What about these Covid-19 learning losses?
Lessons from the 2005 Pakistan Earthquake
3 September 2021
As educational institutions across Pakistan slowly reopen for a new academic year, an underlying anxiety is taking root: by how much has the learning of Pakistani students been set back? If lessons from recent scholarship are anything to go by, our youngest children are likely to pay throughout their lives when a crisis of this magnitude occurs. Through an analysis of a natural experiment conducted four years after the 2005 earthquake that devastated the northern areas of Pakistan, Tahir Andrabi, Ben Daniels and Jishnu Das discover an intriguing paradox. Although state relief efforts were assessed through complete recovery at the household and adult levels as key indicators of stability, the longest lasting effects were on children.
In the Andrabi et al. study, children living close to the fault line in the 2005 scenario went on to achieve lower test scores than those living farther from the fault line. These gaps – identified 4 years after the incident - represent learning losses equivalent to missing approximately two years of schooling. In a setting like Pakistan where the out-of-school population is already one of the largest in the world, and where enrollment does not guarantee actual learning, the implications of findings such as these should startle policymakers into action.
Large-scale disruptions beyond the educational context will have significant consequences on the learning abilities of our youngest members: the basic knowledge and skills our children hang on to, and build upon for higher order thinking. This has long-term socio-cultural and economic ramifications for any country, especially one such as Pakistan where at least 1/3 of the population is still below 15 years of age.
So what must a relevant conversation now look like for effectively mitigating learning losses incurred?
Data: gather lots of it and use it sensibly
In the case of the Covid-19 pandemic, there exists tremendous heterogeneity in the disease’s impact on households and their children. The pandemic represents a bundled shock: the effects on health, income, and employment are complicatedly interwoven, widespread, and produce a multitude of typologies depending on the type(s) of household under consideration. Cross-sectional variation is therefore one of the most important aspects to new datasets that will have to be examined to develop a realistic understanding of how the pandemic has affected Pakistani students.
This raises important questions around what the Pakistani state and its citizens want to know about Covid effects, whether in socio-economic or other categories that mediate student opportunity and performance. Talk of variation also suggests a need for large-scale surveys to become more nuanced and creative down to household levels for more meaningful data collection. This could imply paying more attention to the female (student) condition because of Covid lockdowns or explaining outcomes on learning through the use of both numerical and qualitative variables.
What we are able to gauge so far is that students have been affected by an inequality of resources, whether of basic infrastructure, digital avenues or mental support. For instance, a simulation run by the World Bank predicts learning losses of 0.3 to 0.8 years for an average student in Pakistan, and almost 930,000 children dropping out once schools fully reopen. But what might mediate such losses on the academic front? One answer is that Pakistan's students have suffered - and are likely to continue to - from mental and physical constraints due to school closures, which will affect the rate, and comfort level of their learning. Additionally, state-led interventions to counter school closure (such as TeleSchool and TaleemGhar) have struggled to make effective outreach on this front: less than 1 in 4 students watched government TeleSchool; approximately 40% did not have a TV; and parental wealth had a direct relationship with the extent of a child’s engagement with digital education.
Beyond some of this scholarship, still little can be said about the scale of learning loss amongst Pakistani students - especially at primary levels, and in government schools. For instance, although the Government of Khyber Pakhtunkhwa’s effort to respond to lockdowns was recently woven into its public teachers’ continuous professional development programming in the form of back-to-school diagnostics, the item bank and pedagogic strategy chosen did not embed as well as intended in real classroom settings (the subject of a forthcoming blog on this forum).
The bottom line is that an urgent need persists to engage education researchers for targeted investigations into the design, delivery and analysis of diagnostic assessments. Without rapid insight to the nature of learning loss incurred by students, Pakistan will struggle to protect recent modest improvements to basic learning outcomes.
In contrast to current debates across the country around whether to move towards a more centralised curricular approach, one way to resolve Covid-induced educational setbacks might be to pursue tailored approaches. In the earthquake paper, Andrabi et. al emphasise the need for system-wide changes that are creatively delivered through stakeholders with whom children are more immediately familiar. The lesson here is that parents, teachers, school leaders, and other trusted community members can become co-creators of knowledge sets with researchers. The latter, in particular, can help iteratively and continuously process datasets to keep remedial instruction contemporary, hence responsive, to student needs.
Of course, this is an ambitious approach for an educational market as large as that of Pakistan. But it isn’t impossible.
One way to break down such a significant task is to attempt to meet students where they need help the most. Also known as Teaching at the Right Level (TaRL), this approach highlights the importance of assessing baseline skills and conducting remedial instruction in accordance with the student's base level. TaRL is a pedagogic approach that targets literacy and numeracy skills through two techniques: it groups students according to their learning level rather than grade level; and it tailors educational material to the student's needs and engagement level.
In Pakistan, the flexibility of TaRL has helped ensure that children learn better at higher levels, suggesting potential for the philosophy’s success in our local context. But teaching at the right level isn’t just about academic instruction. Resources can be tailored to emphasise joy and cooperation at levels a student understands to support their interest in learning, keeping them not just engaged, but also likely to advocate their return to school as the sector reopens.
Another such way is to reimagine how education happens for a majority of Pakistani students by tapping into, or recreating, existing marketplaces from a different angle. Take WhatsApp: one of the most prevalent low-tech digital applications used across Pakistan, which when used by inventive teachers to distribute learning packs to locked down students has been found to be twice as impactful to learning gains than traditional instructional methods. Other attempts to create digital personalisable education marketplaces have met with similar successes: bringing together either students and hard-to-reach learning content or supplying members of school management chains with the conversations and skills required to co-construct Communities of Practice.
In the Andrabi et al. study, one of the most curious outcomes is that children with educated mothers experienced lower learning losses after the earthquake, a finding echoed in other contexts across Pakistan. If mothers can, indeed, tap into a maternal instinct to ‘protect the kids’, this suggests a socio-emotional skill can be instrumented into state policy. This is a thought process that has been tested by a headteacher in Sindh to successfully engage parents as continuous partners in the implementation of Covid SOPs.
Although an unusual interpretation of ‘resilience’, this framing of mothers as champions of their children’s education should come as no surprise. Instead, it represents avenues for social insurance of children and their learning opportunities by designing policy that is not just ‘child-centred’, but instead ‘mother-child-centred’: what is achievable by not just children, but also by their mothers for them.
And once again, what this finding reinforces is that the protection and education of women is central to the success of any education system’s aspirations. What the Pakistani state must do differently now is to recognise that the trade-off between investing in human capital and immediate aid disbursements is a false dichotomy, most often evidenced by a returning reliance on the girl child’s educational stipend. Put simply, if Pakistan is to return to any track of productivity, it will not – and cannot – happen without a strategy for rapidly reviving the education system from its current ‘medically induced coma’.
Soufia A. Siddiqi, Assistant Professor, LUMS School of Education
Momina Idrees, Research Coordinator, MHRC
Musharfa Shah, MPhil Education Leadership and Management 2022, LUMS