Session Submission Summary

Direct link:

Learning from learning profiles: implications for learning goals and equity

Mon, April 15, 1:30 to 3:00pm, Hyatt Regency, Floor: Atrium (Level 2), Waterfront C

Group Submission Type: Formal Panel Session


Learning profiles from diverse datasets present a shared message: learning per year of schooling varies massively across countries, and learning profiles in many countries are remarkably flat. Learning profiles provide critical implications for achieving learning goals, including the Sustainable Development Goals of quality schooling and universal literacy and numeracy. They also give new evidence on what it will take to achieve equity of learning outcomes for segments such as girls and the poor.

Learning profiles illustrate the empirical relationship between schooling attainment (e.g. years of schooling) and learning (e.g. skills or capabilities) achieved. In Nigeria for example, learning profiles show that only 19 percent of young adults who completed primary school can read three simple sentences, while in Uganda and Bangladesh only about a quarter can. Demographic and Health Survey data across more than 50 countries show that only half of young adult women who had completed grade six could read a simple sentence. While these are descriptive learning profiles, rather than causal, the typical concern is that descriptive results will over-estimate impacts, making the low learning per year even more concerning.

The papers in this panel use data from more than 50 low- and middle-income countries. Each includes an assessment of a skill, such as the ability to read a simple sentence or solve a math problem, and a measure of number of years or highest level of schooling. The DHS data have included a skill assessment – a simple literacy test – only in recent years, making this a novel, comparable measure of learning across a large number of countries. The ASER and Uwezo surveys assess children’s reading and mathematics ability in India, Pakistan, Kenya, Tanzania, and Uganda. The Indonesian Family Life Survey (IFLS) includes a set of eight math questions, allowing comparisons of math skills with curricular expectations.

All data sources use household-based sampling, rather than school-based sampling, and therefore assess both children who are in and out of school, or adults across the full range of schooling attainment outcomes. The datasets include representative or nearly representative samples, and in the case of ASER cover more than half a million children in India.
The papers present three key findings:

1. Achieving schooling attainment goals such as universal primary completion, or even universal secondary completion, will leave many children without basic skills such as literacy and numeracy.

2. Achieving equality of learning outcomes across segments, such as the poor and rich, or girls and boys, will fall short of global goals of learning equity for all as even the more advantaged segments in many countries have low learning levels.

3. Low and heterogeneous learning per year mean age-based classrooms contain a wide range of student capabilities, potentially making effective teaching difficult and suggesting a need to reorient around skill levels.

The papers collectively show that reaching learning goals in most countries will require dramatic steepening of learning profiles. Further, the papers illustrate the ways the policymakers themselves can use learning profiles to assess their own education systems and inform policy action.

Sub Unit


Individual Presentations