Young children’s dietary recalls generally are conducted with a guardian present, but how this contributes to data accuracy is unclear. Furthermore, some assumptions underlying the preference for guardian presence may be unfounded. To investigate the range of guardian-child interactions within the diet recall setting, we examined transcriptions of guardian-assisted recalls conducted with 34 children aged 7-11 whose households were enrolled with the San Diego site of the Olestra Post-Marketing Surveillance Study (OPMSS). Although guardians did add to the breadth of data collected, children were quite knowledgeable about their diets. Moreover, they sometimes rejected guardian suggestions, and guardians generally could not provide assistance when children requested it. Guardian-child negotiations reflected cultural understandings regarding children, caretaking, and guardian-child power structures as well as food, and sometimes interviewers had to make very subjective data classification decisions. Suggestions for improving dietary recall methods are provided. Findings are relevant for other research methods involving children.