Several study limitations must be acknowledged. The crosssectional nature of this study precludes drawing conclusions regarding causality, and bias may have been introduced as a result of the FEAHQ respondents acting as proxies for their spouses and children. To validate parental reports as proxies for child dietary behaviours, we assessed the concordance between child and parent responses to a food frequency questionnaire in a comparable sample of children from the Niagara Region. In this analysis, we found moderate agreement between child and parent responses (r=0.40 to 0.69), suggesting that use of child reports would perhaps have yielded similar results.33 Nevertheless the possibility remains that differences in findings might have been observed had the children completed the FEAHQ on their own. The agreement between spousal reporting has been reported previously to be high across all items assessed in the FEAHQ (r=0.81 to 0.94), suggesting that having one spouse respond to the survey on behalf of the other was also appropriate