Wednesday, September 28, 2005

Religion and social ills

There's been a minor kerfuffle over a paper that came out with the remarkable finding that:

In general, higher rates of belief in and worship of a creator correlate with higher rates of homicide, juvenile and early adult mortality, STD infection rates, teen pregnancy, and abortion in the prosperous democracies.
DarkSyd and Dr. Myers took this as a sign that "religious belief … is not good for a culture."

Ed Brayton's take is a lot closer to my own. The best that can be said of this study is that it shows no support for the claim that religious belief correlates with better social practices. Many of those correlations are driven by a particular outlier, the point labeled "U" for United States. Exclude that point, and some of those correlations would be much less significant. For instance, syphilis and gonorrhea infections are fairly constant in the other countries studied, but are vastly higher in the US.
Clearly, the dominant factor there is not religious belief, but inadequate sex education and stigmatization of condom use. Religion plays a role in that, but inspecting figures 6 and 7 shows that the US numbers bear no relation to those in other countries with data available. Lower life expectancy in the US (like higher infant mortality) is probably more easily attributed to the lack of affordable healthcare in the US compared to a bunch of countries with nationalized health care (STD rates may also depend on inadequate healthcare). By that standard, it's remarkable how close to the mean the US is in life expectancy, though infant mortality is shockingly high for a developed nation.

The most striking graphic is figure 2, illustrating homicides. Again, the United States is an outlier, with 3 times the number of homicides per capita of the other nations. What distinguishes the US from those nations? If you said that most of the other nations have stricter gun control laws, you'd be closer to a relevant statistic than religion.

It's interesting that the very religious Portugal also has a high homicide rate, but the US dominates that trend. To treat religion without considering other social factors, including wealth, access to healthcare, gun ownership, educational level (as measured on a standardized scale), etc. is just bad methodology.
As Ed says, correlation isn't causation. It would be interesting to look at these data after appropriate statistical controls, or to look at a time series for a nation that is becoming more secular with a nation becoming more religious and a nation with constant religiosity. That'd get you closer to causation.
So, don't go claiming that religiosity causes social ills. That's not justified by this study and won't win you friends.

Regardless of some methodological problems, the study does reject the central hypothesis offered by IDolators and other culture warriors, that religiosity will solve social ills. The correlation between teen abortion and religious belief is stunning, and doesn't appear to be dominated by the US. More religious nations have more teen abortions per capita. Acceptance of evolution shows no obvious correlation (excluding the US), nor does Biblical literalism.

To the extent that one can test simple hypotheses about the impact of religious belief on abortion, there is either a positive impact – more religious nations seem to have more abortions, or no effect – if you want to argue for no causal interaction. One cannot, given these data, argue for a negative impact of religious belief on abortion rates. If anyone tries to do that, slap them down.

A point that emerges from The Republican War on Science by Chris Mooney is that it's easy to manipulate a few bad studies to try to justify bad policy which you want for entirely unrelated reasons. To me, that argues for honesty in describing motives. If people wants to make the country more religious, that's their right. If IDolators want to argue that kids should be taught that they were designed by something (possibly with Noodly Appendages) that's their right, too. If you think it's just wrong to teach kids to use condoms, that's your right. But don't claim that you're just following where the science leads you. It's dishonest and unfair. Your desire not to teach kids about condoms isn't driven by considerations of the effects of sex education on pregnancy rates or STDs or abortions. Talk about your motives honestly.

Religious authoritarians will either ignore this study or tout some other study showing different correlations (perhaps religious belief vs. degree of superpowerness). Their motive isn't reducing abortion, STDs, homicide, adult or infant mortality, or teen pregnancy. Their motive is to make people more religious. It's a fine motive, but the fact is, not everyone agrees with them. To gain support they make claims about the effects of religious belief on this or that social ill. That leads to trouble.

The same goes for everyone. Don't argue against IDC because you think it's bad for the economy unless you'd support teaching IDC if it would help the economy. Don't argue for a more secular society based on this study, unless you'd be out doing missionary work if it got the opposite results. That's the only way to avoid politicizing science and making it impossible for honest scientists to follow their research where it leads them, and for honest policymakers to follow where the science and the societal concerns lead them.

Technorati Tags: , ,