Fig. 5From: The effect of high prevalence of missing data on estimation of the coefficients of a logistic regression model when using multiple imputationRatio of estimated to empirical SE (when cholesterol is subject to missingness)Back to article page