Degree Name

PhD (Doctor of Philosophy)


Early Childhood Education

Date of Award


Committee Chair or Co-Chairs

Amy Malkus

Committee Members

Jane Broderick, William Andrew Clark


The literature indicates a lack of validated scales to measure nutrition knowledge, attitudes and behaviors in preschool-aged children. Reliable and valid assessment tools are critical in the development of nutritional interventions for young children; careful design, including psychometric evaluation of reliability and validity, is the best approach to scale development. Best practice points to an efficient, yet comprehensive look at the constructs of knowledge, attitudes, and behaviors. This multi-phase study included 298 preschool children ages 2-6 years (151 girls, 147 boys; mean age 3.936). A 45-item scale was created, with 4 subscales to measure the nutrition knowledge, attitudes, and behaviors of preschool children, and piloted among 54 of these children (30 girls, 24 boys; mean age 4.3). Cronbach’s alpha (α) for the total scale with predicted behavior excluded was .775, and also determined for each subscale, including food identification (0.642), food group categorization (0.644), and classification of foods as healthy and unhealthy (0.576). Behavior was predicted, as children created a virtual plate of preferred food and beverage selections for breakfast, lunch, and snack. Internal consistency was not calculated for this stage of scale development. ANOVA and post-hoc testing identified age-related differences between 3 and 5-year olds, in overall performance on the scale, F(3, 43) = 6.183, p = .001, partial ŋ2 = .301 but not among other age groups (3, 4, 5, and 6 included).

Through multiple iterations of the scale, it became clear that the healthy versus unhealthy subscale was problematic. An adaptation of the Traffic-light diet, a variation of this subscale, was created and piloted with 74 items, and then narrowed to 25 items (Cronbach’s α .924). This resulted in a revised final version of the scale, with 52 items including 8 food identification tasks, 15 food group categorization tasks, 25 Go, Slow, Whoa tasks, and 4 virtual plate food and beverage preference tasks, for delivery on interactive technology (iPad). Future research will include pre-post testing with a nutrition intervention to determine further validity and test-retest reliability. It is predicted that Cronbach’s α will be similar if not improved by further testing.

Document Type

Dissertation - unrestricted


Copyright by the authors.