Table 1.
Type I Error Rate | Power (q=0.1) | |||||
---|---|---|---|---|---|---|
Method |
|
|
Model 1 | Model 2 | ||
One step | 0.051 | 0.010 | 0.019 | 0.061 | ||
Screening testing, marginal assoc. filter | 0.052 | 0.011 | 0.381 | 0.244 | ||
Screening testing, gene‐env. correl. filter | 0.051 | 0.010 | 0.000 | 0.590 | ||
SPUR, marginal assoc. filter | 0.062 | 0.013 | 0.697 | 0.312 | ||
SPUR, gene‐env. correl. filter | 0.057 | 0.012 | 0.000 | 0.665 | ||
SPUR, hierarchical FDR | NAa | NAa | 0.659 | 0.636 | ||
SPUR, global | NAa | NAa | 0.903 | 0.828 |
These SPUR variants reflect the results from FDR or hierarchical FDR control so lack comparable type I error control values.