Validation of a food frequency questionnaire forsoybeans consumption: Three statistical

The use of soy protein products by the food industry has increased due to its relatively low cost, and mainly for its functional characteristics. The problem is that the consumer is unaware of soy presence in processed products and that it can be transgenic. To validate a food frequency questionnaire (FFQ), using three different statistical methodologies, to identify the consumption of food products that contain soy as an ingredient from a Collective Feeding menu. A FFQ was created according to the food products and preparations, which contained soy declared, as an ingredient, on the label, and which were identified on the menu of a Collective Feeding Unit. The consumers were split into two groups (G1 and G2) of fifty subjects in each. The consumption frequency was defined by six gradients of score’s categories. Three methods were used for validation: Spearman, Bland&Altman and analysis of agreement-Kappa.The correlation coefficient of Spearman (r), achieved a value of 0.77 with variation of 0.6136 to 0.8642 (p<0.0001). The coefficients of Cronbach were 0.88 and 0.79 for G1 and G2, respectively. According to Bland&Altman, it is possible to verify agreement in the responses of both groups, because the differences from the modal scores of frequency’s consumption for products were distributed on the zero difference. The general Kappa value was 0.614. Conclusion: The FFQ was validated by the methods of Spearman’s correlation, Bland&Altman and analysis of agreement-Kappa. These enabled us to draw the consumption profile of the interviewed group.