Z-scores for height, weight, and BMI of pre-schoolers with picky eating habits of different durations, and adjusted associations between z-scores and duration of picky eating behaviours.

† All of the models were constructed by using multilevel (two levels) mixed-effects liner regression with the iterative generalised least-squares estimation method. Results of the z-scores from the regression models with adjustment for child’s birth weight and feeding pattern during the first four months after birth, mother’s education, and family’s per capita monthly income. Results of intelligence from the regression models with adjustment for child’s age, gender, birth weight, and feeding pattern during the first four months after birth, mother’s education, and family’s per capita monthly income.‡ β represents the difference in mean z-scores between the non-picky eating group and the picky eating group after adjusting for the covariates listed above.a SE = standard error,b CI, confidence interval.Z-scores for height, weight, and BMI of pre-schoolers with picky eating habits of different durations, and adjusted associations between z-scores and duration of picky eating behaviours.