The relationship between women's empowerment and women's nutrition is understudied. We aimed to elucidate this relationship by quantifying possible pathways between empowerment and dietary diversity among women in rural Bangladesh. In 2015, we conducted a cross-sectional survey of 2,599 married women ages 15-40 (median: 25) living in 96 settlements of Habiganj District, Bangladesh, as a baseline for the Food and Agricultural Approaches to Reducing Malnutrition trial. We collected data on women's empowerment (highest completed grade of schooling and agency), dietary diversity, and demographic factors, including household wealth. We used exploratory factor analysis and confirmatory factor analysis on random split-half samples, followed by structural equation modelling, to test pathways from schooling, through domains of women's agency, to dietary diversity. Factor analysis revealed 3 latent domains of women's agency: social solidarity, decision-making, and voice with husband. In the adjusted mediation model, having any postprimary schooling was positively associated with voice with husband (β 41 = .051, p = .010), which was positively associated with dietary diversity (β 54 = .39, p = .002). Schooling also had a direct positive association with women's dietary diversity (β 51 = .22, p < .001). Neither women's social solidarity nor decision-making mediated the relationship between schooling and dietary diversity. The link between schooling and dietary diversity was direct and indirect, through women's voice with husband but not through women's social solidarity or decision-making. In this population, women with postprimary schooling seem to be better able to negotiate improved diets for themselves. Few studies have examined whether women's empowerment influences their dietary intake. One study in Bangladesh found a positive association between women's empowerment and household dietary diversity, but did not examine women's dietary diversity (Sraboni, Malapit, Quisumbing, & Ahmed, 2014). Other studies have assessed the relationships between women's empowerment and child nutrition in South Asia (Bhagowalia, Menon, Quisumbing, & Soundararajan, 2012;Cunningham, Ploubidis, et al., 2015;Malapit, Kadiyala, Quisumbing, Cunningham, & Tyagi, 2015;Richards et al., 2013;Shroff, Griffiths, Adair, Suchindran, & Bentley, 2009 2 | METHODS
| Participants and proceduresWe conducted a secondary analysis of data from the baseline survey and all 2,895 eligible women were targeted for inclusion in the study.In total, 2,624 women (91%) were found at home (at repeat visits), consented to participate, and were enrolled. A total of 2,599 enrolled women (90% of the 2,895 eligible women) completed the household and women's sections of the survey.
| VariablesThe achievement or outcome of interest was women's dietary diversity, measured by asking the woman about her food consumption in the previous day through open recall followed by list-based probes, categorizing diets into 21 food categories. When women reported consuming a food ite...