Figure 3
From: Sample size calculations for cluster randomised controlled trials with a fixed number of clusters
Minimum detectable difference (effect size) at 80% power for continuous outcomes: limiting values as the cluster size approaches infinity.