Hyeseon Seo

In many cases an independent variable affects the outcome not only directly but also indirectly through another variable. We call the in-between variable a mediator.

Multiple comparison procedures are fundamental in experimental research. Dunnett’s test, which compares multiple treatments to a single control, is particularly common in laboratory studies. When multiple comparisons are made, proper statistical methods are essential to control false positives. This article demonstrates how the number of comparisons affects p-values in Dunnett’s test, implements the procedure in R, and discusses strategies to improve statistical power.