Abstract
Data from the 2016-17 Multiple Indicator Cluster Survey from Nigeria are used to study the relationship between child stature, mother's years of education, and indicators of early childhood development (ECD). The relationships are contrasted between two empirical approaches: the conventional approach whereby control variables are selected in an ad-hoc manner, and the double machine-learning (DML) approach that employs data-driven methods to select controls from a much wider set of variables and thus reducing potential omitted variable bias. Overall, the analysis confirms that maternal education and the incidence of chronic malnutrition have a significant direct effect on measures of early childhood development. The point estimates based on the ad-hoc specification tend to be larger in absolute value than those based on the DML specification. Frequently, the point estimates based on the ad-hoc specification fall inside the confidence interval of the DML point estimates, suggesting that in these cases the omitted variable bias is not serious enough to prevent making causal inferences based on the ad-hoc specification. However, there are instances where the omitted variable bias is sufficiently large for the ad hoc specification to yield a statistically significant relationship when in fact the more robust DML specification suggests there is none. The DML approach also reveals a more complex picture that highlights the role of context. In rural areas, mother's education affects early childhood development both directly and indirectly through its impact on the nutritional status of both older and younger children. In contrast, in urban areas, where the average level of maternal education is much higher, increases in a mother's education have only a direct effect on child ECD measures but no indirect effect through child nutrition. Thus, DML provides a practical and feasible approach to reducing threats to internal validity for robust inferences and policy design based on observational data.
Highlights
In an ideal world, recommendations for social policies aimed at improving welfare and especially maternal and child outcomes would be based on well-established evidence on the causal relation between the variables of interest
Given the rapid growth faltering in the first two years of a child’s life and the fact that measures of early childhood development (ECD) are available only for children 36–59 months old, this study investigates whether there is a causal relationship in the full sample of children 0 to 59 months of age, and whether there are any structural differences in the causal relation between maternal education and child stature separately in the sample of younger (0–35 months) and older (36–59 months) children
The double machine-learning (DML) estimates are preferred because they are estimates of the causal effect with a more reasonable set of assumptions, they minimize potential omitted variable bias, and they allow for the inclusion of different confounding factors influencing the outcome variable (ECD measure) and the variables for which inferences are desired and maternal education, in this paper
Summary
Recommendations for social policies aimed at improving welfare and especially maternal and child outcomes would be based on well-established evidence on the causal relation between the variables of interest. Evidence of a direct effect of maternal education on ECD, separate from its potential effects on child stunting, would suggest that maternal education has the advantage to compensate for the potentially negative effects of stunting on ECD and provide even stronger justification of the need to increase the quantity and quality of female education as an especially important factor in the development and economic growth of SubSaharan Africa countries undergoing a demographic transition With this background in mind, the analysis in this paper takes advantage of the variation across households in indicators of early childhood development and applies recently developed methods for more robust inferences based on observational data [25]. Mothers in the North-Central zone have about half the years of education than mothers in the southern zones, and mothers in the North-East and North-West zones have less than one-third the years of education than mothers in the South
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
