> “You don’t have to worry about having a hypothesis, a conclusion, beforehand.” Sadly, in my world of biology and this has been a factor in the epidemic of ba

> “You don’t have to worry about having a hypothesis, a conclusion, beforehand.” Sadly, in my world of biology and this has been a factor in the epidemic of bad papers and unreproducible results. Most of this is due to the treatment of p-values as a magic. If you run 20 hypotheses over p < .05, one is going to show up as true, regardless of the data. Keeping your experiments in narrowly defined scopes is still important to producing good science.