“In sociology, fewer than 10% of recent articles in leading journals employ designs that allow for credible causal inference. In econ, it’s more than 70%” Here's a quick and dirty test of the Great Feminization thesis: