Distribution of the type I error probability under selection bias for an increasing number of treatment groups, block and sample sizes.

<p>Each scenario is based on a sample of <i>r</i> = 10,000 sequences, assuming the selection effect <i>η</i> = <i>f</i>_{<i>m</i>,<i>K</i>} equal Cohen’s size <i>f</i>_{<i>m</i>,<i>K</i>}, which depends the group size <i>m</i> = <i>N</i>/<i>K</i> (small: <i>m</i> = 4, medium: <i>m</i> = 8, large: <i>m</i> = 32), and on the number of treatment groups <i>K</i> ∈ {3, 4, 6}. A red dot marks the mean type I error probability in each scenario. The red dashed line marks the 5% significance level. The axis range is (0, 0.25).