Further results on the explained variation in proportional hazards regression

摘要: SUMMARY Two competing measures of the proportion variation possibly censored survival times explained by a given proportional hazards model are compared in Monte Carlo study. It is shown that validity very quick likelihood based measure (Magee, 1990) depends on proportionality assumption. The computationally more involved V2 Schemper (1990) robust this respect. Both produce close results if Cox assumptions perfectly met.