Roy Baumeister's most famous theory, ego depletion, has been hit by the social science replication crisis. Baumeister has been featured in this space before. Baumeister's rebuttal to his failed replicators is that they didn't actually follow his methods. I kind of see where he is coming from, insofar as I think close attention to detail is critical in methods, but if you don't publish all the details of your method that matter, you haven't really done your job as a scientist. I am used to the physical sciences and engineering, where replication is simply a matter of course. I his ideas seemed reasonable, but if the data isn't there, the data isn't there.
This is a perfect story for Lent. The Commandant of Auschwitz was a cradle Catholic who had abandoned his faith. After the war, he was captured and sentenced to death. He expected to be treated as badly in prison as he had treated his own prisoners, but the Poles who held him unexpectedly treated him with mercy. He was prepared for anything but that. Moved, he ask for a priest to hear his confession before his execution. No priest was willing to hear his confession, but then he remembered the name of a Polish Jesuit he had surprisingly released from custody. Fr. Władysław Lohn was duly located, and heard Höss' confession. Höss was then hanged by the neck until dead a few days later.
A risk-based argument that we could afford to make buses more attractive by relaxing safety rules, and still come out ahead by reducing car travel, which is far riskier than bus travel. I see what he means, but this is a tough argument to make, because you have to deliberately reduce safety in pursuit of a greater good. Current risk analysis practice frowns on that, for good reason.
Medical devices are unlike pharmaceuticals, insofar as double-blind studies are nearly impossible. Well, here is one. Pacemakers are unusual in that the actual pacing can be controlled by someone other than the surgeon or the patient, allowing for a true double-blind study. It turns out that the placebo effect occurs in medical devices, too.
I snarked on Twitter that it is almost as if there were some unseen force making these stereotypes. Abstracting from the causes, there is as much empirical evidence as you could want that men and women behave differently. It turns out that people notice that. In that light, if your metric of choice doesn't show any change in 30 years, you may want to try something different.
Unsurprisingly, a lot. Even among women. Unless you are a General.