Skip to main content


Fig. 1 | BMC Medical Research Methodology

Fig. 1

From: Multiple imputation methods for handling missing values in a longitudinal categorical variable with restrictions on transitions over time: a simulation study

Fig. 1

a) Causal diagram for the association between maternal smoking and subsequent body mass index (BMI) for age z-scores developed based on the literature; m_age, maternal age at child birth; m_education, maternal education; sex, study child’s sex; birthweight, study child’s birth weight; ses, family socio-economic status z-score; breastfed, breastfeeding patterns; BMIz2-BMIz6, study child’s BMI for age z-scores at waves 2 to 6; m_smoking0-m_smoking6, maternal smoking at waves 0 to 6; m_depression0-m_depression6, maternal depression at waves 0 to 6; b) Causal diagram for MAR missingness. Rj is an indicator variable of missingness where maternal smoking at wave j were assigned to missing if Rj = 1. Only variables required to model the MAR missingness are shown in the figure

Back to article page