Cognition of Women about the Combination of Foods to Make Iron Bioavailable in the Diet: A Confirmatory Factorial Study

Objective: to validate an instrument designed to assess the knowledge of adult women about combining foods to make dietary iron more bioavailable.

Design Study: cross-sectional, Confirmatory Factor Analysis (CFA).

Place and Duration: Rio de Janeiro, Brazil, May 2018 and February 2019.

Methodology: sample: n=222 women 15 to 49 years old, not pregnant. The questionnaire was designed on Likert scale. Kaiser-Meyer-Olkin (KMO) and Doornik-Hansen tests were applied, as well as the AFC models with Promax rotation. Convergent validity (CV) and discriminant validity (DV) and reliability (Cronbach's alpha) were evaluated.

The Quality of Model Fit was Assessed by: Root Mean Square Error of Approximation (RMSEA), Comparative Fit Index (CFI) and the Tucker-Lewis Index (TLI). The dimensional structure of the questionnaire was reevaluated by Structural Equation Modeling (Modification Indices (MI) and Expected Parameter Changes (EPC).

Results: Initial model: total KMO 0.75 (min = 0.63, max = 0.84), Doornik-Hansen (chi2 (30) = 208.91, P = .00), eigenvalue (F1 = 2.87 and F2 = 1.98), CV not reached (F1: AVE = 0.30 and F2: AVE = 0.21) and DV corroborated, Conbrach alphas reached for 0.75 inhibitory items ( 95% CI = 0.69 to 0.81), for stimulators: 0.62 (95% CI = 0.51 to 0.70), RMSEA = 0.10 (90% CI = 0.08 to 0.11 ), CFI (0.70), TLI (0.65), items with MI>10 (i16I, i23E, i25E, i28E). Re-specified Model: eigenvalue (F1 = 2.44 and F2 = 1.42), CV not reached (F1: AVE= 0.29 and F2: AVE=0.20) and corroborated DVl, Conbrach alphas: for 0.75 inhibitory items (95% CI = 0.69 to 0.81); stimulators items: 0.55 (95% CI = 0.44 to 0.64), RMSEA = 0.03 (90% CI = 0.00 to 0.06), CFI (0.95), TLI (0, 94). The items did not show MI>10.

Conclusion: The reliability was achieved in both models. Validities: convergent was not achieved and the discriminant attested two-dimensionality.