Skip to main content
. 2021 Jun 9;9:133. doi: 10.1186/s40168-021-01034-9

Fig. 4.

Fig. 4

Simulation results for the matched-pair data with a set-level confounding covariate of scenario (4). “no ID” means not adjusting for the set ID (second strategy in scenario (4) of Table 1); “no ID, cov” means not adjusting for the set ID but adjusting for the covariate Xset (third strategy). The LDM and permanovaFL with “no ID” or “no ID, cov” had identical results as their counterparts with the proposed strategy and are thus not shown here