Objective To describe patterns of caretaker-child communication during children's caretaker-assisted dietary recall interviews. Design Data are from transcriptions of audiotaped, caretaker-assisted, 24-hour dietary recalls. Statements were coded to provide quantified measures of caretaker-child interaction. Subjects/setting Thirty-four children aged 7 to 11 years who were enrolled (with their caretakers) at the San Diego site of the Olestra Post-marketing Surveillance Study. Statistical analyses Measures of participation for caretaker-child pairs were compared using Mann-Whitney tests; χ2 tests were conducted to examine within-group differences. Mean numbers of statement types were compared with t tests. Differences between contributions of caretakers and children within caretaker-child pairs were examined using Wilcoxon signed rank tests. Results Caretakers contributed primarily by adding food details and, secondarily, by prompting children. Children rejected a notable proportion of items added by caretakers, and children's knowledge of food details was considerable. Gender made little difference, although the presence of a male caretaker was associated with a reduced proportion of details contributed by children. Applications In assessing children's dietary intake, questions should be directed toward children, even when caretakers are present. Nutrition professionals must clarify expectations for caretakers from the outset, intervene as needed to limit caretaker participation, probe for children's disagreement with caretakers' additions, and question the basis for caretakers' additions. If feasible, caretakers should be excused toward the end of the interview so children may privately make alterations to the record.