Practitioners of many skills face the need to make some realistic statement about the likely outcome of a future 'experiment of interest' on the basis of observed variability of outcomes in previously conducted related experiments.