To assess how women’s dietary diversity is influenced by enabling resources and to quantify possible pathways through women’s agency.
We analyzed cross-sectional data collected in early 2015 from 2,600 women during the baseline survey for a cluster-randomized controlled trial in northeastern Bangladesh. Dietary diversity was measured through a standard 10-Food Group Indicator (FGI-10R). Resources were measured as years of schooling. We applied factor analysis to explore and then test the factor structure of women’s agency. The final agency model had two factors: women’s voice with husband and social solidarity with other women. We used a linear mediation model to estimate indirect, direct, and total effects, adjusting for household wealth and women’s age.
In adjusted models, the direct effect of schooling on dietary diversity was 0.0001 (p=.026) and the total effect was .0001 (p=.037). Schooling did not have significant direct effects on voice with husband (p=.077) or social solidarity (p=.169). The direct effect of voice with husband on dietary diversity was 0.388 (p<.001). Social solidarity had no significant effect on dietary diversity (p=.191).
In our study population in Bangladesh, higher levels of schooling had very small direct and total effects on dietary diversity; the association was not mediated by women’s agency (as measured by voice with husband or social solidarity). Women with higher levels of voice with husband had higher dietary diversity. Research is needed to examine intra-household communication and other possible pathways to improved dietary diversity, including through other dimensions of women’s agency.