Table 6.

Statistical Rejection Rates for B_X2Y1, Simulation A

	Full data		Listwise deletion		Log weights		CART weights		Prune weights		RF weights		MI
	30%	50%	30%	50%	30%	50%	30%	50%	30%	50%	30%	50%	30%	50%
N = 100
Linear	.045	.055	.060	.070	.055	.095	.040	.070	.050	.080	.050	.075	.040	.115
One split	.035	.050	.035	.050	.055	.085	.035	.060	.070	.075	.075	.095	.060	.120
Two splits	.095	.045	.045	.040	.075	.100	.060	.055	.090	.140	.085	.140	.110	.135
Three splits	.095	.045	.095	.055	.095	.095	.095	.050	.075	.105	.075	.115	.095	.170
N = 250
Linear	.045	.035	.025	.050	.040	.050	.045	.060	.050	.050	.050	.075	.040	.075
One split	.055	.075	.055	.065	.100	.085	.080	.085	.095	.100	.090	.105	.110	.105
Two splits	.060	.060	.040	.035	.055	.065	.050	.060	.080	.065	.100	.075	.095	.125
Three splits	.075	.035	.085	.075	.075	.085	.060	.065	.110	.085	.095	.105	.085	.105
N = 500
Linear	.025	.055	.045	.040	.050	.050	.055	.055	.055	.045	.055	.055	.055	.055
One split	.050	.075	.030	.040	.065	.065	.060	.045	.050	.065	.050	.050	.070	.065
Two splits	.070	.055	.090	.050	.085	.045	.085	.065	.085	.050	.095	.055	.095	.055
Three splits	.035	.060	.060	.060	.070	.055	.060	.090	.055	.070	.065	.065	.055	.085

Note. Log = logistic regression; CART = classification and regression trees; Prune = pruned CART analysis; RF = random forests; MI = multiple imputation.