[Selectivity of tests, on which standardized mortality quotients rest]

This paper presents a simple way of calculating the (approximate) power of a test based on the SMR. By means of an example, the use of this quantity, both for planning a study and for analysis is illustrated. Postulating an effect of a given magnitude on mortality, one can calculate the chances of success of a planned study. After data collection, one can, in the case of a nonsignificant result, at least give an upper limit to the effect: larger effects would likely have produced a significant result.