A blogger called Outside the Beltway writes about the validity of the study, and unleashes an intelligent barrage of comments from his readers about the methodology because he assumed I was doing standard regression analysis instead of matching. It's amazing how many methodological details people are able to pull from the media reports, and it's neat to see a political argument turn into a discussion of internal validity. I'm curious whether the phrase "parametric assumptions" will appear.
A few people --- Investors Business Daily and Valerie Huber cited at the end of the Post article --- have made accusations of ideological bias, which is a particularly ridiculous accusation to make about a statistician. If studies were determined by ideological bias, statistics would suddenly become much easier, and entire subfields of statistics such as procedures to adjust for multiple comparisons could disappear.
Ideological bias is also funny to raise about me, in particular. I vote straight blue and am firmly pro-choice, but I'm also frustratingly moderate. In discussions with reproductive health researchers, I stick up for the abstinence folks whenever it's warranted, and I read evangelical sex and dating books to the point that I know these better than some evangelical clergy, and I've even recommended evangelical dating books to friends.
The IBD editorial is particularly funny because it refers to the "fat wallet" of Mayor Bloomberg as having financed my study, which was not financed at all beyond the grants that paid for my PhD education, and then it links my study to the psilocybin study at Bayview, and repeats the same fallacy correlating teen pregnancy decline with rise of abstinence-only education.
They do get in a good line, "Abstinence-ed can't work as a one-night stand."