Testing for a mass extinction without selecting taxa

Statistical inference about mass extinction events is commonly based on the pattern of fossil finds among a group of tare. An important issue for exis ting methods is the selection of taxa for inclusion in the analysis. A comm on approach is to select taxa on the basis of the stratigraphic height of t heir uppermost finds. This approach creates a bias in favor of detecting a mass extinction event. This paper describes and illustrates an approach tha t avoids this problem.