Fig. 6
From: Statistical Inferences Using Effect Sizes in Human Endothelial Function Research
[42]
Relationship between sample size (n = number of pairs) and effect size for a paired samples t test (Panel A), and sample size (n = number in each group) and effect size for independent samples t test (Panel B), across three different levels of statistical power. Fewer participants per group are required for a given effect size using the paired samples t test. Figures assume a two-tailed analysis, with α = 0.05. Adapted from Yu and Yagle