The parents’ weight and height were assessed though self-report. Only 75.3% of the fathers reported this information. Even though the results of univariate analysis showed the father’s BMI was significantly related to preschool overweight, the variable was excluded from the multivariate analysis. In this regard, the risk factors obtained from multivariate analysis might be different if less than 20% of the data were missing. Therefore, in future studies, researchers need to measure participants’ weights and heights so that all the data are collected and biases are eliminated. Future studies also should assess the mesosystem (eg. interaction between home and school), and the exosystem (eg. parents’ work setting, safety of the neighborhood) as suggested by the EcologicalEpidemiological model. Lastly, longitudinal studies might better measure the predictors before children become overweight and control measurement bias.