The effects of transformations on category membership boundaries for natural and processed foods

• In Event: Towards a Holistic Developmental Investigation of Children’s Eating Behaviors

Abstract

Children’s categorization strategies differ based on an item’s origins, for example, natural kinds and artifacts. However, it is unclear if and how foods fit in these categories since these can be both natural and human-made (processed), and both types of foods have a history of human intervention. Across two studies, this research examines the effect of transformations on artifacts, non-food natural kinds, and natural and processed foods.
In Study 1, ninety-seven 4- and 5-year-olds, 6- and 7-year-olds, and adults viewed pictures of artifacts (ball, pencil), natural kinds (branch, cactus), natural foods (apple, broccoli) and processed foods (bread, chocolate) transformed by being cut or crushed. Participants were asked whether the transformed item was part of the original category or not (e.g., “Is it an apple or is it not an apple?”). A linear mixed-effects model using a Binomial distribution was conducted to analyze the probability of giving same-item responses, which indicated the transformed item was part of the original category. Results revealed a main effect for Item Type, Age, and Transformation; however, these were qualified by two 2-way interactions: Item Type x Age interaction, (χ2 (6) = 45.26, p < .0001, and Transformation x Age (χ2 (2) = 16.45, p = .00027. The interactions revealed more same-item responses for foods compared to non-foods for the two older age groups, but not for the youngest age group, with more same-item responses for cut items than crushed for all age groups. See Figure 1.
Study 2 included a broader array of foods and non-foods and two new transformations to examine the generalizability of Study 1 results. There were ninety-two 4- and 5-year-olds, 6- and 7-year-olds, and adults who were told about artifacts (ball, pencil, car, clock), natural kinds (branch, cactus, fly, worm), natural foods (apple, broccoli, egg, shrimp), and processed foods (bread, chocolate, cheese, chicken nugget) undergoing three transformations: taking the insides out, crushing it, and melting it. The test question and type of response recorded were the same as in Study 1, as was the statistical analysis. The results revealed the same main effects, which were qualified by three 2-way interactions: Item Type x Age interaction, (χ2 (6) = 35.32, p < .0001, Transformation x Age (χ2 (4) = 77.50, p < .0001, and a Transformation x Item Type (χ2 (6) = 38.06, p < .0001. It revealed for all age groups, there were more category membership responses for foods compared to non-foods, however only the older age groups differentiated between processed and natural foods. In addition, older age groups provided fewer same-item responses for melted items compared to the other two transformations. See Figure 2.
The results support that even young children do not apply identical categorization strategies for foods compared to non-foods. Moreover, older children and adults do not apply a single, stable strategy for foods with different origins. Theoretically, it implies that foods may be a hybrid of both natural and artifactual domains, and practically it informs developmental trends on specific properties that guide food selection and rejection.

Authors

• Helana Girgis, Stockton University
  Presenting Author

• Simone Nguyen, University of North Carolina at Wilmington

• Jérémie Lafraire, Institut Paul Bocuse

• Jean-Pierre Thibaut, Université de Bourgogne Franche-Comte

• Camille Rioux