Abstract: To test the reliability of a testing apparatus which tests a large number of objects of the same type for one feature by generating a feature signal for each object and checking the feature signal for fulfilment of a first condition, a test signal is derived from the feature signals of several objects and the test signal is checked for fulfilment of a second condition. When checking if the test signal fulfils the second condition, the test signal can be compared with a reference value. The test signal can be the average of the feature signals of several objects.