Last week, I criticized claims that religious belief increases happiness. I pointed out that these claims are based on data that don't actually compare religious believers with atheists and agnostics. Instead, they are based on comparisons between churchgoers and non-churchgoers, and people who pray regularly with those who don't. At most, such comparisons show that religious believers who attend services and/or pray regularly are happier than religious believers who don't. They say nothing about differences between religious believers and nonbelievers. In addition, the argument in question conflated correlation with causation.

This week, columnist Mary Eberstadt uses the same data (collected by Albert Brooks) to argue that religious belief increases charitable donations and volunteering:

This one's by econo-brain Arthur C. Brooks and is called Who Really Cares: America's Charity Divide: Who Gives, Who Doesn't, and Why it Matters. Geeking over what he calls "the fruit of years of analysis on the best national and international datasets available on charity, lots of computational horsepower, and the past work of dozens of scholars who have looked at various bits and pieces of the charity puzzle," numbers nerd Brooks shows beyond a doubt one fact that our Side should not want out — i.e., that American believers are more "generous" in every sense than the enlightened likes of Us.

Brooks says that religious people give more to charity than non-religious people — in fact, much more: "an enormous charity gap," he reports, "remains between religious and secular people."

To see this, imagine two women who are both forty-five years old, white, married, have an annual household income of $50,000, and attended about a year of college. The only difference between them is that one goes to church every week, but the other never does. The churchgoing woman will be 21 percentage points more likely to make a charitable gift of money during the year than the non-churchgoer, and she will also be 26 points more likely to volunteer. Furthermore, she will tend to give $1,383 more per year to charity, and to volunteer on 6.4 more occasions.

Brooks goes on to test the charity gap up, down, and sideways. The results are always the same:

"People who pray every day (whether or not they go to church) are 30 percentage points more likely to give money to charity than people who never pray (83 to 53 percent). And people saying they devote a 'great deal of effort' to their spiritual lives are 42 points more likely to give than those devoting 'no effort' (88 to 46 percent). Even a belief in beliefs themselves is associated with charity. People who say that 'beliefs don't matter as long as you're a good person' are dramatically less likely to give charitably (69 to 86 percent) and to volunteer (32 to 51 percent) than people who think that beliefs do matter."

The flaws in this argument are exactly the same as those I pointed out in the claim that Brooks' data shows that religious belief increases happiness: the analysis doesn't actually compare religious believers and atheists or agnostics. It compares religious believers with different levels of observance. For reasons I noted in my earlier post, the vast majority of Brooks' respondents who say they never pray or go to church are still religious believers. It stands to reason that people with less commitment to their ethical beliefs (as indicated by lower attendance at services, or failure to devote "a great deal of effort" to them) are less likely to contribute to charity based on those beliefs. A religious person who doesn't go to services or pray probably has less commitment to his belifs than one who does, and this may be reflected in their level of charitable giving. The same cannot be said of an atheist or agnostic who doesn't go to services or pray. They simply have a different set of ethical beliefs, one that doesn't require attendance at religious services, but may still require charitable donations, volunteering, and so on.

Some of Brooks' comparisons are even less relevant to to the religious-secular comparison than his data on church attendance. For example, you can disbelieve in God, yet still devote "a great deal of effort" to your "spiritual life," especially given the broad, amorphous definition of what counts as "spiritual" in today's popular culture. Similarly, it is perfectly possible to think that "beliefs" matter a great deal, yet also be nonreligious. Indeed, some of the most strongly committed atheists - people like Richard Dawkins or Christopher Hitchens, who think that nearly all religion is evil - would surely say, in a survey, that "beliefs" matter a great deal.

Finally, it's worth noting that even if Brooks and others can show a correlation between religious belief and giving, that doesn't necessarily prove causation. If believers do indeed give more than atheists and agnostics, it may be because of variables exogenous to religious belief that are correlated with it.

As in the previous post on this subject, I note two caveats. First, I am not claiming that I have proven the opposite of Brooks' and Eberstadt's argument: that religion doesn't increase charitable giving. I merely point out that they haven't proven their own case with the data they rely on. As far as I know, we don't yet have a truly reliable comparison of charitable giving by atheists and agnostics on the one hand and religious believers on the other. This gap in the literature represents an opportunity for enterprising scholars.

Second, the validity of religious belief and of atheism don't depend on their effects on charitable giving. Religion could be true, yet fail to stimulate charity. It could also be false and yet increase giving.