Raising adult literacy without raising the literacy of adults? A cross-national analysis of literacy trends from a cohort perspective

• In Event: Adult literacy in multiple contexts

Abstract

There is a potential disconnect between adult literacy initiatives, campaigns, and programmes on the one hand, and the indicators typically employed to frame global targets and measure their progress. While much of the policy discourse on raising adult literacy is framed in terms of turning illiterate adults into literate adults, the main international indicator, the overall adult literacy rate, is influenced not just by such individual transitions from illiteracy to literacy, but also, or even mainly, by the process of cohort replacement: young, possible better educated, individuals cross into the adult age bracket, while older individuals depart. This leads to an improvement in the adult literacy rate that is perfectly real, but without necessarily benefitting any individual illiterate adult.

Differential mortality and migration may further affect the composition of the adult population in ways that affect the adult literacy rate. This would be the case if literates and illiterates are subject to different risks of dying in a given time period, or to different rates of migration. Neither of these are implausible, indeed assumed benefits to health are one of the rationales for investing in adult literacy in the first place. As a result, the trajectory of adult literacy differs substantially depending on whether we look at a fixe aged group in cross-section or at fixed cohorts over age. Analysing literacy trends along cohort lines therefore provides an important complemententary perspective if we seek to understand changes in adult literacy and their possible causes.

The present study, a background paper for the upcoming 2015 Global Monitoring Report, performs an analysis along these lines. I compare fixed-age cross-sectional and fixed-cohort longitudinal adult literacy rates; for the latter, I estimates bounds on the distorting effect of differential mortality and migration. HIV/AIDS mortality is treated separately, because the direction of the education gradient may be less consistent in this context.
In addition, a decomposition is performed to asses the contributions to the changing cross-sectional adult literacy rate of literacy conditional on school attainment on the one hand, and changing educational attainment composition on the other. For the subset of countries where literacy did not change substantially over adult ages, this observation is exploited to reconstruct the literacy outcomes among school graduates over time. Finally, the role of adult literacy programmes is investigated, as is the issue of partial vs full literacy.

This study uses Demographic and Health Survey (DHS) data on the education and measured (rather than self-reported) literacy status of women aged 20 to 49 for 30 countries. An additional perspective on the literacy of adult males, who are less well covered in DHS, is provided based on census data for five of the ten countries with the largest absolute numbers of illiterates, namely Bangladesh, Brazil, India, Indonesia, and Egypt. Unfortunately, these data only contain self-reported literacy.

Despite its limitations, its results provide an important, nuanced perspective on efforts to improve adult literacy in many countries. In particular, it shows that observed gains in the adult literacy rate overstate the degree to which individual women have become literate. In some countries there are, however, changes in literacy along cohort lines, both upwards and downwards, that cannot be easily attributed to selective migration or mortality and may, therefore, represent ‘true’ gains or losses in individual literacy. Nepal is found to appear as a particularly strong outlier in terms of improvements in literacy at adult ages that cannot be accounted for by any of the other modelled effects, and that do not merely reflect a shift from partial to full literacy. At the present level of analysis, there is no evidence, unfortunately, that such gains can be related to reported participation in adult literacy programmes. Indeed, a tentative conclusion is that some past exposure to primary schooling is necessary to derive sustainable benefit from them. If corroborated, this would have important implications for policy, in terms of helping the most disadvantaged who never received any schooling. For countries in which literacy gains at adult ages were found to be marginal, consistent time series of youth literacy rates can be reconstructed from current cross-sectional data. Notably, those results suggest that the literacy outcomes of primary schooling in general were not "diluted" at times of rapid expansion of school participation.

Author

• Bilal Barakat, Wittgenstein Centre