Skip to main content

Table 5 Estimates from the full selection model

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

Parameter

Estimate

Standard error

β 1,1 (treatment)

-0.095

0.103

β 1,2 (age)

0.026

0.006

β 1,3 (sex)

0.160

0.118

β 1,4 (qualifications)

0.072

0.049

β 1,5 (deprivation)

-0.028

0.069

β 1,6 (conscientiousness)

-0.077

0.073

β 1,7(determination)

-0.010

0.084

β 1,8(support)

0.053

0.046

β 1,9(dependence)

-0.014

0.037

β 1,10(previous)

-0.219

0.114

β 2 (Y)

1.561

3.535

  1. Despite our reservations, and the criticisms in literature of attempting this, the full MNAR selection model is identifiable and resulted in the following parameter estimates. See section Fitting the full model for a discussion of the difficulties in fitting the MNAR model.