6 Comparing Groups: the Log-rank Test
Kaplan–Meier curves for two groups, two treatments or men and women, usually differ. The question is whether the difference is larger than chance would produce. The log-rank test (Mantel, 1966; Peto and Peto, 1972) answers it, and is the standard test in clinical trials with a survival endpoint. It compares, at every event time, the number of events observed in each group with the number expected if the groups had the same hazard, and adds up the differences. This section derives it, extends it to several groups and to strata, and explains the weighted versions used when the curves cross.
6.1 The test
6.2 Several groups, trends and strata
6.3 Weighted log-rank tests
6.4 The log-rank test in software
6.5 Practice Problems
6.2 Several groups, trends and strata
6.3 Weighted log-rank tests
6.4 The log-rank test in software
6.5 Practice Problems
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.