The analysis of the quality of the diet allows evaluating the complexity of the individuals’ food practices. The used questionnaires should provide valid and accurate information, even when they are available in computerized form. The Item Response Theory (IRT) classifies a parameter of interest within a scale of values, allowing the evaluation of the quality of the diet. This study aims to develop a diet quality assessment scale using the Item Response Theory analyses. It is a methodological study carried out in three stages. In the theoretical stage, the items related to the behavioral eating practices of the adolescents and young adults were identified on the literature review. Items and their response options were prepared for each item following the new Food Guide for the Brazilian Population and the recognized national and international recommendations. In focus groups, food and nutrition experts evaluated the relevance of the items to measure the diet quality, and the writing clarity for adolescents and young adults. The discussions were recorded and later transcribed. The transcripts were analyzed by the exploratory content analysis. A codebook with the descriptions of the themes and sub-themes was elaborated to identify the suggestions of the experts. The writing of the items was also evaluated by the individuals with similar characteristics to those participating of the study. The items were organized in small sets using the Incomplete Balanced Blocks design and were transferred to an online questionnaire on the Survey Monkey platform. In the empirical stage, the questionnaire was made available on the World Wide Web in order to collect the answers. In order to participate in the study, the individuals should meet the inclusion criteria and register their agreement. Finally, in the analytical stage, the dimensionality of the items was analyzed using the full information factor analysis. The discrimination and difficulty parameters for each item were estimated using the Item Response Theory analysis. The items with adequate parameters were maintained on the final scale and used to calculate the individuals’ scores. The anchor items, those with more information at each level, are being identified to define the levels of interpretation of the scale.
Betzabeth Slater – Principal investigator (FSP/USP)
Thanise S. S. Santos – Principal investigator (FSP/USP)
Maria Alice Altenburg de Assis – Collaborator (UFSC)
Dalton Francisco de Andrade – Collaborator (UFSC)
Fernanda Baeza Scagliuzi – Collaborator (FSP/USP)
Priscila de Morais Sato – Collaborator (FSP/USP)