By Andreas Schleicher
Director, OECD Directorate for Education and Skills
I first visited Italy’s National Institute for Evaluation (INVALSI) in 1989. In those days, when it was called the CEDE, it was a place where academics debated educational research and contributed to international comparative studies. Back then, few would have thought the institute would build a comprehensive national assessment of the Italian school system. But two decades later, Italy has done just that. The country’s state-of-the-art assessment culture provides broad national diagnostics and tests the performance of students in multiple subjects and grade levels in all Italian schools.
But that’s still the easy part. A much harder task is to convert test results into meaningful feedback that can help improve teaching and learning, and enable schools to become more effective. Publishing league tables of schools doesn’t do the trick, because the performance of students and schools can depend on many things beyond their control – including, most notably, the social and economic background of student populations.
Converting results into meaningful feedback requires looking beyond aggregate school results and tracking individual student learning outcomes alongside relevant contextual information that can help explain performance differences. Not many countries have been able to do that well and systematically, but Italy is one of the few that have.
The most promising path toward building a better, fairer, more effective and more inclusive school system in Italy.
Since 2016, Italy has provided every school with an assessment of not only the quality of its learning outcomes, but also the value it adds relative to other schools for 5th, 8th and 10th grades. Schools can use those metrics to compare themselves with regional and national comparators and with “statistical neighbours”, i.e., schools operating in a similar social and economic context. At individual and school levels, these “value added” estimates account not just for the social background of students and basic demographics, but also for the learning trajectories and test results of students at previous stages, using sophisticated multi-level regression analyses.
This month, 30 years after I first visited the CEDE, Italy published its first comprehensive analysis of the data. Its analysis provides an amazing resource for school improvement and policy development, and I am sure policy makers and researchers around the world will read it with great interest. The results show that, for individual students, preceding test results tend to be the most powerful predictor of learning outcomes, well ahead of social background. For schools, this implies that performance-related interventions targeting students with learning difficulties can be very important, together with socio-economically targeted interventions that focus on students from disadvantaged backgrounds. At the school level, the effect of prior performance tends to become even more important as students grow older, underlining the importance of early interventions to protect students from falling behind.
Importantly, the data allow us to compare the results that schools achieved with what could be expected of them, given their student population and context. This gives educators and policy makers a powerful tool to target support for school development. To this end, INVALSI groups schools by their actual performance compared with their predicted performance. Over two-thirds of schools fall within their expected value, but some do better and can therefore provide inspiration for other schools.
Will Italy and its schools learn from these data? It may take another decade before this new evidence-informed culture will take root among teachers, schools and the education system. But it is the most promising path toward building a better, fairer, more effective and more inclusive school system in Italy.