Approaches to understanding and measuring women’s empowerment and its relationship with women’s and children’s dietary diversity in sub-Saharan Africa.

Abstract

Background

Women’s empowerment has been identified as a key component of development policies since in many cases, there are existing and widening gender gaps in wellbeing including in access to improved nutrition especially in low-and-middle- income countries (LMICs). In addition, in many LMICs especially in those found in sub-Sahara Africa (SSA), there are existing gender-based barriers including norms, attitudes, and patriarchal societies that undermine the opportunities given to women. While there are many interventions, programmes, and field experiments that aim to redress these gender gaps in many LMICs, there is the need to understand and examine the concept and role of women’s empowerment in access to welfare-improving inputs including improved dietary diversity for themselves and their children.

Study objective

To add to this growing literature, this thesis sought to examine the concept of women’s empowerment and its relationship with women’s and children’s dietary diversity using a mixed-methods approach of qualitative focus group discussions (FGDs) and in-depth interviews (IDIs), and quantitative secondary data. Methods Studies 1 and 2 - For the qualitative studies, 89 married participants (64 women of reproductive age and 25 men) who were cohabiting with their spouses were recruited in Southeastern Nigeria in April, 2019. Of this number of women, 38 participated in IDIs while the rest took part in 2 FGDs of 13 participants each. The 25 men participated in 2 FGDs of 12 and 13 men respectively. In the IDIs, participants were grouped into two based on their primary economic activity. One group engaged only in farming and the other engaged in other economic activities (mainly hairdressing and food vending) in addition to farming. For the FGDs, men and women participants were recruited using a form of snowball sampling strategy from the IDIs participants and were also grouped according to their primary economic activity; namely farming only, or food trading as well as farming.

Study 1 reports findings from the IDIs and FGDs concerning the concept of empowerment and, in particular, women’s empowerment using domains contained in two quantitative measures and one qualitative guideline. These measures were respectively; the Women’s Empowerment in Agriculture Index (WEAI), the Survey-based Women's emPowERment (SWPER) index for women's empowerment in Africa, and the Food and Agriculture Organization guide to measuring women’s empowerment and social protection. Men were recruited to permit gauging their perception of empowerment and women’s empowerment.

Study 2 used IDIs to explore women’s consumption of food items contained in the Minimum Dietary Diversity for women of reproductive age (MDD-W) measure and assessed important household dynamics that could influence food consumption decision-making.

Analyses for the IDIs and FGDs were done in excel and NVivo and followed the principle of constant comparison.

Studies 3 and 4: – Study 3 used secondary quantitative data from the Demographic and Health Surveys (conducted on 14,688 respondents) and Study 4 used the Feed the Future (on 10,041 respondents in the FTF) baseline studies. Data was extracted from each of the DHS and FTF for five sub-Sahara Africa countries (Mozambique, Rwanda, Malawi, Uganda, and Zambia). Study 3 explored the relationship between women’s empowerment measured by the SWPER index (which contains three domains of empowerment; attitude towards violence, autonomy (social independence), and decision-making) and children’s dietary diversity measured by using the Infant and Young Children Dietary Diversity Score (IYCDDS). IYCDDS was the outcome of interest and was examined as a continuous and count variable while the seven food groups contained in it were treated as dichotomous variables (i.e. consumed or not consumed). Analyses were restricted to children between 6 and 23 months in line with the IYCDDS guidelines. The three domains of the SWPER index were examined as the key independent variables and other important demographic, economic, and geographic covariates were specified as controls. Interaction effects between the three SWPER domains of empowerment, gender of index child and wealth index were also explored to determine if women’s empowerment has differential effects on IYCDDS for boys and girls, and based on women’s socioeconomic status. Study 4 examined the relationship between women’s empowerment and women’s dietary diversity using the FTF dataset. Empowerment was measured using the WEAI index which contains 10 indicators of empowerment within 5 broader domains of empowerment. The 10 indicators and the aggregate empowerment score were treated as key independent variables while Women’s Dietary Diversity Score (WDDS) was treated as the outcome of interest. The WDDS was examined as a continuous and count variable and the food groups contained in the WDDs were also treated as dichotomous outcome variables. Analyses were restricted for women of reproductive age in line with the WDDS guidelines and those that engaged in agriculture in line with the WEAI index, and important demographic, economic, and geographic variables were specified as controls. Ordinary least squares (OLS) regression verified with marginal effects from Poisson regression analyses were used for the continuous outcome variables and linear probability models (LPMs) verified with marginal effects from logistic regression analyses for the dichotomous outcome variables. Analyses of the two quantitative datasets were appropriately weighted and cluster-adjusted, and significance was established at 95% and 99% confidence intervals.

Findings

Study 1 findings suggest that local understanding of women’s empowerment dynamics does not always resonate with external definitions and ideologies. Contextual factors play a significant role in determining the extent of positive effects of empowerment. It was challenging to identify a direct translation of “empowerment” into the local language with the same meaning when back-translated to English. Different local terminologies elicited different responses from participants with a gender divide where men pushed back on one particular local terminology. This suggests that terminologies used in presenting empowerment interventions might impact on their acceptance and success and make this a challenging term to use in cross-cultural and cross-country research.

Study 2 findings suggest that the consumption of nutrient-rich food groups, especially certain legumes, nuts and seed, flesh protein, and eggs, would benefit from improved women’s economic empowerment, and the consumption of fruits and vegetables might benefit from improved agricultural practices in addition to economic empowerment. By contrast, staple food-items including grains and root-tubers that are consumed by all women irrespective of their income-earning status would not benefit so much. Dietary diversity is influenced by food production and purchase, where factors including seasonal variation in food production and prices are important determinants. Economic empowerment improved women’s autonomy in food purchase and consumption. However, limited income restricted women from exhibiting full autonomy in consumption decisions and access.

Study 3 findings suggest that there were significant associations between women’s autonomy and improved child’s diet diversity in analyses pooled across five countries, and that Uganda and Zambia might account for these significant associations. There were no significant associations between the three domains of empowerment and improved IYCDDS in Mozambique, Rwanda and Malawi. Results from Uganda suggested a differential impact for boys and girls where women’s improved autonomy was protective for a female child’s dietary diversity. Only Mozambique exhibited a significant positive association in the interaction between the domains of empowerment and wealth index. Findings suggest that women in poorer households who viewed violence as disempowering might practice improved dietary consumption for their children.

Findings also suggest that legumes, dairy and dairy products, and other vitamin A-rich fruits and vegetables account for the association between improved autonomy and IYCDDS in the pooled analysis and that Uganda and Zambia might account for these significant associations.

These findings suggest that improvement in women’s autonomy might confer the most dietary benefit for infant and young children through improved consumption of dairy and dairy products, grains and fruits and vegetables. However, these significant associations were only found in two (Uganda and Zambia) out of the five countries examined further suggesting that the benefits from improved autonomy on improved food consumption for infant and young children differs, and this difference in benefits also differ across countries.

Study 4 – Four empowerment indicators (autonomy, input in production decisions, empowerment in public speaking, and working less that 10.5 hours in a day) out of the 10 indicators used in the WEAI were positively and significantly associated with improved WDDS and food consumption in the pooled and disaggregated regression analyses. There were different significant associations between the four indicators of empowerment and women’s food consumption in three out of the five countries examined (no significant associations were found in Malawi and Zambia).

Autonomy and input in production decisions were significantly associated with improved WDDS and findings suggest that Uganda and Rwanda might account for these associations. Autonomy in production was associated with the likelihood of consumption of grains and root-tubers, dairy and dairy products, flesh proteins and vitamin A-rich vegetables and fruits in Uganda. In Rwanda, input in production decisions was associated with the consumption of other fruits and vegetables including vitamin A-rich produce.

Empowerment in public speaking was significantly associated with improved WDDS and consumption of other fruits and vegetables in the pooled analyses and Mozambique and Rwanda might account for these significant associations. In Mozambique, empowerment in public speaking was associated the consumption of other vitamin A-rich fruits and vegetables and in Rwanda, empowerment in public speaking was associated with the consumption of grains/tubers, flesh protein, vitamin A-rich leafy greens, and other fruits and vegetables. Non-excessive workload (i.e. working for less than 10.5 hours in 24 hours) was significantly but negatively associated with improved WDDS and Mozambique accounted for this association, where women who worked below 10.5 hours in a 24- hour cycle were less likely to consume flesh proteins.

The differential performance of the four indicators in the WEAI index further suggests that different empowerment strategies might confer different benefits towards consumption of different food items and these benefits might vary across countries. In summary, this thesis found that local knowledge, perceptions and norms play a significant role in women’s empowerment discourse and this will impact on how women’s empowerment is conceptualised and how different empowerment measures perform. In addition, conceptualizing women’s empowerment ought to be done in a manner that recognizes the multifaceted linkages between the different domains of empowerment and how they dictate women’s participation in rural economic activities. Economic empowerment might be beneficial for the consumption of nutrient-dense food items including flesh proteins that are expensive for women to purchase, while autonomy, including in having input in production, might be beneficial for the consumption of nutrition-vital dairy and dairy products, fruits and vegetables for women and children. Leadership empowerment through empowerment in public speaking might be important for the consumption of grains and tubers and legumes in addition to nutrition-vital fruits and vegetables for women. However, these benefits might differ across different SSA countries.

Conclusion Interpretation of the qualitative findings should be done with caution due to the limitations inherent in the study design including its limited generalisability. However, the qualitative studies contribute to a very limited area of knowledge and are the first to qualitatively examine two known measures of women’s empowerment, and household-level dynamics between women’s empowerment and food consumption. The quantitative studies are prone to limitations including recall bias of women’s and children’s food consumption and other retrospective data that rely on the memory of past events, however, the quantitative studies advance knowledge in the measurement and analyses of women’s empowerment using two different but somewhat complementary measures of empowerment across multiple countries in two regions of sub-Sahara Africa. This enables a comparison of the performance of the two indexes in determining women’s and children’s dietary diversity.

This study provides a few key contributions to knowledge. Firstly, women’s empowerment needs are diverse and largely context-specific: hence, local understanding of empowerment concepts and terms remains an important determinant of the rate of acceptance and successes of women’s empowerment interventions in developing countries. Secondly, economic empowerment in the form of income earning improves women’s consumption of certain food items including flesh proteins and some legumes, nuts and seeds, which are usually nutrient-rich and more expensive however, limited economic prospects for women still limit women from expressing full empowerment in food consumption. Thirdly, women’s autonomy including in production is important for improving dietary diversity for women and children through the consumption of important products especially dairy, and vitamin A-rich fruits and vegetables. However, these benefits differ across countries examined. In addition, while empowerment in public speaking improved women’s dietary diversity, it appears that women might have to work excessively to improve their dietary diversity since non-excessive workload reduced women’s dietary diversity and consumption of flesh proteins. An overall observation is that an explanation of the variation in effect of women’s empowerment measures in different country contexts, perhaps lies in the difficulty of measuring women’s empowerment in a non-context-specific way.