The comparison of three groups is often used in pharmacy and medicine, especially with one-way ANOVA and closed testing procedures. One-way analysis of variance (ANOVA) is based on testing the ratio between variances of different sources. After ANOVA, which shows if there is any statistical difference between groups, we conduct post-hoc tests to determine which of the groups are different. This method can be further improved by using methods based on closed testing. With closed testing methods, we control the type 1 error by adjusting p-values. However, these methods can be conceptually complicated in the context of multiple comparisons. Nevertheless, we argue that the comparison of three groups is a special case for these methods. If we do not meet the assumptions, we can explore alternative approaches.
|
