“…Variables were imputed using the following models: maternal age, education, housing tenure, HCI, vitamin C, vitamin E, and carotenoids using an ordinal logistic regression model; and SGA, binge drinking, smoking, parity, ethnicity, child gender, and breastfeeding (auxiliary variable), using binary logistic regression models. We conducted a sensitivity analysis to test the MAR assumption (Héraud‐Bousquet et al., ) by exploring the extreme effect estimates compatible with the observed data (Leacy et al., ). We hypothesized that, due to social‐desirability bias, pregnant women who do binge drink are less likely to report this (Meiklejohn et al., ; Wild et al., ).…”