Assessing Validity and Reliability of a Nutrition Knowledge Questionnaire with High School-Aged Adolescents

To assess validity and reliability of a nutrition knowledge questionnaire originally developed for adults when used with high-school aged adolescents. The nutrition knowledge questionnaire utilized in this study was originally developed for adults and was shown to have content and construct validity as well as internal consistency and test-retest reliability. Two sections of the adult questionnaire, 50 questions total, were administered to adolescents at four high schools: three in California and one in Hawaii. Participants were convenience samples selected by teachers at each school based on known previous nutrition education (n = 174) or no known previous nutrition education (n = 136). Adolescents self-reported their current age and received a $10 gift card for completing the questionnaire. Questionnaire data were assessed for item difficulty, item discrimination with item-total correlation, and construct validity using an independent samples t-test. The overall questionnaire and individual sections were assessed for internal consistency reliability with Cronbach's alpha. Given the difference in nutrition education offered at each site, a post-hoc regression analysis was conducted to assess % correct by group when controlling for school. Assessment of item difficulty showed that one question exceeded the suggested 80% threshold, however this question was retained due to acceptable item discrimination. One question was found to negatively correlate with total score and was removed from the questionnaire resulting in 49 questions. The overall questionnaire was shown to have internal consistency reliability (Cronbach's α = 0.83) as were the individual sections (Cronbach's α > 0.70). The questionnaire also demonstrated construct validity with the adolescents that had known previous nutrition education having a significantly higher mean % correct compared to the adolescents without known nutrition education (P < 0.0001). The significant association between groups and % correct remained when controlling for school (P < 0.0001). Results suggests that the questionnaire was shown to reliably assess general nutrition knowledge in high school-aged adolescents with validity. UC ANR, USDA NIFA NNF, USDA NIFA Hatch, Jastro Research Award.