BACKGROUND: Food insecurity is a major public concern that occurs when nutritional needs are not met, incorporates psychological and physiological coping mechanisms, and can range from basic concern over obtaining food to severe malnutrition. This study was performed to explore differences in female and male respondent psychometric characteristics of a locally adapted Brazilian Household Food Security Scale (Escala Brasileira de Medida da Insegurança Alimentar--EBIA). METHODS: The 16-item EBIA was incorporated into the 2004 Brazilian National Household Sample Survey 2004 (Pesquisa Nacional por Amostra de Domicílios--PNAD; n = 108 606). Rasch Modelling was used to evaluate survey one-dimensionality, construct and independence through analysis of infit and relative item severities of adult and children items by gender. Differences in estimated item severities between male and female respondents were assessed using Differential Item Functioning (DIF) models. RESULTS: The scale presented good fitness and most item infit values were within adequate range (0.8-1.2), being practically identical when comparing female and male responses. Both female and male respondents presented similar relative item severities for adult and children items and followed the same pattern of increasing relative item severities with each item in the questionnaire. None of the items presented substantial DIF. CONCLUSIONS: This research demonstrates that the psychometric properties of the EBIA are not affected by respondent gender in Brazil. The results of this study support the validity of the proposed scale, suggesting that the scale will provide accurate information regardless of respondent gender for governments, researchers and agencies concerned with reducing epidemic levels of food insecurity and the resulting health disparities.
BACKGROUND: Food insecurity is a major public concern that occurs when nutritional needs are not met, incorporates psychological and physiological coping mechanisms, and can range from basic concern over obtaining food to severe malnutrition. This study was performed to explore differences in female and male respondent psychometric characteristics of a locally adapted Brazilian Household Food Security Scale (Escala Brasileira de Medida da Insegurança Alimentar--EBIA). METHODS: The 16-item EBIA was incorporated into the 2004 Brazilian National Household Sample Survey 2004 (Pesquisa Nacional por Amostra de Domicílios--PNAD; n = 108 606). Rasch Modelling was used to evaluate survey one-dimensionality, construct and independence through analysis of infit and relative item severities of adult and children items by gender. Differences in estimated item severities between male and female respondents were assessed using Differential Item Functioning (DIF) models. RESULTS: The scale presented good fitness and most item infit values were within adequate range (0.8-1.2), being practically identical when comparing female and male responses. Both female and male respondents presented similar relative item severities for adult and children items and followed the same pattern of increasing relative item severities with each item in the questionnaire. None of the items presented substantial DIF. CONCLUSIONS: This research demonstrates that the psychometric properties of the EBIA are not affected by respondent gender in Brazil. The results of this study support the validity of the proposed scale, suggesting that the scale will provide accurate information regardless of respondent gender for governments, researchers and agencies concerned with reducing epidemic levels of food insecurity and the resulting health disparities.