Last time, I concluded that the evidence on this question is mixed. Religious people come off better in some studies, and secularists come off better in others. Some evidence indicates that atheists are less likely to cheat on exams, for instance, whereas other evidence finds that religious people come off better in terms of charitable donations and volunteering.
Several readers objected to my characterization of the evidence, pointing to books such as Arthur C. Brooks' Who really cares. Brooks concluded that conservatives really are compassionate. More to the point, he claims that their good works are determined more by their religion than their politics. Readers felt that Brooks had closed the issue by unambiguously showing that religious people are far more generous. I decided to take a look at Brooks' evidence.
Religious causes, or secular
You cannot compare apples with oranges. When researchers compare the amount of money given to charity, they need to distinguish between donations to religious institutions versus secular charities.
Why is that distinction so important? Well, membership in a religion can be compared with membership in a golf club - at least from an economic perspective. Just as members of golf clubs pay (hefty) membership fees to underwrite the cost of maintaining a golf course, so members of a local church must provide for the upkeep of buildings, payment of staff, and so forth. Religious fundraising is complicated but effective. Members of the congregation contribute in dues, plate collections, fundraising suppers, funeral offerings, bring-and-buy sales, and so forth.
Such contributions are classified as charity by the IRS but they are not really voluntary for the same reason that golf club memberships fees are not voluntary. If members did not support their club, it would go out of business. Similarly, if parishioners cannot put up enough donations the local church may be forced to close its doors.
Suppose that we were to include golf fees as a charitable contribution, then golfers would seem to contribute more to charity. Yet, this is not a fair comparison because people who do not play golf have no reason to support golf clubs.
For similar reasons, people who never go to church contribute around zero dollars to religious institutions and charities. This means that any fair comparison must look at what religious and non religious people give to secular charities.
To his credit, Brooks makes this comparison but, once again, finds that religious people contribute more to secular charities. According to the Social Capital Community Benchmark Survey (2000-2001), 71 percent of religious people contribute to secular charities each year compared to 61 percent of those who rarely, or never, go to church. Religious people volunteer more on an annual basis (60 percent vs. 39 percent) and donate more to secular charities ($532 vs. $467).
Brooks concludes, and many of his readers agree, that the greater generosity of religious people to secular charities is caused by religion. But correlation is not the same as causation.
Brooks overlooks one very simple alternative explanation for why religious people are more generous of their time and money. That reason is that they are older (due to the decline in religious belief among the young) but otherwise demographically, and politically similar to the rest of the nation, according to Brooks. In our society, people over 65 are the ones who command much of the disposable income and who, if they are retired, have plenty of time for volunteering.
Consistent with this thesis, when age is statistically controlled, there is no difference between religious and nonreligious people in the value of their gifts to secular charities. This conclusion emerges from Brooks' own analysis of the General Social Survey for 1996 (Appendix, Table 14). (Among other factors, he also controls income, education, and attitudes toward redistribution of wealth. Volunteering and donation rates were not analyzed).
What superficially seems like clear evidence that religion makes people more charitable turns out, on detailed analysis, to be further evidence that religion has little to do with ethical conduct. There is no consistent difference between religious and nonreligious people in their ethical conduct - at least if one compares apples with apples.
1. Brooks, A. C. (2006). Who really cares: The surprising truth about compassionate conservatism. New York: Basic.
