As I understand it, using Bayesian statistics, you can report the posterior probability given various reasonable priors. It seems useful to know how sensitive the conclusion is to the choice of prior. (With a strong result, it should make relatively little difference.)