Fig. 3From: Constraining PERMANOVA and LDM to within-set comparisons by projection improves the efficiency of analyses of matched sets of microbiome dataSimulation results for the matched-pair data with a sample-level confounding covariate of scenario (3). “no ID” means not adjusting for the set ID but adjusting for the confounder in this scenario. The LDM and permanovaFL with the “no ID” strategy had inflated type I error and thus are not shownBack to article page