Fig. 11From: The effect of high prevalence of missing data on estimation of the coefficients of a logistic regression model when using multiple imputationEstimated standard error (when smoking is subject to missingness)Back to article page