Skip to main content

Table 3 The error between simulated alpha and nominal alpha for the equivalence tests of percentile H0: θ – θT ≤ − δ or θ – θT ≥ − δ versus H1: - δ < θ - θT < δ with δ = 0.6, θT = zp, σ = 1, and α = 0.05

From: Comparison of alternative approaches for difference, noninferiority, and equivalence testing of normal percentiles

μ

p

Exact approach

Chakraborti-Li method

Bland-Altman method

N

Error

N

Error

N

Error

0

0.1

47

− 0.0006

45

0.0196

45

0.0257

0.2

35

0.0003

34

0.0197

34

0.0162

0.3

30

−0.0015

29

0.0121

29

0.0143

0.4

27

−0.0039

26

0.0070

26

0.0056

0.5

26

−0.0006

26

−0.0006

26

−0.0033

0.6

27

−0.0018

26

−0.0071

26

−0.0111

0.7

30

−0.0033

29

−0.0144

29

−0.0154

0.8

35

−0.0047

34

−0.0188

34

−0.0191

0.9

47

−0.0034

45

−0.0205

45

−0.0212

0.3

0.1

110

−0.0024

121

0.0117

123

0.0205

0.2

84

0.0000

91

0.0097

92

0.0103

0.3

73

0.0024

78

0.0063

78

0.0069

0.4

69

0.0001

72

0.0023

72

0.0043

0.5

71

0.0015

71

−0.0014

71

0.0012

0.6

76

0.0008

74

− 0.0081

74

− 0.0069

0.7

87

−0.0050

82

−0.0068

81

−0.0075

0.8

106

−0.0041

99

−0.0112

97

−0.0112

0.9

144

−0.0005

133

−0.0123

131

−0.0152