highly significantp < 0.001
weakly significantp < 0.05
Predictors that were selected consistently [at least weakly significant ( p < 0.05) in two out of three selection processes and at least highly significant ( p < 0.001) in one] were used for consecutive analyses.