Development and test-retest reliability of a nutrition knowledge questionnaire for primary-school children.

<bold>Objective: </bold>To evaluate a web-based nutritional knowledge questionnaire for primary-school children.<bold>Design: </bold>Children's nutritional knowledge was assessed in five domains: healthy choices (twenty-seven items), estimated recommended portions/servings (eight items), nutrient content (five items), main food function (five items) and categorization of food items (eight items).<bold>Setting: </bold>The questionnaires were completed in school.<bold>Subjects: </bold>A convenience sample of 576 Belgian children (aged 7-12 years) from fourteen primary schools completed the questionnaire once, 386 completed the questionnaire twice.<bold>Results: </bold>Healthy choices could be answered correctly by 73 % of the children, nutrients by 59 %, food categorization by 49 %, main function by 38 % and portion estimation by 36 %. Children's test-retest intra-class correlations were 0·75 for healthy choices, 0·33 for nutrients, 0·61 for food categorization, 0·44 for main function, 0·47 for portion estimation and 0·76 for the total scale. The intra-class correlation was lower in the youngest age group (grade 2: 0·51, grade 4: 0·65, grade 6: 0·66). The total score was significantly lower in the retest. The instrument was in general positively evaluated by the children.<bold>Conclusions: </bold>The instrument is a promising, practical, inexpensive tool with acceptable test-retest reliability in fourth and sixth graders. [ABSTRACT FROM AUTHOR]