Table 2.

Power from four multiple comparison procedures MCP_B, MCP_P, MCP_S and $MC{P}_S^{MV}$

				Continuous outcome				Binary outcome
m00	m10	m01	m11	MCPB	MCPP	MCPS	$MC{P}_S^{MV}$	MCPB	MCPP	MCPS	$MC{P}_S^{MV}$
100	0	0	10	0.54	0.68	0.72	0.81	0.30	0.49	0.49	0.58
85	15	0	10	0.54	0.61	0.68	0.80	0.28	0.37	0.40	0.49
60	40	0	10	0.55	0.54	0.64	0.79	0.28	0.30	0.34	0.43
85	0	15	10	0.54	0.58	0.68	0.77	0.28	0.42	0.44	0.65
70	15	15	10	0.54	0.54	0.64	0.77	0.25	0.33	0.36	0.66
45	40	15	10	0.55	0.50	0.60	0.76	0.29	0.30	0.33	0.66
1000	0	0	10	0.28	0.69	0.61	0.77	0.11	0.47	0.34	0.44
985	15	0	10	0.28	0.60	0.58	0.74	0.10	0.37	0.31	0.41
690	310	0	10	0.27	0.33	0.45	0.66	0.10	0.15	0.18	0.23
985	0	15	10	0.27	0.57	0.57	0.77	0.11	0.43	0.35	0.65
970	15	15	10	0.26	0.53	0.54	0.75	0.11	0.36	0.32	0.63
675	310	15	10	0.27	0.33	0.44	0.76	0.08	0.12	0.16	0.42

Note: The first four columns show the number (m₀₀) of biomarkers associated with neither exposure nor outcome, the number (m₁₀) associated with only the exposure, the number (m₀₁) associated with only the outcome and the number (m₁₁) associated with both exposure and outcome. The remaining columns show the power, defined to be the mean proportion of true mediators identified, when $\alpha =0.05$. Details of the simulation can be found in Section 2.