Table 4: Estimates of FWE when testing at the α = 0.05 and using Method Q to choose the covariate values.
n1 | n2 | a = 1 | a = 2 | ||
0 | 0 | 30 | 30 | 0.038 | 0.066 |
0 | 0.2 | 30 | 30 | 0.042 | 0.063 |
0.2 | 0 | 30 | 30 | 0.04 | 0.063 |
0.2 | 0.2 | 30 | 30 | 0.039 | 0.062 |
0 | 0 | 30 | 60 | 0.036 | 0.048 |
0 | 0.2 | 30 | 60 | 0.034 | 0.047 |
0.2 | 0 | 30 | 60 | 0.035 | 0.047 |
0.2 | 0.2 | 30 | 60 | 0.034 | 0.047 |
0 | 0 | 150 | 200 | 0.021 | 0.036 |
0 | 0.2 | 150 | 200 | 0.02 | 0.031 |
0.2 | 0 | 150 | 200 | 0.021 | 0.034 |
0.2 | 0.2 | 150 | 200 | 0.021 | 0.03 |
= skewness parameter for the g-and-h distribution; = kurtosis parameter for the g-and-h distribution; n1 = first sample size; n2 = second sample size; a = 1 = a straight regression line, a = 2 = a quadratic regression line.