Learning Profiles: The Learning Crisis Is Not (Mostly) about Enrollment

The differential patterns of grade progression have direct implications for the calculation of learning profiles. Researchers measure learning in primary school using survey data on reading and math skills of a nationally representative, population-based sample of children in India, Pakistan, Kenya, Tanzania, and Uganda. Research demonstrates that learning levels among primary and secondary students in much of the developing world are extremely low and lag far behind the rich world, constituting what is referred to as a "learning crisis". Beyond demonstrating the learning crisis, this study seeks to estimate how much children learn--measured by reading and mathematics tests--as they progress through primary school. The main methodological challenge is that while the ASER and Uwezo surveys test children at all relevant ages and grade levels, they do not track the same children over time. Researchers are left to infer a longitudinal learning trajectory from cross-sectional data. Results from conducting some simple regressions of scores on grade level to determine how learning changes as children progress through school suggests similar or more rapid learning in East Africa than South Asia. However, the authors hypothesize that this finding may be biased by non-random dropout and grade repetition. This paper explores this hypothesis and discusses strategies for creating learning profiles with cross-sectional data. Tables are appended.

an: ED567233; source: Society for Research on Educational Effectiveness; docTypes: Reports - Research; pubTypes: Report;