Use data from a randomized experiment to compare two treatments; use simulations to decide if differences between parameters are significant.