Abstract: In a method and apparatus for verifying that an occupant sensing system reliably classifies vehicle seat occupants as adults or children, a small number of systems is tested using test objects that have physical characteristics representative of a lower limit adult and an upper limit child. Average measured parameter values are calculated for each test object, as well as standard deviation. Margins are then calculated for each of the two test objects, using separate statistical formulas. If the margins for both objects are non-negative, it is concluded that the system demonstrates a minimum reliability, with a desired confidence level.