You can pretty much show any kind of effect from a treatment by P hacking, not adequatly designing a study or accidentally contaminating your experiment. This i

You can pretty much show any kind of effect from a treatment by P hacking, not adequatly designing a study or accidentally contaminating your experiment. This is exactly the kind of thing that really doesn't need attention until it's replicated.