I wasn't specifically referring to her analysis, indeed as you point out with those sample sizes and the number of potential conflating factors it's unlikely this data is statistically worth anything -

I was commenting more on the general approach, of evaluating ones actions or choices within the context of a statistical framework.