A simple pooling method for variable selection in multiply imputed datasets outperformed complex methods

Table 1 Means and variances used for the simulated dataset

	Varname	Coefficients	Mean	Variance	Standard Deviation (SD)	Distribution
Cat1	Xcat1	0.5 / 1.5 / 1.5^a	2	0.9	0.95	Normal
Cat2	Xcat2	1.5 / 1.5 / 1.5^a	3	1	1	Normal
Dich1	X1	-0.5	0.8	0.2	0.45	Normal
Dich2	X2	-1	0.7	0.3	0.55	Normal
Cont1	X3	0.5	7	2	1.41	Normal
Noise	X4	0	7	2	1.41	Normal
Cont2	X5	-0.1	40	90	9.5	Normal
Cont3	X6	-0.1	26	15	3.9	Normal
Cont4	X7	-0.1	34	23	4.8	Normal

Cat1 Categorical variable 1, Cat2 Categorical variable 2, Dich1 Dichotomous variable 1, Dich2 Dichotomous variable 2
Cont1 Continuous variable 1, Cont2 Continuous variable 2, Cont3 Continuous variable 3, Cont4 Continuous variable 4, Noise Noise variable
^aCoefficients belonging to the dummies of the categorical variable

ISSN: 1471-2288