Via Prof. Leiter, one learns of an obscure Journal of Religion and Society article purporting to show that the more religious a society is, the more "dysfunctional" it is. You can save a lot of reading if you just trust me on this: the report is crap.

If you don't want to trust me, read on. Much like the New Politics Institute's study on blogs (also linked from Leiter--notice a trend? Yeah--why do you still read it?--ed), the study is an exercise in deciding what you want to prove and then choosing the statistics to get you there. The point of such academic work is transparently obvious: they get you into the media (such as The Times of London), who don't think very much, or bother asking anyone else for a contrary view.

(Oh, yeah: a lot of blogs link to you, too.)

The first sign that the report's not very robust is the selection of what is included in "societal dysfunction." Nowhere is the term defined, and nothing that fails to support the thesis is included as dysfunctional. I would have thought one measure of societal dysfunction was long-term unemployment rates. After all, a functioning society is going to be providing its people with measurable employment. However, according to the Journal of Religion and Society, it seems that a society can be both completely functional and completely out of work. Furthermore, while the report considers homicides (which occur more frequently in the U.S.) to be a sign of social dysfunction, it seems that car theft, contact crimes, and burglary (in which the U.S. does not lead at all, see p. 17) are signs of a healthy and functional society. Either that or atheism promotes car theft, which puts a whole new spin on "property is theft," I guess.

So far, so silly. Then there's the choice of what countries qualify. "Prosperous democracy" doesn't include Israel, even though its GDP/head compares favorably with Portugal. Given the religious nature of Israel, one would have thought it would be a good case study, but the study inexplicably excludes it. What the report actually covers, it seems, is Europe, some of its former colonies, and Japan.

Even better is the data for abortion, where the report is simply wrong. It states: "Abortion data (Panchaud et al.) was accepted only from those nations in which it is as approximately legal and available as in the U.S." (This, of course, makes no sense: if a society is religious enough to limit abortion, why not include the lack of teenage abortions on the graph, especially if you will consider their teen pregnancies?) It then continues, "Increasing adolescent abortion rates show positive correlation with increasing belief and worship of a creator, and negative correlation with increasing non-theism and acceptance of evolution; again rates are uniquely high in the U.S. (Figure 8)." But the report's data is questionable. The abortion rate given for 15-19 year olds for the U.S. (higher than 28) is considerably higher than that given by the CDC (about 24, or equivalent to Australia). Further the source cited for the abortion figure in the report (available here doesn't mention abortion at all. (Say what you want about peer review, at least law reviews make someone check the numbers.)

But the other interesting question about Figure 8 is why Japan isn't listed. Abortion is legal in Japan, and has been since well before Roe. And statistics a're not hard to find, even in English: they're reported in English language news and show up as hits on Google. (The Japanese report referenced, with handy pop-up translation help, is here.)

But including data from Japan subtly undermines the thesis. Japanese abortion rates in the 15-19 age cohort are low (about 12%), but have doubled in the last ten years, while U.S. rates have been falling. (Also, Japanese age of majority is 20. This probably makes some difference, as the U.S. rate increases sharply between 17 and 18.) And Japan's overall abortion rate (which, it seems, is not a function of "societal disfunction") is in the middle of those for the JR&S's countries despite its lower rate of theism non-theism. (see p. 28)

So what's happening? Is Japan getting more theistic, and hence more likely to have underage abortions? Not bloody likely. The Japanese abortion rate has always been pretty high, while at the moment they've been undergoing a bit of a sexual revolution: they may not have been theistic, but Japan has remained far more traditional than most western countries until recently, and in that I'd include the U.S. It's likely that rates of teen sex are rising. In the meantime, it's only been five years since the pill's been legalized and takeup has been slow, but if that becomes more common (and condoms less so), I'd expect to see the abortion rate among the young continue to rise.

I'd wager that for any given statistic in the piece, you'd be able to tell a similar story: the factors and trends change over time, with very little to do with any given preference in messiah. In the end, not even the report pretends it can show causation. You can drive a truck through the holes in the logic. So why bother?

It's worth criticizing because some people find that this kind of half-assed scholarship is acceptable if it confirms their preconceptions about religion. Let's imagine a study that compared factors other than theism. For instance, Japan is remarkably racially homogenous, the European countries are a bit less so, and the United States is rather remarkably mixed. This mixture is a good thing, just as in general religion is probably a good thing. (Let me stress that before someone misinterprets what I'm making obvious: OUR DIVERSITY=GOOD THING.) Yet if you're only meaning to "spark debate" and have no intention of decent analysis, see what you could do with the CDC charts on teenage abortion above. The non-hispanic white numbers aren't much more than the Japanese (especially adjusting for a differing age of majority), after all.

If you were to regraph the charts in the JR&S report with "degree of racial homogeneity" along the bottom instead of a belief in God, I can't believe Prof. Leiter would repost it with only the glib comment "Of course, correlation is not causation..." (Or put it in a section giving the racial equivalent of "Texas Taliban Alerts," for that matter.) Does anyone suspect the Times would headline a story "Societies worse off when racially diverse"? Or think that no other expert would be quoted to give perspective?

Of course not. On the topic of race, no one wants to look an idiot.

If one were to look for U.S.-based differences that explain the statistics given better than religion, it doesn't take a genius. For the health-care related ones, for instance, I'm pretty certain the United States is the only nation listed without a national health care system. (Actually, I wonder if the infection statistics differ between curable and non-curable STDs: presence of a national health system could well affect those numbers, as the spread of curable vs. non-curable diseases should differ with access to health care.) Japan has a much lower murder rate, and also makes it incredibly difficult to get a gun. But health care and gun ownership are not matters of theism, any more than atheism makes people likely to steal cars or be unemployable.

In the end, you can create this kind of report by finding one area in which the U.S. is exceptional and graphing it against any number of others in which we're exceptional. Then you call that dysfunction, blame the cause you chose, and call it a day. Assuming you can find a journal to publish you, a professor who won't critique you, and a newspaper too lazy to read what you've written.

UPDATE: Best line I've seen concisely summarizing the silliness of this study? "Rejection of secular humanism gives your kids gonorrhea!"