Part of the problem is the “publish or perish” model of higher education. It causes scientists (at least in my field) to use treatments totally unlike what would be done in real life in order to maximize the chance of showing a statistical difference. However, as long as they are honest about it, there’s nothing wrong…