Good catch! I'm more inclined to use a Bayesian approach here because we can't be certain about the probability of either population becoming infected. The 80%

Good catch! I'm more inclined to use a Bayesian approach here because we can't be certain about the probability of either population becoming infected. The 80% Credible Intervals for the two Beta distributions overlap: [1] and [2]. So indeed, the chance that either a monk or a normal villager would get infected by a parasite could be about the same. I'm not sure if using Credible Intervals in this way is kosher, so someone feel free to critique. Ideally, if p is the unknown probability of a villager getting infected, and q is the unknown probability of a monk getting infected, then we need the posterior distribution of p-q. We have all the data needed to estimate this; you may use either simulations or characteristic functions to do it. [1] - https://www.wolframalpha.com/input?i=Beta%2812%2C9%29+distribution https://www.wolframalpha.com/input?i=Beta%2812%2C9%29+distri... [2] - https://www.wolframalpha.com/input?i=Beta%289%2C18%29+distribution https://www.wolframalpha.com/input?i=Beta%289%2C18%29+distri...