Statistical watchouts

I hope by now it has been impressed upon you that statistical inference is fragile. There are many failure modes, and much of doing good inference involves using techniques to avoid, or at least mitigate, those failure modes.

Here, we present some statistical watchouts to further solidify this notion.