Two Methods for reconstructing the empirical mortality profile

This paper goes one step beyond the age determination of the individual skeleton. It presents two methods for the reconstruction of the distribution of the age at death in a sample of skeletons. It is shown that the maximum likelihood method, in spite of certain weaknesses, is superior to the traditionally used proportional method, especially in situations when the analyzed sample consists of skeletons with very different conditions of preservation and thus age intervals of differing lengths. The maximum likelihood method eliminates a bias towards increased average age at death which is introduced by the proportional method. An analysis of an empirical example indicates that the difference between results obtained using the two methods increases with increasing age.