CHILDHOOD EATING BEHAVIORS: PREVENTION OF CHILDHOOD OBESITY AND CHRONIC DISEASES
Location: Children Nutrition Research Center (Houston, Tx)
Title: Children's Categorization of Foods: Clusters for Food Search Strategies
| Baranowski, Thomas |
| Watson, Kathleen |
| Beltran, Alicia - BAYLOR COLLEGE OF MED |
| Knight-Sepulveda, Karina - BAYLOR COLLEGE OF MED |
| Missaghian, Mariam - BAYLOR COLLEGE OF MED |
Submitted to: International Society for Behavioral Nutrition and Physical Activity
Publication Type: Abstract Only
Publication Acceptance Date: June 20, 2007
Publication Date: June 20, 2007
Citation: Baranowski, T., Watson, K., Beltran, A., Knight-Sepulveda, K., Baranowski, J., Missaghian, M. 2007. Children's categorization of foods: Clusters for food search strategies [abstract]. Sixth Annual Conference of the International Society of Behavioral Nutrition and Physical Activity, June 20-23, 2007, Oslo, Norway. p. 204-205.
The purpose of this study was to identify categories of similar foods that are meaningful to children to facilitate their food search in a computer-administered self-completed 24-hour dietary recall (24hdr). One hundred forty-eight 8- to 13-year0old children sorted 62 cards with food pictures, from 18 professionally identified food groups, into piles they considered similar, and named each pile. Descriptive statistics, multidimensional scaling (MDS), and Robinson matrices (RM) (a form of cluster analysis) were used in analyses. Children sorted the 62 cards into a mean of 11.1 (+ 4.5) piles. The distributions of the food items on the ends of the first 4 dimensions from MDS were similar to RM clusters. As a result, clustering became the primary method of analysis. Eleven somewhat overlapping clusters emerged. Only minor differences in cluster solutions were detected in half samples split by age, language, ethnic group. BMI, or SES. Children used mostly taxonomic (e.g., food groups) or script (e.g., names of meals) categories to label their piles. Younger children sorted the cards similarly to the older children, but had more difficulty generating names/labels for piles. Food clusters emerging from children's categorization should facilitate search and identification of diverse foods in a computer-administered 24hdr. These clusters should provide intuitive and transparent groupings as the first level of food search, since they reflect how the foods were organized in children's memory.