From Assessment Data to Better Evidence
From Assessment Data to Better Evidence
Why learning assessment data should be treated as long-term public assets
Over the past two decades, African governments and their partners have invested substantially in measuring foundational learning. National assessments, regional programmes and household-based surveys have generated an increasingly rich body of evidence on what children know and can do.
Yet the value of an assessment is often judged primarily by the report it produces. Once the headline findings have been published and disseminated, attention shifts to the next assessment cycle. The underlying microdata can become difficult to find, poorly documented or inaccessible to researchers who could use them to answer new questions.
Our new report argues for a different way of thinking about assessment data: not as a by-product of an assessment exercise, but as a long-term public asset.
The assessment report should be the beginning, not the end
An assessment is designed to answer a particular set of questions at a particular point in time. But the microdata it generates can answer many more questions than could ever be addressed in a single report.
Once data are carefully documented, curated and made accessible, researchers can revisit them to examine inequalities, test the robustness of headline findings, explore methodological questions, compare results with other datasets and investigate issues that were not priorities when the assessment was first designed.
This creates additional value without requiring another round of data collection. Every high-quality secondary analysis increases the return on the original investment in the assessment.
Preservation and documentation matter
Open access is not simply a matter of putting a dataset online. Data need to be preserved in a trusted repository, accompanied by the questionnaires, codebooks, sampling information, weighting documentation and technical materials required to understand how the data were generated.
Our report frames this using the FAIR principles: data should be Findable, Accessible, Interoperable and Reusable. Findable and Accessible speak to whether a dataset can be located and obtained at all. Interoperable and Reusable speak to something easily overlooked — even a published dataset is of limited use without the documentation that allows it to be understood, combined with other data and analysed correctly.
Without this supporting information, data may technically be available but remain difficult — or unsafe — to reuse. Good stewardship is therefore part of the research infrastructure required to turn assessment data into evidence.
Access currently varies considerably across assessment programmes. Household surveys such as MICS and school-based programmes such as PASEC, AMPL and PIRLS now make their microdata publicly available, and our own work has extended this to the PAL Network's ICAN-ICAR datasets. Nationally commissioned assessments remain the biggest gap: even where these are technically robust, their microdata are rarely released beyond the commissioning ministry, limiting independent verification and secondary analysis.
A virtuous cycle
The relationship between assessment and secondary analysis is not one-way. Reanalysis can identify problems in sampling, measurement, language, timing or reporting that were not apparent in the original assessment report. Those insights can then inform the design of future assessments.
In this way, open microdata can support a virtuous cycle: better access enables better analysis; better analysis generates better evidence; and that evidence can strengthen both policy and future assessment practice.
A long-term investment in evidence
Africa has already made substantial investments in foundational learning measurement. Maximising the value of those investments requires attention not only to collecting high-quality data, but also to what happens afterwards.
When assessment datasets are preserved, documented, shared and reused, their useful life extends far beyond the publication of a national report. They become part of a cumulative evidence base that can support research, policy and methodological improvement for years to come.