Multiple comparisons tests for comparing survival curves are simpler. By quantifying scatter from all groups - not just the two you are comparing - you gain some degrees of freedom and subsequently some additional power. Multiple comparisons tests after ANOVA are complicated because they not only use a stricter threshold for significance, but also include data from all groups when computing variance (scatter), and use this value with every comparison. How multiple comparisons of survival curves work To protect yourself from making this mistake, you should correct for multiple comparisons. If you perform many pairwise comparisons, the chances are high that one or more pair of groups will generate a P value below the specified threshold (and thus suggesting a rejection of the null hypothesis) purely by chance. However, if you don’t adjust for multiple comparisons after doing this, it is easy to obtain P values that fool you into believing that the null hypothesis that these groups were sampled from a single population can be rejected. This can be done manually by copying the data for each group into a new Survival data table (or duplicating the existing table and adjusting which groups should be included in the analysis). However, instead of looking at the null hypothesis that all of the groups were sampled from a single population, you may be interested in comparing two specific groups at a time. This P value is used to test the null hypothesis that all of the subjects in each of the different groups were sampled from a single population with a single survival profile, and that any differences in the survival of each of the groups was due to random sampling. When you compare three or more survival curves at once, you get a single P value.
0 Comments
Leave a Reply. |