highly significantp ≈ 6.7 × 10 − 1
A one-sided Wilcoxon signed-rank test confirmed this reduction as highly significant (W = 354, p ≈ 6.7 × 10 − 1 4 ), suggesting that the refinement process does not introduce an artificial positivity bias.
A one-sided Wilcoxon signed-rank test confirmed this reduction as highly significant (W = 354, p ≈ 6.7 × 10 − 1 4 ), suggesting that the refinement process does not introduce an artificial positivity bias.