Table 2. Empirical type I error of MMTs and SMT for different nominal α levels.

	MMTs			SMT
Nominal α	SKAT	SKAT-O	Burden	Linear Regression Bonferroni correction	Linear Regression BH correction
5 ∙ 10−2	4.94 ∙ 10−2	5.22 ∙ 10−2	5.01 ∙ 10−2	4.39 ∙ 10−2	4.93 ∙ 10−2
1 ∙ 10−2	0.97 ∙ 10−2	1.10 ∙ 10−2	1.00 ∙ 10−2	0.90 ∙ 10−2	0.99 ∙ 10−2
1 ∙ 10−3	0.94 ∙ 10−3	1.12 ∙ 10−3	1.00 ∙ 10−3	0.88 ∙ 10−3	0.97 ∙ 10−3
1 ∙ 10−4	0.92 ∙ 10−4	1.17 ∙ 10−4	1.03 ∙ 10−4	0.91 ∙ 10−4	0.98 ∙ 10−4
1 ∙ 10−5	0.92 ∙ 10−5	1.10 ∙ 10−5	1.12 ∙ 10−5	1.16 ∙ 10−5	1.25 ∙ 10−5
2.5 ∙ 10−6	2.40 ∙ 10−6	2.40 ∙ 10−6	3.50 ∙ 10−6	2.40 ∙ 10−6	2.90 ∙ 10−6

Data was generated from the null model with size n = 1,000 for m = 10,000,000 replicates.