The amazing Kristy M recently tweeted about the Data Colada blogg, which is fascinating! And filled with numbers! And useful! Run by business school profs at Berkeley and Penn, the authors discuss important statistical stuff in a very approachable way (although I doubt they’d call their stats stuff ‘stuff’).
This post, along with at least one other at the blog, argues that for experiments, cell sizes of 20 is no longer enough. Along with that, they suggest that we can no longer afford to conduct experiments in a lab if we need larger cell sizes. And there’s good support for the size needed to get a good effect in this in this interesting post which will forever be tattooed on my brain as ‘the egg salad experiment’.