Tuesday, June 2, 2009

Tangled up in (red and) blue

Liberals and libertarians often argue that criminalizing abortion would do more to drive the inevitable abortions underground than to decrease their number. There's also a widespread belief that e.g. the Catholic church's anti-contraception attitude indirectly increases the number of abortions by causing more conceptions. Such arguments are usually irritatingly aprioristic, and (in the last case) somewhat unhinged; as Ross Douthat said at the time:
[Will Saletan] makes it sound like the long shadow of Humanae Vitae and the
malign influence of the Quiverfull movement are a big part of America's abortion problem. But if religious-conservative objections to contraceptive use were actually a big part of the cultural background to our abortion and out-of-wedlock birth rate, you'd expect to see some actual evidence of it.

I'm very sympathetic to Douthat's argument (see also here) that the data lend no support to any of these claims. Of course, I don't think they support his claims either -- it's extremely difficult to separate causation from correlation in abortion data because, e.g. out-of-control teen pregnancy rates correlate with poverty and religiousness; poverty sort of correlates with religiousness; places where abortion clinics are sparse, or abortion laws restrictive, tend to be places where people strongly disapprove of abortion; etc. Nate Silver makes this point very nicely with a set of graphs here. The only inference I feel like I could possibly draw from the data is that ceteris paribus, those who disapprove of abortions are less likely to abort, and even this seems a little iffy. (It is consistent, however, with Margaret Talbot's old NY'er article on the far-reaching differences between "red sex" and "blue sex.")

I think that in cases like this a more productive approach than running regressions is to conduct detailed case studies of the kind that Atul Gawande did in the NY'er about Texan healthcare. The debate will, um, shift aimlessly back and forth until someone lays a steady hand on the tiller.