Chi-square results from logistic regression models with eating-related variables at age 19–20 as the dependent variable and pubertal development, sex, and zygosity as the independent variables. Generalized estimating equation modeling corrections applied to account for non-independence of the data.