Survival-time patterns should be compared properly in their entirety rather than at isolated points only. Such overall comparison would require a value function for rating particular durations of survival, but no such function exists. A chi-square procedure is proposed for comparing two sets of life-table data in their entirety. The implicit value function for the pro cedure is reasonable in that it gives greater weight to earlier deaths. By considering the case in which the life-table intervals are arbitrarily short, it is seen to be essentially a rank order procedure. Further, the procedure is defined for general or individual right censorship, for left truncation, and for tied ranks. A second rank order procedure or statistic is developed by considering the case in which the first statistic is computed after each death in a comparative survival-time study. This second statistic is the maximum chi square computed over the course of the study. Such a sta tistic can be used pseudosequentially and significance or nonsignificance ascertained before the end of the study. It is suggested that the procedures may be used in the analysis of other than survival-time data. The possi bility for evaluating the statistical power of the procedures for alternatives when there is a constant force of mortality ratio between the two survival time distributions is indicated.

The results of a medical investigation fre quently take the form of two sets of survival patterns for patients who have been subjected to two different therapeutic regimens. Analysis of such data, because of incomplete observa tion on some or perhaps most patients, is fre quently by the life-table or actuarial method