An exploration of the missing data mechanism in an Internet based smoking cessation trial

Table 6 Estimates from the model for the repeated attempts

Parameter	MAR estimate	MAR standard error	MNAR estimate	MNAR standard error
β _1,1 (treatment)	-0.099	0.081	-0.103	0.081
β _1,2 (age)	0.021	0.004	0.021	0.004
β _1,3 (sex)	0.103	0.086	0.110	0.087
β _1,4 (qualifications)	0.034	0.040	0.033	0.040
β _1,5 (deprivation)	-0.026	0.043	-0.018	0.044
β _1,6 (conscientiousness)	-0.009	0.050	-0.001	0.050
β _1,7 (determination)	-0.002	0.055	-0.008	0.055
β _1,8 (support)	0.096	0.034	0.093	0.034
β _1,9 (dependence)	0.020	0.027	0.018	0.028
β _1,10 (previous)	-0.072	0.085	-0.076	0.085
β ₂ (Y)	-	-	0.215	0.222

The model for the repeated attempts incorporates more data, and hence makes more assumptions, but provides much more satisfactory estimation of β ₂, and hence the role of the outcome Y in the missing data model, than the selection model.

ISSN: 1471-2288