line there is that unless you’re very sure you understand the causal structure of the domain, and have modeled it quite correctly, any small correlations you measure—including highly statistically significant ones—are likely not to mean at all what they seem to mean.
Confidence in small correlations should be inversely proportional to your confidence in your model.
I’ve given a number of IMHO good reasons for thinking that most of these studies contain measurement and modeling errors.
That means that the standards of evidence should be correspondingly high—small correlations should be pretty much ignored, even if they have high statistical significance, because the model has not been validated, and very likely contains unknown systematic errors of significant magnitudes.
Paul, I too think that you’ve given very good reasons to think that studies of happiness contain systematic modeling and measurement errors. Hence — so much for happiness studies!
I don’t really have any precise defence for Clark and Lelkes’s preference for satisfaction over happiness. The only defence I would offer is that I can think of some pretty compelling reasons to prefer satisfaction as a measure over happiness. But ultimately they should speak for themselves.
In hindsight, I think the stuff about the opium model is a separate subject. In order to explore those avenues, I think it we would need to consciously transition away from Clark and Lelkes and leap into a new conversation with both feet.
Maybe someone’s reading because they subscribed….
I just heard back from Clark, who says a revised version of the paper is coming out in a journal. So we can expect to hear more about this in the future, I would think–considering how provocative their thesis is.
Also (Paul W)–thanks for that reference. (O-K now teaches at the school right down the street from me!)