> What if your doctor said ‘I’ve reviewed your test results, and I’ve narrowed it down to two options. Either you need to have a kidney stone out, or you need a haircut. You can either choose Surgeon A or Surgeon B. Here is a combined percentage for the success rates of their operations and their haircuts.’ I think this time you really would ask for these two statistics to be split out. It’s immediately obvious that you shouldn’t be choosing your medical treatment based on your surgeon’s barbershop prowess.

This is probably the most coherent lay explanation of simpson’s paradox I’ve ever heard.

If I was to see the second table with Treatments A and B I would trust more where it has been tested on more people. So A for small stones (270) and B for large stones (263)