| dc.description.abstract | This dissertation embraces three topics for Ghana, Liberia, and Senegal: agricultural production and household food security, family planning communications, and empowering rural women.
From 2012-2013 field survey data representing agroclimatic and farming conditions, our Chi-Square Automatic Interaction Detection identified community support and gender-based farming practice as the most significant predictors of Liberian and Senegalese food insecurity. In Liberia, almost all severely food-insecure households had no village-wide food support in the worst food-insecure (rainy) season. More decision criteria, including crop selling locations, natural irrigation amid inadequate farming technologies, agricultural information, off-farm income, land conflict, and informal labor, followed. In Senegal, female-headed households produced less food than male-headed households, so was the amount of Zakat (involuntary almsgiving) and Sadaqah (donation). For Ghanaians, the random forest algorithm identified crop income and farming inputs as more important predictors than exogenous assistance, gender-based farming practice, and land conflict.
Despite more outside farmworkers, technologies, and investment in Ghana than Liberia or Senegal, rural youth often leave their villages. Ironically, sub-Saharan Africa will only grow the population steadily in this century caused by unmet needs among teenagers.
We applied two seemingly unrelated methods— Instrumental Variable (IV) estimations and Coarsened Exact Matching (CEM)-weighted multiple regressions for aggregated information of Senegal Demographic and Health Surveys (DHS). The result was consistent that to 15-to 49-year-old women who were married, living with partners, fecund, and pregnant, radio communications in family planning and contraception independently reduced the optimal number of children desired. Joint decisions of female and male partners, or wives and husbands, on family planning are essential.
We then found from multiple years of DHS the effects of wives’ and husbands’ education, literacy, and interaction of age and urban residence on positive household wealth. Our ordinal categorical regressions also found evidence that more children five years or below, higher intimate partner violence (IPV) tended to lower wealth. Average marginal effects implied more exposure of rural women to IPV, and they had fewer educational opportunities than their urban sisters. Our study suggests that the mutual support of husbands and wives for equal education is of paramount importance. | en |
