It is well established that mothers' education has positive effects on child nutrition in developing countries. In school, girls can acquire skills which are later used to access modern health services and comprehend health messages. Less explored, however, is the effect exerted by the education of other individuals —the mothers' friends, neighbours and family — which may influence child nutrition directly or modify the effect of maternal education. Furthermore, questions remain about the mechanisms underlying the adult education-child nutrition relationship, especially the role of mothers' health and nutrition (HN) knowledge which has been debated in recent literature. Future research into the impact of adult education on child nutrition should therefore widen the focus from mothers only to others in her household, community and communication network and specifically examine the mediating role of mothers? HN knowledge.
The thesis has two main components. In the first part I analyse data collected in 2002 from the Young Lives (YL) study in Vietnam and Andhra Pradesh (AP) in India to assess the association between child height- and weight-for-age z-scores and adult education in the household and community. Adult education was measured as the level of formal schooling completed and community-level education was measured by aggregating individual-level data. Data were collected on 6019 households in 20 sites and 133 communities with children aged either 1 or 8 years old.
In the second part of the thesis I present the results of a cross-sectional study undertaken in 2004 of a sub-sample of YL mothers in AP, the Knowledge and Networks (KN) study, which was designed to explore mediators of the education effect, specifically the role of HN knowledge. Data were collected on the education, HN knowledge, general awareness, media, communication ability and communication networks of 302 mothers. I used multi-level regression modelling in Stata and MLwiN to adjust simultaneously for confounding and the hierarchical structure of the data-sets. The instrumental variable approach was used in an attempt to adjust for "endogeneity" of knowledge.
The results in part 1 show that the education of mothers, fathers and grandmothers has independent positive effects on child nutrition. A linear effect of parental education was found, with the greatest effect observed for a higher level of education. There was a positive effect of community-level maternal literacy on child height over and above parental education in Vietnam though not in AP. However, the distribution of educated individuals in the community did not have an additional impact. There was no difference between the countries in the effect of maternal education on child nutrition, although the effect was stronger in Vietnam for paternal education and stronger in AP for grandmother education. Overall, adult education was more strongly associated with child nutrition in households and communities that were relatively wealthy and well-educated than among the poorer and less well-educated. Furthermore, grandmothers' education was more strongly associated with boys' nutrition than girls'. Drawing on findings from previous research I explore possible reasons behind the social and spatial differences in education effect and hypothesize about underlying causal mechanisms.
Part 2 revealed that a mother's HN knowledge was associated with her education, the size of her communication network and her ability to communicate. The results were compared with another type of knowledge —referred to here as "general awareness" —which was determined by education, ability to communicate and use of telephones and newspapers. There was no evidence that HN knowledge or general awareness were endogenous variables in the causal model predicting child nutrition, although this may be explained by the weakness of the instruments used. Ordinary least squares regression results suggest that general knowledge may mediate the education effect, but that HN knowledge does not. Mothers? communication ability was identified as an important positive effect-modifier of the education effect and a potential mediator or confounder. The difficulties with the application of the instrumental variables approach are discussed in depth, along with the difficulties in disentangling the effect of mothers' identity and skills in shaping child nutrition.
In this thesis I demonstrate the important role of education among "influential others" beyond the mother in determining child nutrition, thereby calling for a less individualistic approach to research and policy. I challenge the attention given in recent years to the role of mothers' HN knowledge in mediating the education effect and argue instead that mothers' communication skills are more important, perhaps because they enable her to develop social networks and access support which benefit the health and well-being of herself and her family. An important issue worthy of further debate is the extent to which innate capabilities can be enhanced through education. I conclude by outlining key recommendations for future research which address both the methodological weaknesses of the present study and the remaining knowledge gaps in the literature.