Figure 5

Comparison between sam2.20 and the symmetric cutoff method under Setup 2(i). The histogram and the parallel boxplots in Figure 5 are defined the same as in Figure 4. The histogram shows that sam2.20 produces significantly smaller number of true FP than the symmetric cutoff method among all 100 simulations. The first two boxplots at the right of Figure 5 show that sam2.20 under-estimates the true FDR and FP correction (12) made the under-estimation even worse. The last two boxplots at the right of Figure 5 show that the symmetric cutoff method over-estimates the true FDR and the over-estimation has been corrected by FP correction (12).