A comparison of missing data procedures for addressing selection bias in HIV sentinel surveillance data

Table 2 Bias in ${\hat{β}}_{1} = E ({\hat{β}}_{1}) - β_{1}$ , $β_{1} = 0$

Level of selection bias	Complete-case (Fixed-effects)	Complete-case (Mixed-effects)	Model 1	Model 2	Model 3
High	−0.080	−0.002	−0.173	−1.883	−0.411
Moderate	−0.001	0.002	−0.043	−0.887	−0.186
Low	−0.001	−0.001	−0.048	−0.715	−0.162

ISSN: 1478-7954