If the study population contained a subgroup with a much lower than expected mortality, the calculations would be wrong (mortality in the placebo group would be lower) and a true eff ect of the intervention might be missed (a type 2 error).