Skip to main content

Table 2 Specifications of the logistic regression models used to impose missing data under the MAR scenarios

From: A comparison of multiple imputation methods for handling missing values in longitudinal data in the presence of a time-varying covariate with a non-linear association with time: a simulation study

Variable

Odds Ratio

MAR (weak)

MAR (strong)a

 

Model A/ Eq. 5 b (exp(ν i ))

Model B/ Eq. 6 b (exp(ω i ))

Model A/ Eq. 5 b (exp(ν i ))

Model B/ Eq. 6 b (exp(ω i ))

1 Sleep problem at wave 1Yes

1.67

1.61

2.80

2.60

2 Sleep problem at wave 5Yes

1.64

1.58

2.70

2.50

3 Maternal smokingYes

1.61

1.58

2.60

2.50

  1. exp exponential, MAR missing at random
  2. aOdds ratio for MAR (Strong) = square of the Odds ratio for MAR (Weak)
  3. bModel A/ Eq. 5 and Model B/ Eq. 6 represent the logistic regression models used to generate missingness in BMI for age z-scores from waves 2–5 under MAR, to denote bmiz missing for all subsequent waves and intermittent missingness respectively