Todd's excellent recent post on academics and happiness also raises the much broader question of whether religious belief causes happiness. Some studies, including Arthur Brooks' recent important work, do claim to show a correlation between the two. However, the argument that this proves that being religious makes you happier has two serious flaws.

I. Correlation vs. Causation.

First, even if we prove that there is a correlation between religious belief and happiness, that is not the same thing as proving causation. It could be that people who are happy for reasons unrelated to religion are more likely to be religious. There are a number of plausible scenarios under which this theory would be true. For example, it may be that a tendency towards social conformity makes people happier because they clash less with social norms and the people around them. And conformists are more likely to be religious than nonconformists (at least in an overwhelmingly religious society such as the US). An interesting test of this hypothesis would be to see whether religious believers are (controlling for other variables) happier than atheists in majority-atheist societies such as Denmark, Japan, and the Czech Republic.

II. Measuring Religious Belief vs. Measuring Attendance at Religious Services.

Second, and much more important, the studies do not in fact find even a correlation between happiness and religious belief. What they show is a correlation between happiness and attendance at religious services. For example, Brooks, in the article linked above, shows that "religious" people are much more likely to report being "very happy" in surveys than "secular" people. However, he defines "religious" people as those who say they "attend houses of worship at least once per week" and "secular" as those who say they "never" attend houses of worship.

This is a crucial distinction. It is highly likely that all Brooks' work and other similar studies have shown is that religious believers who go to services regularly are happier than those believers who never do so. Brooks' "secular" category includes some 20 percent of the American population. Yet other survey data shows that atheists and agnostics make up only about 3 to 9 percent of the population. Even if all the atheists and agnostics in Brooks' survey were counted as "secular," it would still be the case that the vast majority of his "secular" respondents (at least 55%), are in fact religious believers who don't go to services. Moreover, many atheists and agnostics do attend religious services at least occasionally (e.g. - for family or social reasons), and so would not be included in Brooks' "secular" category. Some would even be categorized as "religious." I have an atheist friend who regularly attends religious services with her believer husband. In Brooks' study, she would be considered "religious," even though she denies the existence of God and doesn't believe that the precepts of her husband's religion (or any religion) are actually true.

Why would believers who attend services be happier than those who never do? There are many possible reasons, and some of the most plausible ones do not apply with equal force to nonattendance by atheists and agnostics. For example, attendance at religious services is a social activity. We know from a great deal of social science evidence that people who build up "social capital" by participating in social and community activities tend to be happier than those who do. Understandably, religious people with high social capital will tend to participate in religiously-oriented groups. Equally understandably, atheists and agnostics will tend to focus on secular ones. For a religious believer, never attending services is a strong indication of low participation in social activities more generally. For an atheist or agnostic, it might just be an indication that he participates in secular activities instead.

Similarly, many religious people believe that they have a duty to attend services. Those who believe they have such a duty but never live up to it may well be down on themselves for what they perceive to be their own immoral conduct. Almost by definition, atheists and agnostics do not believe they have any moral duty to attend religious services. So they are extremely unlikely to engage in self-recrimination for failing to do so.

III. Limitations of the Argument.

It's important to be clear about the limitations of my argument. I'm not saying that the evidence shows that atheists are happier than religious believers. I'm not even saying Brooks' hypothesis that religious belief makes one happier is provably false. All I'm suggesting is that the evidence he presents doesn't substantiate it.

I'm also not suggesting that the lack of a connection between religious belief and happiness proves that religion is false. The validity of belief in God is independent of whether nor not such belief makes people happy. The same is true for the validity of atheism. I am an atheist because I think logic and evidence support the conclusion that God doesn't exist, not because I think that being an atheist will make me happy.

In fact, it's perfectly possible for belief in imaginary beings to increase happiness. For example, many children are probably happier because of their belief in Santa Claus or the Tooth Fairy. When they learn the truth, they tend to be disappointed, and their level of happiness declines (at least for a time). By contrast, belief in real beings is sometimes more unpleasant than blissful ignorance of their existence. For example, many Americans might be happier if they didn't know about Osama Bin Laden and therefore didn't have to worry about the threat he poses. The truth may set you free. But it won't always make you happy.

UPDATE: In his article linked above, Brooks also notes that "people who pray every day are a third likelier to be very happy than those who never pray, whether or not they attend services." To my mind, this means of measuring religiosity is not sufficiently distinct from attending services. Like attendance at services, regular prayer is also often a social activity (e.g. - many people do it with their families and friends), and is thus likely to be correlated with "social capital." Similarly, many of those who say they "never" pray are likely to be religious believers rather than atheists or agnostics. They may simply belong to religions that don't require prayer; or they doubt its effectiveness despite belonging to denominations that hold otherwise. Thus, Brooks' use of the prayer variable likely proves only that religious people who pray regularly tend to be happie than those religious people who don't. Brooks himself offers a plausible explanation for this result when he notes (in a different context) that "what makes some religious people unhappy is an image of God as severe, unloving or distant . . . regular churchgoers who feel 'very close to God' are 27% more likely to be very happy than churchgoers who do not feel very close to God." If you are religious and believe that God is likely to answer your prayers, that may well make you happier than you would be if you believed in God but thought that he is - in Brooks' words - "severe, unloving, or distant." But that fact says little about the effects on happiness of not believing in God at all.