Check this box to use EigenCorr scores1 to determine which principal components to include in the regression.
The p-value for testing for nonzero correlation between each principal component (PC) and the trait variable is calculated. PCs with significant p-values, based on the Multiple Testing Method and alpha level for selecting a PC that are specified, are included in the model to adjust for population stratification.
Note: This method can be performed only when a single variable is specified in the Trait Variables field.
Lee, S., F.A. Wright, and F. Zou. 2010. Control of Population Stratification by Correlation-Selected Principal Components. Biometrics doi: 10.1111/j.1541-0420.2010.01520.x