Dixon's Q test for a single outlier (Minitab Outlier Test › Dixon), 3 ≤ n ≤ 30, with the Dean–Dixon
ratio r10 / r11 / r21 / r22 chosen by sample size as Minitab does. Critical values and p-values come
from a seeded 20 000-sample simulation of the null distribution (agrees with Rorabacher's 1991 tables
to ±0.01), cached per (n, alternative).
Dixon's Q test for a single outlier (Minitab Outlier Test › Dixon), 3 ≤ n ≤ 30, with the Dean–Dixon ratio r10 / r11 / r21 / r22 chosen by sample size as Minitab does. Critical values and p-values come from a seeded 20 000-sample simulation of the null distribution (agrees with Rorabacher's 1991 tables to ±0.01), cached per (n, alternative).