Heckman imputation models for binary or continuous MNAR outcomes and MAR predictors

Table 3 Binary Y and logit selection model: Simulation results for β₁=1 estimates

Methods	\(\beta _{Y}^{sl}\)	% R b i a s	S E _cal	S E _emp	RMSE	Cover
Before	0	0.9	0.109	0.110	0.110	95.3
deletion	1	1.0	0.109	0.103	0.103	96.4
	2	1.2	0.109	0.115	0.115	94.4
CCA	0	1.5	0.135	0.137	0.138	95.8
	1	-7.2	0.133	0.131	0.149	90.6
	2	-15.4	0.134	0.146	0.212	74.0
Heml	0	-2.4	0.167	0.170	0.171	93.5
	1	-2.5	0.153	0.152	0.154	96.0
	2	-3.5	0.144	0.152	0.156	94.9
MIHEml	0	-4.0	0.163	0.163	0.168	93.9
	1	-3.7	0.152	0.152	0.156	95.2
	1	-4.2	0.145	0.155	0.160	94.9

%Rbias: % relative bias; SE_cal: Root mean square of the estimated standard error; SE_emp: Empirical Monte Carlo standard error; RMSE: Root mean square error; Cover: % coverage of the nominal 95% confidence interval; CCA: Complete case analysis; HEml: Heckman’s one-step ML estimation; MIHEml: Multiple imputation using Heckman’s one-step ML estimation

ISSN: 1471-2288