Fig. 3From: The effect of high prevalence of missing data on estimation of the coefficients of a logistic regression model when using multiple imputationStandard deviation of estimated coefficient across simulation replicates (when cholesterol is subject to missingness)Back to article page