Skip to main content
Figure 3 | BMC Medical Research Methodology

Figure 3

From: Practical considerations for sensitivity analysis after multiple imputation applied to epidemiological studies with incomplete data

Figure 3

Analysis of the variable genotype 3 with δ = 0.15. Left panel: normalized weights (wm) versus θ ^ m (estimated logistic regression coefficient of genotype 3 in the imputed dataset m), for each imputed data set. The dash line represents θ ^ MAR (mean of θ ^ m over the 1000 imputed datasets). Right panel: running estimate, calculated as the moving average of the θ ^ MNAR according to the number of imputed datasets. On the right axis is plotted the ‘rug’ of the 1000 estimates θ ^ m for each imputed dataset. The dash line represents θ ^ MAR (mean of θ ^ m over the 1000 imputed datasets).

Back to article page