Table 1.

Results of simulation study showing the selection bias in estimating an association that is null in the intended study population

Simulation settings		Results—association between allele score and outcome
Association between missingness and both phenotype and outcome (OR)	Association between allele score and phenotype (r)	Mean regression coefficient (SD)	Mean z-score (SD)	Number of 95% CIs containing zero
OR = 1.8	0.05	−0.001 (0.001)	−1.04 (1.00)	83
	(0.25% variance)
	0.10	−0.003 (0.001)	−2.06 (0.98)	45
	(1.00% variance)
	0.15	−0.004 (0.001)	−3.07 (0.98)	9
	(2.25% variance)
	0.20	−0.006 (0.001)	−4.10 (0.98)	0
	(4.00% variance)
	0.30	−0.008 (0.001)	−6.18 (1.06)	0
	(9.00% variance)
OR = 1.5	0.05	−0.001 (0.001)	−0.42 (0.95)	94
	(0.25% variance)
	0.10	−0.001 (0.001)	−0.80 (0.96)	89
	(1.00% variance)
	0.15	−0.001 (0.001)	−1.22 (0.96)	77
	(2.25% variance)
	0.20	−0.002 (0.001)	−1.64 (0.97)	61
	(4.00% variance)
	0.30	−0.003 (0.001)	−2.44 (0.94)	35
	(9.00% variance)
OR = 1.2	0.05	−0.0002 (0.001)	−0.16 (0.92)	97
	(0.25% variance)
	0.10	−0.0003 (0.001)	−0.25 (0.94)	97
	(1.00% variance)
	0.15	−0.0005 (0.001)	−0.38 (0.95)	93
	(2.25% variance)
	0.20	−0.0006 (0.001)	−0.47 (0.95)	91
	(4.00% variance)
	0.30	−0.0009 (0.001)	−0.66 (0.96)	89
	(9.00% variance)

Each scenario was simulated 100 times.