Unequal access to quality education can have wide-ranging effects on the lives of children, influencing everything from educational attainment to labor market outcomes. While much research on this topic has looked at individual factors, such as household environment or school quality, comparatively little work has considered these factors jointly.

In their recent HCEO working paper, ECI/HI network member Jere Behrman and co-authors Paul Anand, Hai-Anh H. Dang, and Sam Jones analyze the joint effects of household resources and schooling quality in the context of low-income countries.

“There’s been a big discussion for many developing countries around low learning outcomes,” Jones says. “It’s called a learning crisis. Motivated by that, we’re interested not just in the average level of outcomes, but how learning outcomes are varying across kids. We also want to look at inequalities, particularly from a social justice point of view. Because if you have a poor or weak opportunity for a good education, this is likely to have significant implications in adult life. The idea being that having a good education is a key way to improve social mobility.

“Our motivation was really to say: can we decompose or disaggregate the contribution of these different factors that have often been looked at separately, and get a sense of not just how important schools and households are, but also, critically, how they might be co-varying.”

To study this, the authors developed a framework to explain how both factors and their interactions affect variations in learning outcomes. They then applied this framework to a rich micro data set for over one million children from three East African countries: Kenya, Tanzania, and Uganda. The data include test scores from large-scale household surveys conducted since 2010 by the Uwezo initiative. The authors note that the surveys were “designed to be representative at both national and district levels, based on the administrative classifications in the most recently available population census.” The surveys also collect information on household characteristics and the demographic and educational details of the children. During the survey, school-age children also take basic oral literacy and numeracy tests, which are “anchored to skills that should be achieved by the majority of pupils after two years of completed schooling.”

The authors find that while household factors are an important source of inequality in educational opportunity, school quality and sorting also play a significant role. Family effects contributed about 15 percent to the difference in outcomes. But by looking at the factors together, the authors write, “we find that inequality in educational opportunity is substantial, accounting for almost half of all variation in test scores. However, given the importance of schools and sorting within this total, it follows that educational (school) reforms that alter the distribution of school quality, such as via the allocation of teachers across schools, can enhance opportunities for the most disadvantaged.”

“What we find is that schools do contribute a significant variation in learning outcomes,” Jones says. “If you think of the tenth-worst school to the tenth-best school, it’s associated with a very large difference in learning outcomes of around .66 standard deviation. That’s a large difference in the kinds of magnitudes we’re talking about.”

The authors also found evidence of positive sorting, meaning that better schools are more accessible to households with higher socioeconomic status. Jones notes that this finding suggests that policymakers should look to implementing changes that level the playing field regarding access to quality schooling.

“Minimally, provide children from more disadvantaged circumstances with access to better quality teachers and schools,” he says. “Because they seem to be doubly disadvantaged. They’re disadvantaged by their family circumstances, and they’re disadvantaged by having poorer quality schools.”

While this paper helps to fill a significant gap in the literature, the authors note that more research is needed to be able to identify what aspects of school quality are driving the effects. They also point out that there is regional variation among the factors, with areas near each country’s capital displaying comparatively low overall test score inequalities, “as well as larger contributions from the household effects and smaller contributions from both school and sorting affects.”

“There seems to be big differences across regions of these countries,” Jones says. “That would be something to dig into as well in the future.”

These findings make clear the importance of identifying and analyzing the drivers of educational inequalities. Making quality education available to all children not only strengthens the economic potential of a nation, it also ensures that the right to an education is as equitable as it can be.  As the authors write, “If able children do not achieve their educational potential, countries face potentially significant losses against the counterfactual where all have an equal opportunity to develop their talents and skills.”