In the life of any species, extinction is the final evolutionary process. It is a common one at present, as the world is entering a major extinction crisis. The pattern of extinction and threat is very non-random, with some taxa being more vulnerable than others. Explaining why some taxa are affected and some escape is a major goal of conservation biology. More ambitiously, a predictive model could, in principle, be built by integrating comparable studies of past and present extinctions. We review progress towards both explanatory and predictive frameworks, comparing correlates of extinction in different groups at different times. Progress towards explanatory models for the current crisis is promising, at least in some well-studied taxa, but the development of a truly predictive model is hampered by the formidable difficulties of integrating studies of present and past extinctions.